Investigation of the Heterogeneous Ice Nucleation Potential of Sea Spray Aerosol
by
Lilian A. Dove
Submitted to the Department of Earth, Atmospheric and Planetary Sciences
in Partial Fulfillment of the Requirements for the Degree of
Bachelor of Science in Earth, Atmospheric and Planetary Sciences
at the Massachusetts Institute of Technology
May 18, 2018 D 7a ?019
Copyright 2018 Lilian A. Dove. All rights reserved.
The author hereby grants to MIT permission to reproduce
and to distribute publicly paper and electronic copies of this
thesis document in whole or in part in any medium now
known or hereafter created.
Signature redacted
Certified by
Accepted by
Department of Earth, Atdiespl (ic a'nd Planetary Sciences
May 18, 2018
Signature redacted
/ / Daniel J. Cziczo
(MASSA HU~
,,]hesis Supervisor
Bignature redacted
Richard P. Binzel
Chair, Committee on Undergraduate Program
)MTS INSTITUTE
CHNOLOGY 0)
LU
062018
ARIES
OF TE
JUN
LIBR
Author
Investigation of the Heterogeneous Ice Nucleation Potential of
Sea Spray Aerosol
Lilian A. Dove
May 15, 2018
Abstract
Bubble bursting at the ocean surface generates smaller film-burst particles and larger jet
drop particles that differ in composition. The chemical composition of sea spray aerosols
is an important parameter for the evaluation of their impact on the global climate system.
This study investigates the role of particle chemistry on the heterogeneous ice nucleation
potential of laboratory-generated sea spray aerosols. Cultures of Procholorococcus, a highly
abundant marine phytoplankton species, were used as a model source of organic sea spray
aerosols. Results show that smaller particles generated from the lysed Procholorococcus
cultures were organically enriched and effectively activated as ice nucleating particles at
warmer temperatures and lower supersaturations than larger particles. The role of chemical
composition in the activation of the particles was studied by measuring the nucleation abili-
ties of single component organic molecules that mimic proteins, lipids, and carbohydrates in
Procholorococcus. Amylopectin, agarose, and aspartic acid exhibited nucleation behaviors
similar to particles generated from Procholorococcus cultures. Therefore, carbohydrates and
proteins with numerous and well-ordered hydrophilic functional groups may determine the
ice nucleation potential of organic sea spray aerosols.
1
Contents
1 Introduction
2 Methods
2.1 Chemical Representation and Culturing of Procholorococcus
2.2 Instrum entation . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.1 Aerosol Generation - Frit Bubbler . . . . . . . . . . . .
2.2.2 Aerosol Generation - Atomizer . . . . . . . . . . . . .
2.2.3 SPectrometer for Ice Nuclei . . . . . . . . . . . . . . .
2.2.4 Particle Analysis by Laser Mass Spectrometry . . . . .
2.2.5 Differential Mobility Analyzer . . . . . . . . . . . . . .
2.2.6 Condensation Particle Counter . . . . . . . . . . . . .
2.3 Experiniental Setup . . . . . . . . . . . . . . . . . . . . . . .
3 Results
3.1 Chemical Composition of SSA Samples . . . . . . . . . . . . .
3.2 Ice Nucleation Properties of Organic Components . . . . . . .
4 Discussion
4.1 SSA Formation Mechanism Impact on Chemical Composition
4.2 Effect of SSA Composition on Ice Nucleation Potential . . . .
4.3 Broader Implications of SSA on Global Ice Nucleation . . . .
5 Conclusion
6 References
Appendix A
Appendix B
2
3
5
5
5
5
6
7
8
9
9
9
10
10
13
15
. . . . . . . . . 15
. . . . . . . . . 16
. . . . . . . . . 17
18
19
24
25
1 Introduction
Atmospheric aerosols have a profound impact on climate by directly interacting with incoming
solar radiation, as well as by serving as the seeds that nucleate clouds (Haywood and Boucher,
2000). Atmospheric aerosols come from a wide variety of natural and anthropogenic sources and
have typical radii ranging from 0.001 to 10 micrometers (Haywood and Boucher, 2000). Regional
air quality and global climate are both impacted by the presence of these particles. Unlike
greenhouse gas emissions which have exacerbated the warming trends of climate change, aerosols
are predicted to have a net cooling effect on the climate (IPCC 2014). However, there exist
large uncertainties in the magnitude and the direction of the forcing that aerosol indirect effects
have on Earth's radiation budget (IPCC 2014). The subjective confidence level of IPCC Fifth
Assessment Report was "low" for cloud adjustments due to aerosols (IPCC 2014). Understanding
aerosol-cloud relationships is vital to reducing the uncertainty present in climate models and
forecasts (IPCC 2014).
Atmospheric aerosol particles act as centers for water droplet or ice crystal nucleation.
Many types of particles found in the atmosphere act as cloud condensation nuclei, which are
the surfaces upon which water makes the transition from a vapor to a liquid to form a cloud
droplet. A small subset of the atmospheric aerosol population can induce ice formation at
conditions under which ice would not form without them (Hoose and M6hler, 2012). Only
approximately one in 105 aerosol particles can act as these ice nucleating particles (INPs),
which are the foundation for the formation of ice crystals (Pruppacher and Klett, 2011; DeMott
et al., 2003; Rogers et al., 1988). Activated INPs can go on to form high-altitude tropospheric ice
clouds, referred to as cirrus clouds, and additionally contribute to mixed-phase cloud glaciation,
which are important factors in global climate (DeMott et al., 2003; Rossow and Schiffer, 1999).
The potential of different types of aerosols to act as INPs is not well documented (Hoose and
M6hler, 2012). Until empirical tests on nucleation ability are done, predicting which particles
will nucleate ice is impossible.
Temperature and saturation ratio with respect to ice are the main environmental factors
which determine the mechanism of ice nucleation (Hoose and M6hler, 2012). Ice can form in
clouds without the presence of INPs through homogeneous freezing of water and solution droplets
at high ice supersaturation and at temperatures below -38'C (Koop, 2000). The specific onset
point of nucleation is a function of the water activity of the solution under consideration (Koop,
2000). Homogeneous freezing requires soluble components of a given aerosol to deliquesce, or
take up water below 100% relative humidity. After deliquescence occurs, the liquid water can
freeze spontaneously given the low temperatures.
Alternatively, heterogeneous ice nucleation requires preexisting INPs in order to occur
(Cziczo et al., 2013). Heterogeneous ice nucleation is thermodynamically assisted by the pres-
ence of INPs so that nucleation takes place at lower supersaturations or warmer temperatures
than are required for homogeneous ice nucleation (Vali et al., 2015). Ambient temperature,
relative humidity, and the physical and chemical properties of aerosol particles impact the con-
centrations of heterogeneous INPs in the upper troposphere (DeMott et al., 2003). Depositional
nucleation occurs when ice forms directly from water vapor onto INPs, a process which occurs
in cirrus clouds (Prupaccher and Klett, 2011). Other heterogeneous nucleation processes occur
in supercooled liquid environments, such as mixed-phase clouds. These modes include conden-
sation freezing, in which water vapor condenses on INPs at below freezing temperatures to form
a liquid droplet which freezes instantaneously; immersion freezing, in which INPs act as cloud
condensation nuclei and the formed droplet freezes when the temperature is lowered; and contact
freezing, in which a particle comes into contact with the supercooled water phase and induces
freezing (Prupaccher and Klett, 2011).
Like all aerosols, some INPs have natural sources, such as desert dust, pollen, or marine
plankton, while others have anthropogenic sources (Vali et al., 2013). However, the largest
3
natural source of atmospheric aerosol by emissions is sea spray (Erickson and Duce, 1988). Sea
spray aerosols (SSA), which have total global emissions estimated to be 1-3x 1016 g per year,
exhibit a wide variety of chemical and physical properties. There are multiple modes of aerosol
production through sea spray, but two produce aerosol particles small enough to last in the
atmosphere: bubble bursting and jet drops. The particles generated in both of these production
methods can act as cloud condensation nuclei in the lower troposphere, and some as INPs if
they are transported into the upper troposphere (Feingold et al., 1999, Carrio et al., 2007).
