Probing Heterogeneous Efflorescence in an Optical Levitator for Earth and Mars Systems

by Shuichi Ushijima B.S., Colorado School of Mines, 2013

A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Doctor of Philosophy Chemistry 2020

Committee Members: Margaret Tolbert Raina Gough Joost de Gouw Eleanor Brown Ryan Davis Ushijima, Shuichi (Ph. D., Chemistry)

Thesis directed by Distinguished Professor Margaret A. Tolbert

ABSTRACT

The phase transition of soluble inorganic salts has great importance to both Earth’s atmosphere and the Martian surface. On Earth, the phase state of an aerosol will affect their impact on the climate. On Mars, the formation and stability of liquid brines could be key in understanding present day geological features and potentially life on Mars. While the homogeneous phase transitions have been studied in detail, heterogeneous efflorescence has not. Heterogeneous efflorescence occurs when a heterogeneous nucleus induces efflorescence, or crystallization, of an aqueous salt droplet at the surface upon a collision, or from within the droplet after it has become immersed inside the droplet. Heterogeneous efflorescence can change our current understanding of how many aerosols in Earth’s atmosphere are solid particles, and the time that a brine is stable for on the Martian surface.

In this study, heterogeneous efflorescence of salt aerosols in Earth’s atmosphere by mineral dust particles were probed. Specifically, the effect of three mineral particles, illite, Na- montmorillonite, and NX Illite on the efflorescence of ammonium sulfate and sodium chloride was studied. The study showed that mixing of salt and mineral aerosols can occur quite often and some salt and mineral dust pairing caused efflorescence to occur under significantly more humid conditions than without the mineral dust. The effect that organic molecules, such as secondary organic aerosols, have on the efflorescence behavior of atmospheric aerosols was also investigated.

Two organics, polyethylene glycol 400 and raffinose were mixed with ammonium sulfate for the study. Polyethylene glycol 400 phase separates from ammonium sulfate and coats the aqueous

ii droplet whereas raffinose stays homogenously mixed. The study showed that the effect of organics on efflorescence seems to be strongly linked to how the organic affects the mixed aerosols viscosity. Finally, heterogeneous efflorescence of brines believed to be on Mars were investigated.

While some brines were completely unaffected by the presence of Martian dust analogues, the stability of other brines were drastically reduced when exposed to mineral particles.

iii

To Mom, Dad, Miki, Aya, and Shiho.

Thank you for always supporting me.

ACKNOWLEDGEMENTS

I would like to first thank my advisor Maggie Tolbert for giving me the opportunity to work with such a unique instrument. When I first met with Maggie, I had no clue what optical levitation was nor did I have any idea what heterogeneous efflorescence was. But Maggie gave me a chance to work with the optical levitator and through it I have worked on projects that span our solar system. I truly appreciate all the guidance that you have given me during my time in your lab.

I would also like to Ryan Davis who taught me everything about the optical levitator. With

Ryan’s help I learned how optical levitation works, how to run experiments and maintain the optical levitator. Also, to all the past and current members of the Tolbert Research Group I thank you for all the times you helped me out in lab, the times you listened to my practice talks, and for just being a great group of people.

Finally, I would like to thank my family. Thank you, mom, for always supporting me.

Without your support I would not have made it here. For my three amazing sisters who have and always will inspire me to be better and work harder. And to my dad who I know would be proud of me if he were here today.

This work was supported by a National Science Foundation grant, a NASA Earth and Space

Science Fellowship, and a Cooperative Institute for Research in Environmental Sciences

Fellowship.

TABLE OF CONTENTS

Chapter 1. Introduction ...... 1

1.1. Research Overview ...... 1

1.2. Homogeneous Efflorescence and Deliquescence ...... 1

1.3. Heterogeneous Nucleation: Contact and Immersion Mode ...... 3

1.4. Relevance to Earth ...... 5

1.5. Relevance to Mars...... 9

1.6. Thesis Focus...... 12

Chapter 2: Immersion and Contact Efflorescence Induced by Mineral Dust Particles...... 14

2.1. Introduction ...... 14

2.2. Experimentation ...... 16

2.2.1. Preparation of Aqueous Solution Droplets ...... 16

2.2.2. Preparation of Mineral Dust Particles ...... 16

2.2.3. Experimental Arrangement ...... 17

2.2.4. Imaging Efflorescence and Collisions ...... 20

2.2.5. Determining Heterogeneous ERH ...... 22

2.3. Results ...... 23

2.3.1. Heterogeneous Efflorescence by Mineral Dust ...... 23

2.3.2. Immersion ERH as a Function of Surface Area of Immersed Illite ...... 28

2.4. Discussion ...... 30

2.4.1. Crystal Lattice Match ...... 30

2.4.2. Active Site ...... 33

2.4.3. Ion-Specific Effects ...... 33

2.4.4. Collisions Lifetimes of Salt Aerosol in a Dust Plume ...... 36

2.5. Conclusion ...... 39

Chapter 3: Contact Efflorescence of Internally Mixed Organic – Ammonium Sulfate

Aerosols: A Glassy Organic vs An Organic Coating ...... 41

3.1. Introduction ...... 41

3.2. Experimentation ...... 44

3.2.1. Preparation of Droplets for Trapping ...... 44

3.2.2. Experimental Setup ...... 44

3.2.3. Observing Efflorescence ...... 47

3.2.4. Conducting Contact Efflorescence Experiments ...... 48

3.2.5. Measuring Viscosity and Immersion Times ...... 48

3.3. Results ...... 51

3.3.1. Raman Signal for the Components ...... 51

3.3.2. Dehumidification of Mixed (NH4)2SO4:Organic Droplets ...... 51

3.3.3. Contact Efflorescence of Mixed (NH4)2SO4:Organic Droplets ...... 56

3.4. Discussion ...... 60

3.4.1. PEG-400: Coating Thickness and Diffusion Times ...... 60

vii

3.4.2. Raffinose: Viscosity and Diffusion ...... 68

3.5. Conclusion ...... 72

Chapter 4: Probing Heterogeneous Efflorescence of Mars Relevant Salts with an Optical

Levitator...... 74

4.1. Introduction ...... 74

4.1.1. Background and Choice of System for Study ...... 76

4.2. Experimentation ...... 78

4.2.1. Optical Levitation and Flow Cell Arrangement ...... 78

4.2.2. Generating Droplets of Aqueous Solutions ...... 80

4.2.3. Generating Contact Nuclei ...... 80

4.2.4. Imaging Efflorescence and Collisions ...... 81

4.2.5. Determining Contact and Immersion Heterogeneous ERH ...... 85

4.3. Results ...... 86

4.3.1. Homogeneous Efflorescence of Brines ...... 86

4.3.2. Heterogeneous Efflorescence of Brines ...... 87

4.4. Discussion ...... 95

4.4.1. Heterogeneous Efflorescence on Mars ...... 95

4.4.2. Crystal Lattice Match ...... 97

4.5. Conclusion ...... 101

Chapter 5: Summary of Conclusions ...... 103

viii

Chapter 6: Bibliography ...... 107

Chapter 7: Appendix ...... 116

Tables

2- Table 3.1. Diffusion times of H2O and SO4 in 1:1 (NH4)2SO4:raffinose at two RH ...... 71

Table 4.1. Crystal lattice mismatch of select brines and minerals ...... 100

Table A1. Crystal lattice constants and systems for all salts and minerals analyzed ...... 118

Figures

Figure 1.1. Hysteresis of efflorescence and deliquescence ...... 2

Figure 1.2. Homogeneous and heterogeneous efflorescence ...... 4

Figure 1.3. Climate impacts of liquids and crystals ...... 7

Figure 1.4. TEM images of internally mixed aerosol particles ...... 8

Figure 1.5. Image of RSL on the walls of Garni Crater ...... 11

Figure 2.1. Experimental setup of the optical levitator ...... 18

Figure 2.2. Diagram comparing efflorescence of a droplet upon contact with a heterogeneous nucleus and the immersion of the particle into the droplet ...... 21

Figure 2.3. Contact efflorescence experiment results for (NH4)2SO4 and NaCl by mineral dust particles ...... 24

Figure 2.4. Contact efflorescence experiment results for MgCl2 and NH4Cl by montmorillonite ...... 26

Figure 2.5. Summary of results for heterogeneous efflorescence of (NH4)2SO4 and NaCl by mineral dust particles ...... 27

Figure 2.6. Heterogeneous ERH of (NH4)2SO4 vs the calculated total surface area of immersed illite ...... 29

Figure 2.7. Heterogeneous σERH of (NH4)2SO4 and NaCl as a function of crystal lattice mismatch(δ) ...... 32

Figure 2.8. Comparing contact σERH of various chlorides by Na-montmorillonite ...... 35

Figure 2.9. Calculated collisions lifetimes for a single salt aerosol to collide with a mineral dust particle in a plume ...... 38

Figure 3.1. Arrangement of CCD cameras, LED, and optical components to record far field and bright field imaging and conduct Raman spectrometry ...... 46

Figure 3.2. The dual-balance linear quadrupole electrodynamic balance (DBQ-EDB) ...... 49

Figure 3.3. Raman spectrum of the mixed (NH4)2SO4:organic droplets and of the individual components ...... 52

Figure 3.4. Raman spectrum and scattering images for dehumidification of 1:1

(NH4)2SO4:PEG-400 ...... 53

Figure 3.5. Raman spectrum and scattering images for dehumidification of 1:1

(NH4)2SO4:raffinose ...... 55

Figure 3.6. Defect intensity with elapsed time for (NH4)2SO4:PEG-400 efflorescence at

34.9% RH and for (NH4)2SO4:raffinose at 4.7% RH ...... 57

Figure 3.7. Defect intensity with time elapsed for three contact efflorescence experiments ...59

Figure 3.8. Summary of results for contact efflorescence of 1:1 (NH4)2SO4:organic by crystals of (NH4)2SO4 ...... 61

Figure 3.9. Calculated thickness of PEG-400 and diffusion times for a molecule of H2O and a heterogeneous nucleus that is 300nm in size ...... 62

Figure 3.10. Viscosity of PEG-400 and calculated diffusion coefficient of H2O and a 300 nm sized heterogeneous nucleus as a function of RH ...... 67

xii

2- Figure 3.11. Viscosities of (NH4)2SO4:raffinose and diffusion coefficients for H2O and SO4 as a function of RH ...... 69

Figure 4.1. Far field and bright field image of Mg(ClO4)2 liquid and crystal ...... 82

Figure 4.2. Sequence of images for contact efflorescence experiment of Mg(ClO4)2 by a crystal of Mg(ClO4)2 and NaCl ...... 84

Figure 4.3. Homogeneous dehydration and hydration of Mg(ClO4)2, Ca(ClO4)2, MgCl2, and

CaCl2 ...... 88

Figure 4.4. Contact efflorescence experiment results for Mg(ClO4)2 ...... 90

Figure 4.5. Summary of heterogeneous efflorescence of Mg(ClO4)2 ...... 91

Figure 4.6. Summary of heterogeneous efflorescence of MgCl2, Ca(ClO4)2, and CaCl2 by montmorillonite ...... 93

Figure 4.7. Contact efflorescence experiment results for Ca(ClO4)2 and CaCl2 by montmorillonite ...... 94

Figure 4.8. Contact ERH of Mg(ClO4)2 vs crystal lattice mismatch ...... 99

Figure A1. Particle size distribution for heterogeneous nuclei made by the nebulizer ...... 116

Figure A2. Viscosity of pure raffinose and raffinose:(NH4)2SO4 vs RH ...... 117

xiii

CHAPTER 1: INTRODUCTION

1.1: Research Overview

The focus of the research presented in this thesis was to investigate heterogeneous efflorescence of inorganic salts relevant to Earth’s atmosphere and Mars’ surface and subsurface in a laboratory setting. For Earth’s atmosphere the results from this thesis can affect our current understanding on the phase state of aerosols in the atmosphere which in turn will help to better understand the climate system to improve climate models. On Mars, the results further our understanding of how water interacts with the Martian regolith which could help explain current day geologic features. Additionally, water on present day Mars is of interest as life as we know it is closely linked to liquid water. The remainder of this chapter will further discuss efflorescence and deliquescence as well as heterogeneous nucleation, and how these processes are relevant to

Earth and Mars.

1.2: Homogeneous Efflorescence and Deliquescence

Soluble inorganic salts can undergo transitions between two phases, aqueous liquid and crystalline solid, as relative humidity (RH) is changed. A schematic of the two transitions along a

RH axis is shown in Figure 1.1. When the RH is high enough, the crystalline salt will absorb water vapor from the surrounding environment and form an aqueous droplet in a process termed deliquescence at a characteristic relative humidity, the deliquescence relative humidity (DRH).

The reverse process, where the dried aqueous droplet releases water and crystallizes to a solid, is termed efflorescence, occurring at the efflorescence relative humidity (ERH). Typically, the DRH is higher than the ERH resulting in a region between the two RH values where an aqueous solution is metastable. The hysteresis between DRH and ERH is due to a nucleation barrier that kinetically hinder homogeneous efflorescence. Due to the kinetic barrier to nucleation the ERH must be

Figure 1.1: Hysteresis of efflorescence and deliquescence and the metastable region in between

ERH and DRH.

determined experimentally (Martin, 2000; Seinfeld & Pandis, 2006) whereas the DRH of a salt can be predicted thermodynamically.

