Atmospheric Research 51 Ž1999. 1–14
Nanoparticle formation in marine airmasses:
contrasting behaviour of the open ocean and
coastal environments
A.G. Allen a , J.L. Grenfell b, R.M. Harrison
M.J. Evans c
a
a,)
, J. James a ,
Institute of Public and EnÕironmental Health, UniÕersity of Birmingham, Edgbaston, Birmingham B15 2TT,
UK
b
NASA-GISS, Columbia UniÕersity, 2880 Broadway, New York, NY 10025, USA
c
Centre for Atmospheric Sciences, UniÕersity of Cambridge, Chemistry Department, Lensfield Road,
Cambridge CB2 1EW, UK
Received 28 April 1998; accepted 14 December 1998
Abstract
Massive and rapid increases in nanoparticles have been observed at a remote coastal site in
western Ireland. The same phenomenon was not detected aboard a ship situated approximately 160
km off-coast. On-shore nanoparticles correlated remarkably well with the march of the tide,
peaking at low-water. This suggests a link between marine biogenic gas emissions Žas yet
unidentified. and nanoparticle formation events. This paper examines the contrasting behaviour
observed at the coast and in the open ocean, with respect to nanoparticle formation. q 1999
Elsevier Science B.V. All rights reserved.
Keywords: Nanoparticles; Nucleation; Formation mechanisms; Troposphere
1. Introduction
The formation of new particles in the marine atmosphere is believed to be important
in the regulation of climate ŽCharlson et al., 1987.. Whilst such particles when formed
)
Corresponding author
0169-8095r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.
PII: S 0 1 6 9 - 8 0 9 5 Ž 9 8 . 0 0 1 1 1 - 2
2
A.G. Allen et al.r Atmospheric Research 51 (1999) 1–14
are too small to act as cloud condensation nuclei ŽCCN. they grow by condensation and
coagulation to sizes where they are effective agents in the nucleation of cloud water
droplets. Sources of fine particles in the atmosphere Žfor both low, sustained nucleation
or sporadic bursts. in the marine atmosphere remain poorly defined although observational studies cite homogeneous gas to particle formation Žorrand condensation onto
precursor embryos. as possible sources ŽAyers and Gras, 1991; Hegg et al., 1991;
Hoppel et al., 1994.. Entrainment from the free troposphere Že.g., Hudson, 1993. is a
further possibility, while submicron particles may also be generated by physical means
during bubble-bursting at the ocean surface ŽPollak, 1955; Mason, 1957; Cipriano et al.,
1983; Resch and Afeti, 1991; Despiau et al., 1996.. Mihalopoulos et al. Ž1992. made
particle measurements using a TSI model 3020 CPC at a coastal site in NW Brittany in
France near a large algae field during June 1989 and observed regular, sharp maxima in
particles 1–2 h after low tide. Similar rapid ‘bursts’ in such aerosol have also been
recorded, e.g., by Davison et al. Ž1996.. The possibility that local sources of ultrafine
particles might exist along the western coast of Ireland was first suggested by Georgii
and Metnieks Ž1958. during a series of measurements conducted near the Valentia
Observatory. Nanoparticle formation in the coastal–marine and marine environments
reported in this paper has been investigated as part of the ACSOE ŽAtmospheric
Chemistry in the Oceanic Environment. UK community research project. Data were
collected between April 27th and May 28th 1997 at a remote site ŽMace Head. on the
west coast of Ireland and aboard a ship, the RRS Challenger, which sailed approximately 160 km west of the coastal site. We discuss the different processes controlling
nanoparticle formation in the coastal–marine and marine environments. In particular the
role of the tide in coastal regions is highlighted.