Figure 1 shows an idealized view of SSA generation in the ocean. When particles are
produced through bubble bursting, bubbles created by cresting waves rise to the surface of the
water and pop, sending submicron pieces of the film of the bubble up to 20 centimeters into
the air (Blanchard, 1963). Meanwhile, jet drop formation occurs when the crater in the ocean
surface left by the bursting bubble fills, producing a central jet which ejects droplets which are
larger than the aerosols produced by bubble bursting. Both of these aerosol production methods
may result in the presence of marine organic matter in the atmosphere. For bubble bursting,
the bubbles typically have an organically-enriched film of small hydrophobic organics from lysed
cells and vesicles. When the bubble bursts, fragments of the organic film are ejected into the
atmosphere (Zhuang et al., 1993). In jet droplet aerosol formation, larger portions of cells and
vesicles can be sent into the atmosphere along with other inorganic components present in bulk
seawater (Bigg and Leck, 2008; Collins et al., 2014).
Wave Jet Drop
Breaking
Film Burst
Bubble OrganicBubbleMatter%
Entrainment ,t
Organic /
Partitioning
Figure 1: Idealized schematic of SSA generation. Breaking waves entrain bubbles below the
sea surface. These bubbles take up organic matter which is present in the sea surface microlayer
and subsurface water. At the surface, the organically enriched film of the bubble bursts, resulting
in small film burst particles. The crater left by the popped bubble fills with subsurface water,
resulting in the discharge of jet drop particles.
Current research focuses mainly on the role of SSA as INPs through condensation and im-
mersion freezing (DeMott et al., 2016; Junge and Swanson, 2008). Limited data exist regarding
the role of deposition freezing for the same particles. SSA can homogeneously nucleate ice well
because the presence of salts allows the particle to deliquesce, taking up water before the relative
humidity reaches above 100%. Additionally, the porosity of organically enriched SSA may enable
these particles to nucleate ice depositionally (Adler et al., 2013). The increased porosity of the
aerosols in high-altitude clouds results from phase separation upon freezing followed by a glass
4
transition of the organic material. A porous structure can be preserved after ice sublimation
(Adler et al., 2013). The observed increase in ice nucleation efficiency of glassy organic particles
present in high-altitude clouds may be explained by the presence of the porous structures.
We are studying how chemical composition of SSA impacts the ability of these aerosols
to act as INPs. A continuous flow diffusion chamber is used to study how particles related to
the marine cyanobacteria Prochlorococcus activate as INPs. Prochlorococcus is one of the most
widespread and abundant marine primary producers and is used as a representative source of
organic SSA (Flombaum et al., 2013). In addition, we use a single particle mass spectrometer to
investigate how the chemical composition of SSA impacts ice nucleation potential. Ice nucleation
potential is the likelihood of a particle to nucleate ice under certain conditions and is character-
ized by temperature and supersaturation at nucleation onset. By exposing the lysed cells and
analogues of Prochlorococcus's organic components to a range of tropospherically-relevant tem-
peratures and ice supersaturations consistent with heterogeneous freezing, we begin to determine
the ice nucleation potential of laboratory-generated SSA.
2 Methods
2.1 Chemical Representation and Culturing of Procholorococcus
Inorganic components of SSA were replicated using solutions of NaCl ( 99.9%, Macron Chem-
ical) and synthetic seawater (SSW; Paragon Scientific). The major inorganic salts of natural
seawater (NaCl, MgCl 2 , KC1, CaSO 4 , and NaHCO 3 ) are present in SSW. Particles of single
component organics were generated from compounds representative of proteins, lipids, and car-
bohydrates found in natural SSA. Proteins were represented by bovine serum albumin (BSA;
>99%, protease and fatty acid free), L-threonine ( 99%), L-arginine ( 99%), and L-aspartic
acid (>98%). Lipids were represented by stearic acid ( 98.5%), oleic acid ( 99%), elaidic
acid ( 98%), and 1,2-decanediol ( 99%). The carbohydrates were represented by amylopectin
( 99%), D-(+)-trehalose (>99%), D-(+)-raffinose ( 99%), and agarose ( 99.5%).
The phytoplankton Prochlorococcus was used to represent the sources of organic carbon
in the global ocean. One of the most abundant marine cyanobacteria, Prochlorococcus are found
across many ocean regions in global mean abundances of 2.9 x 1027 cells (Flombaum et al.
2013). The organisms are typically found in the upper water column, making them suitable for
this study's interests in the organic components of the ocean surface microlayer (Flombaum et
al., 2013). To grow cultures of the phytoplankton, two strains of phenotypically distinct high-
light adapted Prochlorococcus (MED4 HLI and MIT9312 HLII) and one strain of a low-light
adapted ecotype (NATL2A LL) were cultured axenically, or free from the presence of other
organisms. The culturing occurred in a natural seawater based PRO99 liquid medium with a
13:11 Light:Dark incubator at 24'C (Moore et al., 2007). The cultures were harvested in the
late exponential growth phase. In order to mechanically break apart the cells, some cell cultures
were lysed by immersion in a pulsing sonicator for five minutes before aerosolization.
2.2 Instrumentation
2.2.1 Aerosol Generation - Frit Bubbler
A bubbler allows for the generation of aerosols by bubbling filtered air through sintered glass into
a solution. The rate of the airflow and the sinter pore size determine the size of the bubbles. It
is not always possible to examine SSA in its natural state, so the bubbling method is one way to
simulate seawater aerosolization in a laboratory setting. A wave tank is another aerosolization
method. Our wave tank was constructed using a large flat-bottomed flask attached to a 0.5 -
5 Hertz shaker table. Splashing motions create aerosol particles, yet the total concentration of
particles produced is an order of magnitude less than with the bubbler.
5
The employed generation method for this study had to accurately recreate the production
of aerosols through both the film burst and jet drop modes. Figure 2 shows size distributions of
SSA collected from field studies in the North Atlantic Ocean (Koponen et al. 2002), the tropical
Atlantic Ocean (Atmospheric Radiation Measurement Climate Research Facility, 2014), and the
tropical North Pacific Ocean (Marine Atmospheric radiation measurement Global energy and
water cycle experiment cloud system study pacific cross section intercomparison Investigations
of Clouds (MAGIC) campaign (Lewis et al., 2012)). Overlaid are size distributions from this
study's particle generation using a frit bubbler, a wave tank, and from the Scripps Institute
for Oceanography (SIO) Hydraulics Laboratory generation using a wave tank (Stokes et al.,
2013). The bubbler was able to most accurately recreate the size distribution, and likely the
chemistry, of natural SSA. We attribute the bimodality of the size distributions to the film burst
particle (100-200 nm) and jet drop particle (400-600 nm) production mechanisms (Wang et al.,
2017). The wave tank method was not able to recreate the two desired mechanisms and did not
show a bimodal distribution which is seen in natural SSA. Out of the tested methods of aerosol
generation, the bubbler most accurately recreated the size distribution of the natural SSA. We
therefore used the bubbler to aerosolize cultures of Prochlorococcus.
100
C?
E
x Bubbler\
V- 
-
-
SWave Tank
- -.. SIO Hydraulic Lab Wave Tank
- - Tropical South Atlantic
-Tropical Pacific
LJTropical Pacific Interseasonal Variability
---- North Atlantic
=North Atlantic interseasonal Variability
10-3
10 10 10
Particle Diameter (nm)
Figure 2: Size distributions of particle concentration versus particle diameter for field
and laboratory data. Field measurements are reported from the Tropical South Atlantic (Ko-
ponen et al., 2002), Tropical Pacific (Lewis et al., 2012), and the North Atlantic (Atmospheric
Radiation Measurement Climate Research Facility, 2014). Wave tank data are reported from the
Scripps Institute for Oceanography Hydraulic Lab Wave Tank (Stokes et al., 2013). Reproduced
from Wolf et al. (submitted).
2.2.2 Aerosol Generation - Atomizer
The atomizer also produces aerosol from a solution. Compressed air draws bulk liquid from a
reservoir as a result of the Bernoulli effect. The high-velocity air breaks apart the liquid solution
and the resulting droplets are suspended to form an aerosol. By varying the pressure of the
compressed air or the dilution ratio of the solution, one can modify the resulting particle size
distribution. (Baron and Willeke, 2001) The aerosolization method is typically used in the case
when a given solution cannot be aerosolized using the bubbler because of its viscosity.
