1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
In-Situ Measurements of Pinonaldehyde Gas/Particle Phase Partitioning
Brent J. Williams, Allen H. Goldstein, Dylan B. Milleta, Rupert Holzingerb
Dept. of Environmental Science, Policy, & Management, University of California,
Berkeley, CA, 94720, USA
Nathan M. Kreisberg, Susanne V. Hering
Aerosol Dynamics Inc., Berkeley, CA, 94710, USA
Douglas R. Worsnop
Aerodyne Research Inc., Billerica, MA, 01821, USA
James D. Allan
School of Earth, Atmospheric and Environmental Science, University of Manchester,
M60 1QD, UK
Jose L. Jimenez
Department of Chemistry, University of Colorado, Boulder, CO, 80309, USA
Short title:
Pinonaldehyde Phase Partitioning
Submitted to:
?
Index terms:
Author keywords: phase partitioning, mass spectrometry, pinonaldehyde
________________________________________________________________________
a Now at the Department of Earth and Planetary Sciences, Harvard University, Cambridge,
Massachusetts, USA
b Now at Institute for Marine and Atmospheric Research, Utrecht University, Utrecht, The Netherlands
39
40
41
42
Abstract
1
43
44
We examine the phase partitioning of pinonaldehyde, a terpene oxidation product, using
45
in-situ gas and particle phase measurements and show periods where observed
46
concentrations in the particle phase are far higher than expected using traditional
47
partitioning theory. This is evidence for either efficient secondary organic aerosol (SOA)
48
uptake of compounds such as pinonaldehyde which have relatively high vapor pressures
49
through formation of oligomers that can decompose during analysis giving back the
50
original pinonaldehyde, or further evidence for aldehyde production through the thermal
51
decomposition of unstable peroxide radicals.
52
53
Introduction
54
Many organic compounds found in aerosols are semivolatile, and thus may be present in
55
both the gas phase and in particles. The partitioning of semivolatile organics is driven by
56
their volatility which is temperature dependent, and by the organic aerosol mass present
57
to which these compounds can partition [Seinfeld and Pankow, 2003]. The partitioning of
58
such compounds between the gas and particle phase has rarely been quantified using field
59
measurements, but it has raised concerns over potential gas phase artifacts in aerosol
60
measurements and has created significant confusion in the literature with regard to how
61
particle phase measurements of semivolatile species should be interpreted.
62
63
The authors may wish to note that in a modeling study of SOA formation from the
64
reaction of ozone with a-pinene [Jenkin, Atmos. Chem. Phys. 4, 1741-1757 (2004)]
65
the product vapor pressures needed to be reduced by a factor of 120 relative to their
2
66
pure component values in order to achieve reasonable agreement between modeled
67
and measured SOA.
68
69
During the summer of
70
71
During our field study simultaneous aerosol measurements with TAG and gas phase
72
measurements with PTR-MS [Holzinger et al., 2006] provided the opportunity to
73
examine the gas/particle partitioning for pinonaldhyde, a biogenic terpene oxidation
74
product.
75
76
Calibration
77
Calibration for particle phase pinonaldehyde was performed following the ICARTT study
78
in order to compare measured concentrations to those observed in the gas phase by PTR-
79
MS. Particle phase pinonaldehyde concentrations were estimated using a two part
80
comparison of manually injected liquid standards. A comparison was made between
81
“off-line” injections of decanal and 4,4’-dimethoxybenzophenone, an oxygenated
82
standard that was also used in the field, to determine what the decanal response would
83
have been “on-line” at Chebogue Point. Pinonaldehyde (a C10-aldehyde) quantification
84
was calculated based on this decanal (a C10-aldehyde) response.
85
86
Gas/Particle Partitioning of Pinonaldehyde
87
Figure xx shows a plot of gas-phase and particle-phase pinonaldehyde concentrations,
88
along with air temperature, over a 24 hour period. Our observations show pinonaldehyde
3
89
partitioning favors the gas phase during midday elevated temperatures. The ratio of gas-
90
phase to particle-phase pinonaldehyde for the whole measurement period is shown in
91
Figure xx as a function of temperature, and ranges from 4.0 to 170. Only data greater
92
than 0.5 standard deviations above the detection limit is included for both gas and
93
particulate concentrations to avoid errors due to uncertainty in low concentration
94
measurements. We use these data to estimate the apparent enthalpy of vaporization
95
(Hvap) for pinonaldehyde from this slope (s = -19,000 ± 5400) using the Clausius-
96
Clapeyron relationship, to be 160 ± 45 KJ/mol. The vapor pressure (Pvap) for
97
pinonaldehyde was reported as 5.1 Pa at 298 K (3.8 × 10-2 Torr) [Hallquist et al., 1997].