1.3: Heterogeneous Nucleation: Contact and Immersion Mode

While the homogeneous efflorescence and deliquescence for many salts have been examined in detail (Cziczo & Abbatt, 1999; Gough et al., 2011; Martin, 2000; Nuding et al., 2014), the effect of a heterogeneous nucleus on efflorescence is not well established. In heterogeneous nucleation, the presence of a foreign nucleus can lower the energy barrier needed for a crystal seed to form, thus aiding the crystallization process. For heterogeneous efflorescence, the effect manifests as a decrease in the stability of the brine through the increase in the ERH. Heterogeneous efflorescence occurs via two distinct pathways, immersion mode and contact mode as shown in

Figure 1.2. During immersion efflorescence, the heterogeneous nucleus induces crystallization from within the droplet. Contact efflorescence occurs when the nucleus collides with the surface and induces crystallization from the outside. The RH at the time when the heterogeneous nucleus and the liquid droplet come together will drive the pathway of efflorescence. If the RH is lower than the contact ERH, contact efflorescence will occur. However, if the RH is higher than the contact ERH, a collision between the liquid aerosol and insoluble particle will not induce effloresce and the heterogeneous nucleus will become immersed. The droplet with an immersed particle will remain liquid until the RH becomes low enough for immersion efflorescence to occur.

Contact and Immersion freezing of ice has been studied in detail (Broadley et al., 2012;

Durant & Shaw, 2005; Hoffmann et al., 2013; Kiselev et al., 2017; Murray et al., 2012; Niehaus et al., 2014; Niehaus & Cantrell, 2015; Shaw et al., 2005), but much less is known about heterogeneous efflorescence. Overall, very few studies have looked at heterogeneous

Figure 1.2: Homogeneous efflorescence and the two pathways for heterogeneous efflorescence, immersion and contact. The effect of heterogeneous nucleus on the energy barrier that kinetically hinders crystallization.

4 efflorescence (Davis et al., 2015a, 2015b; Davis & Tolbert, 2017; Han et al., 2002; Han & Martin,

1999; Martin et al.,2001; Pant et al., 2006; Primm et al., 2018; Ushijima et al., 2018) and the first study to report on contact efflorescence was a study from Davis et al. (2015). The study focused on heterogeneous efflorescence of inorganic salts such as ammonium sulfate ((NH4)2SO4), sodium chloride (NaCl), and ammonium nitrate (NH4NO3) by contact with other soluble inorganic crystals.

The effect of heterogeneous efflorescence has yet to be fully understood and there is no comprehensive model that can explain all the results. Additionally, laboratory experiments have shown that for heterogeneous nucleation, the same nucleus can be more effective at inducing nucleation when it is in contact with the surface of the droplet rather than when it has become immersed inside the droplet (Davis & Tolbert, 2017; Ladino Moreno et al., 2013; Ushijima et al.,

2018). The difference has not yet been fully explained theoretically further complicating the narrative in understanding heterogeneous nucleation.

1.4: Relevance to Earth:

In Earth’s atmosphere, soluble inorganic compounds have been detected globally in both terrestrial and marine environment (Boucher et al., 2013; Fitzgerald, 1991; Zhang et al., 2007).

Above land, anthropogenic emissions of SOx and NOx from burning fossil fuel and NH3 from agricultural emissions act as precursors for these inorganics. When these gasses mix, they combine to form secondary aerosols such as ammonium sulfate ((NH4)2SO4) and ammonium nitrate

(NH4NO3). Above marine environments, Sea Spray Aerosols (SSA), whose primary inorganic component is sodium chloride (NaCl), are the dominant aerosol species. SSA can be emitted from the ocean surface through wave crashing actions that then lead to bubble bursting at the surface which emits film and jet droplets with film droplets typically being smaller of the two (Fitzgerald,

1991).

In addition to the composition, size, and concentration of aerosols, their phase state is another key factor to determining their effect on climate. Figure 1.3 shows several ways how the phase state of aerosols in Earth’s atmosphere alter the aerosol’s impact on global and regional climate (Andreae & Rosenfeld, 2008; Boucher et al., 2013; Ramanathan et al., 2001; Seinfeld &

Pandis, 2006). Liquid aerosols change size when the RH changes by taking up and losing water.

The size of the aerosol affects how it interacts with solar radiation, thus affecting climate. Liquid aerosols also act as surfaces for heterogeneous chemistry, impacting the transformation of atmospheric emissions. The chemical reactions can affect the sinks and reservoirs of species such as NOx shown in the figure and will affect their transportation. Finally, liquid droplets and crystal solids nucleate ice differently. A solid could act as a depositional ice nucleus, often causing ice formation at warmer temperatures when compared to the homogeneous freezing of the liquid droplets. The cloud type and conditions necessary for cloud formation is affected based on the phase state (Andreae & Rosenfeld, 2008). For example, ice nucleation on crystalline (NH4)2SO4 has been shown to be a viable pathway for the formation of cirrus clouds (Abbatt et al., 2006).

Cloud interactions of aerosols is of interest as clouds and aerosols contribute the largest uncertainty for Earth climate models (Boucher et al., 2013). Thus, to accurately model the influence of aerosols on climate, the conditions under which phase transitions occur must be established.

For soluble inorganic salts in the atmosphere, the homogeneous phase transitions are well known (Martin, 2000). However, in the atmosphere these aerosols will at times collide with other aerosols and become internally mixed particles. TEM images of captured aerosol particles, shown in Figure 1.4, show how salt aerosols can become mixed with other soluble inorganic salts or insoluble inorganic particles such as soot and silicate (Buseck & Pósfai, 1999; Li et al., 2003). In the process of forming mixed aerosols, the collisions between an aqueous salt and a solid

Figure 1.3: Examples of the impacts to climate for liquids and crystals. Liquids change size with humidity and participate in heterogeneous chemistry. Crystals can undergo depositional ice nucleation which often occurs at a warmer temperature.

Figure 1.4: TEM images of internally mixed aerosol particles. Images from Buseck et al., (1999)† and Li et al., (2003)‡.

heterogeneous nucleus could result in contact efflorescence. If the heterogeneous nucleus was insoluble and it did not cause contact efflorescence, the solid heterogeneous nucleus would then become immersed. The immersed particle could then induce efflorescence from within once the

RH has become low enough for immersion efflorescence. Additionally, secondary organic aerosols are also a major component of the atmosphere and aerosols can also be a mixture of both organic and inorganic components (Zhang et al., 2007). Organics that become mixed with the aqueous salt aerosols exhibit different behavior based on the organic’s composition. Some organics will become homogeneously mixed into the aqueous aerosol and can inhibit efflorescence

(Robinson et al., 2014). Others have been shown to phase separate from the aqueous fraction to create a coating that will have various morphologies such as core shell or partially engulfed

(Ciobanu et al., 2010; Freedman, 2017). All these various interactions between the different aerosol types need to be studied to further understand how these collisions and mixed states affect the phase state of inorganic salt aerosols.

1.5: Relevance to Mars:

Liquid water is currently the only biochemical solvent that is known to be used by life

(Cockell et al., 2016) and thus has been a key factor in determining the habitability of planets

(Cockell et al., 2016; Ehlmann et al., 2016). Determining whether Mars is currently sustaining life or had so in the past has always been an integral part of missions to Mars. One of the main missions for the most recent rover to leave for Mars, Perseverance, is to look for signs of life in Mars’ ancient rocks. Mars is widely believed to have had liquid water in its ancient past based on geomorphological features and the presence of clay minerals (Bibring et al., 2006; Ehlmann et al.,

2011; Mangold et al., 2004). While water ice and water vapor have been observed and measured, on current day Mars the instability of liquid water at low Martian pressures has made its detection

9 challenging. Although pure liquid water may be unstable, it has been hypothesized that liquid water may be present in the form of a brine (Gough et al., 2011, 2014; Martín-Torres et al., 2015;

McEwen et al., 2011; Nuding et al., 2015, 2014; Ojha et al., 2015; Primm et al., 2017). Landers on the surface and satellites orbiting Mars have shown that inorganic salts, such as perchlorates and chlorides, are present in the Martian regolith and are believed to be found globally (Clark &

Kounaves, 2016; Glavin et al., 2013; Hecht et al., 2009; Ming et al., 2014). Some of these salts are highly hygroscopic and can deliquesce to form briny solutions at low RH. Recently, the Mars

Advanced Radar for Subsurface and Ionosphere Sounding (MARSIS) has detected a subglacial pond which could potentially be comprised of brine (Orosei et al., 2018). Additionally, hydrates of salts that could potentially form brine may have been detected near recurring slope lineae (RSL) by the Compact Reconnaissance Imaging Spectrometer for Mars (CRISM) (Ojha et al., 2015).

RSL are geographic features that are characterized by dark streaks that slowly grow on steep slopes seasonally. An example image of an RSL found on the walls of the Garni Crater taken by the

High-Resolution Imaging Science Experiment (HiRISE) aboard the Mars Reconnaissance Orbiter

(MRO) is shown in Figure 1.5. The long dark streaks on the image are the RSL. While the formation mechanism of RSL have yet to be determined, brines have been considered as potentially having a role in RSL formation (Martín-Torres et al., 2015; McEwen et al., 2011; Ojha et al., 2015). To fully understand the role of brines on present-day Mars, it is essential to study how salts interact with water under Mars conditions.

Studies on Mars’ brines have mainly focused on the homogeneous phase transitions

(Gough et al., 2011, 2014; Nuding et al., 2015, 2014; Primm et al., 2017). However, for brines that form on the surface and in the subsurface of Mars, they will inevitably interact and become mixed with the Martian regolith. In the surface and subsurface, the soil particles that have mixed

Figure 1.5: Image of RSL on the walls of Garni Crater taken by the High-Resolution Imaging

Science Experiment camera aboard the Mars Reconnaissance Orbiter. Photo credit: NASA/JPL-

Caltech/Univ. of Arizona

into brines can act as heterogeneous nuclei for immersion efflorescence. On the surface where dust fall is common on Mars, the falling dust can act as contact nuclei, and induce efflorescence.

Thus, the effect of heterogeneous efflorescence on the stability of Mars relevant brines are necessary to further our understanding on the interaction of inorganic salts and water on Mars.

1.6: Thesis Focus

The focus of this thesis is to investigate the effect that heterogeneous efflorescence may have on the phase state of soluble salts. Through these laboratory studies, the aim is to better our understanding of processes in the Earth’s atmosphere that may influence how often a salt aerosol is in the crystalline phase. The results can be applied to climate models to more accurately depict how Earth’s aerosols interact with the climate. Additionally, the work looking at Mars’ surface and subsurface will investigate the effect that the Martian regolith has on the stability of brines believed to be on Mars. Knowing what minerals destabilize brines on Mars could help map regions on Mars with the highest probability for brine formation and stability which could aid future missions to Mars in search of water and life.

The next three chapters will cover three different studies using the optical levitator and will be presented in the following format. A brief introduction which will supplement this introduction with discussion of specific topics for that study. An experimental section that will describe the instruments and techniques used to conduct the study. Finally, the results from the study will be reported followed by a discussion of the results which will include their impacts to either Earth’s atmosphere or Mars’ surface.

Specifically, Chapters 2 and 3 will first cover Earth systems. Chapter 2 studies the effect that mineral dust particles have on the efflorescence of two atmospheric salts, (NH4)2SO4 and NaCl.

Contact and immersion efflorescence by mineral dust, NX Illite, Na-montmorillonite, and illite

12 will be discussed. Chapter 3 explores the effect of mixing two organics, Polyethylene glycol 400 and raffinose, into (NH4)2SO4 will have on its efflorescence behavior. Results from contact efflorescence of the mixed organic-inorganic droplets by crystalline (NH4)2SO4 will be shown and discussed. Chapter 4 will move onto a Mars system, where heterogeneous efflorescence of

Mg(ClO4)2 by a crystal of itself, NaCl, (NH4)2SO4, montmorillonite and Mojave Mars Simulant

(MMS) will be discussed. Additionally, heterogeneous efflorescence of MgCl2, Ca(ClO4)2, and

CaCl2 by montmorillonite was also probed. Finally, the thesis will end with a summary of conclusions in Chapter 5.

Chapter 2: Immersion and Contact Efflorescence Induced by Mineral Dust

Particles

2.1: Introduction

Heterogeneous nucleation has been studied in detail for the heterogeneous freezing of ice in both contact and immersion mode (Broadley et al., 2012; Durant & Shaw, 2005; Hoffmann et al., 2013; Kiselev et al., 2017; Murray et al., 2012; Niehaus et al., 2014; Niehaus & Cantrell, 2015;

Shaw et al., 2005). However, much less is known about heterogeneous efflorescence. A recent study from Davis et al. (2015b) was the first to report observations of contact efflorescence. That study examined the ERH of ammonium sulfate ((NH4)2SO4), sodium chloride (NaCl), and ammonium nitrate (NH4NO3) by contact with other soluble inorganic crystals demonstrating that soluble salt crystals can induce efflorescence upon contact at an RH that is significantly higher than the homogeneous ERH. In another study, Davis and Tolbert (2017) demonstrated that contact-mode efflorescence can be induced by collisions with charged amorphous organic aerosols that do not induce immersion-mode efflorescence. That study provided evidence for a hydration- mediated ion-specific nucleation pathway that is relevant to the non-equilibrium conditions following a collision, and thus unique to contact-mode efflorescence. Other studies have shown that immersed particles of metal oxides and mineral dust can initiate efflorescence of (NH4)2SO4 and NH4NO3 at a higher RH than homogeneous ERH (Han et al., 2002; Han & Martin, 1999;

Martin et al., 2001; Pant et al., 2006). However, a comparison between contact and immersion efflorescence has yet to be established for heterogeneous nuclei composed of mineral dust. In the present study, efflorescence of salt droplets interacting with mineral aerosols are probed in both immersion and contact mode nucleation.