6
WK
2.2.3 SPectrometer for Ice Nuclei
The SPectrometer for Ice Nuclei (SPIN; Droplet Measurement Technologies, Boulder, CO) is a
continuous flow diffusion chamber (CFDC) instrument (described in Rogers, 1988) for studying
the ice nucleation properties of aerosols. SPIN quantifies the concentrations of ice nucleating
particles by controlling the temperature and relative humidity to which aerosol particles of
interest are exposed (Garimella et al., 2016). Particles enter SPIN through an inlet and then
pass between two walls coated with approximately 1 mm ice and separated by 1 cm. A sheath
flow of about 9 liters per minute (lpm) constrains a laminar particle flow of 1 1pm to the center
of the chamber. As shown in Figure 3, the walls are held at different temperatures to cause
water vapor and heat to diffuse from the warmer wall to the colder wall. The magnitude
of the temperature gradient determines the degree of ice supersaturation, with the maximum
in supersaturation occurring near the center of the chamber (Rogers, 1988). For a typical
experiment, the supersaturation within SPIN changes at a rate of approximately 2% relative
humidity with respect to ice per minute. A full air and water flow diagram of SPIN can be
found in the Appendix.
Within the chamber, some INPs will nucleate ice while other particles will pass through
unactivated. Particles then enter the evaporation section of the instrument which is held at ice
saturation. Here, ice particles stay as ice but liquid water is evaporated from droplets which have
not nucleated ice. At the bottom of SPIN is an Optical Particle Counter which outputs data
about the size and polarization properties of the individual particles which pass through it. The
values of the polarization properties describe whether a particle has nucleated ice or has remained
unactivated. Garimella et al. (2016) describes a constructed machine learning algorithm which
trains on the four polarization parameters to create probability density functions for ice crystals,
water droplets, and unactivated aerosols. Individual particles can then be classified into their
respective phases.
Sheath Sheath
flow Y flow Figure 3: Representation of an idealized CFDC. AParticles flow of particles passes between two ice-coated walls which
1 are held at different temperatures. The difference in tem-
perature between the two walls results in water vapor and
* ' heat diffusion from the warm wall to the cold wall. Super-
saturation exists at a maximum near the centerline of the
- chamber. The particle flow is between two sheath flows
along the walls, restricting the particles to the supersat-
Flow profile uration maximum at the central lamina. Figure adapted
from Garimella et al. (2017).
Garimella et al. (2017) describes the tendency for particles within a CFDC instrument
to spread outside the center lamina, where ice supersaturation is greatest. Through timed
pulse experiments under conditions comparable to the ones being tested, the fraction of aerosols
constrained to the center lamina were calculated. Figure 4 shows the fraction of aerosols in
the lamina flow as a function of the ice saturation ratio in SPIN at five different temperatures.
In the ideal case, when all particles are constrained within the laminar flow, all of the data
7
i
points would form a horizonatal line at 1.0. However, the fraction of particles in the lamina
is much lower than 1.0 at the given temperatures and total flow of 9.8 L min-1. As particles
outside the lamina experience a lower supersaturation and are unlikely to activate, a correction
factor between 1.86 and 7.96 should be applied to fractional activation data derived assuming all
particles are constrained to the lamina. A correction factor is calculated by taking the inverse
of the fraction of particles in the lamina at a specific temperature and supersaturation.
308
04/
C 03\ /
00 0400 1 12 14 1 12 14
-56CC 48 c -46'C -44'C -42'C
Ice Saturation Ratio (S,,)
Figure 4: Particle lamina spreading in SPIN. The fraction of particles constrained to the
SPIN aerosol lamina (fiam) as a function of ice saturation ratio (Sic) is represented across
temperatures relevant to this study (-56 to -42 C). One standard deviation in variability is rep-
resented by the shaded regions and points represent individual measurements. A correction factor
for fractional activation is obtained by taking the inverse of fiam at a specified temperature and
Sice. Adapted from Wolf et al. (submitted).
2.2.4 Particle Analysis by Laser Mass Spectrometry
The Particle Analysis by Laser Mass Spectrometry (PALMS) instrument is a single particle mass
spectrometer described in detail by Thomson et al. (1997) and Cziczo et al. (2006). Aerosol
particles are drawn by vacuum into PALMS, with a typical size sampling range between 0.1-1.5
micrometers (Cziczo et al., 2006). A fraction of the indrawn particles is sized by two 532 nm
neodymium yttrium aluminum-garnet lasers using the particle transit time between the two
laser beams. A 193 nm excimer laser then vaporizes and ionizes the particles and the atomic
and molecular ions travel into a time of flight mass spectrometer which records their mass-
to-charge ratio. Since multiple charges are rare with the PALMS excimer laser, the resulting
mass-to-charge ratios are equivalent to particle mass. The spectrometer collects either positive
or negative ion spectra for the particles (DeMott et al., 2003). Particle ionization within PALMS
is not quantitative, meaning that individual particles from the same stream may show widely
contrasting chemical compositions. Therefore, large numbers of spectra have to be collected in
order to perform statistical analysis on the chemical composition of particles. A full schematic
diagram of PALMS can be found in the Appendix.
8
2.2.5 Differential Mobility Analyzer
A Differential Mobility Analyzer (DMA; Brechtel Manufacturing Inc., Hayward, CA) uses the
electrical mobility of particles to guide particle size-selection. A particle's electrical mobility is
primarily a function of its size and charge and is used to measure the ability of a particle to
travel across an electric field. Within the DMA, aerosols pass through a neutralizer which gives
the stream of aerosols a known electrostatic charge distribution. The stream of charged particles
then travels through sheath air (5-8 lpm) and enters into an electric field (0-6000 V). Changing
the electric field strength and the sheath air flow rate allows for the separation of particles of a
particular electrical mobility, and therefore of a known size range.
2.2.6 Condensation Particle Counter
A Condensation Particle Counter (CPC; Brechtel Manufacturing Inc., Hayward, CA) is used to
count aerosols when they are too small to be detected using traditional optical techniques. The
working liquid (typically butanol) rests on top of a heated saturation block, generating a high
vapor content of the solution in the chamber. The particles entering the CPC come into contact
with the vaporized working liquid and then enter a cooled chamber, in which the difference
in temperature results in supersaturation. The supersaturation causes the particles to act as
condensation nuclei and to grow in size as they attract the vapor. The grown particles can then
be detected by shining a light emitting diode through the sample and detecting and recording
light scattering. The greater the difference in temperature between the heated chamber and the
cooled chamber, the smaller the particles that can be counted.
2.3 Experimental Setup
Figure 5 shows a schematic of the experimental setup for a typical experiment. Initially, particles
were generated with either a bubbler or an atomizer with a flow of 1.3 lpm provided by dry
nitrogen air. The aerosol stream was further dried by passing through driers which uptook
excess water from the particles. The stream then passed through the DMA, which size-selected
particles. Particles thought to result from film burst generation were size selected at 200 nm, jet
drop particles at 500 nm, and pure organic molecules at 100 nm. Because a large concentration
of particles is necessary for there to be ice nucleation with these organic materials, the DMA may
not work as effectively on a single-particle basis and therefore potentially allowed some doubly-
charged large particles through. In addition, in some cases in which aerosol generation was low,
a polydisperse (non-size selected) distribution of aerosols was necessary. In these polydisperse
cases, the DMA was not used to size-select the generated particles.
A 0.3 1pm stream of the size-selected particles then traveled through the CPC, providing
a total count of the number of particles of all sizes. SPIN sampled the remaining 1.0 slpm of
the particle stream. From the CPC and SPIN data, the fractional activation of particles which
activated as INPs was calculated. In order to obtain chemical composition data, the stream of
size-selected particles could also be sent to PALMS at a flow rate of 0.3 lpm.
9
0.3 Ipm CondensationParticle
Counter
(CPC)
1.3 1pm N2  Differential
Mobility 1.0 1pm
Analyzer
Spectrometer
for Ice Nuclei
Frit Bubbler (SPIN)
Figure 5: Schematic of the typical experimental setup. 1.3 lpm of dry lab air entered the
frit bubbler, which contained the solution to be aerosolized. The resulting particle stream passed
through the DMA, which size selected the appropriately-sized particles. 0.3 lpm of the stream
passed to the CPC, where the total number of particles were counted. The remaining 1.0 lpm of
the stream passed through to SPIN. Alternatively, the flow from the DMA could travel directly
to PALMS.
3 Results
3.1 Chemical Composition of SSA Samples
Characteristic negative PALMS mass spectrometry spectra for synthetic seawater (SSW), the
Prochlorococcus culturing medium (Pro99), and 200 and 500 nm particles from lysed Prochloro-
coccus cultures are shown in Figure 6. The spectrum for SSW shows large peaks at ion
mass/charge (m/z) ratios of 35 and 37, corresponding to chlorine isotopes. Most of the ion
peaks in the SSW spectrum are inorganic in nature. There is very little carbon signal in the
SSW, and that which may be present is indistinguishable as a signal in the background noise
from the spectra.