98
Using the Clausius-Clapeyron equation with the value of the enthalpy of vaporization
99
determined here, we estimate Pvap = 1.5 × 10-2 to 0.16 × 10-2 Torr for pinonaldehyde over
100
the range of ambient temperatures in this study (20.5 ºC to 11 ºC, respectively).
101
According to partitioning theory, a species should have a Pvap of 1.5 × 10-5 Torr (at 18 ºC)
102
to 3.2 × 10-6 Torr (at 12 ºC) (i.e. about 500 - 600 times lower) to partition to the particle
103
phase at the levels we measured pinonaldehyde in aerosols at Chebogue Point (see
104
Figures xx,xx). Thus it is very likely that pinonaldehyde was present in the particle-
105
phase not only as its parent compound, but also as a thermal decomposition product of
106
unstable peroxides [Tobias et al., 2000] and/or the decomposition product of larger
107
oligomers [Jang et al., 2003; Barsanti and Pankow, 2004].
108
109
Based on chamber studies, Tobias et al. [2000] suggest that the major products (i.e.
110
molecularly small ketones, aldehydes, and alkanoic acids) in aerosol formed by the
111
ozonolysis of 1-octene and 1-decene (as performed by Forstner et al. [1997b]) were too
4
112
volatile ( > 2 × 10-3 Torr) to form much aerosol, but were detected instead probably as
113
decomposition products of a-hydroxyalkyl hydroperoxides, dihydroxyalkyl peroxides, a-
114
acyloxyalkyl hydroperoxides, and possibly, secondary ozonides. Whereas, Jang et al.
115
[2003] demonstrated in the lab that particle growth occurs from the acid-catalyzed
116
heterogeneous reactions of small carbonyls (e.g. pinonaldehyde) on preexisting seed
117
aerosols. Our discovery that pinonaldehyde contributes more to SOA production than is
118
predicted possible by traditional partitioning theory provides field-based evidence that
119
supports the findings of these chamber studies.
120
121
While our estimate of the enthalpy of vaporization is highly uncertain due to the low R2
122
value (0.16) of the correlation between partitioning ratio and temperature, to our
123
knowledge this is the first estimate of the enthalpy of vaporization for pinonaldehyde
124
based on field measurements of its partitioning in the atmosphere. This value of Hvap is
125
within the range of those from single organic compounds, but it is much higher than the
126
enthalpy of vaporization recently reported by Offenberg et al. [2006] for the bulk SOA
127
from -pinene (33-43 kJ mol-1), which was the largest of the enthalpies determined by
128
these authors for various SOA precursors and reaction conditions.
129
130
Conclusions
131
Phase partitioning of semivolatile compounds measured in the field can provide novel
132
insight into their behavior under real atmospheric conditions. Our measurements showed
133
that processes such as oligomerization of pinonaldehyde transform more volatile into less
134
volatile species, increasing their contribution to SOA far above that predicted by
5
135
traditional partitioning theory. Collocated gas and particle phase in-situ measurements of
136
a broad range of semivolatile compounds should be pursued in future field campaigns to
137
search for evidence of chemistry in organic aerosols in the ambient atmosphere which
138
favors organic material staying in the particles rather than portioning back into the gas
139
phase.
140
141
Acknowledgements.
142
TAG deployment during ICARTT 2004 was funded by a U.S. Department of Energy
143
SBIR Phase II grant (DE-FG02-02ER83825), with in kind infrastructure support from the
144
National Oceanic and Atmospheric Administration. In addition, a graduate research
145
fellowship was provided by DOE’s Global Change Education Program (BJW). JLJ’s
146
participation in this research was supported by NOAA OGP (grant NA05OAR4310025).
147
We thank Megan McKay and Jennifer Murphy of the University of California at
148
Berkeley and David Parrish of the NOAA Earth System Research Laboratory for
149
logistical support.
150
151
References
152
153
154
155
156
157
158
159
160
161
162
163
Barsanti, K.C., and JF Pankow (2004), Thermodynamics of the formation of atmospheric
organic particlulate matter by accretion reactions – Part 1: aldehydes and ketones,
Atmos. Environ., 38 (26), 4371-4382.
Forstner, H.J.L., R.C. Flagan, and J.H. Seinfeld (1997b), Molecular speciation of
secondary organic aerosol from photooxidation of the higher alkenes: 1-octene and 1decene, Atmos. Environ., 31 (13), 1953-1964.
Hallquist, M.; Wangberg, I.; Ljungstrom, E. Atmospheric Fate of Carbonyl Oxidation
Products Originating from -Pinene and 3-Carene: Determination of Rate of Reaction
with OH and NO3 Radicals, UV Absorption Cross Sections, and Vapor Pressures.
