Click
Here
GEOPHYSICAL RESEARCH LETTERS, VOL. 35, L22801, doi:10.1029/2008GL035650, 2008
for
Full
Article
Phase changes of ambient particles in the Southern Great Plains of
Oklahoma
Scot T. Martin,1 Thomas Rosenoern,1 Qi Chen,1 and Donald R. Collins2
Received 11 August 2008; revised 23 September 2008; accepted 8 October 2008; published 18 November 2008.
[1] A new instrument, a 1 3 tandem differential mobility
analyzer (1 3-TDMA), was deployed in June 2007 in the
Southern Great Plains, Oklahoma, USA to study the phase
of ambient particles. Its primary measurement, the
irreversibility of the hygroscopic growth factor, is
obtained by reversibly cycling relative humidity (RH) by
±8% and testing for irreversible changes in diameter. In 101
runs, efflorescence occurred 72% of the time for particles
sampled at ambient RH. Deliquescence occurred in 13% of
the runs. The more frequent occurrence of efflorescence
compared to deliquescence was explained at least in part by
the distribution of ambient RH, which had a median of 80%
and quartiles of 65% and 93% RH. The deliquescence and
efflorescence events were nearly exclusive from one another
and could be separated by Min[RH Ambient, Inlet RH] <40%
for deliquescence and Max[RH Ambient, Inlet RH] >50% for
efflorescence. In outlook, the data set from the 1 3-TDMA
regarding the phase and hence water content of ambient
particles can be used for validating regional chemical
transport models of particle phase. Citation: Martin, S. T.,
T. Rosenoern, Q. Chen, and D. R. Collins (2008), Phase changes
of ambient particles in the Southern Great Plains of Oklahoma,
Geophys. Res. Lett., 35, L22801, doi:10.1029/2008GL035650.
1. Introduction
[2] Numerous studies have observed the deliquescence
and the crystallization of atmospheric particles. Using a
nephelometer, Rood et al. [1989] found that ambient aerosol
particles were metastable 50% of the time for ambient RH
of 45 to 75% in the Grand Canyon, Mojave Desert, and
Riverside, California. Using a hygroscopic-growth tandem
differential mobility analyzer (HTDMA), Pitchford and
McMurry [1994] observed deliquescence phase transitions
of ambient particles in the Grand Canyon. Additional
examples are reviewed in section VI (A) of Martin
[2000]. More recent observations include those of Day
and Malm [2001], Malm et al. [2003], Santarpia et al.
[2004], and Khlystov et al. [2005]. A purpose-designed
instrument to study the phase transitions of ambient particles, however, has not been deployed previously, and such
an instrument could be expected to provide more sensitivity,
new statistics, and further insights on this topic.
1
School of Engineering and Applied Sciences and Department of Earth
and Planetary Sciences, Harvard University, Cambridge, Massachusetts,
USA.
2
Department of Atmospheric Sciences, Texas A&M University, College
Station, Texas, USA.
[3] This paper describes the deployment in June 2007 of a
1 3 tandem differential mobility analyzer (1 3-TDMA)
in the Southern Great Plains, Oklahoma, USA. The primary
measurement is a test of the reversibility of the hygroscopic
growth factor of aerosol particles. The growth factor is the
ratio of a particle diameter measured at a high RH to a
reference diameter conditioned at a low RH [Biskos et al.,
2006]. Compared to a nephelometer or an HTDMA,
although the 1 3-TDMA does not provide information
on the magnitude of the growth factor, it provides increased
sensitivity to irreversible changes of this factor. Irreversible
changes are associated with phase changes. The greater
sensitivity is especially important for measurements at low
RH for which the diameter change between aqueous and
effloresced particles is small.
[4] The basic principles of phase transitions and their
effects on the reversibility of the growth factor can be
illustrated by a test aerosol of ammonium sulfate (AS)
particles. On the one hand, when going from low to high
RH, solid AS particles do not become aqueous (i.e.,
deliquesce) until 80% RH. On the other hand, when RH
is decreasing, aqueous AS particles do not crystallize by
homogeneous nucleation (i.e., effloresce) until 35% RH.
