Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
!"#$%&'(
)*+,*-.,/0(
!"#$%&'()#!"#*)+)(+%'#
1*+(2344(
0-1(25$678(57(1*+(29:&'$4(
1;9(<5=(>?@,0A-BC(
!"#$%&'#()#$*%+)!,#'*)-#%.)/0)/$%)/0)+./)
1%(0/('+%2)341/5%46))'+)%$2)/0)"/-'7*%)8*'44)#$*%+)
9:;)1(/7%)
<=>/-'7*%)4+'#$*%44?)4+%%*)4'"1*%)#$*%+)
@#,%2)1%(0/('+%2)#$*%+)(#$84)
A%$+)+/)B//2)
CD9)E'+%()F'G5%+)
;H"#2#I%2)'#()#$*%+)
>J!)#$*%+)+/))"/4+)
2/.$4+(%'")41/5%4)))
K$*%+)+/)"#,#$8)'#()41/5%4)
L/)2%+%G+/()
0.3
0.2
0.1
0.0
-0.1
Diffe
renc
e
1000800600400200m/z
X
X 500-nm Part.
250-nm Part.
Monomers
Dimers
Oligomers
Uptake -‐ liquid PEG
Uptake -‐ SOA
Loss -‐ liquid PEG
Loss -‐ SOA
Nitrate absorbance @
1280 cm
-‐1
Time (min)
OH
DMA OH
DMA
Véronique Perraud,1 Ma1hew L. Dawson,1 Carla Waring-‐Kidd,1 Michael J. Ezell,1 Mychel E. Varner,1 R. Benny Gerber,1 Andrew S. MarInez,2 Donald Dabdub2 and Barbara J. Finlayson-‐Pi1s1
1Department of Chemistry, University of California, Irvine, CA, United States; 2Department of Mechanical and Aerospace Engineering, University of California, Irvine, CA, United States
New ParIcle FormaIon and Growth From the ReacIon of Methanesulfonic Acid with Amines and the Ozonolysis of Terpernes
1x104
0
Tota
l Par
ticle
Con
cent
ratio
n (p
artic
les
cm-3
)
2.0x106
1.5
1.0
0.5
MSA+
TMA
MSA+
H2O
MSA+TMA+H2OExcess TMA
MSA+TMA+H2OExcess MSA
% RH
% RH
4% 9% 20% 3%
8%
19%
18%0% 1x10
4
0
Tota
l Par
ticle
Con
cent
ratio
n (p
artic
les
cm-3
)
1.0x106
0.5
MSA+
DMA
MSA+
H2O
MSA+DMA+H2OExcess DMA
MSA+DMA+H2OExcess MSA
% RH
% RH
3%
8%
18%
8%12%
21%
18%
0%
TMA DMA
No Am
ine
No Water
No Am
ine
No Water
Figure 2: Complex dependence of parKcle formaKon on precursor concentraKon in the MSA-‐amine-‐water system, for trimethylamine (TMA) and dimethylamine (DMA).13
• Atmospheric aerosols negaKvely impact human health, reduce visibility and affect the climate by scaUering and absorbing solar radiaKon and changing cloud properKes.1-‐4 • Models typically consider sulfuric acid (H2SO4) nucleaKon as the major source of new parKcles in the atmosphere. However, these methods consistently underpredict parKcle formaKon, indicaKng that other sources and/or co-‐nucleaKng species may play a role.2,3,5 • Ammonia and amines have recently been idenKfied as important co-‐nucleaKng species in parKcle formaKon from sulfuric acid.6-‐12 Also, recent research at AirUCI using a flow tube reactor (Fig.1) has idenKfied methanesulfonic acid (MSA, CH3SO3H) and amines as a potenKally important source of parKcles.13
• ParKcle formaKon in these mulKcomponent systems shows a complex dependence on precursor concentraKon (Fig. 2), making them difficult to model using parameterizaKons based on nucleaKon theory. Two kineKcs-‐based nucleaKon mechanisms, one for sulfuric acid, amines and water, and one for MSA, amines and water (Fig. 3), have been proposed as accurate, computaKonally-‐inexpensive methods for predicKng parKcle formaKon from these systems.12,13
Figure 1: Flow tube uKlized for parKcle formaKon experiments
IntroducIon / Previous Work on MSA + Amines + H2O
Modeling
New Mini Flow tube Experiments
150 m
1100 m
40 m 0 m
310 m
670 m
80 Cells 30 Cells
123 Gas Species 296 Aerosols: 37 species, 8 sizes 361 ReacIons
Each Cell: 5 x 5 km2
Figure 4: UCI-‐CIT SpecificaKons
• UCI-‐CIT solves the Diffusion-‐AdvecKon-‐ReacKon equaKon in three dimensions in a 30x80x5 grid encompassing the South Coast Air Basin of California (SoCAB).