2. Experimentalr
r research platforms
Two continuous Condensation Particle Counters ŽCPC. were employed ŽTSI instruments, models 3022A and 3025. on the shore. One 3022A instrument was employed on
the ship. Before the 1997 campaign the shore-based 3022A instrument was calibrated by
the UK TSI agents. The authors subsequently adjusted the 3025 instrument signal to that
of the 3022A using monodisperse aerosol of diameter 100 nm. The 3022A instrument
has 50% efficiency in sampling particles having diameters ) 7 nm, 90% efficiency for
) 15 nm and almost 100% for ) 20 nm. The 3025A instrument has 50% efficiency in
sampling particles having diameters ) 3 nm, 90% efficiency for ) 5 nm and almost
100% for ) 10 nm ŽWiedensohler et al., 1997.. For the sake of conciseness, however,
these are hereafter referred to as total particles Žcmy3 . ) 3 nm and ) 7 nm for the
3025A and 3022A, respectively. On both ship and shore, air was sampled into the CPC
via a 1.5 m length of copper tubing of internal diameter 0.64 cm. The inlets were
situated 3 m above the oceanrground on both ship and shore. Inlet losses of 10 nm
particles are likely to have been small ŽHinds, 1982. and similar for both sites. Ozone
photolysis rate coefficients were measured with a jO1 D radiometer ŽJunkermann et al.,
1989., and aerosol Fuchs surface area with an epiphaniometer ŽGaggeler et al., 1989..
A.G. Allen et al.r Atmospheric Research 51 (1999) 1–14
3
The coastal site, Mace Head Atmospheric Research Station, County Galway Ž53819X 34Y N;
9854X 14Y W; 10 m above sea level. is situated on the headland of a peninsula in western
Ireland ŽFig. 1.. The surrounding terrain is mainly flat and rocky.
Fig. 1. Location of the Mace Head coastal site.
4
A.G. Allen et al.r Atmospheric Research 51 (1999) 1–14
3. Results and discussion
Particle count data from Mace Head showed frequent massive excursions in particle
number density. Large divergences between the 3022A and 3025 instruments were
observed during the particle bursts, indicative of particles in the 3–7 nm diameter size
range arising from homogeneous nucleation processes. The rapid concentration fluctuations and lack of a corresponding phenomenon on the ship are strongly suggestive of
local production of new particles at Mace Head. In the following sections we describe
the experimental observations and examine the possible causes of new particle production.
Various factors able to affect nanoparticle concentrations were considered and
include the following.
Ža. Solar intensity, which affects OH and hence the rate of sulphate andror
methanesulphonate production from dimethyl sulphide ŽDMS. oxidation ŽYin et al.,
1990; Ayers et al., 1991., and from anthropogenic sulphur dioxide ŽSO 2 ., as well as of
condensible products from biogenic hydrocarbon oxidation ŽGlasius et al., 1997; Hoffmann et al., 1997.. Fig. 2 shows 10 min averaged particle Ž) 3 nm. number concentrations and jO1 D radiometer measurements at Mace Head. The onset of sunlight does not
always correspond to an enhancement in total particles ) 3 nm. However, sunlight does
play a role in particle production, as previously indicated ŽMcWilliams and Morgan,
1955; Pollak, 1955; McGovern et al., 1996., because no bursts were recorded on shore
during the night. The lack of formation at night also shows that production of
nanoparticles did not occur via purely physical processes such as bubble bursting. It
Fig. 2. Ten-minute averaged jO1 D and particles ) 3 nm at Mace Head, April 27th–May 28th, 1997.
A.G. Allen et al.r Atmospheric Research 51 (1999) 1–14
5
appears that the mechanism controlling the burst phenomena requires sunlight in order
to proceed.
Žb. Background aerosol, which is a sink for both nanoparticles and some possible
precursor species via scavenging. Fuchs surface area ŽFSA. is plotted alongside the
aerosol number concentration in Fig. 3, for the same timescale as Fig. 2. The epiphaniometer data inversion procedure adopted was based on that described by Pandis et al.
Ž1991. and Rogak et al. Ž1991.. No FSA data were available between May 3rd–10th,
due to instrumental failure; however, low background particle counts in the westerly air
masses sampled at this time indicated the presence of clean marine air Žwhich at other
times always possessed low FSA..
The data in Fig. 3 do not offer clear evidence that background aerosol concentrations
exerted any major controlling influence over the burst phenomenon, which would have
been the case if periods of enhanced particles corresponded to low FSA. There is a
general suppression of particle production during the pollution event characterised by
high FSA between May 15th–21st, and during this period rather poorer correlation of
particle numbers with tidal condition suggests that alternative sources of nanoparticles
were present within the prevailing continental air masses which experienced prolonged
transits over land.