Due to this hysteresis, AS particles may, therefore, be either
aqueous or solid between 35 and 80% RH: the growth factor
is not a single-valued function of RH. For example, a cycle
from 77% to 85% to 77% converts solid AS particles to
aqueous, with a corresponding irreversibility of the growth
factor for the described cycle. The 1 3-TDMA takes
advantage of this effect to infer the initial physical state of
the particles (i.e., solid) prior to the RH cycle. Similarly, the
growth factor of aqueous AS particles is irreversible for a
cycle such as 40% to 32% to 40% RH.
[5] In contrast to AS particles, atmospheric particles have
much more complicated chemical compositions [Seinfeld
and Pandis, 1998]. They are internal mixtures of sulfate,
nitrate, ammonium, other inorganic and organic ions, undissociated organic molecules, and insoluble inclusions,
such as soot or mineral dust. Because particles crystallize
and deliquesce individually, the phase transition behavior of
real atmospheric aerosols is complex, and purpose-designed
instruments such as the 1 3-TDMA are needed for their
study. This paper introduces the results of the first field
measurements with this instrument.
2. Experimental
[6] The 1 3-TDMA is described by Rosenoern et al.
(The 1-by-3 Tandem Differential Mobility Analyzer for
measurement of the irreversibility of the hygroscopic
growth factor, submitted manuscript, 2008). In brief, after
Copyright 2008 by the American Geophysical Union.
0094-8276/08/2008GL035650$05.00
L22801
1 of 5
L22801
MARTIN ET AL.: PHASE CHANGES OF AMBIENT PARTICLES
charging at the inlet, aerosol particles are equilibrated to a
setpoint RH0 (cf. Figure S1 of the auxiliary material).1 A
first differential mobility analyzer DMAmono selects the
fraction of the aerosol particles lying within a band of
electric mobility, equivalent to 150 nm for this study. The
aerosol flow is split into three parallel arms, including a
‘‘reference’’ (denoted 0), a ‘‘deliquescence test’’ (denoted
+d), and an ‘‘efflorescence test’’ (denoted d). In the +d and
d arms, the aerosol particles are perturbed to RH0 ± d
before returning to the set RH0. In this study, we use d =
8%. The specific RH histories imposed are RH0 ! (RH0 + d)
! RH0 (e.g., 72% ! 80% ! 72%) as a deliquescence
test, RH0 ! (RH0 d) ! RH0 (e.g., 72% ! 64% !
72%) as an efflorescence test, and RH0 ! RH0 ! RH0
(e.g., 72% ! 72% ! 72%) as the reference. Exiting the
RH tests, the three flows pass through separate DMAs,
labeled DMA0, DMA+d, and DMAd, which are tuned to
the same electric mobility as DMAmono. Condensation
particle counters CPC0, CPC+d, and CPCd measure the
concentration passing through each DMA. An irreversible
change in particle diameter in either test arm caused by the
RH cycle is detected as a reduction in the transmitted
concentration. The transmission ratios CPC+d:CPC0 and
CPCd:CPC0 respectively define the deliquescence and
efflorescence tests. A phase transition abruptly changes
the transmission ratio as RH0 is scanned. Accompanying
the phase change the diameter must change by a factor of
1.17 for the flow rates used to obtain complete separation of
solid particles (DMA0) and aqueous particles (DMA+d); the
factor is 0.85 for complete separation of aqueous (DMA0)
from solid particles (DMAd).
[7] Measurements were conducted from June 4 to 16,
2007, in the Guest Trailer of the Central Facility at the DOE
Atmospheric Radiation Measurement (ARM) site in the
Southern Great Plains (SGP) [Sheridan et al., 2001]. The
site, located at N36° 370, W97° 300, 320 m asl outside of
Billings, Oklahoma (5 hours UTC), is instrumented with
multiple trailers to provide a climatological record of light
scattering properties of aerosol particles, radiative balance,
and meteorological variables.
[8] The aerosol particles were introduced to the 1 3-TDMA
through a 10-m mast outfitted with a 1200 (OD) copper tube.
A PM-2.5 inlet (URG-2000-30ED, URG Corp., North
Carolina) with a design flow of 3 Lpm was used. It
consisted of a 180° loop at the top, an insect screen, and
a cyclone. The actual flow was 3.4 Lpm. The inlet was
connected to the 1 3-TDMA in the Guest Trailer through
a 3-m horizontal connection of 1200 copper tubing. The trailer
was air-conditioned to 23°C, and the inlet temperature
inside the trailer was heated between 24.3 and 26.4°C.