• Emissions of species besides MSA and amines follows 2005 emissions inventories.
Table 1: OxidaKon reacKon rates
• The species MSA, methylamine (MMA), dimethylamine (DMA) and trimethylamine (TMA) were included in the UCI-‐CIT model along with their oxidaKon reacKons with OH and ozone (O3).
Choose Emissions Rate Factor
EDMA = 15 pptv*m/min
Define Emissions Scenario
Set Chemistry Switch
AMINE + O3
AMINE + OH
AMINE + O3
AMINE + OH O
P Run Airshed Model
Choose SpaIal Metric
Specific LocaKons Domain-‐Wide Values
Evaluate ConcentraIon
100 pptv Other O
P
Figure 6: IteraKve model execuKon and analysis process
• Amine emissions were set to track those of ammonia. • The magnitude of the amine emissions was iteraKvely adjusted unKl average concentraKons were 100 pptv14,15 between 6 and 9 am. • Model results with and without oxidaKon chemistry were performed to assess the impact of oxidaKon reacKons on amine concentraKon.
Goal: Study the formaKon and growth mechanism/kineKcs of SOA from MSA + Amines +H2O at shorter reacKon Kme.
Results
Figure 8: Long Beach hourly variaKons in OH and DMA. Solid: OxidaKon Disabled. Dashed: OxidaKon Enabled.
Goal: Use a regional air quality model (UCI-‐CIT) to asses the impact of kineKcs-‐based nucleaKon mechanisms in the South Coast Air Basin of California
References: 1. Forester et al., IPPC Report (2007); 2. Finlayson-‐PiUs & PiUs (2000); 3. Seinfeld & Pandis (2006); 4. Pope & Dockery, W. J. Air Waste Manage. Assoc,. 56, 709–742 (2006); 5. Sipila et al., Science, 327, 1243–1246 (2010); 6. Angelino et al., Environ. Sci. Technol., 35, 3130–3138 (2001); 7. Berndt et al., Atmos. Chem. Phys., 10, 7101–7116 (2010); 8. Kirkby et al., Nature, 476, 429–U77 (2011); 9. Smith et al., PNAS, 107, 6634–6639 (2010); 10. Yu et al., Geophys. Res. LeJ., 39, (2012); 11. Zollner et al., Atmos. Chem. Phys., 12, 4399–4411 (2012); 12. Chen et al., PNAS, 109, 18713-‐18718 (2012); 13. Dawson et al., PNAS, 109, 18719-‐18724 (2012).
!"#$%&'(''
'
Figure 3: Calculated energy diagram and proposed kineKcs based nucleaKon mechanism13
Figure 5: Co-‐locaKon of amine parKcle precursor species in SoCAB
• Modeled amine concentraKons are a linear funcKon of emission factors (Fig.7)
• Fast oxidaKon rates lead to chemistry dominaKng over advecKon in determining amine concentraKons (Fig. 7)
• Magnitude of required emission factors varies slightly between amine species but by several orders of magnitude by geographical locaKon (as low as 45 in the domain’s peak, to 460 in Long Beach, and 1900 in Anaheim, all in units of pptV-‐m/min)
• Low amine concentraKons suggest a smaller role for amines in parKcle formaKon during the day
• CompeKKon for amine chemistry is driven mostly by oxidaKon via OH. Amines and OH are observed to have inversely correlated temporal variaKons (Fig. 8 and Fig. 9).