Overall, the culturing medium ion spectrum resembles that of the SSW. The limited carbon
(peak at C 2- m/z = 24) present in the culturing medium may relate to chelating agents used to
keep metallic nutrients suspended in the solution (Moore et al., 2007). However, the culturing
medium shows reduced sulfate and sulfite ion signals compared to the SSW.
In contrast, the spectra of the lysed and aerosolized Prochlorococcus cultures show strong
ion signals corresponding to carbon-containing compounds. The 200 nm film burst particles
demonstrate strong carbon signals (C- m/z = 12, 13; C 2- m/z = 24, 25; CN- m/z = 26, CNO-
m/z = 42, 44; and C4- m/z = 48). Meanwhile, the 500 nm jet drop particles showed lower
carbon, phosphorus (PO2- m/z = 63; P03- m/z = 79; and P0 4- m/z = 95), and nitrogen (NO-
m/z = 26) signals. Nitrogen is found in amine functional groups while phosphorus is present in
the phosphate backbones of nucleic acids and headgroups of phospholipids. The nitrogen and
phosphorus present in the jet drop particles may result from the ionization of these biochemical
components.
10
10"
10-1
0
10-3
10
10
1 1
0
10-2
10-3
20 40 60 80 100
Ion Mass/Charge (m/z)
10 0S SynthticSeawaer
Cl-
NaC1I SO
503,
0 20 40 60 80 100 12C
CpCturing Mediuip
CI
2 
NaC NaCI2
10-3
100
101
102
120
10'1
10-2
0 20 40 60 80 100
Ion Mass/Charge (m/z)
120
Figure 6: Particle mass spectra. Negative characteristic spectra from PALMS are shown for
synthetic seawater, Prochlorococcus culturing medium (Pro99), and 200 and 500 nm particles
lysed Prochlorococcus cultures.
Carbon and chlorine ion signal comparisons from over 750 PALMS spectra are shown for
Prochlorococcus culturing medium and 200 and 500 nm particles from lysed Prochlorococcus
cultures in Figure 7. The carbon signal is expected to correlate to organic mass fraction while
the chlorine signal corresponds to inorganic components of the sample. The carbon ion signal
for the culturing medium is low to zero in almost every sample. However, the aerosols generated
from the lysed cultures, especially the 200 nm particles, demonstrate a larger signal of carbon.
As the signal of the organic carbon increases in the samples, the signal of the inorganic chlorine
decreases.
11
200 9M LysqdProc loroopcus CIture
CN- Cl-
C, CNO*
0- P03
C C4- P02-
lSoi 03 P04liii ~ C~aCII I
11 504
0 20 40 60 80 100 12C
500 rn Lys d Proc4Iorocoqcus C4Iture
CI
NaC 2
C2- F)04
j ,N'C4- NaC1 53 50C P
0
200 nrn Lysed Prochlorococcus Culture 500 nm Lysed Prochlorococcus Culture
0.9 0.91
0.8 0.*1
-0.71 r 
- 0.71
O A CA
.'0.3.
0.2
0.1
0'
0
0.2
0.1
0.2 0A 0.6
Carbon Ion Signal
01
0.8 0 0.2 0A 0.6
Carbon Ion Signal
I I
0.9
0.8
-0.7
10.6
0.5
o 0.40A
0 0.3
0.2i
0.1
0'
0
Culturing Medium
Figure 7: Carbon and chlorine ion signal comparisons of lysed Prochlorococcus cul-
tures. Comparisons of ion signals organic carbon (ion m/z of 12, 13, 24, 25, 26, 28, 38, 42,
44, and 48) and chlorine (ion m/z of 35 and 37) are shown for >750 samples of Prochlorococcus
culturing medium (Pro99) and 200 and 500 nm particles from lysed Prochlorococcus (NATL2A)
cultures.
Figure 8 shows carbon and chlorine ion signal comparisons of water samples collected
from the Straits of Florida in March 2018. The particles generated from the water sampled from
the organically enriched sea surface microlayer demonstrate stronger carbon signals than those
generated from the water sampled from the subsurface, two meters down. These results correlate
to those shown in Figure 7. The smaller particles suspected to form from bubble bursting at
the ocean surface, where the sea surface microlayer exists, have more organic carbon than the
larger particles formed from the jet drop mechanism, which aerosolizes subsurface water. This
comparison suggests that the laboratory analogue to seawater using Prochlorococcus and the two
SSA generation mechanisms employed in this study were effective in mimicking the chemical
composition natural seawater.
1 -Microlayer water
0.9
0.*
.0.7
$0.6
0.5i
040.3t.
0.2.
0.1
0 0.2 0.4 0.6
Carbon Ion Signal
Subsurface wateir
0.7
0.*
0.7
10.5
0.4
0.3
0.2
0.1
0.8 0 0.2 04 0*6
Carbon Ion Signal
Figure 8: Carbon and chlorine ion signal comparisons of field samples. Comparisons
of ion signals organic carbon (ion m/z of 12, 13, 24, 25, 26, 28, 38, 42, 44, and 48) and chlorine
(ion m/z of 35 and 37) are shown for samples of a) sea surface microlayer and b) subsurface
water collected from the Straits of Florida.
12
0.2 0.4 0.6
Carbon Ion Signal
0.8
0.8
3.2 Ice Nucleation Properties of Organic Components
The supersaturation with respect to ice (Sice) at the onset of ice nucleation for inorganic, protein,
lipid, carbohydrate, and Prochlorococcus particles is shown in Figure 9. The onset of nucleation
is defined as when 1 in 10 4 particles activated as an INP. Correction for the lamina spreading in
SPIN (Figure 5) was implemented. Each condition was replicated at least twice and the points
represent the average of the trials conducted at that temperature. The black line corresponds to
homogeneous nucleation, as defined in Koop et al. (2000), the Sice at which ice crystals nucleate
due to chance aggregation of liquid molecules. None of the groups of particles followed uniform
patterns of nucleation and all of the tested samples nucleated homogeneously above -43 C.
The inorganic particles generated form SSW and from the culturing medium solution
primarily activated homogeneously. Some heterogeneous nucleation occurred around -55'C but
particles again nucleated homogeneously below -58'C. Kong et al. (2018) provides an explanation
for this phenomenon. Between -55 and -58'C, depositional freezing and deliquescence compete
kinetically. Deliquescence leads to homogeneous freezing; however, between these temperatures,
deposition nucleation is more favorable, even for purely inorganic substances (Kong et al., 2018).
Other studies which tested the heterogeneous nucleation abilities of inorganic substances found
similar results (Schill and Tolbert, 2014; Ladino et al., 2016).
Particles generated from proteins and amino acids demonstrated two modes of nucleation.
The protein mixture (a combination of all of the tested protein substances in equal parts) and the
aspartic acid were effective heterogeneous INPs. As the temperature got lower, the Sice required
to activate the particles lowered as well. The critical ice saturations for aspartic acid and the
mixture were similar, suggesting that aspartic acid determined the nucleation potential of the
mixture. Meanwhile, the threonine and bovine serum albumin did not activate heterogeneously
until below -47 C while the arginine solely activated homogeneously.
Some carbohydrates followed a similar pattern to the aspartic acid. Particles produced
from two polysaccharides (amylopectin and agarose) were effective INPs below -46'C while
the oligosaccharide samples only nucleated homogeneously. The particles generated from lipids
generally weakly nucleated ice below homogeneous freezing conditions. Oleic acid and the lipid
mixture nucleated ice at Sice values of 1.21 and 1.27 at approximately -56 and -54'C, respectively.
Different forms of particles generated from Prochlorococcus cultures were tested for their
ice nucleation abilities. Jet drop (500 nm) particles demonstrated the lowest ice nucleation
potential of the particles generated from cells, nucleating homogeneously until -49'C, below
which they nucleated between Sice values of 1.32 and 1.36. Particles generated from whole cells
nucleated homogeneously until -48 C with only slightly more ability to nucleate heterogeneously
than jet drop particles. Meanwhile, the film burst (200 nm) particles demonstrated effective
INP properties below -44'C. At -50'C, the film burst particles activated at Sice = 1.12. The
polydisperse stream of particles demonstrated highly efficient INP properties. The behavior of
the film burst and the polydisperse samples follow similar patterns between -40 and -50'C.