The RH in the inlet varied from 32.5%, to 64.7%, to 79.1%
for the 10%, 50%, and 90% points of the cumulative
distribution function (CDF).
[9] The seven-day back-trajectories for the site were
analyzed [Schoeberl and Newman, 1995]. For June 4 –6
(17:00 local) the local air mass arrived from the north out of
Canada, for June 7 from the west from the Pacific across
southern California, for June 8 and 9 from complex trajec-
1
Auxiliary materials are available in the HTML. doi:10.1029/
2008GL035650.
L22801
Figure 1. Observation of a deliquescence event. (a) Scan
of RH0 and RH+d from 19:08 to 20:46 on 7 Jun 2007 using
RH+d = RH0 + 8%. (b) Measurements of CPC0 and CPC+d.
(c) The transmission ratio CPC+d:CPC0 for an upscan of
RH+d. The ambient RH was 29% and the inlet RH was 35%.
tories from several directions including Canada and the
Pacific, for June 10, 11, and 12 from the south across
Texas, and for June 13, 14, and 15 from a stagnant air mass
generally from the south – southeast. The wind directions
measured on the site by the meteorological station were
consistent with the local directions indicated by the trajectories. There was precipitation on the evening of June 10
and in the night and day on June 13 and 14. The median
daily temperature was 23.3°C, with diel swings of ca.
10°C. Excluding rainy days, the ambient RH had afternoon
2 of 5
MARTIN ET AL.: PHASE CHANGES OF AMBIENT PARTICLES
L22801
L22801
lows between 15 and 40% and nighttime highs of 80 to
100%.
3. Results and Discussion
Figure 2. Observation of an efflorescence event. (a) Scan
of RH0 and RH-d from 08:14 to 09:40 on 12 Jun 2007 using
RHd = RH0 8%. (b) Measurements of CPC0 and CPCd.
(c) The transmission ratio CPC-d:CPC0 for a downscan of
RHd. The ambient and inlet RHs were 79%.
3.1. Observations of Deliquescence and Efflorescence
[10] A deliquescence event is shown in Figure 1 for 19:08
to 20:46 on June 7. Figure 1a shows that RH0 was scanned
upward from 22 to 82% while RH+d was simultaneously
scanned from 30 to 90%, corresponding to d = +8%.
Figure 1b shows the corresponding measurements by
CPC0 and CPC+d during the RH scan. At 20:16 (RH0 =
70%, RH+d = 78%), the concentration measured by CPC+d
dropped precipitously. The growth factor thus changed
irreversibly, indicating deliquescence, for the RH cycle of
70% ! 78% ! 70%. Figure 1c, showing the transmission
ratio CPC+d:CPC0 for increasing RH, tracks the entire
deliquescence event. Irreversible growth is apparent for
78% < RH+d < 90%. Given that d = +8%, the implication is
that the ensemble of particles had deliquescence phase
transitions ranging from 78 to 82% RH.
[11] For comparison, Charlson et al. [1974] in St. Louis
and Pitchford and McMurry [1994] in the Grand Canyon
observed similar RH values for deliquescence at times when
ammonium sulfate particles were prevalent. ten Brink et al.
[1996] in the Netherlands observed deliquescence at 60%
RH when ammonium nitrate particles were prevalent.
Santarpia et al. [2004] observed deliquescence between
70 and 75% RH in southeastern Texas. In the remote Pacific
Ocean, Berg et al. [1998] observed deliquescence between
75 and 80% RH.
[12] An efflorescence event is shown in Figure 2. RH0
was scanned downward from 82 to 22% while RHd was
simultaneously scanned from 74 to 14%, corresponding to
d = 8% (Figure 2a). At RHd = 27%, the concentration
measured by CPCd dropped abruptly (Figure 2b). The RH
cycle of 35% ! 27% ! 35% thus caused an irreversible
change in the growth factor, indicating efflorescence. The
transmission ratio CPCd:CPC0 for decreasing RH shows
that the irreversible growth occurred across 16% < RHd <
27% (Figure 2c). Given that d = 8%, the implication is
that the ensemble of particles had efflorescence RHs ranging from 24 to 27% RH.