• Predicted emission factors are in reasonable agreement with global average measurements of NH3:amine raKos (Fig. 10). Differences are likely due to local variaKon in NH3 and amine emission rates.
SO2 NH3
Methodology
• Amines and reduced sulfur species (precursors for sulfuric acid and MSA) were not previously included in the UCI-‐CIT model. However, SO2 (an important source of sulfuric acid) and ammonia are included and show areas of overlapping concentraKon (Fig. 5).
Conclusions
• Under this emissions scenario, there appears to be a potenKal for parKcle formaKon in the morning hours.
• OxidaKon of amines by OH presents the greatest compeKKon to parKcle formaKon through amine-‐acid reacKons.
References:
14. Ge et al., Atmos. Environ., 45, 524-‐546 (2011) 15. Facchinni et al., Environ. Sci. Technol., 42, 9116-‐9121 (2008)
Figure 7: Long Beach concentraKon-‐emission trends
!"!!
!"!#
!"$!
!"$#
!"%!
!"%#
!
#!
$!!
$#!
%!!
%#!
! # $! $# %! %#
!"#$%&
'(&)*+),%&
#-..
)/
012#$%&
'(&)*+),%&
#-..
)/
" %3*
&'( &'( )* )*
!"#$%&'(&)*&)+,-./'0&
Figure 9: Top: OH and DMA at Kme of low OH oxidaKon BoUom: OH and DMA during Kme of high OH oxidaKon
Figure 10: Comparison of model emission rates to literature values
!"#$%&
!"#$%'
!"#$%(
!"#$%!
!"#) %%
**+ ,*+ -*+!"#$%&'()*+"
#,,#-$
,*./0%*
12%#34
0*5/,#,6
./012/3*45627 *054/3-082/
x4.8 x21 x5.6 Global Median14 UCI-CIT Model Total
Figure 14: Organic nitrate uptake on α-‐pinene/O3 SOA and Poly(ethylene glycol) seed parKcles (PEG)
1 -‐ ATR-‐FTIR Experiments
Goal: InvesKgate SOA phase and uptake of VOCs
-‐ IR spectrum of SOA coaKng -‐ IR signal monitored as a funcKon of Kme -‐ EvaporaKon rate and IR spectrum of evaporated material obtained
SOA evaporaKon
-‐ RONO2 uptake rate -‐ RONO2 loss rate
RONO2 uptake
• Very liUle evaporaKon, less than 20% loss in 20 hrs • Good agreement with results from Vaden et al., (2011) for lab generated and ambient SOA parKcles • EvaporaKon behavior is not consistent with that predicted based on instantaneous equilibrium parKKoning of known ozonolysis products
References:
Normalize
d pe
ak area
Time (hrs)
• Recent studies show that aerosols are semi-‐solid or glassy,16-‐27 which will affect the way SVOCs parKKon into SOA. • The uptake of SVOCs may be beUer represented by a condensaKon mechanism.28-‐30
• IBN uptake and loss much slower in SOA than liquid PEG model system • SOA has a higher viscosity than PEG (0.1 pa s) • Loss of IBN from SOA is very slow considering the high vapor pressure of IBN (10 Torr) • SOA behaving like a semi-‐solid and hindering diffusion • Aim to extend method to other organic nitrates/ SVOCs/model systems