Mineral dust particles are of interest as heterogeneous nuclei due to their abundance in

Earth’s atmosphere (Usher et al., 2003) and their ability to act as effective ice nuclei (Broadley et al., 2012; Murray et al., 2012). Field studies indicate mineral dust particles are the dominant ice residual under many atmospheric conditions (Daniel J. Cziczo et al., 2013; Richardson et al., 2007).

Mineral dust is one of the most abundant species in the atmosphere by mass, with a size distribution that spans from submicron to supermicron particles. Compositionally, the particles mainly contain feldspar, quartz, and clay mineral species (Usher et al., 2003). A study by Broadley et al. (2012) determined that illite, a clay mineral, was a major component of mineral dust particles and that

NX illite, a mixture of minerals including illite, feldspar, quartz, and kaolinite, is a good analogue of atmospheric dust. Furthermore, there is evidence that as a dust plume travels and undergoes atmospheric aging, clay minerals become more prevalent (Usher et al., 2003). For example, comparisons between the mineralogy of aerosols collected at Sal Island off the coast of Africa and

Miami, FL, show that the clay mineral’s weight percent increases over time (Glaccum & Prospero,

1980). Thus, for the present study, NX illite and clay minerals were chosen as possible heterogeneous nuclei.

Heterogeneous efflorescence was examined by optically levitating single droplets of aqueous salts and exposing them to mineral dust particles. Three different mineral dust particles were used as heterogeneous nuclei for (NH4)2SO4 and NaCl: illite, montmorillonite, and NX illite.

To study ion-specific effects, montmorillonite was further used as contact nuclei for a range of chloride salts (NaCl, NH4Cl, and MgCl2). Although similar in structure, illite is a non-swelling clay while montmorillonite is a swelling clay and responds differently to humidity (Broadley et al.,

2012). For each salt aerosol and mineral dust pair, contact and immersion ERH was determined.

Additionally, the immersion ERH of (NH4)2SO4 by multiple particles of immersed illite was also probed.

2.2: Experimentation

2.2.1: Preparation of Aqueous Solution Droplets

Aqueous droplets were prepared using 5 wt % solutions of (NH4)2SO4 (SigmaAldrich),

NaCl (Fisher Scientific), NH4Cl (SigmaAldrich), and MgCl2 (Mallinckrodt) in HPLC grade water

(SigmaAldrich). The aqueous salt solution was filtered through a 0.22 μm pore nylon filter and used to fill a droplet generator. The droplet generator (Microfab) ejected droplets of the filtered solution from a piezo-driven tip with a 20 μm orifice. By applying a controlled alternating positive and negative voltage to the piezo device, droplets were produced with a reproducible size range of

10−15 μm in diameter. While aqueous aerosols in the atmosphere are smaller than the droplets used for this study, the experiments should still be relevant for atmospheric particles large enough so that the Kelvin effect is minimal (>100 nm) (Biskos et al., 2006).

2.2.2: Preparation of Mineral Dust Particles

Illite (IMt-2) and montmorillonite (SWy-2b) were obtained from the Clay Mineral Society.

The mixture NX Illite was obtained from Arginotec. Montmorillonite and NX Illite were used as received. Because the illite was a mixture with non-uniform grain sizes, the sample was first coarsely ground in a ceramic mortar and pestle. The coarsely ground illite was then ground further with a stainless-steel capsule and ball pestle (Wig-L-Bug) for 5 minutes to produce a fine powder.

Each mineral was then mixed into HPLC grade water to make a slurry that was aerosolized with a nebulizer (Omron NE-U22). It has been shown that generating mineral dust aerosols from a slurry can change the surface properties of the particle due to a redistribution of the soluble species

(Garimella et al., 2014). The effect that the change may have on heterogeneous efflorescence has

16 not been studied and is not known. To prevent the nebulizer from clogging, supermicron particles from the slurry were allowed to settle out before nebulization. Settling velocities (w) used to calculate settling times for each mineral dust were determined based on Stokes’ law:

2 2(휌푝 − 휌푓)𝑔푟 푤 = (퐸푞 2.1) 9휇 where ρp and ρf are the particle and fluid densities respectively, g is the acceleration due to gravity, r is the particle radius, and µ is the fluid’s dynamic viscosity. Densities of 2.2 g/cm3 and 2.7 g/cm3 were used for illite and montmorillonite, and NX Illite respectively. A density of 0.998 g/cm3 and a viscosity of 1.002 mPa∙s was used for water at 20 °C. For a particle with a diameter of 1 µm, the settling time to fall 3 cm was 12.7 and 9 hours for illite and montmorillonite, and NX Illite respectively. The size distributions of the aerosolized mineral dust particles using the method described above and measured with a scanning mobility particle sizer (SMPS TSI model 3010) are shown in Figure A1 in the Appendix.

2.2.3: Experimental Arrangement

The optical levitation apparatus has been described in detail previously (Davis et al., 2015a) and is shown schematically in Figure 2.1. Particles are trapped inside a flow cell with observational windows. To control and measure the RH inside the flow cell two N2(g) flows are used. One flows directly into a mass flow controller then into a diffusion drier, while the other passes through a water bubbler filled with HPLC water to humidify the flow before going to a mass flow controller. The two flows then are mixed and enters the flow cell from below the trapping site and exits from above the trapping site. Two probes (Vaisala HMP60) are placed at the inlet and outlet of the flow cell to measure RH. The RH at the trapping region is calculated as the average of the two probes (±1 S.D.).

Figure 2.1: The experimental setup of the optical levitator. The flow system controls RH and flows particles into the optical cell. The optical system traps droplets and the light scattering is imaged onto CCD cameras. The two beam profiles, Bessel and Gaussian, are shown.

Heterogeneous nuclei of mineral dust particles are made by diverting the dry flow after the mass flow controller towards the nebulizer. The nebulizer is filled with 3 – 4 mL of the mixture of mineral dust particles in water and creates a mist of the suspension. The mist is carried into a diffusion drier to create a particle stream of dried mineral particles. The gas carrying the particles of mineral dust particles then mix with the wet flow before entering the flow cell.

The entire flow cell is attached to a sliding stage that moves in both horizontal directions.

The droplet’s position is determined by the lasers and controlled by optics placed outside of the flow cell, and thus is independent of the flow cell’s position. In contrast, the mineral dust particles flow through the center of the cell. Thus, by moving the flow cell, the particle stream can be moved into and out of the laser’s focus. When the particle stream is misaligned to the laser, no collisions between heterogeneous nuclei and droplet occur and when aligned collisions occur.

Two counter propagating 532 nm ND:YAG lasers are used to optically trap droplets of aqueous solutions inside a closed flow cell with windows. The laser directed from below has a

Gaussian profile while the laser directed from above has a Bessel profile. For a scattering particle, a laser beam exerts radiation pressure on the illuminated particle. Radiation pressure has two components, scattering and gradient force. The scattering force is a force in the direction of the laser propagation. The gradient force pulls the particle towards the region of highest laser intensity, typically at the center of a laser. The laser from below in combination with drag force from air flowing upwards counters the force of gravity to levitate the particle. The gradient force traps the droplet by drawing it towards the center of the lasers, stabilizing the trap in the horizontal direction.

While it is possible to trap multiple droplets at different positions in the trap, this study probed single levitated droplets.

2.2.4: Imaging Efflorescence and Collisions

Efflorescence is monitored using light scattering focused onto two charge coupled device

(CCD) cameras and recorded using a LabView program, allowing both far field and near field scattering to be imaged at the same time. The scattering pattern of a liquid droplet is distinctly different from a crystalline solid, allowing for a visual identification of efflorescence. Upon efflorescence, the droplet loses water, resulting in a significant change in mass and upward movement in the trap. The sudden change in position is used as a second determination of efflorescence. Finally, the scattered laser light is used to observe collisions between the droplet and mineral dust particles. Scattering from mineral dust particles can be imaged onto the same

CCD camera that is imaging the droplet. Methods to monitor collisions between a heterogeneous nucleus and aqueous droplet are discussed in detail in Davis et al. (2015a). Figure 2.2 illustrates how a collision with a heterogeneous nucleus can result in an effloresced salt aerosol or an aqueous droplet with a particle immersed inside. Far field images are shown for an experiment with

(NH4)2SO4 contacted by illite. When the collision induces crystallization, the light scattering from

(NH4)2SO4 changes distinctly from linear fringes to a mosaic pattern. The particle also moves up in the field of view of the image as it loses mass due to loss of water. When the collision does not cause efflorescence but instead becomes immersed inside the droplet, the linear fringes remain, and the scattering does not move in the field of view. However, the immersed particle will occasionally disrupt the light and causes ripple-like features to appear in the image. In the figure, to highlight the disruptions the far field light scattering images, in green, were self-correlated. The far field image is filtered with Gaussian smoothing to remove noise. Then the absolute difference between the image and the image shifted by 20 pixels to the right is calculated. The resulting image of the defect is further enhanced by converting the result into a binary image by choosing a

Figure 2.2: Diagram outlining the difference between efflorescence of an aqueous droplet upon contact with a heterogeneous nucleus (top) and the immersion of the particle into the droplet

(bottom). The example is of (NH4)2SO4 droplet and illite particle. The original light scattering is shown in green with the defects of the self-correlation shown in red

21 threshold intensity, where only the pixels that are brighter than the threshold will be indicated by red. As seen in the figure, the pure droplet exhibits very few defects whereas the droplet with an immersed particle shows a cluster of defects in the area where the ripples appear. With this process, we are able to determine whether a collision occurred, and whether that collision resulted in efflorescence or whether the mineral dust particle became internally mixed with the aqueous droplet.

2.2.5: Determining Heterogeneous ERH

Contact and immersion ERH values were determined through different experimental methods. For contact efflorescence, an aqueous particle was trapped at a RH significantly higher than the homogeneous ERH but lower than the DRH. For each trial, the trapped droplet was introduced to a stream of mineral dust particles for 60 seconds. During the period of exposure to dust particles, the droplet was exposed to an average of a collision every 9.5±1.6 seconds. The droplet was monitored for whether it effloresced or not during the trial. After 60 seconds, if the droplet had not yet effloresced it was ejected from the trap and a new droplet was caught. The process was repeated to calculate a probability of efflorescence (PEff) as the ratio of observed contact efflorescence events to the number of trials done at that RH. Thus, a PEff of 1.0 implies that efflorescence was observed during every trial of 60 seconds of dust exposure conducted at that RH. Each trial had on average ≤6.3 collisions, and thus over 16% of collisions led to efflorescence. Contact ERH was determined as the RH where PEff is equal to 0.5.

To determine the immersion ERH an aqueous droplet was exposed to mineral dust particles at a RH significantly higher than both homogeneous ERH and contact ERH. Once a mineral dust particle became immersed in the droplet, the droplet with the immersed mineral dust was then isolated from the particle stream, using the sliding stage, to prevent any additional collisions. The

22 recorded video of the droplet being exposed to the mineral dust was reviewed to ensure that collisions had occurred and verify the number of immersed mineral dust particles. Droplets that had been exposed to more than three collisions were discarded. The RH inside the cell was then lowered at a rate of ~1% RH/min until efflorescence occurred. The immersion ERH was determined as the average RH (±1 S.D.) for all immersion efflorescence events observed.

To further explore the immersion ERH as a function of total surface area of immersed mineral dust, multiple particles of illite were immersed into aqueous droplets and the immersion

ERH was determined in a similar manner as above. However, in contrast to the immersion experiments described above, the droplet was exposed to the mineral dust particle stream for an extended period of time. By varying the exposure time, the number of immersed particles was changed. The number of particles immersed was determined by recording and analyzing videos taken during dust exposure and counting the collisions. To calculate the total surface area of the immersed illite, the particle size distribution shown in Figure A1 of the Appendix was used. Based on the measured size distribution of three samples, the geometric mean was 380 nm with a geometric standard deviation of 1.63.

2.3: Results

2.3.1: Heterogeneous Efflorescence by Mineral Dust

Experimental results for contact heterogeneous efflorescence of (NH4)2SO4 and NaCl are shown in Figure 2.3. The data was fit with a sigmoid curve to determine the RH at PEff = 0.5 and thus contact ERH. The sigmoid curve was constrained to have a maximum value of 1 and a minimum value of 0 when the fit was performed. For both (NH4)2SO4 and NaCl the contact ERH for all three heterogeneous nuclei was well above the homogeneous ERH. Furthermore, the probability of contact efflorescence decreased relatively rapidly with RH above the contact ERH.

Figure 2.3: Contact efflorescence experiment results for (NH4)2SO4 (top) and NaCl (bottom) by different mineral dusts. The data for each dust (filled circles) was fitted with a sigmoid curve

(dashed line) to calculate contact ERH. PEffloresence is the ratio of observed contact efflorescence events to the number of trials done at that RH range.

For (NH4)2SO4, the transition was steepest for montmorillonite while the most gradual was for illite. For NaCl the steepest transition was for NX Illite and the most gradual was for montmorillonite. Results for the complementary experiments of MgCl2 and NH4Cl contact efflorescence with montmorillonite contact nuclei are shown in Figure 2.4. The contact ERH for

MgCl2 was 10.9±0.6% and for NH4Cl was 54.0±0.3%. Homogeneous ERH and DRH values for

MgCl2 were determined as 3.7±0.4% and 13.7±0.5% using the optical levitator. For NH4Cl, literature values for homogeneous ERH and DRH values were used as 45% and 77% respectively

(Martin, 2000).