13
a b__ _ c
1. a .1.6 b1.6
- -- - ---- - - -
11.4 -1.44
C
11.3 -X1.3 1.3
+
1.2 1.2 1.2X SSW
O Cuitudng Medium A Protein MU 2 Lipid Mixture
SKong et al. d .i 1 A Aspartic Acid 1.1 E Oleic Acid
-+ ChinoaUt aL A T 1,2 Dmanodlol1 Lacino et al. A Bovine Serum min N Elaidic Acid
X 1 snine . .1 . e Stearic Acid
-60 -55 -50 -45 -0-80 -55 -50 -4 -40 -60 -5 -60 -45 -40
d
1.6 '1.6
-Homogeneous Fresing (Koop et al.)
1.5.1.5
1.4 1.4 -- Water Saturation
S.o Inorganic
1 Proteins
ri1.2 0 Film BurstSPolydaperse 0 Lipids
Bu 8
Agyrope 1.1 Jet Drop 0 Carbohydrates
1.1 ateMixure- eWilson at al. Micirolvior
o Wilson et al. Subsurface o Plankton
1 Trehelose 114 Knopf at al. Diaoms
-o -55 -60 -45 -40 -60 -55 -50 -48 -40
Temperature (*C) Temperature (0C)
Figure 9: Onset of ice nucleation. The supersaturation with respect to ice at the onset of ice
nucleation is shown for a) inorganics and culturing media, b) proteins, c) lipids, d) carbohydrates,
and e) cultures of Prochlorococcus. Amylopectin, agarose, aspartic acid, and the protein mixture
exhibited nucleation behaviors similar to the smaller particles generated from Prochlorococcus
cultures. Reproduced from Wolf et al. (submitted).
The fractional activation as a function of Sice at -46'C is shown in Figure 10. Fractional
activation demonstrates what percentage of particles have activated as ice at a given temper-
ature. Higher fractional activation at lower ice supersaturation suggests the substance to be a
good INP. The possible underactivation of particles due to lamina spreading is accounted for
using the correction factor determined from Figure 5.
Particles derived from inorganics are ineffective INPs, demonstrating no activation until
homogeneous conditions are reached. Aspartic acid is the only individual protein sample which
exhibits heterogeneous nucleation. The protein mixture follows a similar trend. Lipids as a
whole demonstrate high Sice upon nucleation onset. Agarose and amylopectin demonstrated
strong heterogeneous nucleation potential; however, the carbohydrate mixture did not. This
idiosyncrasy is atypical, as the mixtures usually follow the potential of the most effective com-
ponent. For the cells, the film burst and polydisperse particles exhibited strong heterogeneous
nucleation potential, activating at Sice values of 1.2 and 1.18, respectively.
14
I
Inorganics Proteins Upids Carbohydrates Cells
1.45. - - - 101
1.4-
.1.35 10-2.
E 1.3 -
1.25 -013'1
S1.23
- 1.15 10
1.1
1.05 10
-- Homogeneous Freezing -46'C jRepresentative Lamina Ske Certainty
Figure 10: Fractional ice activation. The fraction of particles activated as ice as a function
of Sice is shown for the experiments run at -4 C. Inorganics are poor INPs while aspartic acid,
the protein mixure, agarose, and amylopectin effectively activated as INPs. The smaller particles
produced from Prochlorococcus also demonstrated strong ice nucleation abilities. Reproduced from
Wolf et al. (submitted).
4 Discussion
4.1 SSA Formation Mechanism Impact on Chemical Composition
The compositional differences between the film burst and the jet drop particles suggests that the
differences in generation of the two particle types result in different values of organic enrichment.
The synthetic seawater and culturing medium have very little organic content, reflected by the
low organic signal of aerosols generated from these solutions (Figure 7). Meanwhile, both of the
particle types generated from lysed Prochlorococcus cells contain organic material. The smaller
film burst particles are more organically enriched than the larger jet drop particles. The organic
microlayer across the ocean surface from which film burst particles arise contains more organic
material which is relevant for ice nucleation than the deeper subsurface water from which jet
drop particles form (Wilson et al., 2015).
The organic components present in the jet drop and film burst particles cannot be identified
directly, only in terms of relative abundance. However, the relative organic enrichment of the
film burst particles demonstrates that the different modes of generation are important in creating
distinguishing differences within particles considered bulk SSA. Overall, the different sizes and
compositions of these particles suggests that SSA generation must be considered and replicated
accurately when testing the ice nucleation properties of natural sea spray aerosols in a laboratory
setting. The organic enrichment of film burst particles may also be important in contexts outside
of ice nucleation. For instance, Salter et al. (2016) suggests that calcium enrichment in SSA may
15
I
be important for the pH-dependent aqueous chemistry of marine aerosols due to contributions
to excess alkalinity.
4.2 Effect of SSA Composition on Ice Nucleation Potential
Amylopectin, agarose, aspartic acid, and the protein mixture most closely mimicked the nu-
cleation behavior of the lysed Prochlorococcus film burst particles. This trend suggests that
the presence of certain carbohydrate and protein molecules determine the heterogeneous ice
nucleation onset of natural sea spray.
Amylopectin and agarose are both polysaccharides and show a steep decline of supersatu-
ration at nucleation onset below -45'C. The large chain-building structure and atomic orientation
of these molecules may explain their ability to effectively act as INPs. Both of these molecules
have many exposed hydroxyl groups which can act as sites for hydrogen bonding with water
and other molecules. The liquid water which is attracted to these sites can form a critical ice
embryo, with the particle acting as an INP. The hydroxyl groups contribute to the surface area
of the particle which can act as a critical ice embryo (Graether and Jia, 2001). Unlike other
particles which bind to ice and inhibit further growth, the amylopectin and agarose present
large ice-nucleating surfaces. If a particle's molecular structure contains sequences which align
with ice lattice structure, a stable ice embryo can form, decreasing the activation barrier for ice
nucleation (Pummer et al., 2015).
Some other saccharide compounds may not act as INPs because they too readily adsorb
water (Peng et al., 2001; Chan et al., 2008). If a molecule attracts water and forms a thin liquid
coat, depositional freezing will not occur. Although there are no literature values for the water
adsorption qualities of raffinose and trehalose, the other tested carbohydrate compounds, they
may have formed a water coat, causing them to deliquesce and to be incapable of freezing de-
positionally. The deliquesence may explain why the carbohydrate mixture did not follow the ice
nucleation behavior of the good INPs but instead was nucleation-inhibited until values consistent
with homogeneous freezing. However, some particles generated from the carbohydrate mixture
did nucleate below homogeneous freezing values. Potentially these particles were predominantly
composed of the polysaccharides, which prevented deliquescence and allowed for depositional
nucleation.
Comparatively, the aspartic acid and protein mixture demonstrate a steep decline of su-
persaturation at nucleation onset below -42'C, reaching an Sice below 1.1 at -50'C. Aspartic
acid, like agarose and amylopectin, also has an exposed hydroxyl group. However, both threo-
nine and arginine, other tested nucleic acids, also have exposed hydroxyl groups. It is important
to consider atomic orientation as it relates to ice nucleation. Ice will only nucleate in specific
orientations and incompatible alignments will inhibit or prevent ice crystal growth. INPs have
molecular features, such as well-positioned and distanced hydroxyl groups, which contribute to
their ability to nucleate ice (Pandey et al., 2016; Graether and Jia, 2001; Pummer et al., 2015).
Aspartic acid's positioning of hydroxyl groups may contribute to its ability to nucleate ice. Un-
like for the carbohydrates, the protein mixture followed the good INP ability of the aspartic
acid, suggesting that the threonine and arginine did not limit the mixture with deliquescent
behavior (Chan et al., 2005).
Considering fractional activation, although they were all good INPs, more agarose and
amylopectin than aspartic acid activated at lower supersaturations. This finding suggests that
the polysaccharides played the largest role in causing the SSA to activate as INPs. The long-
chain polysaccharides have more hydroxyl sites, thousands per chain, which have the capability
and orientation to act as hydrogen bonding sites. The organic enriched film burst particles likely
rely on hydroxyl groups to act as sites upon which ice nucleation can occur.