Žc. Wind direction, since some local regions may be more efficient emitters of
nanoparticle precursor gases at low tide than others, depending on phytoplanktonrmacroalgae occurrencerspeciation and the extent of exposure at low tide. Wind direction
data were obtained from a sensor located on the shore at 2 m height, adjacent to the
instrumental installation. Fig. 4 shows that bursts proceeded throughout most wind
sectors. No obvious diurnal variation in wind direction is apparent in this, or any other
Fig. 3. Ten-minute averaged FSA and particles ) 3 nm at Mace Head, April 27th–May 28th, 1997.
6
A.G. Allen et al.r Atmospheric Research 51 (1999) 1–14
Fig. 4. Ten-minute averaged wind direction and particles ) 3 nm at Mace Head, April 27th–May 28th, 1997.
such data for the campaign, tending to rule out influence from local land–sea circulation
cells or micrometeorological transport effects arising from recirculating flows. The
effect of windspeed upon CPC sampling efficiency was not determined; however, based
on our observations of fairly constant clean-air background values for a variety of
windspeeds, this effect is considered to be negligible. Further, no correlation was
apparent between variations in wind speed and the burst phenomena Žnot shown.. Until
May 3rd, air trajectories originated in mid-Atlantic, looping cyclonically over Wales and
SW England before arriving at the site; May 4th was Atlantic south westerly, whilst
May 5th–13th experienced Arctic trajectories sometimes clipping W. Scotland. Trajectory information regarding the latter half of the 1997 campaign is discussed later, with
the ship data.
Žd. Tidal height, which is shown in Fig. 5 together with 10 min averaged on-shore
total particles ) 3 nm. Figs. 6 and 7 show tidal height and total particles ) 7 nm for
Mace Head and the ship, respectively. The tidal height was calculated for Galway Bay,
situated 12.9 km south and 32.2 km west of the measurement site. The data was
obtained using a numerical model ŽMinistry of Defence Hydrographic Office; ‘Tidecalc’
version 1, 1991., which was accurate to within 10–15 min of the times of high or low
tide published in local tide tables current at the time of the measurements.
The shore-based data ŽFigs. 5 and 6. clearly show a persistent relationship between
tidal height and the concentration of total particles ) 3 nm. At low tide during the
daytime the latter are often Žbut not always. enhanced. A closer examination of the data
on an expanded scale Žnot shown. revealed secondary bursts proceeding as the tide
returns. Typically, the number of particles within the 3–7 nm size-range, evaluated by
differencing the two CPC instruments, rose from a few hundred up to ; 60,000 cmy3 .
A.G. Allen et al.r Atmospheric Research 51 (1999) 1–14
7
Fig. 5. Tidal height and particles ) 3 nm at Mace Head, April 27th–May 28th, 1997.
The term, R, is the ratio of particles ) 3 nm to particles ) 7 nm, and hence increases in
R imply new particle formation. R typically rose from ; 1.1 to 4–6 at low tide. The
ship data shown in Fig. 7 do not, as expected, show any obvious relation with the tide.
Fig. 6. Tidal height and particles ) 7 nm at Mace Head, May 13th–28th, 1997.
8
A.G. Allen et al.r Atmospheric Research 51 (1999) 1–14
Fig. 7. Tidal height and particles ) 7 nm on the ship, May 13th–28th, 1997.
The large increases in CPC readings on the ship were due to sampling of pollutants from
the ship’s stack Židentified from NOx and ozone data.. Modest rises in total particles
) 7 nm were recorded on the ship around dawn Žwhen OH begins to build-up and
initiates DMS oxidation. on some days Že.g., May 21st–22nd.. The dawn events could
not, however, be unequivocally attributed to non-anthropogenic effects due to the bustle
of activity on board the ship at dawn. NOx and ozone showed small shifts concomitant
with the dawn bursts, unlike the shore data. Other days, e.g., May 24th–26th suggested
very little new particle formation and did not feature dawn enhancements. Clearly,
day-to-day mechanisms must be invoked to explain this variability Žsee subsequent
discussion.. New particle formation at dawn was not clearly evident on the shore,
although it has been observed at a more polluted coastal site ŽHarrison, 1997.. Weather
on ship and shore was generally very similar, with many sunny days in the period May
14th–27th.
