Embed Size (px)
Citation preview
Acknowledgements
Implications • NOx influence on SOA formation from terpenes is much more variable than the accepted suppression theory. • Map of terpene distribution in the US (left) shows that α-pinene is dominant in many locations, but not everywhere. • Most large scale regional atmospheric models include a single monoterpene as representative of the class of compounds. • If NOx enhances aerosol growth from some monoterpenes but not others, models including only a single monoterpene could be under-predicting SOA formation. • Could help decrease the gap between modeled and measured SOA.
α-pinene
β-pinene
Δ-carene
d-limonene
camphene
β-myrcene
(Geron, 2000, Atmos. Environ.)
Percentage of monoterpene emissions:
Major results • NOx suppresses SOA formation from α-pinene but may enhance SOA formation from other monoterpenes. • SOA yields vary dramatically from different monoterpenes and different oxidants. • This significant variation is not currently included in large-scale atmospheric chemistry
models; its incorporation could improve predictions of SOA formation.
Cavity Ring-‐‑down Spectroscopy
Background Secondary organic aerosol (SOA) forms from the reaction of volatile organic compounds (VOCs) with atmospheric oxidants (OH, O3, NO3). Some of these oxidized organic products are much less volatile than their precursors and are able to form particles.
Estimates indicate that close to 90% of global VOC emissions originate from biogenic sources (i.e. trees), but anthropogenic pollution contributes to the majority of the oxidants in the atmosphere. NOx (NO and NO2) is formed in the process of combustion and contributes to both O3 and NO3 formation:
NO2 + hν NO + O
O + O2 + M O3 + M
NO2 + O3 NO3* + O2
A series of dark chamber experiments were performed in Reed College’s environmental chamber (400 L Teflon bag) beginning December 2012 to assess the effects of NOx on SOA formation from ozonolysis of different monoterpenes. Reagents include:
• 300 ppb monoterpene (vapor from cooled liquid) • 500 ppb O3 (UV photolysis of ambient air) • 0-1700 ppb NO2 (calibrated cylinder)
Reagents are measured using GC-FID, optical O3 detection, and chemiluminescence NOx detection. Aerosol is measured with a scanning electrical mobility sizer (SEMS) and offline HPLC-ESI-MS.
Methods
hydrocarbon source
O3 source
NOx box
clean air source
O3 monitor NO2
Additional experiments were carried out in a 10000 L chamber at NCAR in Boulder, CO using 10-50 ppb monoterpene and comparable concentrations of N2O5 as a NO3 radical source.
Monoterpenes studied:
α-pinene β-pinene Δ-carene limonene These 4 VOCs encompass the dominant monoterpene emissions in the US.
2
3
4
5
6
7
8
9
100
2
3
10.410.09.69.2Time (hrs)
2
3
4
5
6
7
8
9100
2
3
12.412.011.611.2Time (hrs)
2
3
4
5
6789
100
2
3
Diam
eter
(nm
)
6.46.05.65.2Time (hrs)
2
3
4
5
6
789
100
2
Diam
eter
(nm
)
21.521.020.5Time (hours)
2
3
4
5
6
7
89
100
2
3
4
5
6
7
8
Diam
eter
(nm
)
1:20 PM1/11/13
1:45 PM 2:10 PM
Time
2
3
4
5
6
7
89
100
2
3
4
5
6
7
8
Diam
eter
(nm
)
3:25 PM1/9/13
3:50 PM 4:15 PM 4:40 PM
Time
2
3
4
5
6789
100
2
3
4
5
6
78
Diam
eter
(nm
)
4:15 PM1/7/13
4:40 PM 5:05 PM 5:30 PM
Time
2
3
4
5
6789
100
2
3
4
5
678
Diam
eter
(nm
)
3:25 PM1/5/13
3:50 PM 4:15 PM
Time
150x103100500
Total Particles (nm3/cm3)
Chamber experiment results
β-pinene +
O3
Danielle Draper,1 Delphine Farmer,2 Yury Desyaterik,3 Juliane L. Fry1 1Chemistry Department, Reed College, Portland, OR.
2Chemistry Department, Colorado State University, Fort Collins, CO. 3Department of Atmospheric Science, Colorado State University, Fort Collins, CO.
Correspondence: [email protected], [email protected]
SOA yield from ozonolysis of BVOC at varying NO2 concentrations
β-pinene +
NO3
α-pinene +
O3
α-pinene +
NO3
Δ-carene +
O3
Δ-carene +
NO3
limonene + O3
limonene +
NO3
VOCs
O3 NO3 +
OH VOCs
oxidized VOCs
SOA
(Kroll & Seinfeld, Atmos. Environ., 2008)
Once VOCs are oxidized, they undergo rapid radical chemistry (right) including self-reactions and reactions with NO or NO2. A variety of products are formed, but only some of them are able to condense into the particle phase.
Previous studies (Perraud, 2011; Presto, 2005) have observed that NO2 suppresses SOA formation in the reaction of O3 with α-pinene, which is taken to be representative of the NO2 effect on SOA formation from ozonolysis of any monoterpene. In both cases, the observed results are explained by the formation of organic nitrates (RONO2), which are assumed to be less volatile than some other oxidation products. These organic nitrates can be formed either through RO2 + NOx channels or as a direct result of NO3 oxidation (formed from the O3 and NO2 present).
*NO3 is rapidly photolyzed and thus is relevant primarily at night.
(Perraud et al., 2011, PNAS)
(Right) FTIR spectra of SOA from the Perraud study show an increase in intensity of the three characteristic NO3 peaks with increasing NO2 concentration and a simultaneous decrease in intensity of the carbonyl peaks – indicating a shift in products.