Results from both contact and immersion efflorescence experiments for (NH4)2SO4 and

NaCl are summarized in Figure 2.5. The heterogeneous ERH for each salt and dust combination in both immersion and contact mode is higher than the homogeneous ERH. Compared to the homogeneous ERH of 35% for (NH4)2SO4 the heterogeneous ERH was 42 – 57%. For NaCl, which has a homogeneous ERH of 45%, the heterogeneous ERH ranged from 49 – 63%. The observed increases in ERH shows that the hysteresis effect can be dampened in the presence of mineral dust particles. The narrowing of the meta-stable region for an aqueous particle could thus increase the length of time that a salt aerosol stays in the crystalline phase.

In addition to raising the ERH of the salts, a difference between the two modes of heterogeneous efflorescence was observed. As seen in Figure 2.5 the contact ERH for all pairs of dust and salt were higher than its immersion ERH. The difference between the ERH values of the two modes varied depending on the salt and dust pair and was most pronounced for NaCl with montmorillonite as the contact nucleus. The higher ERH for contact mode than immersion mode is consistent with prior work in our laboratory that found amorphous organic aerosol induced salt efflorescence via contact but not when immersed (Davis & Tolbert, 2017). The higher

Figure 2.4: Contact efflorescence experiment results for MgCl2 (top) and NH4Cl (bottom) by montmorillonite (SWy-2). The data for each dust (filled circles) was fitted with a sigmoid curve

(dashed line) to calculate contact ERH. The solid lines indicate homogeneous ERH and DRH.

Figure 2.5: Summary of results for heterogeneous efflorescence of (NH4)2SO4 (top) and NaCl

(bottom) by NX Illite, Na-Montmorillonite, and Illite. Solid bars represent immersion ERH, dashed bars represent contact ERH. Homogeneous ERH and DRH of (NH4)2SO4 and NaCl are marked by solid black lines.

27 heterogeneous efflorescence efficiency in contact mode compared to immersion mode is also consistent with findings from studies on heterogeneous ice freezing (Durant & Shaw, 2005;

Niehaus et al., 2014; Niehaus & Cantrell, 2015; Shaw et al., 2005). A study by Shaw et al. (2005) showed that when the heterogeneous nucleus was in contact with the surface of the droplet, freezing temperatures were significantly higher than when the nucleus was inside the droplet. In another study by Niehaus and Cantrell (2015), contact ice nucleation by soluble salts was examined.

That study showed that soluble species, which are thought to be ineffective immersion mode ice nuclei, can induce freezing at significantly warmer temperatures than homogeneous freezing.

2.3.2: Immersion ERH as a Function of Surface Area of Immersed Illite

The above results show that contact efflorescence can often be more effective than immersion efflorescence where there are less than or equal to three particles immersed inside.

However, previous work has shown that the heterogeneous ERH of atmospheric salts by an immersed metal oxide particle could be enhanced by increasing the surface area of the heterogeneous nucleus (Han et al., 2002; Martin et al., 2001). Thus, immersion ERH of (NH4)2SO4 was examined as a function of the total surface area of the immersed illite by increasing the number of immersed particles and the results are shown in Figure 2.6. To calculate total surface area of the immersed illite each particle was assumed to be spherical with a diameter of 380nm. The surface area of one particle was scaled by the number of collisions and thus the number of particles in the droplet. By varying the number of immersed illite particles the total surface area was changed. It can be seen that the immersion ERH increases with surface area and that it takes many immersed illite aerosols for the immersion ERH to approach the contact ERH.

Figure 2.6: Heterogeneous ERH of (NH4)2SO4 vs the calculated total surface area of immersed illite. Solid black lines on the graph represent the homogeneous ERH and DRH values while the dashed black line is the contact ERH by illite.

2.4: Discussion

2.4.1: Crystal Lattice Match

A factor that is often used to compare a particle’s effectiveness at causing heterogeneous nucleation is the crystal lattice match. It is suggested that when a heterogeneous nucleus has a similar lattice structure as the target crystal, it will be more efficient at initiating nucleation. Thus, when a heterogeneous nucleus has a high lattice match to the salt, the heterogeneous ERH should be higher (Davis et al., 2015b; Martin et al., 2001; van Meel et al., 2010). A study by Davis et al.

(2015b) showed that when the calculated lattice mismatch was lower than 0.12, a strong trend between lattice mismatch and contact ERH was observed. To numerically compare lattice structures of the heterogeneous nucleus to the salt crystal, lattice mismatches (δ) between their crystal faces were calculated,

푎1,퐻푁 − 푎1,푠 푎2,퐻푁 − 푎2,푠 | 푎 | + | 푎 | 훿 = 1,푠 2,푠 (퐸푞. 2.2) 2 where a1 and a2 are the two lattice constants used to define a crystal face for the heterogeneous nucleus (HN) and salt aerosol (s). A lower mismatch value indicates that the two crystal lattice structures are more similar. Each of the three faces (100, 010, 001) of the salt was compared to the three faces of the heterogeneous nucleus, resulting in nine calculated lattice mismatches. The lowest mismatch value of the nine was used for analysis.

Since inorganic salts have distinct homogeneous ERH and DRH values, and thus different hysteresis ranges, when comparing heterogeneous ERH values between multiple salts σERH is used (Davis et al., 2015b). The parameter σERH is calculated as:

(퐷푅퐻 − ℎ푒푡퐸푅퐻) σERH = (퐸푞. 2.3) (퐷푅퐻 − 퐸푅퐻)

30 where ERH and DRH are the homogeneous transitions and hetERH is the heterogeneous ERH value for either immersion or contact mode efflorescence. An efficient heterogeneous nucleus would be indicated by σERH that is closer to zero. Heterogeneous σERH is plotted against the calculated lattice mismatch values for illite and montmorillonite in Figure 2.7. NX Illite was excluded from the analysis because it is a mixture of minerals and has no singular lattice structure.

A table with the lattice constants of all species analyzed in the study is shown in Table A1 of the

Appendix. In addition, heterogeneous ERH of (NH4)2SO4 by immersed particles of three metal oxides are included (Han & Martin, 1999). The trend between contact σERH and lattice mismatch observed by Davis et al. (2015b) is also shown as a dashed line. For combinations with a lattice mismatch less than 0.12 ((NH4)2SO4 with illite and montmorillonite), they seem to follow the trend of a lower lattice mismatch resulting in a lower σERH. For all other combinations analyzed in the present study, the mismatch is larger than 0.12 and does not seem to have a strong correlation between lattice mismatch and heterogeneous σERH. It is important to note that the calculation for lattice mismatch used here does not fully address the complexities of crystal growth on substrates.

A study with a more detailed model on epitaxial growth of (NH4)2SO4 on metal oxide particles has been done by Martin et al. (2001). However, regardless of how crystal lattice mismatch is determined, crystal lattice structure does not explain why contact ERH is higher than immersion

ERH for the same salt and dust pair. The crystal lattice of the nucleus and the salt is unchanged whether the particle comes into contact with the outer surface or is immersed inside the droplet.

However, surface properties could be different and thus factors other than crystal lattice should be considered when analyzing the difference between contact and immersion mode efflorescence.

Figure 2.7: Heterogeneous σERH of (NH4)2SO4 (Square) and NaCl (Diamond) as a function of lattice mismatch with different minerals (represented by different colors). Contact σERH are represented as unfilled markers and immersion σERH are represented as filled markers. The dashed line shows the trend seen by Davis et al. (2015b). *Han and Martin (1999).

2.4.2: Active Site

Another model often used to explain heterogeneous nucleation is the active site model

(Fletcher, 1969; Han et al., 2002; Martin et al., 2001). It is believed that active sites are the result of defects to the mineral’s structure and their specific nature would be dependent upon the particle type and preparation method. Active sites are thought to be where crystallization is most likely to begin (Martin et al., 2001). A study of ice nucleation onto feldspar particles using an environmental SEM showed that freezing initiated exclusively at defects on the mineral surface

(Kiselev et al., 2017). Past work on immersion efflorescence by metal oxides suggested that by increasing the surface area of the immersed particle, the number of active sites increases, and thus a higher immersion ERH is found (Han et al., 2002; Martin et al., 2001). As shown in the results, immersion ERH for (NH4)2SO4 seems to increase with increased surface area of illite and thus the total number of active sites. While the active site model is consistent with the observed trend in efflorescence efficiency with surface area, it does not explain the difference between contact and immersion ERH. Unless the active sites are altered by becoming immersed into solution, the heterogeneous nucleus should have the same number of active sites regardless of whether it causes efflorescence in immersion or contact mode.

2.4.3: Ion-Specific Effects

Ion-specific interactions between the mineral dust and the ions in solution may influence efflorescence and thus heterogeneous ERH. A recent study by Davis and Tolbert (2017) demonstrated that contact efflorescence initiated by charged amorphous organic aerosols

(polystyrene latex spheres) exhibits an ion-specific trend, consistent with the Hoffmeister series, which is a set of ions ordered based on how well they precipitated out proteins from a solution

(Baldwin, 1996). In addition to following the Hoffmeister series, the trend also correlated with

33 hydration strengths of the aqueous ions attracted to the charged surface of the heterogeneous nucleus. Studies have shown that aggregation of montmorillonite, which has a negative surface charge, was influenced by the cation in solution, demonstrating that aqueous cations are attracted to the surface of Na-montmorillonite (Tian et al., 2014). To observe the ion specific effect, contact efflorescence by montmorillonite of two additional chlorides, ammonium chloride (NH4Cl) and magnesium chloride (MgCl2), were studied. Since montmorillonite has a negatively charged

+ surface the cations were altered while keeping the chloride anion constant. Additionally, NH4 and Mg2+ were specifically chosen as the two lie on opposite sides of Na+ in the Hofmeister series.

Heterogeneous σERH were plotted against the three cations in order of decreasing hydration strength in Figure 2.8. As shown in the figure, the negatively charged surface of montmorillonite was more efficient at causing efflorescence of the cation with higher hydration strength. The observed trend is consistent with the study by Davis and Tolbert (2017) where the contact nucleus was a polystyrene latex sphere with a charged surface. The results suggest that the surface charge of the colliding heterogeneous nucleus influences the counter ions in solution, affecting heterogeneous efflorescence efficiencies via a destabilizing ion-specific hydration- mediated pathway, as proposed by Davis and Tolbert (2017). Ion-specific effects could also explain the observed difference between contact and immersion ERH. It has been suggested by

Fukuta (1975) that the spreading of water onto the surface of the heterogeneous nucleus causes the difference between immersion and contact freezing. The forced movement of water creates a region with higher energy that then increases the chances for a critical ice germ to form. However, when the heterogeneous nucleus is already immersed, the liquid is already spread onto the particle surface and no region of higher energy is formed. For efflorescence, a similar effect could occur where upon collision, due to the supersaturation of the aqueous phase, ions in the droplet are forced

Figure 2.8: Comparing contact σERH of various chloride salts by Na-Montmorillonite with respect to the hydration strength of the cations of the salts.

35 to share water molecules with the incoming contact nucleus to create a hydration shell. The forced sharing destabilizes the aqueous phase which is stabilized by uptake of additional water from the surrounding to fully hydrate the contact nucleus or by crystallization (Davis & Tolbert, 2017). The amount of energy increase could depend on the present ion’s hydration strength as well as the surface charge of the heterogeneous nucleus, thus creating the ion-specific trend in which contact efflorescence efficiency is correlated to hydration strengths of the counterions in solution.

2.4.4: Collision Lifetimes of Salt Aerosol in a Dust Plume

Samples of atmospheric particles have shown that salt and dust aerosols are found internally mixed in many cases (Andreae et al., 1986; Li et al., 2003). The mixed aerosols, when wet, could undergo immersion efflorescence. As discussed, immersion efflorescence is not as efficient as contact efflorescence. However, contact efflorescence requires a collision between the aqueous salt droplet and a mineral dust particle.

Previous work has shown that the lifetime of a single aqueous salt particle with respect to collisions with a heterogeneous nucleus could be comparable to its atmospheric lifetime (Davis et al., 2015b). The lifetime (τ) of a single salt particle to collide with a heterogeneous nucleus is calculated as

휏 = (퐾 ∗ 푁)−1 (퐸푞. 2.4) where K is the coagulation coefficient in units of cm3/s and N is the heterogeneous nucleus (HN) concentration in units of cm-3. The coagulation coefficient depends strongly on the particle’s sizes

K = 2 π (2푅푠푎푙푡 + 2푅퐻푁) (퐷푠푎푙푡 + 퐷퐻푁) (퐸푞. 2.5) where R is the radius of the salt and HN and D is their estimated diffusion coefficient in units of cm2/s. Diffusion coefficients are estimated by interpolating data from table 9.5 in Seinfeld and

Pandis (2006). Coagulation coefficients were calculated by using a mineral dust particle size

36 distribution based on in-situ measurements taken at Peking University in Beijing, China an hour after a dust storm event had begun (Wehner et al., 2004). The particle size distribution was broken up into equally spaced intervals on a log scale to estimate the dust aerosol number concentrations at each dust particle radius. For each size and corresponding number concentration (N1, N2, N3,

…), coagulation coefficients (K1, K2, K3, …) and lifetimes (τ1, τ2, τ3, …) were determined. The final collision lifetime was then calculated:

1 1 1 −1 1 τ = ( + + + ⋯ ) = (퐸푞. 2.6) 휏1 휏2 휏3 (퐾1 ∗ 푁1 + 퐾2 ∗ 푁2 + 퐾3 ∗ 푁3 + ⋯ )

To determine lifetimes for various number concentrations of dust, the original size distribution was scaled by 0.002 – 2. Number concentrations were also converted to mass concentrations by assuming spherical particles with a density of 2.6 g/cm3.