The main conclusion we draw from the ship data is a noticeable lack of the
Žnon-anthropogenic. particle bursts which were frequently recorded on the shore. The
shore-based data, however, revealed a series of diurnal ‘bursts’ in nanoparticle concentrations which usually correlated very well with low-tide. Trajectories for the period
during the 1997 campaign when ship and shore measurements overlapped revealed two
ship to shore connected flow periods, May 13th–14th, with westerly flow from ship to
shore, Žair originating from Greenland looped down cyclonically between Scotland and
Iceland., and May 26th–27th Žair from the west coast of Spain, travelling cyclonically to
the mid-Atlantic then turning anti-cyclonically toward the site.. There was one shore to
ship connected flow period, May 23rd–25th, which was due easterly from central
UKrWales.
A.G. Allen et al.r Atmospheric Research 51 (1999) 1–14
9
Non-connected flow was observed between May 15th–20th, when variable flows
originated from the Bay of Biscay then southern EuroperItaly, and during May
20th–23rd air masses were variable but mainly continental in origin and approached the
site along the Irish coast. Thereafter until the end of the campaign airmasses originated
from the mid-Atlantic, then looped anticyclonically southwards traversing northern
Ireland and approaching the site from northerly sectors.
Little daytime burst activity was recorded on either platform during the first period of
ship to shore flow ŽMay 13th.. During the second period ŽMay 26th–27th., a low-tide
burst was recorded on the shore ŽFigs. 5 and 6. but not on the ship ŽFig. 7. consistent
with a local formation mechanism near the shoreline. Bursts recorded on the ship, e.g.,
on May 20th and 27th were anthropogenic in origin. During a brief shore to ship period
on May 24th a large burst was recorded at Mace Head ŽFig. 5. around midday which
correlated well with low tide. Although trajectories were indicative of easterly flow, the
local wind direction was on-shore at this time Žcalculated air mass backward trajectories
were frequently not accurately reflected in local wind directions, which were more
variable due to the headland location of the site.. The coastal region extends to the east
of the site, which is situated on a peninsula near an estuary stretching inland. Therefore,
even easterly air may encounter inter-tidal regions. Little CPC activity was recorded on
the ship during this period. It appears, therefore, that the formation event is local in
nature, and the particles do not survive the ; 12-h journey from shore to ship. Without
full knowledge of the particle size distribution, it is not possible to calculate the rate of
loss of particles by coagulation. However, loss rates calculated for polydisperse coagulation of an aerosol of initial concentration 10 5 cmy3 over a 12-h travel time indicate that
if the burst occurred in the coastal zone, it would not be observed on the ship Žfurther
details of the procedures used for these calculations may be found in Seinfeld and
Pandis, 1998.. Surviving a 12-h journey is definitely not feasible for the bulk of the
nanoparticles for which lifetimes are short in polluted easterly air masses due to rapid
coagulation with high levels of background aerosol.
Further evidence for the importance of the different variables for production of
nanoparticles is provided by plotting particle number concentrations vs. jO1 D and wind
direction, or tidal height ŽFigs. 8–11., for periods when clean marine air masses were
present, as evidenced by low values for FSA andror low background particle number
concentrations. These periods were April 28th–30th, May 5th–9th, May 11th–14th and
May 22nd–28th. Fig. 8 shows that particle production was not associated with any
particular local wind direction, as might be expected due to the location of the site on a
headland with numerous bays and inlets possessing large populations of intertidal biota.
Fig. 9 similarly shows that there was no observable correlation with jO1 D. However,
comparison of Fig. 10, which shows the relationship between tidal height and particle
Ž) 3 nm. number concentration for the period 00:00 GMT–08:00 GMT, with Fig. 11,
which shows data for 08:00 GMT–16:00 GMT, indicates clearly that particle bursts did
not occur at night, and that during the daytime very high production occurred at or near
low tide.