16. Zobrist et al., ACP, 8, 5221-‐5244 (2008); 17. Virtanen et al., Nature, 467, 824-‐827 (2010); 18. Virtanen et al., ACP, 11, 8759-‐8766 (2011); 19. Vaden et al., PNAS, 107, 6658-‐6663 (2010); 20. Vaden et al., PNAS, 108, 2190-‐2195 (2011); 21. Cappa and Wilson, ACP, 11, 1895-‐1911 (2011); 22. Koop et al., PCCP, 13, 19238-‐19255 (2011); 23. Tong et al., ACP, 11, 4739-‐4754 (2011); 24. Mikhailov et al., ACP, 9, 9491-‐9522 (2009); 25. Saukko et al., Atmos. Meas. Tech., 5, 259-‐265 (2012); 26. Zelenyuk et al., EST, 46, 12459-‐12466 (2012); 27. Abramson et al. PCCP, 15, 2983-‐2991 (2013); 28. Perraud et al., PNAS, 109, 2836-‐2841 (2012); 29. Kleinman et al., ACP, 9, 4261-‐4278 (2009); 30. Riipinen et al., ACP, 11, 3865-‐3878 (2011); 31.Bruns et al., Anal. Chem., 82, 5922-‐5927 (2010); 32. Bruns et al., PCCP, J. Phys. chem., 116, 5900-‐5909 (2012)
Figure 11: Design of the mini flow tube.(Ezell et al., 2013)
Step 1: Residence Ime measurements
ParKcle conc. (# cm
-‐3)
ParKcle conc. (# cm
-‐3)
2 -‐ A New Approach for SOA analysis: Atmospheric Solid Analysis Probe Mass Spectrometry (ASAP-‐MS)31-‐32
HeatedN2
H2O
[ ] Coronadischargeneedle3 kV, 5 A
ASAPprobe
To MSanalyzer
Sampleon PEEKtube
M H --
2 -‐ Placement of the wafer on the new ASAP probe
[M+H]+
1 -‐ CollecKon of the SOA on Si-‐wafer
1.0
0.8
0.6
0.4
0.2
0.0
Norm
laiz
ed In
tens
ity
6005004003002001000Scan number
!"#$%
&""#$%
&!"#$%
'""#$%
'!"#$%
(""#$%
(!"#$%
)""#$%
Figure 16: Thermogram
1.2
1.0
0.8
0.6
0.4
0.2
0.0
Norm
aliz
ed In
tens
ity
1000800600400200m/z
167183
199
155
295313
339353
369
383
399
479493
Stage C500-nm particles
Integrated MS spectra (20'C to 350'C)
521
Fig. 17: Integrated Mass Spectrum (20’C to 350’C)
!"#$%&'($)$!"#$%&'($*$!+,$%&'($*$!+,$%&'($)$
!"#$
New ParIcle FormaIon and Growth from Ozonolysis of Terpenes
• Thermal desorpKon of the SOA products using a temperature ramp from 20°C to 450°C. • Sou ionizaKon forming [M+H]+ ions. • DetecKon using a Kme-‐of-‐flight mass spectrometer allowing accurate mass determinaKon.
• New sampling/analysis method: no more transfer of the sample!!! • The new method increases run-‐to-‐run reproducibility. • Able to compare different condiKons, such as those presented here for d > 500-‐nm parKcle vs. 250-‐nm < d < 500-‐nm parKcles (Sioutas impactor, stage C and D respecKvely; sampling at 9 Lpm, from Port 5 [31min rxn Kme])
(X = impurity peaks)
Figure 12: Experimental design
Figure 13: α-‐Pinene/O3 SOA EvaporaKon experiments
Results
Fig. 18: Comparison Spectrum
Figure 15: Sampling/Analysis sequence
!"#$%&'()*+&,-&.)/!012))34)5567)834)99:6+*)
Air
Results
Total Flow (Lpm) 6 11 17
Distance d (cm) tr(s) tr(s) tr(s) 3.5 1.2 0.8 0.5 0.9 2.5 1.5 0.9 12 4.2 2.6 1.6 22 7.7 4.9 3.0 32 11.2 7.1 4.3 42 14.7 9.3 5.7 52 18.2 11.5 7.1
• Laminar flow/residence Kme tested with the measurement of NO2 in air by UV-‐vis spectroscopy).
• First experiments with MSA + amines + H2O are underway (see B.J. Finlayson-‐PiUs’ talk on Tuesday/NPF breakout session)
d
Port 5 (31 min rxn Ime) p C=O n C–H
Port 1 (7 min rxn Ime) p C=O n C–H
Amine Oxidant Rate Constant (ppm/min)
MMA OH 3.4E+04 O3 1.1E-‐05
DMA OH 1.0E+05 O3 2.6E-‐03
TMA OH 9.3E+04 O3 1.2E-‐03
Table 2: Residence Kme in the flow tube
Amine:NH 3 Emissions Rate
(weight b
asis)