16
4.3 Broader Implications of SSA on Global Ice Nucleation
The chemical composition and properties of marine aerosol can have important consequences for
the climate properties of these particles. The partitioning of water-soluble and water-insoluble
organic matter and the diversity of molecules within the organic fraction is dependent on the
size of the particle (Facchini et al., 2008). Considering the multiple generation methods of
SSA, the organic composition of particles is therefore likely related to the mechanism of particle
generation: film burst or jet drop. Differences in organic enrichment can have impacts on the
ice nucleation properties of a particle. Wang et al. (2015) found a correlation between aliphatic-
rich species, or non-aromatic hydrocarbons, and concentrations of INPs in an experiment with
natural marine phytoplankton blooms. These aliphatic-rich organics were most likely aerosolized
via the bubble film burst mechanism. When testing a different bloom, there was no enrichment
of aliphatic species and no associated rise in INP concentration, suggesting that the presence of
these organic compounds was vital to the increase in concentration of INPs (Wang et al., 2015).
Changes in the marine system, such as ocean acidification and eutrophication resulting
from a lack of dissolved oxygen, may impact the chemistry of SSA. The molecular diversity
of the compounds which make up the organic component of SSA is linked to the presence of
marine bacteria and phytoplankton, which can vary in time and space (Cochran et al., 2017). In
addition, a mesocosm study performed by Galgani et al. (2014) suggests that ocean acidification
may impact the chemical composition of the ocean's surface microlayer. Changes include higher
concentrations of hydrophilic amino acids and lower concentrations of carbohydrates (Galgani et
al., 2014). Such changes in organic composition at the surface microlayer can impact bacterial
composition and concentration at the ocean surface by impacting pH and habitability. These
changes would impact the potential of SSA film burst particles to activate as INPs. Rapid shifts
in the conditions of the marine system resulting from anthropogenic impacts and emissions may
cause a range of impacts on a variety of climate-related processes, including ice nucleation (Wurl
et al., 2017).
Models demonstrate great uncertainties when it comes to ice nucleation by SSA. Although
models take differences in sub-micron and fine emissions into account, the importance of organic
matter enrichment in these models is still being defined (O'Dowd et al., 2008; Vignati et al.,
2009). Current models assume particles arising from sea spray to be internally mixed, meaning
the particles are chemically and physically homogeneous and reflect the average of all of the
components, inorganic and organic (Lesins et al., 2002). However, our data suggest that SSA
assemblages are less internally mixed than these models assume. It appears that the presence
of a specific organic compound in a particle can highly impact that particle's ability to nucleate
ice. Individual particle composition is important to consider in an externally mixed assemblage
of particles as each particle can demonstrate different properties, effecting characteristics such as
the ability to impact direct radiative forcing, as studied in Lesins et al. (2002) or ice nucleation.
Vergara-Temprado et al. (2017) demonstrates that global distribution of concentrations
of INPs are better aligned with ambient concentrations of feldspar and marine organic aerosols
than with theoretical descriptions of INPs by temperature or size. Similarly, Burrows et al.
(2013) found strong regional differences in the importance of INPs coming from marine and dust
sources. These studies suggest that INPs from different sources must be better understood in
order to effectively model the aerosol-cloud-climate system; treating all INPs as one assemblage
is imprecise and results in errors in climate sensitivity in models. This study attempted to
present a more cohesive view of INPs arising from SSA. A better understanding of the ambient
concentrations of the particles arising from sea spray with respect to chemical composition will
contribute to better modeling of the population of INPs. In addition, it is important to consider
that despite arising from the same source, film burst and jet drop particles demonstrated highly
disparate heterogeneous ice nucleation properties. In order to accurately represent the role of
SSA in global ice nucleation processes, models must begin to accommodate for the differences
in size and chemical composition of particles generated from the same source.
17
5 Conclusion
In this study, we investigated the role of SSA chemical composition on heterogeneous ice nu-
cleation behavior. SSA particles were generated in a laboratory setting using the naturally-
abundant Prochlorococcus cyanobacteria as model source of marine organic matter. Particle
generation from a frit bubbler resulted in the production of two kinds of particles, which were
designated as smaller film burst particles and larger jet drop particles. The smaller film burst
particles were organically enriched compared to the larger jet drop particles, as confirmed by
mass spectrometry. The difference in chemical composition between the particles suggests fur-
ther differences between the particles' abilities to activate as INPs.
The heterogeneous ice nucleation potentials of the SSA and relevant carbohydrates, pro-
teins, and lipids were analyzed. Inorganic components were ineffective as INPs. Larger jet
drop particles from the lysed Prochlorococcus cultures acted as INPs below -49'C, while the
smaller film burst particles were effective INPs at warm temperatures and low supersaturations
at colder temperatures. Comparing the results of the SSA generated from lysed cell cultures to
pure organic molecules found that the large polysaccharides agarose and amylopectin as well as
aspartic acid (a nucleic acid) most closely mimicked the heterogeneous ice nucleation behavior of
film burst particles. Nucleation was initiated at high temperatures and low supersaturations for
colder temperatures for these molecules, suggesting an explanation for the good INP behavior
documented in film burst particles. However, other organic substances were moderate or poor
INPS. The presence and favorable orientation of hydroxyl groups in organic molecules appears
to be an important factor in considering ice nucleation potential of film burst SSA particles.
Future work in this area should focus on confirming the difference in and importance of
organic enrichment between different modes of SSA. Individual organic components, specifically
carbohydrate and protein substances, should be isolated from natural sea spray and have their
ice nucleation properties tested. Further data will help develop an understanding of the role SSA
play in the larger aerosol-cloud-climate system. In addition, field research and further laboratory
studies will establish how the nucleation behaviors of SSA may change as a result of the marine
impacts of Earth's changing climate.
Acknowledgements
I would like to thank my advisor Dr. Dan Cziczo for his support as a research supervisor.
I would also like to thank Martin Wolf for his daily supervision and joy in sharing knowledge.
Additionally, I would like to thank Jane Connor for her assistance in writing and preparing
this report and its accompanying presentation. Finally, thank you to all of the members of
the Cziczo Group and the MIT Department of Earth, Atmospheric, and Planetary Science for
their encouragement and assistance throughout my undergraduate career. The MIT David
P. Bacon Fund provided financial support for this project. The culturing of Prochlorococcus
by the Chisholm Lab was supported by a grant from the National Science Foundation (OCE-
1356460).
18
6 References
Adler, Gabriela, Thomas Koop, Carynelisa Haspel, Ilya Taraniuk, Tamar Moise, Ilan Koren,
Reuven H. Heiblum, and Yinon Rudich. 2013. "Formation of Highly Porous Aerosol Parti-
cles by Atmospheric Freeze-Drying in Ice Clouds." Proceedings of the National Academy of
Sciences 110 (51): 20414-19. https://doi.org/10.1073/pnas.1317209110.
Bigg, E. Keith, and Caroline Leck. 2008. "The Composition of Fragments of Bubbles Bursting
at the Ocean Surface." Journal of Geophysical Research: Atmospheres 113 (D11): D11209.
https://doi.org/10.1029/2007JD009078.
Blanchard, Duncan C. 1963. "The Electrification of the Atmosphere by Particles from Bubbles
in the Sea." Progress in Oceanography 1 (January): 73-202. https://doi.org/10.1016/0079-
6611(63)90004-1.
Burrows, S. M., C. Hoose, U. P6schl, and M. G. Lawrence. 2013. "Ice Nuclei in Marine Air: Bio-
genic Particles or Dust?" Atmos. Chem. Phys. 13 (1): 245-67. https://doi.org/10.5194/acp-
13-245-2013.
Carrio, G. G., S. C. van den Heever, and W. R. Cotton. 2007. "Impacts of Nucleating Aerosol
on Anvil-Cirrus Clouds: A Modeling Study." Atmospheric Research 84 (2): 111-31.
https://doi.org/10.1016/j.atmosres.2006.06.002.
Chen, Y. L., P. J. DeMott, S. M. Kreidenweis, D. C. Rogers, and D. E. Sherman. 2000. "Ice
Formation by Sulfate and Sulfuric Acid Aerosol Particles under Upper-Tropospheric Condi-
tions." Journal of the Atmospheric Sciences 57 (22): 3752-66. https://doi.org/10.1175/1520-
0469(2000)057<3752:IFBSAS>2.0.CO;2.
Cochran, Richard E., Olga Laskina, Jonathan V. Trueblood, Armando D. Estillore, Holly S.
Morris, Thilina Jayarathne, Camille M. Sultana, et al. 2017. "Molecular Diversity of Sea
Spray Aerosol Particles: Impact of Ocean Biology on Particle Composition and Hygroscop-
icity." Chem 2 (5): 655-67. https://doi.org/10.1016/j.chempr.2017.03.007.