DMS oxidation has been suggested to play a causative role in new particle formation
in the marine atmosphere Že.g., Charlson et al., 1987.. Other possible candidates in this
respect are dimethyl disulphide ŽDMDS. ŽPersson and Leck, 1994. and terpene oxida-
10
A.G. Allen et al.r Atmospheric Research 51 (1999) 1–14
Fig. 8. Scatter plot for particles ) 3 nm and wind direction at Mace Head.
tion ŽHatakeyama et al., 1989.. The relative contribution of these various chemical
species to particle formation over the open ocean is still subject to uncertainty. Sources
in the intertidal zone may be different. Macroalgae exposed at low tide tend to emit
more DMS precursor in order to combat the effects of desiccation ŽCharlson et al.,
1987.. Also, biogenic gases are emitted directly into the atmosphere at low tide, rather
Fig. 9. Scatter plot for particles ) 3 nm and jO1 D at Mace Head.
A.G. Allen et al.r Atmospheric Research 51 (1999) 1–14
11
Fig. 10. Scatter plot for particles ) 3 nm and tidal height Ž00:00 GMT–08:00 GMT. at Mace Head.
than diffusing across the air–sea surface interface, and the biota are agitated as the tide
advances and retreats. Although CPC bursts tended to peak at low tide during periods of
uncontaminated marine air, smaller secondary bursts were observed soon afterwards,
coinciding with re-agitation of biota by the advancing tide, or the displacement of gases
Fig. 11. Scatter plot for particles ) 3 nm and tidal height Ž08:00 GMT–16:00 GMT. at Mace Head.
12
A.G. Allen et al.r Atmospheric Research 51 (1999) 1–14
from intertidal sediments due to the advancing waters. Marine phytoplankton have been
found to emit a variety of halocarbon compounds, precursors to species such as IO and
BrO which may enhance the rate of DMS oxidation ŽAtkinson et al., 1996 and
references therein.. This release of several potential precursors and oxidants may be
enhanced in the coastal zone, and many additional measurements would be needed to
define the chemical processes responsible for new particle production at this site.
During previous field experiments, our research group conducted measurements of
particle number concentrations at two other locations.
Ž1. South Uist, an island in the Outer Hebrides ; 100 km off the northwest coast of
Scotland. Particle data was collected at a coastal site with a shallow beach during two
winter and two summer campaigns in 1993r1994 ŽLowe et al., 1996.. An initial
analysis of CNrtidal data has revealed a weaker correlation than at Mace Head; CN
concentrations correlated with low tide on some days but on others they related to
insolation. The poorer relationship may have reflected errors in the calculation—tide
data were only available for Stornoway, situated on the Outer Hebrides ; 100 km north,
or may have reflected the variation in macroalgae speciation and exposure to the air.
Ž2. Santa Maria, an island in the Azores in the mid-North Atlantic. Particle data were
collected during a summer campaign which took place in 1992 ŽHarrison et al., 1996..
Total particle concentrations Ž) 3 nm. were generally quite low Ž- 1000 cmy3 ..
Occasional daytime peaks between 1000–4000 cmy3 were observed which appeared to
be due to subsidence of air enriched in nanoparticles mixing down from aloft. The site
was situated on a cliff overlooking deep water and the outgoing tide exposed a much
smaller surface area of macroalgae compared with a shallow beach. This situation was
more representative of the open ocean, compared to Mace Head.
4. Conclusions
The observations presented in this paper suggest that the connection between marine
biogenic gas emissions and particle formation may be considerably more complex than
envisaged hitherto. The observed temporal connection between massive bursts of new
particle formation and the state of the tide at Mace Head suggests that such particle
formation is an extremely local phenomenon which, therefore, requires very rapid
chemistry. This fact, and the lack of any obvious relationship with ozone photolysis,
other than the fact that a certain minimum amount of sunlight is required for the bursts
to proceed, suggests strongly that unlike some other locations oxidation of DMS andror
SO 2 is unlikely to be the mechanism of new particle formation unless an oxidant is
involved which can oxidise sulphur gases extremely rapidly and whose formation is
strongly associated with low tidal periods. Whilst the phenomenon observed appears to
be very local, it clearly warrants further study. It seems likely that such massive
production of new particles is limited to a few coastal locations such as Mace Head
which have precisely the appropriate local conditions. However, it is probable that any
process occurring so efficiently at one site may well be occurring on a far more
widespread basis at a lesser rate, and hence if the processes observed at Mace Head are
occurring across large areas of ocean at a rate which enhances the background
A.G. Allen et al.r Atmospheric Research 51 (1999) 1–14
13
concentration of particles without creating noticeable bursts, this may prove to be of
considerable importance in global climate regulation. Whilst the tide appears to be the
controlling factor on many days, the reason for lack of particle production on others
remains unclear. We suggest that particle production at this location may be controlled
by a complex interaction between tidal condition, insolation, and the chemical composition of the air.