[13] The value of 0.7 for the ratio CPC+d:CPC0 during
deliquescence, compared to 1.3 outside of the event
(Figure 1c), indicates that a lower limit of 50% of the
ambient aerosol particles deliquesced. The balance of particles either had chemical compositions that did not have
deliquescence behavior (such as high acidity or high organic
mole fraction) or were already aqueous when entering the
1 3-TDMA. The latter would imply efflorescence RHs
Table 1. Statistics of the Observations From June 4 to 16
Scans
Min[Inlet RH, Ambient RH] < 40%
Max[Inlet RH, Ambient RH] > 50%
10%/50%/90% distribution of the fraction
of particles that deliquesce or effloresce
Number
(Fraction)
Deliquescence
(Fraction)
Efflorescence
(Fraction)
101 (1.00)
20 (0.20)
90 (0.89)
n/a
13 (0.13)
13 (0.65)
n/a
0.14/0.30/0.43
73 (0.72)
n/a
72 (0.80)
0.07/0.12/0.21a
a
Lower limit.
3 of 5
L22801
MARTIN ET AL.: PHASE CHANGES OF AMBIENT PARTICLES
for those particles of less than 29% RH, corresponding to
the minimum of the ambient RH and the inlet RH. The
value of 50% as a lower limit can be understood by the
example of Figure 1 that the ambient particles had a
heterogeneity of deliquescence RHs from 78 to 82%,
implying that particles having a deliquescence RH of
78% did not contribute to the signal for RH+d > 86% (given
d = +8%) while those having a deliquescence RH of 82%
did not contribute to the signal for RHd < 82% RH. Of the
L22801
13 deliquescence events, in median a lower limit of 30% of
the particles deliquesced, with variability between 14% and
43% for the 10% and 90% of the CDF (Table 1). The onset
of the deliquescence occurred between 77% and 79% RH+d
for the observations.
[14] The ratio CPCd:CPC0 = 0.8 during efflorescence
(Figure 2c) compared to CPCd:CPC0 = 1.1 outside of the
event indicates that a lower limit of 30% of the ambient
aerosol particles effloresced. The shape of the efflorescence
response, without a flat bottom, indicates that the diameter
difference between the aqueous particles and their effloresced counterparts was not sufficient at low RH to get full
separation in the 1 3-TDMA (Rosenoern et al., manuscript, 2008). The shape further decreases the lower limit of
the fraction of particles that effloresced beyond the limit
caused by heterogeneity in the efflorescence RHs of the
ambient particles (i.e., in analogy to the explanation of the
lower limit for deliquescence fraction in regard to Figure 1).
An upper limit that 70% of the particles did not effloresce
can likewise be established, explained either by chemical
compositions that did not have efflorescence behavior or by
an insufficient maximum in the ambient and inlet RHs to
cause deliquescence (i.e., the particles were already solid
when entering the 1 3-TDMA). Although the common
atmospheric inorganic species such as mixtures of H2SO4,
(NH4)2SO4, and NH4NO3 deliquesce at RH < 80% [Martin
et al., 2003], many small organic acids, at least in pure
form, deliquesce at higher RH [Parsons et al., 2004]. Of the
73 efflorescence events, a median lower limit of 12% of the
particles effloresced, with variability from 7% to 21% for
the 10% and 90% points of the CDF (Table 1). The onset of
efflorescence occurred between 26% and 30% RHd. For
comparison, Shaw and Rood [1990] observed efflorescence
within a parabolic probability distribution from 10 to 45%
RH in the Grand Canyon, the Mojave Desert, and Riverside,
California.
[15] The auxiliary material shows two additional examples of deliquescence and efflorescence similar to Figures 1
and 2 as well as three scans during which no phase
transitions were observed. In Figure 1c, although the ratio
CPC+d:CPC0 was constant outside of the deliquescence
event, it should theoretically have equaled unity yet a
systematic offset is apparent. The explanation is that the
test and reference arms of the 1 3-TDMA were not
perfectly balanced (Rosenoern et al., manuscript, 2008).