Figure 2.9 shows the results of the calculation as a color plot, where the red colors represent shorter collision lifetimes and the blue colors represent longer lifetimes. Salt particle diameters ranging from 50 nm to 2.5 µm were assumed to represent both submicron and supermicron sized particles. Newly formed particles of (NH4)2SO4 and the subsequent growth of the aerosol into the accumulation mode are represented in the lower end of the size range (Seinfeld & Pandis, 2006).

Sea spray aerosol has a wide range of particles sizes produced through bubble bursting actions.

Each bubble that bursts creates a few jet drops of a larger size and many more film drops that are less than a micrometer in radius (Fitzgerald, 1991). Dust concentrations range from ~2 – 2,000

µg/m-3 to represent regions that are far from the source, such as marine environments and near the source such as East Asia and Northern Africa (Chou et al., 2008; Prospero, 1999; Tegen & Fung,

1994).

In Figure 2.9, it can be seen that salt aerosols will have a shorter lifetime as the dust number increases. At the lower end representing a typical dust concentration far from the source

Figure 2.9: Calculated collision lifetimes (days) for a single salt aerosol to collide with a mineral dust particle in a plume. The particle size distribution of mineral dust used was measured at

3:00AM during a dust storm event at Peking University, China from Wehner et al. (2004, Figure

1d). To calculate lifetimes for a range of mineral dust particle concentrations, the distribution was scaled by factors between 0.002 and 2. A density of 2.6 g/cm3 was assumed to estimate the dust aerosols mass concentration.

(Prospero, 1999), below 10 µg/m3, the salt lifetimes with respect to a collision are on the order of a few days to a few weeks. During low dust periods, aqueous salt particles smaller than 1 µm most likely do not collide with dust particles. At concentrations between 10 – 100 µg/m3 of dust, which are observed occasionally in regions far from the source and is a typical concentration for regions near the source (Chou et al., 2008; Prospero, 1999), the salt lifetimes are a few hours to a few days.

Collisions between salt and dust particle are likely to occur and have some impact on the aqueous stability. During severely dusty and dust storm events, greater than 100 µg/m3, collisions occur within a few seconds to a few hours (Chou et al., 2008; Wehner et al., 2004). During such an event, collisions between salt and dust will be a dominant atmospheric process for salt aerosols of all sizes and could provide a major pathway for contact efflorescence.

Figure 2.9 also shows that the size of the salt aerosol plays a significant role in the collision lifetime. The calculated lifetimes are shorter for larger salt aerosols due to the sensitivity of the coagulation coefficients to particle size. The coagulation coefficient increases when the size difference between the salt and dust particle increases. Due to the high number concentration of the small dust particles in the assumed distribution, the larger salt aerosols have a much shorter collision lifetime and have a higher probability of undergoing efflorescence.

2.5: Conclusion

Mineral dust particles in the atmosphere can have a significant impact on the phase state of soluble inorganic aerosol. Salt particles could exist as a solid more often than currently estimated due to collisions with dust particles. The total effect will depend on whether the two types of particulate are internally mixed or externally mixed. When the salt and dust aerosols are internally mixed, immersion efflorescence will slightly decrease the hysteresis effect. However, if they are externally mixed, collisions between aerosols could cause the more efficient contact

39 efflorescence to occur. As shown in the calculation of coagulation lifetimes, collisions are strongly dependent on the dust concentration. In regions where dust concentrations are typically high, contact efflorescence could play a significant role in shaping the phase state of soluble inorganic aerosols. Crystal lattice and active site models have been useful when comparing heterogeneous efflorescence efficiencies of different particles, but struggle to explain the difference between the two modes of nucleation. When several chlorides with various cations were heterogeneously effloresced by montmorillonite, efflorescence efficiencies were higher for the cations with higher hydration strength. The observed trend validates the mechanism proposed in our previous publication (Davis & Tolbert, 2017) demonstrating that hydration-mediated ion-specific interactions between the ion in the aqueous droplet and surface charge of the heterogeneous nucleus affects contact mode efflorescence. The influence that the heterogeneous nucleus has on the ions in solution could also explain observed differences between immersion and contact mode efflorescence.

Chapter 3: Contact Efflorescence of Internally Mixed Organic – Ammonium

Sulfate Aerosols: A Glassy Organic vs An Organic Coating.

3.1: Introduction

As discussed in Chapter 1, aerosols in the atmosphere impact climate through direct radiative effects as well as indirect effects such as cloud activation through cloud condensation or ice nucleation (Andreae & Rosenfeld, 2008; Boucher et al., 2013; Ramanathan et al., 2001;

Seinfeld & Pandis, 2006). While Chapter 2 focused on the inorganic fraction of atmospheric aerosols, the organic fraction of atmospheric aerosols can represent 50% or more of the aerosol mass depending on the region (Zhang et al., 2007). Organic aerosols can be directly emitted through burning fossil fuels and biomass burning, but the organic fraction in aerosols are often dominated by secondary organic aerosols (SOA). SOA are organic aerosols formed from the chemical reaction of emitted volatile organic compounds that undergo chemical reactions. The chemical transformation changes the volatility of the organic compound, and the low-volatile products condense into the aerosols. The precursor VOCs, such as monoterpenes and aromatic compounds, for SOA formation can be both anthropogenically and biogenically emitted. The anthropogenic impact on SOA formation is largest for more urban regions. Biogenic emissions include species like monoterpene, isoprene, and glyoxal. Characterizing the organic portion of atmospheric aerosol, especially when looking at secondary organic aerosols, is challenging due to the complex mechanism for SOA (Jimenez et al., 2009; Kroll & Seinfeld, 2008). Due to the large variety of molecules that make up atmospheric SOA understanding their effects on climate has been more challenging.

Atmospheric aerosols can be found in various phase states including liquid, solid, and glass depending on temperature and relative humidity. As discussed in Chapter 1, soluble inorganic

41 salts can be found as an aqueous liquid droplet or a crystalline solid, transitioning between the two phases based on humidity. The previous chapter probed heterogeneous efflorescence by mineral dust particles and showed that the rise in ERH depends on the aqueous droplet’s composition and the identity of the heterogeneous nucleus (Davis et al., 2015b; Davis & Tolbert, 2017; Martin et al., 2001; Ushijima et al., 2018). In a different study where the super saturated solutions were contacted by a crystal of itself, seeded crystal growth takes over and efflorescence can occur at an

RH right below the DRH (Davis et al., 2015b). Contact efflorescence of an aqueous droplet by a crystal version of itself will effectively shut down the hysteresis (Davis et al., 2015b). While seeded crystal growth has been observed for several inorganic droplets (Davis et al., 2015b), it is not clear if such a mechanism would be viable for more complex droplets composed of inorganic/organic mixtures.

Organic aerosols tend to not crystallize and remain liquid or amorphous even when the humidity and or temperature has been lowered (Marcolli et al., 2004). Laboratory studies where soluble inorganics were mixed with organic molecules have shown that some organics will inhibit efflorescence by lowering the ERH value when compared to the pure inorganic droplet (Bodsworth et al., 2010) or can even prevent efflorescence all together (Robinson et al., 2014). Further, as the organic aerosols are dried and/or cooled, they can transition into a glassy particle (Zobrist et al.,

2008). Glasses are particles with amorphous structure whose viscosities have become high enough that they are essentially solid (Angell, 1995). The temperature and humidity conditions necessary for an organic molecule to transition into a glass are different for each organic. While laboratory studies measuring the viscosity of SOA from different precursors have been conducted (Grayson et al., 2016; Petters et al., 2019), the diversity of SOA compounds have made it difficult to fully understand their impacts. Models which estimate the glass transition of a compound have been

42 suggested based on molecular characteristics such as atomic O:C ratio and molar mass (Derieux et al., 2018; Koop et al., 2011). In the atmosphere where the temperatures are colder than Earth’s surface, some studies have suggested that organic glasses are more common in the atmosphere, especially in the middle to upper troposphere (Shiraiwa et al., 2017; Zobrist et al., 2008).

The mixing state of atmospheric aerosols describes how the chemical components are arranged in an aerosol population. The mixing of organic compounds in with inorganic species can change the phase transition properties of the aerosol. As mentioned above, some organics has been shown to suppress efflorescence of the inorganic fraction (Bodsworth et al., 2010; Robinson et al., 2014). In addition, depending on the chemical composition and humidity, the organic and inorganic fractions can undergo liquid-liquid phase separation (LLPS) (Freedman, 2017). Often the phase separation results in an organic coating around an aqueous inorganic core (Freedman,

2017), but other times the organic only partially engulfs the inorganic core (Freedman, 2017).

In this Chapter we examined the effect of two different model organics, raffinose and polyethylene glycol 400 (PEG-400) on the efflorescence behavior of (NH4)2SO4. Specifically, we probed whether the two organics would inhibit seeded crystal growth when contacted with a crystal of (NH4)2SO4. Raffinose is a trisaccharide which has been detected in atmospheric aerosols

(Decesari et al., 2007) and has a high glass transition temperature. Additionally, a 1:1 by weight solution of (NH4)2SO4 and raffinose was shown to not effloresce at room temperature when using a droplet on a plate technique (Robinson et al., 2014). PEG-400 is a polymer whose average molecular weight is 400 amu and can also be described as a polyol. Polyols are organic molecules containing multiple hydroxy groups which have also been detected in atmospheric aerosols (Wan

& Jian, 2007). PEG-400 was specifically chosen due to it undergoing LLPS with (NH4)2SO4 at

1:1 weight ratio below 90% RH (Ciobanu et al., 2010). Through the two model organics we will

43 address the following two questions: 1. At what RH values do glassy organics such as raffinose impact the seeded crystal growth of (NH4)2SO4 and 2. Will a liquid coating of organic prevent a heterogeneous nucleus from inducing seeded crystal growth of the unexposed aqueous core. To answer the questions, droplets of 1:1 weight ratio of either (NH4)2SO4 and Raffinose or (NH4)2SO4 and PEG-400 were trapped in an optical levitator. The trapped droplets were then exposed to a stream of crystalline (NH4)2SO4 to observe contact efflorescence.

3.2: Experimentation

3.2.1. Preparation of Droplets for Trapping

For droplet generation, solutions of 5 wt % total 1:1 (NH4)2SO4:Raffinose and

(NH4)2SO4:PEG-400 in HPLC grade water (SigmaAldrich) were made. (NH4)2SO4, Raffinose, and PEG-400 were all obtained from SigmaAldrich. The solutions were filtered through a nylon filter with 0.45 µm pores. A droplet generator (Microfab MJ-APB-20) was used to create droplets by applying an alternating positive and negative voltage to a piezo electric ring near the tip of a glass capillary with a 20 µm diameter orifice.

3.2.2. Experimental Setup

Optical levitation coupled with Raman spectroscopy was utilized to characterize the levitated droplets as well as observe contact efflorescence. The optical levitation set up without the Raman has been described in detail previously (Davis et al., 2015a) and in Section 2.2.3, thus here we only briefly describe the setup. Particles are trapped inside a rectangular flow cell with observational windows on all four sides. To control and measure the RH inside the flow cell two

N2(g) flows are used. One flows directly into a mass flow controller, while the other passes through a water bubbler filled with HPLC water to humidify the flow before going to a mass flow controller.

The two flows then are mixed and enters the flow cell from below the trapping site and exits from

44 above the trapping site. Two probes (Vaisala HMP60) are placed at the inlet and outlet of the flow cell to measure RH. The RH at the trapping region is calculated as the average of the two probes

(±1 S.D.).

Heterogeneous nuclei of (NH4)2SO4 crystals are generated by diverting the dry flow after the mass flow controller towards a nebulizer (Omron NE-U22). The nebulizer is filled with 3 – 4 mL of a solution of 10 wt. % (NH4)2SO4 and creates droplets of the solution. The droplets are carried into a diffusion drier causing efflorescence of the (NH4)2SO4 (ERH=35%) The gas carrying the particles of crystalline (NH4)2SO4 then mix with the wet flow before entering the flow cell.

Figure 3.1 depicts the optical trapping setup inside of the flow cell. Two counter propagating 532 nm Nd:YAG lasers enter above and below the droplet. The laser from below has a Gaussian profile to help facilitate droplet capture while the laser from above has a Bessel profile to enhance droplet trapping. The trapped droplet is observed by two charge coupled device (CCD) cameras that are placed perpendicular to the laser’s axis. One of the cameras take images of far field scattering while the other obtains bright field microscopy images. To obtain bright field microscopy images, a light emitting diode (LED) is placed on the other side of the droplet, shining light on the trapped particle and then into the CCD camera.