Aerosol growth curves from chamber experiments for different monoterpenes: • O3 experiments used 500 ppb O3 with 300 ppb VOC • NO3 experiments used 50 ppb each of N2O5 and VOC and the α-pinene experiment included (NH4)2SO4 “seed” aerosol, which did not grow. Observations: • α-pinene makes an especially large amount of aerosol with O3 and NONE with NO3. • vastly different aerosol yields are observed for different monoterpenes • endocyclic double bonds (α-pinene and Δ-carene) exhibit higher yields with O3 oxidant; exocyclic double bonds with NO3. • limonene (2 double bonds) makes copious aerosol with either oxidant.
What if the aerosol suppression observed in previous studies is simply an artifact of α-pinene’s anomolous behavior when oxidized? é [NO2] leads to é[NO3] leads to êSOA
α-pinene limonene
Number concentrations and derived particle volume concentrations are plotted below for all four monoterpenes studied. NO2 is observed to suppress total number of particles and total mass from α-pinene ozonolysis, as observed in other studies. Suppression is observed in total number, but enhancement of total mass is observed for both β-pinene and Δ-carene with increasing [NO2]. Limonene yields are enhanced in both number concentration and mass with the addition of NO2.
Many thanks the Reed College Chemistry Department for support in conducting this research, and to Jim Smith, Steve Brown, and John Ortega for hosting the chamber experiments at NCAR.
1.20.80.40.0Hours From Start
O3 only low NO2
medium NO2 high NO2
1.20.80.40.0Hours From Start
1.20.80.40.0Hours From Start
101
102
103
104
105
Tota
l Par
ticle
s (#
/cm
3 )
107
108
109
1010
1011
Tota
l Vol
ume
(nm
3 /cm
3 )
1.20.80.40.0Hours From Start
β-pinene Δ-carene
20181614121086420Retention Time (min)
Size Distribution Data:
Compositional Data: • Aerosol samples were collected on filters during each chamber experiment
• Filters were extracted and shipped to CSU for offline analysis by HPLC-ESI-MS
• Qualitative compositional differences are observed in SOA formed from different precursors
(Right) Chromatograms shown from SOA generated from β-pinene ozonolysis at varying [NO2]. As é[NO2] new peaks appear in addition to the ozonolysis peaks.
β-pinene + 750 ppb NO2
β-pinene + 400 ppb NO2
β-pinene + O3 only
700
600500400
300200100
Diam
eter
(nm
)
2.01.00.0Hours From Start
O3 only 220 ppb NO2 520 ppb NO2 900 ppb NO2
2.01.00.0Hours From Start
O3 only 280 ppb NO2 560 ppb NO2 930 ppb NO2
700600
500400300200100
Diam
eter
(nm
)
O3 only 300 ppb NO2
O3 only 400 ppb NO2 750 ppb NO2
450
400
350
300
250
200
150
100
50
0
[Pre
curs
or]
(ppb
)
20.019.519.018.518.017.517.016.5Hours From Midnight
1000
900
800
700
600
500
400
300
200
100
0
[Products] (ug/m3)
modeled [b-pinene] modeled [O3] modeled total [O3 products] modeled condensable [O3 products] observed [O3] observed aerosol mass
450
400
350
300
250
200
150
100
50
0
[Pre
curs
or]
(ppb
)
20.019.519.018.518.017.517.016.5Hours From Midnight
1000
900
800
700
600
500
400
300
200
100
0
[Products] (ug/m3)
modeled [b-pinene] modeled [O3] modeled total [O3 products] modeled condensable [O3 products] observed [O3] observed aerosol mass
450
400
350
300
250
200
150
100
50
0
[O3,
NO 2
, N2O
5, b
-pin
ene]
(pp
b)
18.017.016.015.014.013.012.011.0Hours From Midnight
240
220
200
180
160
140
120
100
80
60
40
20
0
[Products] (ppb)
modeled [NO2] modeled [N2O5] modeled total [NO3 products]
400
300
200
100
0[O3,
NO 2
, N2O
5, b
-pin
ene]
(pp
b)
1817161514131211Hours From Midnight
240
200
160
120
80
40
0
[Products] (ppb)
450
400
350
300
250
200
150
100
50
0
[O3,
b-p
inen
e] (
ppb)
20.019.018.017.0Hours From Midnight
1000
900
800
700
600
500
400
300
200
100
0
[Products] (ug/m3)
Kinetics Modeling:
Growth Rates: • Aerosol growth curves were plotted for each experiment, tracing the diameter at which the peak # concentration occurs • Growth rates differ drastically for different terpene and oxidant conditions. • Growth rates should be proportional to the concentration of the gas-phase condensing species ( ) according to the equation: • values provide a kinetically- driven, lower limit constraint on the steady state concentration of condensable vapor, which will be compared to box model simulations.
χ∞
(Brock et al., GRL 2011)
dDp
dt=4Dvχ∞F(Dp )
ρpDp
limonene
β-pinene
Δ-carene
α-pinene
χ∞
Ongoing Analysis
Box model simulation of β-pinene + O3 experiment. Observed aerosol mass can be fit as a fixed 12% of the total O3 oxidation products, accounting explicitly for gas and aerosol dilution losses, suggesting a kinetically-driven condensation mechanism.
Box model simulation of gas-phase chemistry in the β-pinene + O3 + “low” NO2 experiment. β-pinene introduced at hour 14. Comparing O3 products and NO3 products shows the dominance of NO3 chemistry in the NO2 experiments.