Collins, D. B., D. F. Zhao, M. J. Ruppel, 0. Laskina, J. R. Grandquist, R. L. Modini, M. D.
Stokes, et al. 2014. "Direct Aerosol Chemical Composition Measurements to Evaluate the
Physicochemical Differences between Controlled Sea Spray Aerosol Generation Schemes."
Atmos. Meas. Tech. 7 (11): 3667-83. https://doi.org/10.5194/amt-7-3667-2014.
Cziczo, D. J., P. J. DeMott, C. Brock, P. K. Hudson, B. Jesse, S. M. Kreidenweis, A. J. Prenni,
J. Schreiner, D. S. Thomson, and D. M. Murphy. 2003. "A Method for Single Particle Mass
Spectrometry of Ice Nuclei." Aerosol Science and Technology 37 (5): 460-70.
https://doi.org/10.1080/02786820300976.
Cziczo, Daniel J., Karl D. Froyd, Corinna Hoose, Eric J. Jensen, Minghui Diao, Mark A. Zondlo,
Jessica B. Smith, Cynthia H. Twohy, and Daniel M. Murphy. 2013. "Clarifying the Dom-
inant Sources and Mechanisms of Cirrus Cloud Formation." Science 340 (6138): 1320-24.
https://doi.org/10.1126/science.1234145.
Cziczo, Daniel J., David S. Thomson, Thomas L. Thompson, Paul J. DeMott, and Daniel M.
Murphy. 2006. "Particle Analysis by Laser Mass Spectrometry (PALMS) Studies of Ice Nu-
clei and Other Low Number Density Particles." International Journal of Mass Spectrometry,
19
Aerosols/ Microparticles Special Issue, 258 (1): 21-29. https://doi.org/10.1016/j.ijms.2006.05.013.
DeMott, P. J., D. J. Cziczo, A. J. Prenni, D. M. Murphy, S. M. Kreidenweis, D. S. Thomson,
R. Borys, and D. C. Rogers. 2003. "Measurements of the Concentration and Composition
of Nuclei for Cirrus Formation." Proceedings of the National Academy of Sciences of the
United States of America 100 (25): 14655-60. https://doi.org/10.1073/pnas.2532677100.
DeMott, Paul J., Thomas C. J. Hill, Christina S. McCluskey, Kimberly A. Prather, Douglas B.
Collins, Ryan C. Sullivan, Matthew J. Ruppel, et al. 2016. "Sea Spray Aerosol as a Unique
Source of Ice Nucleating Particles." Proceedings of the National Academy of Sciences 113
(21): 5797-5803. https://doi.org/10.1073/pnas.1514034112.
Deng, Chunhua, Sarah D. Brooks, German Vidaurre, and Daniel C. 0. Thornton. 2014. "Us-
ing Raman Microspectroscopy to Determine Chemical Composition and Mixing State of
Airborne Marine Aerosols over the Pacific Ocean." Aerosol Science and Technology 48 (2):
193-206. https://doi.org/10.1080/02786826.2013.867297.
Erickson, David J., and Robert A. Duce. 1988. "On the Global Flux of Atmospheric Sea Salt."
Journal of Geophysical Research: Oceans 93 (C11): 14079-88.
https://doi.org/10.1029/JC093iC11p14079.
Facchini, Maria Cristina, Matteo Rinaldi, Stefano Decesari, Claudio Carbone, Emanuela Finessi,
Mihaela Mircea, Sandro Fuzzi, et al. 2008. "Primary Submicron Marine Aerosol Dominated
by Insoluble Organic Colloids and Aggregates." Geophysical Research Letters 35 (17).
https://doi.org/10.1029/2008GL034210.
Feingold, G., W. R. Cotton, S. M. Kreidenweis, and J. T. Davis. 1999. "The Impact of Gi-
ant Cloud Condensation Nuclei on Drizzle Formation in Stratocumulus: Implications for
Cloud Radiative Properties." Journal of the Atmospheric Sciences 56 (24): 4100-4117.
https://doi.org/10.1175/'1520-0469(1999)056<4100:TIOGCC>2.0.CO;2.
Flombaum, Pedro, Jos6 L. Gallegos, Rodolfo A. Gordillo, Jos6 Rinc6n, Lina L. Zabala, Nianzhi
Jiao, David M. Karl, et al. 2013. "Present and Future Global Distributions of the Marine
Cyanobacteria Prochlorococcus and Synechococcus." Proceedings of the National Academy
of Sciences 110 (24): 9824-29. https://doi.org/10.1073/pnas.1307701110.
Galgani, Luisa, Christian Stolle, Sonja Endres, Kai G. Schulz, and Anja Engel. 2014. "Effects
of Ocean Acidification on the Biogenic Composition of the Sea-surface Microlayer: Results
from a Mesocosm Study." Journal of Geophysical Research: Oceans 119 (11): 7911-24.
https://doi.org/10.1002/2014JC010188.
Garimella, Sarvesh, Thomas Bjerring Kristensen, Karolina Ignatius, Andre Welti, Jens Voigt-
laender, Gourihar R. Kulkarni, Frank Sagan, et al. 2016. "The SPectrometer for Ice Nuclei
(SPIN): An Instrument to Investigate Ice Nucleation." Atmospheric Measurement Tech-
niques 9 (7): 2781-95. https://doi.org/10.5194/amt-9-2781-2016.
Garimella, Sarvesh, Daniel A. Rothenberg, Martin J. Wolf, Robert 0. David, Zamin A. Kanji,
Chien Wang, Michael Rosch, and Daniel J. Cziczo. 2017. "Uncertainty in Counting Ice
Nucleating Particles with Continuous Flow Diffusion Chambers." Atmospheric Chemistry
and Physics 17 (17): 10855-64. https://doi.org/10.5194/acp-17-10855-2017.
20
Graether, Steffen P., and Zongchao Jia. 2001. "Modeling Pseudomonas Syringae Ice-Nucleation
Protein as A-Helical Protein." Biophysical Journal 80 (3): 1169-73.
https://doi.org/10.1016/S0006-3495(01)76093-6.
Haywood, J., and 0. Boucher. 2000. "Estimates of the Direct and Indirect Radiative Forcing
Due to Tropospheric Aerosols: A Review." Reviews of Geophysics 38 (4): 513-43.
https://doi.org/10.1029/1999RG000078.
Hoose, C., and 0. M6hler. 2012. "Heterogeneous Ice Nucleation on Atmospheric Aerosols: A
Review of Results from Laboratory Experiments." Atmos. Chem. Phys. 12 (20): 9817-54.
https://doi.org/10.5194/acp-12-9817-2012.
IPCC, 2014: Climate Change 2014: Synthesis Report. Contribution of Working Groups 1, 11
and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change
[Core Writing Team, R.K. Pachauri and L.A. Meyer (eds.)]. IPCC, Geneva, Switzerland,
151 pp.
Junge, K., and B. D. Swanson. 2008. "High-Resolution Ice Nucleation Spectra of Sea-Ice Bac-
teria: Implications for Cloud Formation and Life in Frozen Environments." Biogeosciences
5 (3): 865-73. https://doi.org/10.5194/bg-5-865-2008.
Ke, Wei-Ren, Yu-Mei Kuo, Chih-Wei Lin, Sheng-Hsiu Huang, and Chih-Chieh Chen. 2017.
"Characterization of Aerosol Emissions from Single Bubble Bursting." Journal of Aerosol
Science 109 (July): 1-12. https://doi.org/10.1016/j.jaerosci.2017.03.006.
Kong, Xiangrui, Martin J. Wolf, Michael Roesch, Erik S. Thomson, Thorsten Bartels-Rausch,
Peter A. Alpert, Markus Ammann, Nonne L. Prisle, and Daniel J. Cziczo. 2018. "A Contin-
uous Flow Diffusion Chamber Study of Sea Salt Particles Acting as Cloud Nuclei: Deliques-
cence and Ice Nucleation." Tellus B: Chemical and Physical Meteorology 70 (1): 1463806.
https://doi.org/10.1080/16000889.2018.1463806.
Ladino, L. A., J. D. Yakobi-Hancock, W. P. Kilthau, R. H. Mason, M. Si, J. Li, L. A. Miller,
et al. 2016. "Addressing the Ice Nucleating Abilities of Marine Aerosol: A Combination
of Deposition Mode Laboratory and Field Measurements." Atmospheric Environment 132
(May): 1-10. https://doi.org/10.1016/j.atmosenv.2016.02.028.