Raman spectroscopy has been coupled to the optical levitator to obtain chemical information about the trapped particle. The light scattered by the aerosol from the trapping lasers is collected and collimated by a plano-convex lens with a focal length of 30 mm. The collimated light is filtered through a 532 nm notch filter to remove the Rayleigh scattering and allow the

Raman signal to pass through. The light is then focused onto a fiber optic cable with another plano-convex lens which carries the signal to a Raman detector (Teledyne Princeton Instruments

Figure 3.1: Arrangement of CCD cameras and LED to observe and record far field and bright field scattering. Arrangement of lenses, notch filter, and fiber optic to obtain a Raman spectrum of levitated droplet is also shown. Sample images of a liquid and crystal particle for far field and bright field are shown. A sample Raman spectrum is also shown. The beam profiles of the trapping lasers are shown. The beam from above has a Bessel profile while the beam from below has a Gaussian profile.

FERGIE). The lenses, filter, and fiber optic adaptor are joined by a cage system to form a single optical assembly. The position of the joined optical assembly can be adjusted in all three directions to maximize the signal from the droplet. The spectra are collected at one second exposures per frame with each spectrum made up of 120 frames that are averaged. As seen in Figure 3.1, spectra are reported over the relative wavenumber range of 500 – 4,000 cm-1. To minimize stray light from entering the detector, a black covering for the optical assembly was applied. Additionally, a curtain is drawn over the entire laser table and the laboratory’s lights are turned off during acquisition to limit ambient light from entering the detector.

3.2.3. Observing Efflorescence

Efflorescence of the droplet can be determined by observing the far field scattering pattern.

As shown in Figure 3.1, the far field scattering pattern is distinctly different for a liquid droplet and a crystalline particle. Because the liquid droplet is spherical, Mie scattering is observed as alternating bright and dark bands which are stable for constant RH and thus size. The crystal on the other hand will scatter light differently, due to the multitude of crystal facets, creating a mosaic pattern. The scattering pattern will also change with time as the crystal rotates inside the trap. The bright field microscopy images of the trapped aerosol can also yield information about the shape of the droplet and thus efflorescence. Example images of a liquid and a crystal are shown in Figure

3.1. The liquid is shown to be spherical while the crystal has a rectangular prism shape. While the crystal does not always form a rigid shape and sometimes looks spherical, inhomogeneities in the image due to light passing through the crystal facets can be used to determine efflorescence in bright field images.

3.2.4. Conducting Contact Efflorescence Experiments

To observe contact efflorescence a droplet of the (NH4)2SO4:organic solution is trapped in the flow cell. The trapped droplet is exposed to a particle stream of crystalline (NH4)2SO4 for a maximum of 25 s at a set RH. During the exposure period, a collision between a crystalline particle of (NH4)2SO4 and the trapped droplet occurs once every 4.2 ± 1.8 s. Thus, the trapped droplet encounters a maximum of six collisions on average within a 25 s exposure period. The videos from the CCD cameras were monitored during the exposure period to determine whether the droplet effloresced or not during the exposure period. If the droplet did not effloresce the droplet was ejected from the trap and a new droplet was captured for a new trial. The process was repeated at different RH values to calculate a probability of efflorescence (PEff) as the ratio of observed efflorescence to the number of total trials as a function of RH.

3.2.5. Measuring Viscosity and Immersion Times

The viscosity of (NH4)2SO4:raffinose droplets and the immersion timescales for (NH4)2SO4 crystals into PEG-400 droplets was studied in a dual-balance linear quadrupole electrodynamic balance (DBQ-EDB), as described extensively elsewhere (Richards et al., 2020) and shown in

Figure 3.2. In the DBQ-EDB technique, two particles can be simultaneously levitated and equilibrated at the same RH, and then subsequently merged to infer viscosity and other physical characteristics of the merged particles. As shown in Figure 3.2a, droplets are generated from 5 wt% stock solutions using piezo-driven droplets dispensers (50 µm orifice; Microfab) and injected into the DBQ-EDB through an induction electrode, which charges the surface of the droplets. In the DBQ-EDB, the linear quadrupole, with an applied oscillating voltage (Vac ±600 V at 300 Hz typical), axially confines the charged droplets. Two independent counterbalance electrodes

Figure 3.2: The dual-balance linear quadrupole electrodynamic balance (DBQ-EDB). a)

Overview of the experimental arrangement and process for merging droplets. In panel 1, two oppositely-charged droplets are simultaneously levitated in the top and bottom balance. Droplets are imaged with far-field laser scatter and bright-field imaging. In panel 2, upon removing the voltage applied to the top balance, droplets are merged. Aspect ratio from the bright field image and the image correlation from the far field is shown as a function of time. b) Merging of crystalline ammonium sulfate (~15 µm diameter) with PEG-400 droplets (~30 µm) shows that the crystalline particle becomes fully immersed instantaneously on the fastest timescales that can be imaged.

(Vdc ±200 V typical) counter the force of gravity and the drag force from a humidified nitrogen gas flow (which controls the RH within the DBQ-EDB). The levitated particles are oppositely charged, with the particle in the bottom balance having a higher charge to mass ratio. When the voltage from the top balance is removed, the two particles merge through electrostatic attraction and the morphology of the merged dimer is tracked as a function of time using far-field laser scatter and bright-field imaging.

Upon merging of two droplets, the shape of the merged dimer deviates significantly from that of a perfect sphere. For viscous liquid droplets, merged dimers exponentially relax to a spherical shape to minimize surface energy. The time it takes to relax to sphericity is related to the viscosity (η), surface tension (σ), and density of the merged droplets as well as the radius (r) of the relaxed sphere [29]. Above viscosities of ~40 mPa s, such as is the case in the present study (where the lower-limit to our measurements is limited by camera imaging frame rate within our experimental setup), merged droplets are in the overdamped regime where the characteristic timescale of coalescence (τ) can be related to viscosity through Equation 3.1:

휂푟 휏 ≈ (퐸푞. 3.1) 휎 where τ is from an exponential fit to the relaxation of dimer aspect ratio (in bright-field images) or far-field defect image correlation value (Davis et al., 2015a; Richards et al., 2020). This technique was used to infer the viscosity of (NH4)2SO4:raffinose droplets with a 30 min equilibration timescale, where σ was approximated as 55 ± 30 mN∙m-1 and r = 18 ± 2 µm (Richards et al., 2020;

Song et al., 2016).

The DBQ-EDB technique was also used to provide insight into the factors influencing the movement of crystalline (NH4)2SO4 through liquid PEG-400. Thus, as shown in Figure. 3.2b, a 15

µm diameter crystalline (NH4)2SO4 particle was merged with a 30 µm diameter PEG-400 droplet at 30% RH. The particles used in these experiments are much larger than those used for contact efflorescence, which provides insight into the relative importance of diffusion at the moment of contact between liquid PEG-400 and crystalline (NH4)2SO4.

3.3: Results

3.3.1. Raman Signal for the Components

Figure 3.3 are a set of Raman spectra taken in the optical levitator for droplets of aqueous

(NH4)2SO4, raffinose, PEG-400, 1:1 (NH4)2SO4:raffinose, and (NH4)2SO4:PEG-400. For each droplet the far field scattering is shown to the right indicating that each are liquid. As seen in the

-1 2- -1 figure, a sharp peak at ~980 cm for SO4 and a broad peak centered at ~3,100 cm for -NH are

-1 observed for the three droplets containing (NH4)2SO4. A broad peak below 3,000 cm indicates organics for all the droplets that contain raffinose or PEG-400. Finally, a broad peak centered at~3,500 cm-1 indicates -OH for water. From the spectra, it is clear that the droplets prepared to be mixtures do indeed contain both ammonium sulfate and the organic of interest, indicating we are creating an internally mixed composition droplet.

3.3.2. Dehumidification of Mixed (NH4)2SO4:Organic Droplets

A series of spectra and far field scattering images for a dehumidification experiment of 1:1

(NH4)2SO4:PEG-400 is shown in Figure 3.4. The droplet was first captured at ~74% RH and the

RH was gradually lowered to 13% RH. The droplet undergoes efflorescence at 36% RH as seen by the change in the far field scattering pattern taken at 43% and 36% RH. The scattering pattern still retains some features of the liquid with interferences suggesting the particle has partially crystallized. The scattering pattern is consistent with a liquid PEG-400 coating around the

(NH4)2SO4 crystal. The pattern resembles the scattering image of an aqueous droplet with a

Figure 3.3: Raman Spectra of (NH4)2SO4 (Red), Raffinose (Yellow), 1:1 (NH4)2SO4:Raffinose

(Purple), PEG-400 (Blue), and 1:1 (NH4)2SO4:PEG-400 (Black). The two mixtures are 1:1 by weight percent. For each spectrum, the scattering image is shown to the right. The relevant peaks

2- -1 have been highlighted with color. The SO4 (red) is a sharp peak at ~980 cm , the organic -CH

(yellow) is a broad peak below 3,000 cm-1, the -NH (green) is a brad peak centered at ~3,100 cm-

1, and the -OH (blue) is a broad peak centered at ~3,500 cm-1.

Figure 3.4: Spectra and scattering images for dehumidification of 1:1 (NH4)2SO4:PEG-400. The droplet began at 74%RH and was lowered to 13% RH. The spectra have been normalized to the

2- 2- -1 SO4 peak and offset from each other. The SO4 (red) peak at ~980 cm and -OH (blue) broad peak centered at ~3,500 cm-1 are highlighted.

53 mineral dust particle immersed inside taken from previous work on the optical levitator (Ushijima et al., 2018). The Raman spectra also shows the change in droplet composition with RH. As the droplet is dried and loses water to the surrounding, the -OH peak becomes less intense. For the effloresced particle at 36% RH, the -OH signal is near zero, consistent with crystallization of the aqueous core. The observed ERH of a 1:1 (NH4)2SO4:PEG-400 and the Raman spectra is also consistent with other studies that investigated this mixture using a droplet on a plate technique

(Ciobanu et al., 2010). The observed ERH of the mixed aerosol is also consistent with the homogeneous ERH of pure (NH4)2SO4 near 35% RH (Martin, 2000).

Figure 3.5 shows data for a similar experiment with a droplet of 1:1 (NH4)2SO4:raffinose.

In this experiment, the droplet was first captured at 62% RH and was dried to 5% RH. Unlike the droplet with PEG-400 as the organic, the droplet with raffinose did not effloresce as humidity was lowered. As mentioned earlier, the efflorescence inhibition at room temperature is consistent with a study done on 1:1(NH4)2SO4:raffinose droplets observed on a plate. As seen by the scattering images the linear Mie scattering pattern remains even as the humidity is lowered to single digit values well below the homogeneous ERH of pure (NH4)2SO4. There are some interferences to

Mie scattering at low RH which could potentially be localized regions of inhomogeneities impacting the refractive index. A study on a droplet of (NH4)2SO4 mixed with sucrose, a disaccharide formed from two of the three monosaccharides that make up raffinose, in an EDB coupled with Raman microscopy showed that under dry conditions an enhancement of sucrose was observed near the surface of the droplet (Chu & Chan, 2017). If raffinose were to respond similarly at low RH, the enhancement at the surface can create differences in refractive index within the droplet causing the observed interferences to the Mie scattering. The Raman spectra are also consistent with the gradual loss of water indicated by the slowly decreasing -OH signal as the

Figure 3.5: Spectra and scattering images for an efflorescence experiment on 1:1

(NH4)2SO4:raffinose by weight. The droplet began at 62%RH and was lowered to 15% RH. The

2- 2- spectra have been normalized to the SO4 peak and offset from each other. The SO4 (red) peak at ~980 cm-1 and -OH (blue) broad peak centered at ~3,500 cm-1 are highlighted.

55 droplet is dried. Note, however, in contrast to the PEG-400 system, water is still present in the droplet at 5% RH.

Homogeneous efflorescence was also analyzed by self-correlating the far field image and calculating its defect as discussed previously in Davis et al. (2015a). Briefly, self-correlation is done by making a copy of the far field image, shifting it 20 pixels to the right and calculating the defect as the absolute difference in pixel intensity between the original and the shifted image when the two images are overlaid. For a liquid’s scattering pattern, the intensity for a light and/or dark band is relatively consistent in the horizontal direction. Thus, shifting the image 20 pixels to the right does not cause a high defect value. However, for a crystal particle, the mosaic pattern causes the shifted image to be significantly different to the original causing a high defect value. Figure

3.6 is a plot of defect intensity vs time for a droplet of (NH4)2SO4:PEG-400 undergoing efflorescence and a droplet of (NH4)2SO4:Raffinose at 4.7% RH. For the droplet with PEG-400 the defect intensity is low until the droplet undergoes efflorescence near 15 s when the defect intensity increases significantly. The defect is highest right after efflorescence due to the change in trapping position. As the droplet settles in the new trap position the defect stabilizes at a higher value than before efflorescence. For the droplet with raffinose, the defect intensity at 4.7% RH remains consistent and low, indicating it has not effloresced despite the droplet being dried to well below the homogenous ERH of pure (NH4)2SO4.

3.3.3. Contact Efflorescence of Mixed (NH4)2SO4:Organic Droplets

Contact efflorescence experiments for (NH4)2SO4:PEG-400 were conducted at RH values ranging between 41.6 and 82.9 % RH. Contact experiments below 41.6% RH were not performed as the mixed droplet homogeneously effloresces at 36% RH. For droplets of (NH4)2SO4:raffinose, contact efflorescence was studied for RH values between 16.9 and 84.2% RH. Since the raffinose

Figure 3.6: A plot of defect intensity with elapsed time for efflorescence of a 1:1 (NH4)2SO4:PEG-

400 at 34.9% RH and a droplet of 1:1 (NH4)2SO4:Raffinose at 4.7% RH.