Acknowledgements
We are grateful to the UK NERC ŽNatural Environment Research Council. for
funding this work under the ACSOE ŽAtmospheric Chemistry Studies in the Oceanic
Environment. programme Žcontract number GSTr02r1274.. We also thank Dr. Paul
Monks of the University of Leicester for providing the jO1 D radiometer data.
References
Atkinson, R., Baulch, D.L., Cox, R.A., 1996. Evaluated kinetic and photochemical data for atmospheric
chemistry. Atmospheric Environment 30.
Ayers, G.P., Gras, J.L., 1991. Seasonal relationship between cloud condensation nuclei and aerosol methanesulfonate in marine air. Nature 353, 834–835.
Ayers, G.P., Ivey, J.P., Gillett, R.W., 1991. Coherence between seasonal cycles of dimethyl sulphide,
methanesulphonate and sulphate in marine air. Nature 349, 404–406.
Charlson, R.J., Lovelock, J.E., Andreae, M.O., Warren, S.G., 1987. Oceanic phytoplankton, atmospheric
sulphur, cloud albedo and climate. Nature 326, 404–406.
Cipriano, R.J., Blanchard, D.C., Hogan, A.W., Lala, G.G., 1983. On the production of Aitken nuclei from
breaking waves and their role in the atmosphere. Journal of the Atmospheric Sciences 40, 469–479.
Davison, B., Hewitt, C.N., O’Dowd, C.D., Lowe, J.A., Smith, M.H., Schwikowski, M., 1996. Dimethyl
sulfide, methanesulfonic acid and physicochemical aerosol properties in Atlantic air from the United
Kingdom to Halley Bay. Journal of Geophysical Research 101, 22855–22867.
Despiau, S., Cougnenc, S., Resch, F., 1996. Concentrations and size distributions of aerosol particles in coastal
zone. Journal of Aerosol Science 27, 403–415.
Gaggeler, H.W., Baltensperger, U., Emmenegger, M., Jost, D.T., Schmidtott, A., Haller, P., Hofmann, M.,
1989. The epiphaniometer, a new device for continuous aerosol monitoring. Journal of Aerosol Science 20,
557–564.
Georgii, H.-W., Metnieks, A.L., 1958. An investigation into the properties of atmospheric freezing nuclei and
sea-salt nuclei under maritime conditions at the west coast of Ireland. Geofisica Pura e Applicata 41,
159–176.
Glasius, M., Calogirou, A., Jensen, N.R., Hjorth, J., Nielsen, C.J., 1997. Kinetic study of gas-phase reactions
of pinonaldehyde and structurally related compounds. International Journal of Chemical Kinetics 29,
527–533.
Harrison, R.M., Peak, J.D., Msibi, M.I., 1996. Measurement of airborne particulate and gaseous sulphur and
nitrogen species in the area of the Azores, Atlantic Ocean. Atmospheric Environment 30, 133–143.
Harrison, R.M., 1997. Proceedings, seventh European symposium on physico-chemical behaviour of atmospheric pollutants. In: Larsen, B., Versino, B., Angeletti, G. ŽEds.., European commission, directorate-general for science, research and development DGXII-D, Brussels, pp. 495–502.
Hatakeyama, S., Izumi, K., Fukuyama, T., Akimoto, H., 1989. Reactions of ozone with alpha-pinene and
beta-pinene in air—yields of gaseous and particulate products. Journal of Geophysical Research 94,
13013–13024.