An additional observation is that for the time periods
outside of deliquescence Figure 1b shows drift in the time
series CPC0 and CPC+d, though not in the ratios. The drift
Figure 3. Relationships of deliquescence and efflorescence
events to inlet and ambient RHs. (a) Line diagram showing
the distribution with respect to the inlet RH of deliquescence
events, efflorescence events, and the absence of both. (b) As
with a but with respect to ambient RH. (ce) Histograms
with respect to ambient RH of (red) deliquescence events,
(blue) efflorescence events, (green) neither, and (black) total
number of observations. (f) Histogram of deliquescence
events with respect to minimum of ambient compared to
inlet RH (red). Histogram of efflorescence events with
respect to maximum of ambient compared to inlet RH
(blue).
4 of 5
L22801
MARTIN ET AL.: PHASE CHANGES OF AMBIENT PARTICLES
arose not from instabilities in the 1 3-TDMA but rather
from changes in the particle source function. Specifically,
the number concentration of 150-nm diameter particles
entering DMA0 changed both as the ambient aerosol
changed during a scan and more importantly as the scanning
of RH0 changed the particles selected from that RH-sensitive
distribution. This latter effect is especially prominent at high
RH, appearing as a decrease in CPC0, because of a shift in
the number size distribution of hygroscopic ambient particles to diameters greater than 150 nm. These issues, which
are highlighted for the deliquescence data in Figure 1, are
equally applicable to the efflorescence data in Figure 2.
3.2. Predictive Factors of Deliquescence and
Efflorescence
[16] In the 101 scans between June 4 and 16, efflorescence occurred 72% of the time (Table 1), where an event is
defined as at least 5% of the particles undergoing a phase
transition. Deliquescence occurred in 13% of the runs.
There was a strong correlation between the occurrence of
deliquescence or efflorescence and ambient RH, and the
relative occurrence of efflorescence compared to deliquescence was therefore influenced by the distribution of ambient RH (median of 80% and quartiles of 65% and 93%).
The clustering of deliquescence, efflorescence, and their
absence is shown in Figure 3a with respect to the inlet RH
and in Figure 3b with respect to the ambient RH. Of the
two, greater clustering is apparent for ambient RH. Deliquescence occurred only when the minimum of the inlet and
the ambient RH was less than 40%, and this condition was
met for 20% of the 101 runs. For these cases, deliquescence
was observed 65% of the time. Excluding one outlier,
efflorescence occurred only when the maximum of the inlet
and the ambient RH was greater than 50%. Efflorescence
occurred in 80% of these scans, which constituted 89% of
all scans.
[17] The histograms in Figures 3c3f provide further
insight into these statistics. Figures 3c3e show that a
univariate analysis with ambient RH as the main predictor
is largely successful in sorting efflorescence and deliquescence events from another. The absence of phase transitions
is most frequent for ambient RH of 70%. The RH of the
inlet, when significantly different from atmospheric RH, can
alter the physical state of the particles prior to their entering
the 1 3-TDMA. The median of the absolute value of the
difference was 14% during the measurement period, as
explained by the differences between ambient temperature
and inlet temperature. Figure 3f shows a bivariate analysis
in inlet RH and ambient RH to further cluster the observations. The deliquescence and efflorescence events are nearly
exclusive from one another and can be separated by
Min[RH Ambient, Inlet RH] <40% for deliquescence and
Max[RH Ambient, Inlet RH] > 50% for efflorescence.
[18] In conclusion, the observations reported herein
employing the 1 3-TDMA imply that a cycling between
physical states can be expected over continental regions,
provided that particles at a given location have appropriate
chemical compositions, because large swings in RH occur
from morning to afternoon through the evening. Furthermore, vertical gradients in RH are even greater, ranging
L22801
from 100% at the lifting condensation level to very low
humidity near the planetary surface in mid-afternoon or
above the inversion, and under many conditions particles
turnover in this layer (experiencing the RH extremes) every
15 to 30 min.
[19] Acknowledgments. This material was supported by the National
Science Foundation under grant ATM-0633840. The DOE Atmospheric
Radiation Measurement (ARM) Program provided access to the SGP
facilities. Data from the Surface Meteorological Observation Station (1-min
avg, E13 Central Facility) were obtained from the ARM Program
(www.arm.gov/data/datastream.php?id=1smos). We thank Tom Kucsera
(NASA/Goddard) for back-trajectories. T.R. received support from the
Danish Agency for Sci Technol and Innov. Advice and assistance from
John Ogren are gratefully acknowledged.