Environ. Sci. Technol.; 1997; 31(11); 3166-3172.
Holzinger, R., D.B. Millet, B. Williams, N.M. Kreisberg, S.V. Hering, and A.H.
Goldstein, VOC emission, oxidation, and heterogeneous chemistry as observed at the
6
ICARTT Chebogue Pt. field site, Journal of Geophysical Research-Atmospheres,
Submitted.
Jang, M., B. Carroll, B. Chandramouli, and R. Kamens (2003), Particle growth and acidcatalyzed heterogeneous reactions of organic carbonyls on preexisting aerosols,
Environ. Sci. Technol., 37, 3828-3837.
Offenberg, J. H., T. E. Kleindienst, M. Jaoui, M. Lewandowski, and E. O. Edney (2006),
Thermal properties of secondary organic aerosols, Geophys. Res. Lett., 33, L03816.
Seinfeld, J.H., and J. F. Pankow, Organic Atmospheric Particulate Material, Annu. Rev.
Phys. Chem., 54: 121-140, 2003.
Tobias, H. J., K. S. Docherty, D. E. Beving, and P. J. Ziemann. Effect of relative
humidity on the chemical composition of secondary organic aerosol formed from
reactions of 1-tetradecene and O3. Env. Sci. Technol. 34, 2116-2125 (2000).
Williams, B.J., A.H. Goldstein, N.M. Kreisberg, S.V. Hering,(2006) An In-Situ
Instrument for Speciated Organic Composition of Atmospheric Aerosols: Thermal
Desorption Aerosol GC/MS-FID (TAG), Aerosol Sci. and Technol., 40, 627-638.
Williams, B.J., A.H. Goldstein, D.B. Millet, R. Holzinger, N.M. Kreisberg, S.V. Hering,
J.D. Allan, D.R. Worsnop, J.L. Jimenez, and A.B. White (2006), Chemical Speciation
of Organic Aerosol during ICARTT 2004: Results from In-Situ Measurements,
Journal of Geophysical Research-Atmospheres, Submitted.
186
0.03
16
1000 0.04 181500
0.05
Temp.
Particle × 40
Gas
337.7
18
16
0.01 14
5000.02
225.1
14
112.6
Air Temperature (ºC)
0.06
20
20
450.3
12
0.0
0.0
12
0
Gas Pinonaldehyde (ng m-3)
Particle Pinonaldehyde × 40 (ng m-3)
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
209.0
0:00
209.2
6:00
209.4
12:00
209.6
209.8
18:00
210.0
0:00
July 27, 2004
187
188
Figure xx. A 24-hour period comparison of pinonaldehyde phase partitioning with
189
varying ambient temperatures. An increase in the gas phase portion (light grey line) is
7
190
observed during elevated temperatures (thick black line). Particle phase concentrations
191
(dark grey line) have been multiplied by a factor of 40 in order to display on the same
192
scale as the gas phase concentrations.
193
0.25
[Particle / (Gas+Particle)] Pinonaldehyde
Hallquist Vapor Pressure
Aerosol Type 1
0.2
Aerosol Type 2
y = 3700x - 13
R2 = 0.99
0.15
0.1
y = 280x - 0.94
R2 = 0.09
0.05
y = 130x - 0.44
R2 = 0.11
0
0.00344
0.00345
0.00346
0.00347
0.00348
1/T (K)
194
195
196
8
0.00349
0.0035
0.00351
0.00352
ln (gas/particle ratio)
6
5
y = -19,000(±5400)x + 70
R2 = 0.16
4
3
2
1
0
0.00342 0.00344 0.00346 0.00348 0.0035 0.00352
1/temperature (K)
197
198
Figure xx. Plot of the natural logarithm of the gas phase to particle phase ratio of
199
pinonaldehyde concentrations verses the inverse ambient temperature (Kelvin) observed
200
at Chebogue Point, NS during July 26 – August 15, 2004. Gas phase concentrations
201
were measured by PTR-MS, and particle phase concentrations were measured by TAG.
202
The Clausius-Clapeyron plot predicts the Hvap for pinonaldehyde to be 160±45 KJ mol-1.
203
Error is calculated using the standard error for this linear regression.
204
9
2
Hallquist Vapor Pressure
1.5
Aerosol Type 1
y = -9100x + 32
R2 = 1.00
1
Aerosol Type 2
ln(P)
0.5
0
-0.5
y = -8100x + 28
R2 = 0.05
-1
y = -31000x + 100
R2 = 0.99
-1.5
1/T (K)
205
10
0.00352
0.00351
0.0035
0.00349
0.00348
0.00347
0.00346
0.00345
0.00344
-2