57 inhibited homogeneous efflorescence of the droplet below 35% RH, contact experiments could be conducted at all RH values below it. Figure 3.7 shows three sample contact efflorescence experiments: an experiment at a high RH for a mixed droplet of (NH4)2SO4:PEG-400 and two experiments for mixed droplets with raffinose at two different RH values with two different results.

For each experiment, a plot of the defect intensity with elapsed time is shown. The plots have been aligned at the collision event marked by the dotted line. Far field scattering images before and after the collision as well as the bright field image after the collision are also shown. In the experiment with (NH4)2SO4:PEG-400 at 77%RH it can be seen that the defect intensity is at a stable and low value, but upon collision increases rapidly and then stabilizes. The variance in defect also increases greatly after the collision. The far field scattering pattern changes upon collision and the bright field image of the droplet after collision indicates a rectangular prism shape.

All of these criteria indicate that the droplet effloresced rapidly upon collision. For the experiment with (NH4)2SO4:raffinose at 65% RH, similar results are observed. Upon collision the defect rises, the far field scattering loses the Mie scattering pattern and the bright field image of the crystal is less spherical, thus indicating efflorescence. In contrast, the experiment shown for

(NH4)2SO4:raffinose at 35% RH has a different outcome. Upon collision, the defect intensity does not rapidly increase although the variance in the data increases. While there are interferences to the far field scattering image, it still maintains the Mie scattering pattern that is distinctive of a liquid. The interference is believed to be due to the crystal of (NH4)2SO4 that collided with the droplet and has become stuck on the surface without causing efflorescence of the droplet. The bright field image after collision further supports that the heterogeneous nucleus did not induce efflorescence because the droplet looks spherical with no inhomogeneities.

Figure 3.7: Defect Intensity with time elapsed for three contact efflorescence experiments. Two experiments with 1:1 (NH4)2SO4:Raffinose at 35% and 65% RH. The third experiment is for 1:1

(NH4)2SO4:PEG-400 at 77% RH. The collision event has been shown as the dotted line on the graph. For each contact experiment the scattering image for the droplet before and after the collisions is shown. The bright field image of the droplet after the collisions is also shown.

Contact efflorescence experiments were repeated for a range of RH values and a summary of the results are shown in Figure 3.8. Figure 3.8 is a plot of PEff to RH for droplets of (NH4)2SO4 mixed with raffinose or PEG-400. The homogeneous ERH and DRH of pure (NH4)2SO4 are shown as vertical dashed lines for reference. For both mixed droplets, above the homogeneous DRH of

(NH4)2SO4 at 80%RH, the Peff was zero, meaning no efflorescence was observed. This result is not surprising since the RH is above the DRH, thermodynamically it is not possible to effloresce

(NH4)2SO4. For every RH tested between the DRH and ERH of (NH4)2SO4, the droplets mixed with PEG-400 effloresced. As seen on the plot, the highest RH that efflorescence was observed for the (NH4)2SO4:PEG-400 droplet was 77.8 ± 1.8% RH. The result is consistent with contact efflorescence of pure (NH4)2SO4 droplets by crystals of itself (Davis et al., 2015b) suggesting that the PEG-400 coating had minimal effect on contact efflorescence. The behavior of the

(NH4)2SO4:raffinose mixtures below 80% RH was similar to that of (NH4)2SO4:PEG-400 until the

RH was lowered to 45.5 ± 1.3% RH. For the experiments conducted above that RH, efflorescence of the (NH4)2SO4:raffinose droplets was observed. For experiments conducted at and below 45.5

± 1.3% RH no efflorescence was observed.

3.4: Discussion

3.4.1: PEG-400: Coating Thickness and Diffusion Times

To determine why the PEG-400 coating did not affected the efflorescence behavior of the aqueous (NH4)2SO4 core, the diffusion times through the coating were estimated. The left-hand plot in Figure 3.9 shows the estimated thickness of the PEG-400 coating in µm as a function of the droplet’s diameter (core plus coating) and humidity. The black dotted line represents an example trajectory for a droplet that begins as a 15 µm droplet at 80% RH as the humidity is decreased.

For a droplet of 1:1 by weight (NH4)2SO4:PEG-400, the thickness of the PEG-400 coating was

Figure 3.8: Summary of results for contact efflorescence of 1:1 (NH4)2SO4:Raffinose (top) or 1:1

(NH4)2SO4:PEG-400 (bottom) by a crystal of (NH4)2SO4.

Figure 3.9: A plot showing the calculated thickness of the PEG-400 coating (in µm) and diffusion times (in seconds) for a molecule of H2O or a heterogeneous nucleus that is 300nm in diameter as a function of RH and droplet diameter. The dotted line represents a trajectory for changing the

RH on a droplet that was 15 µm in diameter at 80% RH to begin with.

62 calculated as a function of relative humidity (RH) and the droplet size. First, we assume that the entire mass of (NH4)2SO4 (AS) was contained inside the aqueous core (Aq) and the entire mass of

PEG-400 (PEG) was contained inside the organic coating (Org) that surrounds the aqueous core.

The aqueous core is composed of (NH4)2SO4 and H2O, and the organic coating is composed of

PEG-400 and H2O. The assumption is not perfect as a study of a 1:1 (NH4)2SO4:PEG-400 droplet on a plate showed that small satellites of aqueous (NH4)2SO4 were observed inside the PEG-400 coating (Ciobanu et al., 2010). However, due to a lack of quantitative data on the satellites, they were not included. Finally, the mass of (NH4)2SO4 is equivalent to the mass of PEG-400.

푀푎푠푠퐴푆 = 푀푎푠푠푃퐸퐺 (퐸푞. 3.2)

We established the following to begin:

4 푉표푙 + 푉표푙 = 푉표푙 = 휋푅3 (퐸푞. 3.3) 퐴푞 푂푟푔 퐷푟표푝 3 퐷푟표푝

Where Vol is the Volume for the aqueous core, organic layer, or the entire droplet (Drop) and R is the radius. Volume is substituted as the quotient of mass by density:

푀푎푠푠퐴푞 푀푎푠푠푂푟푔 4 3 + = 휋푅퐷푟표푝 (퐸푞. 3.4) 퐷푒푛푠퐴푞 퐷푒푛푠푂푟푔 3

The mass of the aqueous core and the organic layer can be expressed by the following:

푀푎푠푠퐴푆 푀푎푠푠퐴푞 = (퐸푞. 3.5) 푊푡%퐴푆퐴푞

푀푎푠푠푃퐸퐺 푀푎푠푠푂푟푔 = (퐸푞. 3.6) 푊푡%푃퐸퐺푂푟푔

Where Wt%ASAq is the mass fraction of (NH4)2SO4 in the aqueous core, and Wt%PEGOrg is the mass fraction of PEG-400 in the organic layer. Substituting in equations 3.5 and 3.6 to equation

3.4 we get:

푀푎푠푠퐴푆 푀푎푠푠푃퐸퐺 4 3 + = 휋푅퐷푟표푝 (퐸푞. 3.7) 푊푡%퐴푆퐴푞퐷푒푛푠퐴푞 푊푡%푃퐸퐺푂푟푔퐷푒푛푠푂푟푔 3

Since the mass of (NH4)2SO4 and PEG-400 are equal we can substitute the mass of PEG-400 with that of (NH4)2SO4. The equation is then rearranged to solve for the mass of (NH4)2SO4.

푀푎푠푠퐴푆 푀푎푠푠퐴푆 4 3 + = 휋푅퐷푟표푝 (퐸푞. 3.8) 푊푡%퐴푆퐴푞퐷푒푛푠퐴푞 푊푡%푃퐸퐺푂푟푔퐷푒푛푠푂푟푔 3

1 1 4 3 푀푎푠푠퐴푆 ( + ) = 휋푅퐷푟표푝 (퐸푞. 3.9) 푊푡%퐴푆퐴푞퐷푒푛푠퐴푞 푊푡%푃퐸퐺푂푟푔퐷푒푛푠푂푟푔 3

−1 4 3 1 1 푀푎푠푠퐴푆 = 휋푅퐷푟표푝 ( + ) (퐸푞. 3.10) 3 푊푡%퐴푆퐴푞퐷푒푛푠퐴푞 푊푡%푃퐸퐺푂푟푔퐷푒푛푠푂푟푔

For a given RH, the density of the aqueous core and the mol fraction of (NH4)2SO4 in the aqueous core is obtained from the E-AIM model (Clegg et al., 1998). The mol fraction is then converted to the mass fraction to be used in Equation 3.10. For the organic layer, a study by Marcolli and

Kreiger (2006) gives the mass fraction of PEG-400 in the organic layer as a function RH to be used in Equation 3.10. The mass fraction is converted to a mol fraction and using data on the density of PEG-400 as a function of mol fraction in water from Han et al. (2008) the density of the organic layer is obtained. With data from literature and a chosen radius of a droplet, the mass of

(NH4)2SO4 is calculated. Then the volume of the aqueous core is calculated as the following:

푀푎푠푠퐴푆 = 푉표푙퐴푞 (퐸푞. 3.11) 푊푡%퐴푆퐴푞퐷푒푛푠퐴푞

From the volume the radius of the aqueous core is calculated as:

1 3푉표푙퐴푞 3 푅 = ( ) (퐸푞. 3.12) 퐴푞 4휋

Finally, the thickness is calculated as the difference between the radius of the droplet and the radius of the aqueous core:

푇ℎ𝑖푐푘푛푒푠푠 = 푅퐷푟표푝 − 푅퐴푞 (퐸푞. 3.13 )

The process was repeated for droplet diameters between 10 and 20 µm in increments of 0.1 µm and for RH values between 35 and 90% RH in increments of 0.55% RH to obtain the left-hand plot in Figure 3.9. As seen in the figure, for droplets between 10 – 20 µm in diameter, the PEG-

400 coating is 1.2 – 2.3 µm in thickness. For a droplet being dried, as shown by a sample trajectory, the droplet diameter decreases while the organic layer’s thickness also thins. The decrease in droplet size and thickness of the organic coating is due to both (NH4)2SO4 and PEG-400 responding to changes in RH.

Based on the estimated thickness of the PEG-400 coating, diffusion times for a molecule of H2O with a molecular diameter of 2.75 Å and a heterogeneous nucleus with a diameter of 300 nm was calculated. The diffusion times were plotted as a function of RH and diameter of the droplet on the two plots on the right-hand side of Figure 3.9. The heterogeneous nucleus size was based on the size distribution of the crystalline (NH4)2SO4 measured in an SMPS (TSI 3010). The distribution is included as Figure A1 in the Appendix and shows that the mode diameter is approximately 300 nm. The diffusion time (τ) was calculated as:

푋2 휏 = (퐸푞. 3.14) 2퐷

Where X is the organic layer’s thickness, and D is the diffusion coefficient (cm2∙s-1). The diffusion coefficient for a molecule or particle traveling through the PEG-400 layer can be estimated as the following

푘 푇 퐷 = 퐵 (퐸푞. 3.15) 6휋휂푟

-23 -1 Where kB is the Boltzmann Constant (1.381∙10 J∙K ), T is temperature (K), η is the viscosity of the PEG-400 coating (Pa∙s), and r is the radius of the traveling molecule or particle. The viscosity of PEG-400 as a function of mole fraction of PEG-400 in water was obtained from Jerome et al.

(1968). A plot of PEG-400 viscosity and the resulting diffusion coefficients for a water molecule and a 300 nm heterogeneous nucleus is shown in Figure 3.10. As seen in the right-hand side of

Figure 3.9, it takes a fraction of a second for water to travel through the PEG-400 layer. The rapid diffusion time for a molecule of water suggests that the coating of PEG-400 is not viscous enough to prevent water from leaving the aqueous core upon efflorescence and thus does not inhibit efflorescence of (NH4)2SO4. The diffusion times for a heterogeneous nucleus such as the crystalline particles of (NH4)2SO4 to travel from the outside to the inner core is much slower, taking tens of seconds to approximately two minutes. This timescale is much longer than the experimental observation of seconds for the droplet of (NH4)2SO4:PEG-400 to effloresce upon collision.

To understand the discrepancy, we measured the immersion time for a crystal of (NH4)2SO4 of ~15 µm in diameter into a droplet of PEG-400 that is ~30 µm in diameter at 30% RH in the

EDB. It took the crystal less than 0.1 s to become fully submerged into the PEG-400 suggesting forces other than diffusion can cause the heterogeneous nucleus to become immersed into the PEG-

400 layer and reach the (NH4)2SO4 core. Additionally, the satellite pockets of aqueous (NH4)2SO4 that are dispersed throughout the PEG-400 layer could also play a role in contact efflorescence.

For example, the crystalline (NH4)2SO4 could contact a satellite first, causing efflorescence that propagates through the shell to the core.

Figure 3.10: A plot of Viscosity of PEG-400 (black) and the calculated diffusion coefficient of

H2O (dotted red) and a 300 nm sized heterogeneous nucleus (solid red) as a function of RH.

3.4.2: Raffinose: Viscosity and Diffusion

Raffinose inhibited effloresce of (NH4)2SO4 both homogeneously and during contact efflorescence at low RH. Previous studies have shown that the viscosity of pure raffinose increases as humidity is lowered (Grayson et al., 2017). Specifically, the viscosity of pure raffinose increases an order of magnitude for every 5% decrease in RH (See Figure A2 in the Appendix) to a point that at ~40% RH the viscosity was too high for an accurate measurement to be made

(Grayson et al., 2017). The viscosity of a 1:1 mixture of (NH4)2SO4:raffinose was determined in the EDB by measuring the coalescence time for two droplets of 1:1 (NH4)2SO4:raffinose to merge into one droplet. The results for the viscosities as a function of RH are plotted in Figure 3.11.