References
Berg, O. H., E. Swietlicki, and R. Krejci (1998), Hygroscopic growth of
aerosol particles in the marine boundary layer over the Pacific and Southern
oceans during the First Aerosol Characterization Experiment (ACE 1),
J. Geophys. Res., 103, 16,535 – 16,545.
Biskos, G., L. M. Russell, P. R. Buseck, and S. T. Martin (2006), Nanosize
effect on the hygroscopic growth factor of aerosol particles, Geophys.
Res. Lett., 33, L07801, doi:10.1029/2005GL025199.
Charlson, R. J., A. H. Vanderpol, D. S. Covert, A. P. Waggoner, and
N. C. Ahlquist (1974), Sulfuric acid-ammonium sulfate aerosol: Optical
detection in the St. Louis region, Science, 184, 156 – 157.
Day, D. E., and W. C. Malm (2001), Aerosol light scattering measurements
as a function of relative humidity: A comparison between measurements
made at three different sites, Atmos. Environ., 35, 5169 – 5176.
Khlystov, A., C. O. Stanier, S. Takahama, and S. N. Pandis (2005), Water
content of ambient aerosol during the Pittsburgh air quality study,
J. Geophys. Res., 110, D07S10, doi:10.1029/2004JD004651.
Malm, W. C., D. E. Day, S. M. Kreidenweis, J. L. Collett, and T. Lee
(2003), Humidity-dependent optical properties of fine particles during
the Big Bend regional aerosol and visibility observational study, J. Geophys. Res., 108(D9), 4279, doi:10.1029/2002JD002998.
Martin, S. T. (2000), Phase transitions of aqueous atmospheric particles,
Chem. Rev., 100, 3403 – 3453.
Martin, S. T., J. C. Schlenker, A. Malinowski, H. M. Hung, and Y. Rudich
(2003), Crystallization of atmospheric sulfate-nitrate-ammonium particles, Geophys. Res. Lett., 30(21), 2102, doi:10.1029/2003GL017930.
Parsons, M. T., J. Mak, S. R. Lipetz, and A. K. Bertram (2004), Deliquescence of malonic, succinic, glutaric, and adipic acid particles, J. Geophys.
Res., 109, D06212, doi:10.1029/2003JD004075.
Pitchford, M. L., and P. H. McMurry (1994), Relationship between measured water vapor growth and chemistry of atmospheric aerosol for Grand
Canyon, Arizona, in winter 1990, Atmos. Environ., 28, 827 – 839.
Rood, M. J., M. A. Shaw, T. V. Larson, and D. S. Covert (1989), Ubiquitous
nature of ambient metastable aerosol, Nature, 337, 537 – 539.
Santarpia, J. L., R. J. Li, and D. R. Collins (2004), Direct measurement of
the hydration state of ambient aerosol populations, J. Geophys. Res., 109,
D18209, doi:10.1029/2004JD004653.
Schoeberl, M. R., and P. A. Newman (1995), A multiple-level trajectory
analysis of vortex filaments, J. Geophys. Res., 100, 25,801 – 25,816.
Seinfeld, J. H., and S. N. Pandis (1998), Atmospheric Chemistry and Physics,
John Wiley, Hoboken, N.J.
Shaw, M. A., and M. J. Rood (1990), Measurements of the crystallization
humidities of ambient aerosol particles, Atmos. Environ., Part A, 24,
1837 – 1841.
Sheridan, P. J., D. J. Delene, and J. A. Ogren (2001), Four years of continuous surface aerosol measurements from the Department of Energy’s
Atmospheric Radiation Measurement Program Southern Great Plains
Cloud and Radiation Testbed site, J. Geophys. Res., 106, 20,735 – 20,747.
ten Brink, H. M., J. P. Veefkind, A. Waijers-Ijpelaan, and J. C. van der Hage
(1996), Aerosol light-scattering in the Netherlands, Atmos. Environ., 30,
4251 – 4261.
Q. Chen, S. T. Martin, and T. Rosenoern, School of Engineering and
Applied Sciences and Department of Earth and Planetary Sciences,
Harvard University, 29 Oxford Street, Pierce Hall, Room 122, Cambridge,
MA 02138, USA. ([email protected])
D. R. Collins, Department of Atmospheric Sciences, Texas A&M
University, College Station, TX, USA.
5 of 5