Lesins Glen, Petr Chylek, and Ulrike Lohmann. 2002. "A Study of Internal and External Mix-
ing Scenarios and Its Effect on Aerosol Optical Properties and Direct Radiative Forcing."
Journal of Geophysical Research: Atmospheres 107 (D10): AAC 5-1.
https://doi.org/10.1029/2001JD000973.
Lewis, E. R., W.J. Wiscombe, B. A. Albrecht; C. N. Flagg; R. M. Reynolds; A. P. Siebesma.
2012. "MAGIC: Marine ARM GPCI Investigation of Clouds." DOE/SC-ARM-12-020, U.S.
Department of Energy, 12 pp.
Moore Lisa R., Coe Allison, Erik R. Zinser, Mak A. Saito, Matthew B. Sullivan, Debbie Lindell,
Katya Frois-Moniz, John Waterbury, and Sallie W. Chisholm. 2007. "Culturing the Marine
Cyanobacterium Prochlorococcus." Limnology and Oceanography: Methods 5 (10): 353-62.
https://doi.org/10.4319/lom.2007.5.353.
Pandey, Ravindra, Kota Usui, Ruth A. Livingstone, Sean A. Fischer, Jim Pfaendtner, Ellen H. G.
21
Backus, Yuki Nagata, et al. 2016. "Ice-Nucleating Bacteria Control the Order and Dynamics
of Interfacial Water." Science Advances 2 (4): e1501630. https://doi.org/ 10.1126/sciadv.1501630.
Prather, Kimberly A., Timothy H. Bertram, Vicki H. Grassian, Grant B. Deane, M. Dale Stokes,
Paul J. DeMott, Lihini I. Aluwihare, et al. 2013. "Bringing the Ocean into the Laboratory to
Probe the Chemical Complexity of Sea Spray Aerosol." Proceedings of the National Academy
of Sciences 110 (19): 7550-55. https://doi.org/10.1073/pnas.1300262110.
Pruppacher, Hans R, and James D Klett. 2011. Microphysics of Clouds and Precipitation.
Dordrecht; Heidelberg; London: Springer.
Pummer, B. G., C. Budke, S. Augustin-Bauditz, D. Niedermeier, L. Felgitsch, C. J. Kampf, R.
G. Huber, et al. 2015. "Ice Nucleation by Water-Soluble Macromolecules." Atmos. Chem.
Phys. 15 (8): 4077-91. https://doi.org/10.5194/acp-15-4077-2015.
Rogers, David C. 1988. "Development of a Continuous Flow Thermal Gradient Diffusion Cham-
ber for Ice Nucleation Studies." Atmospheric Research 22 (2): 149-81.
https://doi.org/10.1016/0169-8095(88)90005-1.
Salter, M. E., E. Hamacher-Barth, C. Leck, J. Werner, C. M. Johnson, I. Riipinen, E. D. Nils-
son, and P. Zieger. 2016. "Calcium Enrichment in Sea Spray Aerosol Particles." Geophysical
Research Letters 43 (15): 8277-85. https://doi.org/10.1002/2016GL070275.
Schill, Gregory P., and Margaret A. Tolbert. 2014. "Heterogeneous Ice Nucleation on Simulated
Sea-Spray Aerosol Using Raman Microscopy." The Journal of Physical Chemistry C 118
(50): 29234-41. https://doi.org/10.1021/jp505379j.
Seinfeld, John H., and Spyros N. Pandis. 2006. Atmospheric Chemistry and Physics: From Air
Pollution to Climate Change. 2 edition. Hoboken, N.J: Wiley-Interscience.
Stokes, M. D., G. B. Deane, K. Prather, T. H. Bertram, M. J. Ruppel, 0. S. Ryder, J. M.
Brady, and D. Zhao. 2013. "A Marine Aerosol Reference Tank System as a Breaking Wave
Analogue for the Production of Foam and Sea-Spray Aerosols." Atmos. Meas. Tech. 6 (4):
1085-94. https://doi.org/10.5194/amt-6-1085-2013.
Vali, G., P. J. DeMott, 0. Moehler, and T. F. Whale. 2015. "Technical Note: A Proposal
for Ice Nucleation Terminology." Atmospheric Chemistry and Physics 15 (18): 10263-70.
https://doi.org/10.5194/acp-15-10263-2015.
Vergara-Temprado, J., B. J. Murray, T. W. Wilson, D. O'Sullivan, J. Browse, K. J. Pringle,
K. Ardon-Dryer, et al. 2017. "Contribution of Feldspar and Marine Organic Aerosols to
Global Ice Nucleating Particle Concentrations." Atmos. Chem. Phys. 17 (5): 3637-58.
https://doi.org/10.5194/acp-17-3637-2017.
Vignati, E., M. C. Facchini, M. Rinaldi, C. Scannell, D. Ceburnis, J. Sciare, M. Kanakidou, S.
Myriokefalitakis, F. Dentener, and C. D. O'Dowd. 2010. "Global Scale Emission and Distri-
bution of Sea-Spray Aerosol: Sea-Salt and Organic Enrichment." Atmospheric Environment
44 (5): 670-77. https://doi.org/10.1016/j.atmosenv.2009.11.013.
Wang, Xiaofei, Grant B. Deane, Kathryn A. Moore, Olivia S. Ryder, M. Dale Stokes, Charlotte
M. Beall, Douglas B. Collins, et al. 2017. "The Role of Jet and Film Drops in Controlling
22
the Mixing State of Submicron Sea Spray Aerosol Particles." Proceedings of the National
Academy of Sciences 114 (27): 6978-83. https://doi.org/10.1073/pnas.1702420114.
Wang, Xiaofei, Camille M. Sultana, Jonathan Trueblood, Thomas C. J. Hill, Francesca Mal-
fatti, Christopher Lee, Olga Laskina, et al. 2015. "Microbial Control of Sea Spray Aerosol
Composition: A Tale of Two Blooms." ACS Central Science 1 (3): 124-31.
https://doi.org/10.1021/acscentsci.5b00148.
Wilson, Theodore W., Luis A. Ladino, Peter A. Alpert, Mark N. Breckels, Ian M. Brooks, Jo
Browse, Susannah M. Burrows, et al. 2015. "A Marine Biogenic Source of Atmospheric
Ice-Nucleating Particles." Nature 525 (7568): 234. https://doi.org/10.1038/naturel4986.
Wolf, Martin J., Allison Coe, Lilian A. Dove, Maria A. Zawadowicz, Keven Dooley, Steven J.
Biller, Yue Zhang, Sallie W. Chisholm, Daniel J. Cziczo. "Investigating the Heterogeneous
Ice Nucleation Potential of Sea Spray Aerosols Using Prochlorococcus as a Model Source of
Marine Organic Matter." Submitted to Environmental Science and Technology.
Wurl, Oliver, Werner Ekau, William M. Landing, and Christopher J. Zappa. 2017. "Sea Surface
Microlayer in a Changing Ocean - A Perspective." Elem Sci Anth 5 (0).
https://doi.org/10.1525/elementa.228.
Zhuang, Y., E.P. Lozowski, J.D. Wilson, and G. Bird. 1993. "Sea Spray Dispersion Over the
Ocean Surface - a Numerical-Simulation." Journal of Geophysical Research-Oceans 98 (C9):
16547-53. https://doi.org/10.1029/93JC01406.
23
Appendix A
S- 
-
FLOW~ Ii
S5AmW FLOW
Air and water flow diagram of the SPIN chamber. Adapted from Garimella et al. (2016).
24
VALW
SAMPLE
INPUT
1tif10R I~ LP2I
AWtE OM I I LPN
WU LPU
LL 2 C"U
A20MP
IN CHAMBER
1 . PARTCLIAIR SA
+ 
4- amnM
4- mi p-m
I I ONE
DEYEC lOP
CC"" WAN
lot2
zC"
iFlI
SWRAIO
|1G AU Fnaw e
Appendix B
Parlcie Inlet Ecmu Sewn YAG Tiftng
5C Bear) Sp4ttcr
Twning
Bean PUT
bI E~dmactlon
PWU I scotw
ollection Optcs YAG Scater Gnd DetectorSc~ Seem
scam Beame
Bam" PUT Alodvv Plate Detector
Poecor-SwnsyvDetector /
II
YAG Monitor
Photodiode
/
YAG Laser
Excirer Laser
Knie Edge Proller Excner Attenumon Fltels
Diagram of the internal workings of the PALMS instrument. Adapted from Cziczo et al. (2006).
25
IIIIIIIII