14
A.G. Allen et al.r Atmospheric Research 51 (1999) 1–14
Hegg, D.A., Ferek, R.J., Hobbs, P.V., Radke, L.F., 1991. Dimethyl sulphide and cloud condensation nucleus
correlations in the northeast Pacific Ocean. Journal of Geophysical Research 96, 13189–13191.
Hinds, W.C., 1982. Aerosol Technology. Wiley, New York.
Hoffmann, T., Odum, J.R., Bowman, F., Collins, D., Klockow, D., Flagan, R.C., Seinfeld, J.H., 1997.
Formation of organic aerosols from the oxidation of biogenic hydrocarbons. Journal of Atmospheric
Chemistry 26, 189–222.
Hoppel, W.A., Frick, G.M., Fitzgerald, J., Larson, R.E., 1994. Marine boundary-layer measurements of new
particle formation and the effects non-precipitating clouds have on aerosol-size distribution. Journal of
Geophysical Research 99, 14443–14459.
Hudson, J.G., 1993. Cloud condensation nuclei near marine cumulus. Journal of Geophysical Research 98,
2693–2702.
Junkermann, W., Platt, U., Volz-Thomas, A., 1989. A photoelectric detector for measurement of photolysis
frequencies of ozone and other atmospheric molecules. Journal of Atmospheric Chemistry 8, 203–227.
Lowe, J.A., Smith, M.H., Davison, B.M., Benson, S.E., Hill, M.K., O’Dowd, C.D., Harrison, R.M., Hewitt,
C.N., 1996. Physicochemical properties of atmospheric aerosol at south Uist. Atmospheric Environment
30, 3765–3776.
Mason, B.J., 1957. The oceans as a source of cloud-forming nuclei. Geofisica Pura e Applicata 36, 148–164.
McGovern, F.M., Jennings, S.G., O’Connor, T.C., 1996. Aerosol and trace gas measurements during the Mace
Head experiment. Atmospheric Environment 30, 3891–3902.
McWilliams, S., Morgan, W.A., 1955. The relation between the concentration of atmospheric condensation
nuclei and other meteorological elements at Valentia Observatory. Geofisica Pura e Applicata 31,
129–146.
Mihalopoulos, N., Nguyen, B.C., Boissard, C., Campin, J.M., Putaud, J.P., Belviso, S., Barnes, I., Becker,
K.H., 1992. Field study of dimethylsulfide oxidation in the boundary layer: variations of dimethylsulfide,
methanesulfonic acid, sulfur dioxide, non sea-salt sulfate and Aitken nuclei at a coastal site. Journal of
Atmospheric Chemistry 14, 459–477.
Pandis, S.N., Baltensperger, U., Wolfenbarger, J.K., Seinfeld, J.H., 1991. Inversion of aerosol data from the
epiphaniometer. Journal of Aerosol Science 22, 417–428.
Persson, C., Leck, C., 1994. Determination of reduced sulfur-compounds in the atmosphere using a cotton
scrubber for oxidant removal and gas-chromatography with flame photometric detection. Analytical
Chemistry 66, 983–987.
Pollak, L.W., 1955. Condensation nuclei, Symposium in Dublin. Nature 175, 1072–1073.
Resch, F., Afeti, G., 1991. Film drop distributions from bubbles bursting in seawater. Journal of Geophysical
Research 96, 10681–10688.
Rogak, S.N., Baltensperger, U., Flagan, R.C., 1991. Measurement of mass transfer to agglomerate aerosols.
Aerosol Science and Technology 14, 447–458.
Seinfeld, J.H., Pandis, S.N., 1998. Atmospheric Chemistry and Physics, From Air Pollution to Climate
Change. Wiley, New York, pp. 656–684.
Wiedensohler, A., Orsini, D., Covert, D.S., Coffmann, D., Cantrell, W., Havlicek, M., Brechtel, F.J., Russell,
L.M., Weber, R.J., Gras, J., Hudson, J.G., Litchy, M., 1997. Intercomparison study of the size-dependent
counting efficiency of 26 condensation particle counters. Aerosol Science and Technology 27, 224–242.
Yin, F., Grosjean, D., Seinfeld, J.H., 1990. Photooxidation of dimethylsulfide and dimethyl disulfide: I.
Mechanism development. Journal of Atmospheric Chemistry 11, 309–364.