Since the linear fit for the viscosities on a log scale as a function of RH for pure raffinose on a log scale was a good fit for RH up to 100% RH, the linear fit for the (NH4)2SO4:raffinose was also extended to RH values higher than the tested RH. Based on the fit, the diffusion coefficient for

2- molecules of H2O and SO4 (ionic radius = 0.242 nm (Marcus, 1988)) were determined and are also included in Figure 3.11. We note that at low water activity and high viscosity, the Stokes-

Einstein relationship may under-predict diffusion of water (Price et al., 2016), and thus the values shown here likely represent the lower limit on D, and thus an upper limit on timescales. The

+ diffusion coefficient for NH4 is not shown since its ionic radius (0.154 nm (Sidey, 2016)) is

2- 2- smaller than SO4 and thus SO4 diffuses slower and would be the limiting factor to crystallization.

As shown in the Figure 3.11, when the mixed droplet reaches an RH of 35% RH, the homogeneous

6 ERH of (NH4)2SO4, the viscosity has already reached 10 Pa∙s. The high viscosity prevents the

(NH4)2SO4 from crystallizing at 35% RH and thus the 1:1 (NH4)2SO4:raffinose aerosol does not effloresce homogeneously at room temperature.

Figure 3.11: Measured viscosities of 1:1 (NH4)2SO4:Raffinose droplets at various RH values and

2- an extended fit for viscosity up to 65% RH. Diffusion coefficients for molecules of H2O and SO4 are shown on the right axis. The extended portion of the fit are expressed as dotted lines.

Contact efflorescence experiments showed that the mixed droplet of (NH4)2SO4:raffinose began resisting contact efflorescence between 45.5 ± 1.3% RH and 53.9 ± 1.2% RH. The viscosities at 45.5 and 53.9% RH are 1.2∙104 and 450 Pa∙s, respectively, which suggests that the viscosity threshold necessary for a droplet to undergo efflorescence is somewhere between those

2- limits. Calculated diffusion time are shown in Table 3.1 for a molecule of H2O and ion of SO4 to diffuse distances of 7.5 µm and 1 µm in a droplet of 1:1 (NH4)2SO4:raffinose at 45.5% and

53.9% RH. The 7.5 µm distance represents the time it takes the molecule to travel out from the center of a droplet that is 15 µm in diameter. The 1 µm diffusion time represents the time for the molecules and ion to move through the outer most layer of the droplet. Efflorescence is initiated by movement of ions from the aqueous to crystalline phase, followed by loss of the water that had been solvating the ions. However, as seen in the table for the droplet of 1:1 (NH4)2SO4:raffinose at 45.5%, movement of ions and water is significantly hindered. For a water molecule to travel 1

µm it would take approximate 1 hour and approximately 2.5 days to travel through 7.5 µm. The viscosity is thus too high for ions to diffuse to the heterogeneous nucleus and for water to evaporate out of the droplet, thus inhibiting efflorescence. For the droplet at 53.9% RH where contact efflorescence was observed, the times for H2O to travel 1 µm and 7.5 µm are 2.5 min and 2.3 hrs respectively. Based on diffusion times, water near the surface of the droplet could evaporate and cause efflorescence, while the water near the center of the droplet probably could not. The observed loss of linearity in the Mie scattering pattern for the (NH4)2SO4:raffinose at 53.9% RH could be caused by the outer surface crystalizing while trapping the remainder of the water in its core. The surface enhancement of sucrose for a mixed particle of sucrose and (NH4)2SO4 observed in a study by Chu and Chan (2017), could be a similar phenomenon, where the surface has lost water creating a crust like feature and trapping the components in the interior of the particle.

2- Table 3.1: Diffusion times for an H2O molecule and an SO4 ion traveling 1 µm or 7.5 µm through a droplet of 1:1 (NH4)2SO4:raffinose at 45.5% and 53.9%RH. 45.5 % RH 53.9% RH 1.2∙104 Pa∙s 450 Pa∙s 2- 2- H2O SO4 H2O SO4

D (cm2/s) 1.3∙10-12 7.4∙10-13 3.4∙10-11 1.9∙10-11

τ (1 µm) 1.1 hrs 1.9 hrs 2.5 min 4.4 min

τ (7.5 µm) ~60 hrs ~106 hrs 2.3 hrs 4.1 hrs

3.5: Conclusion

Internally mixed aerosols composed of both organic and inorganic aerosols are commonly believed to be in the liquid state due to the incorporation of the organic fraction. The assumption has been revisited as studies on organic glasses (Bodsworth et al., 2010; Derieux et al., 2018;

Grayson et al., 2017, 2016; Koop et al., 2011; Petters et al., 2019; Robinson et al., 2014; Shiraiwa et al., 2017; Zobrist et al., 2008) and heterogeneous efflorescence of aerosols (Davis et al., 2015b,

2017; Martin et al., 2001; Ushijima et al., 2018) have shown other processes that can affect the phase state of the aerosol. In this study we examined the effect of organics on the efflorescence of (NH4)2SO4 by mixing with two model organics: PEG-400 and raffinose. PEG-400 phase separates from the aqueous (NH4)2SO4 and creates an organic coating. Raffinose can transition into a glass at room temperature upon lowering RH. By contacting these mixed droplets with a crystal of (NH4)2SO4 the study showed that viscosity is more important than the morphology of the mixed droplet. The organic coating of PEG-400 seemed to have little to no effect on efflorescence of (NH4)2SO4. By calculating the thickness of the coating and using measured viscosities of PEG-400 from literature we showed that the organic coating does not act as a barrier for the transfer of water molecules out from the aqueous core at room temperature. Additionally, the PEG-400 coating also allowed heterogeneous nucleus to pass through with ease, allowing heterogeneous efflorescence to occur. Unless the heterogeneous nucleus is chemically transformed by the interaction with the organic layer, it is unlikely that an LLPS aerosol whose outer coating is not highly viscous will have any effect on homogeneous or heterogeneous efflorescence. Raffinose on the other hand was highly effective at inhibiting the efflorescence of

(NH4)2SO4. The glassy organic caused the viscosity of the mixed aerosol to be highly sensitive to

RH, where ~10% RH difference would result in changes to the diffusion coefficient and diffusion

72 times by ~2 orders of magnitude. In the middle and upper troposphere the where aerosols are believed to be mostly glassy (Shiraiwa et al., 2017) homogeneous and heterogeneous efflorescence are likely to have little impact. However, in the lower troposphere where the aerosols phase will be based on RH (Shiraiwa et al., 2017), homogeneous and heterogeneous efflorescence can still be a major process affecting to the phase state of aerosols.

Chapter 4: Probing Heterogeneous Efflorescence of Mars Relevant Salts with

an Optical Levitator

4.1: Introduction

In the previous two Chapters, heterogeneous efflorescence of aerosols in Earth’s atmosphere was investigated. In this Chapter, heterogeneous efflorescence of brines on the surface and in the subsurface of Mars is studied. As mentioned in Chapter 1, liquid water is currently the only biochemical solvent that is known to be used by life (Cockell et al., 2016). Although pure liquid water may be unstable on present day Mars, it has been hypothesized that liquid water may be stable in the form of a brine (Gough et al., 2016, 2011, 2014; Martín-Torres et al., 2015;

Martínez & Renno, 2013; McEwen et al., 2011; Nuding et al., 2015, 2014; Ojha et al., 2015; Primm et al., 2017, 2018; Toner et al. 2014; Zorzano et al., 2009). The salt and water stability diagram include three general phases: brine ice, liquid brine, and crystalline salt, possibly hydrated. Here we focus on the transitions between the liquid brine and the crystalline phase. Laboratory studies on the phase transitions of salts believed to be on the surface of Mars have been conducted at Mars relevant temperatures and RH using a droplet on a plate technique (Gough et al., 2011, 2014, 2016;

Nuding et al., 2015, 2014; Primm et al., 2017, 2018). When the results are compared to modeled results of temperature and RH at the surface of Mars, the brines could be stable or metastable during certain times of a Martian year and sol (Gough et al., 2011, 2014; Martínez & Renno, 2013;

Nuding et al., 2015, 2014; Primm et al., 2017). In the subsurface where temperature conditions are more suitable for brine stability, brines could be present for much longer (Nuding et al., 2014).

Additionally, a study on bulk samples of Mars relevant salts also showed their supercooling capabilities, further supporting brine stability on Mars’ surface and subsurface (Toner et al., 2014;

Zorzano et al., 2009). However, on the surface and subsurface, the brines will inevitably interact

74 with the Martian regolith, where soil particles could act as a heterogeneous nucleus and potentially induce efflorescence of a brine at a higher humidity than the homogeneous ERH.

On Mars, immersion mode efflorescence could affect all brines that form on the surface or in the subsurface when salts and mineral are internally mixed. Contact efflorescence could occur near the surface of Mars where dust fall is common, and at the liquid front of a flowing brine at the surface or subsurface. Laboratory studies performed on heterogeneous efflorescence (Davis et al., 2015a, 2015b; Han et al., 2002; Martin et al., 2001; Pant et al., 2006; Primm et al., 2018;

Ushijima et al., 2018) show that the effect of a heterogeneous nucleus on efflorescence varies greatly based on the brine and nucleus pair. Some pairs, such as Mg(ClO4)2 and montmorillonite, show minimal to no effect on the brine stability (Primm et al., 2018) when droplets on a plate are investigated. In contrast, other pairs, such as sodium chloride (NaCl) and montmorillonite, result in a significant decrease in brine stability (Ushijima et al., 2018). Studies with immersed particles of heterogeneous nuclei have also shown that the effect on brine freezing temperature varies with brine and nucleus pair (Primm et al., 2018; Toner et al., 2014; Zuberi et al., 2002). Unfortunately, no comprehensive model that can explain the various results exists and thus we must rely on laboratory experiments. An additional observation that has yet to be explained theoretically is the observation that the same nucleus can be more effective in contact nucleation than immersion nucleation (Davis & Tolbert, 2017; Ladino Moreno et al., 2013; Ushijima et al., 2018). The higher efficiency of contact nucleation may disproportionately affect brines near the Martian surface, making the exposed surface less ideal for brine meta-stability than previously believed. Here, we use a novel optical levitation experiment to probe the effect of heterogeneous efflorescence on four

Mars relevant salts by a series of heterogeneous nuclei.

4.1.1: Background and Choice of System for Study

The Martian regolith has been characterized remotely and at the surface through various techniques such as infrared spectroscopy, mass spectrometry, gas chromatography, and X-ray diffraction (XRD) (Bibring et al., 2006, 2005; Blake et al., 2012; Ehlmann & Edwards, 2014;

Ehlmann et al., 2011; Poulet et al., 2005; Vaniman et al., 2014). The upper layer of the Martian crust is mostly basaltic rock, composed of various mineral groups such as olivines, plagioclase feldspar, orthopyroxenes, and silicate glasses. Depending on the region, the actual composition of the basaltic crust can vary. For example, the southern highlands have a higher proportion of plagioclase, while the northern plains have more silicate phases (Ehlmann & Edwards, 2014). In addition to the basaltic species, other secondary species, such as phyllosilicates or clay minerals, sulfates, and oxides, have been measured in localized regions. For example, clay minerals have been observed at Gale Crater where Mars’ earliest terrain has become exposed (Ehlmann et al.,

2011; Vaniman et al., 2014). These clay minerals are believed to have formed early in Mars’ history during which nonacidic aqueous environments interacted with the basaltic crust to alter their mineralogy (Bibring et al., 2006; Vaniman et al., 2014). In their paper from 2006, Bibring et al. termed it the phyllosian period to indicate the formation of clay minerals during this time. It is unclear how these different minerals may impact the stability of brines on current Mars. To examine the effect of mineral particles on efflorescence, two mineral samples were chosen as heterogeneous nuclei in our experiments: montmorillonite and Mojave Mars Simulant (MMS).

Montmorillonite is a clay mineral, while MMS is a Mars soil analogue that is collected from crushing basaltic boulders from the Mojave Desert (Peters et al., 2008). Montmorillonite belongs to the smectite group, which has been detected on the surface of Mars (Poulet et al., 2005; Vaniman

76 et al., 2014), and MMS has similar chemical and physical characteristics to Mars dust (Peters et al., 2008).

The stability of brines for four salts that are believed to be present on the surface and in the subsurface of Mars was investigated using the optical levitator. The four brines were Mg(ClO4)2 magnesium chloride (MgCl2), calcium perchlorate (Ca(ClO4)2), and calcium chloride (CaCl2).

These four salts were chosen due to the actual detection of the salt itself or the individual ions having been detected in the Martian soil (Glavin et al., 2013; Hecht et al., 2009; Ojha et al., 2015;

Osterloo et al., 2010). Perchlorate anion has been found by both the Wet Chemistry Laboratory aboard the Phoenix lander and the Sample Analysis at Mars instrument aboard the Curiosity Rover at Gale Crater (Glavin et al., 2013; Hecht et al., 2009). The Phoenix lander also measured cations and found magnesium and sodium to be the dominant species and calcium was detected at lower concentrations (Hecht et al., 2009). The Thermal Emission Imaging System (THEMIS) aboard the Mars Odyssey orbiter has detected signals that are believed to be from chloride-bearing materials (Osterloo et al., 2010).

Contact efflorescence of aqueous Mg(ClO4)2 by crystalline Mg(ClO4)2, NaCl, and ammonium sulfate ((NH4)2SO4) was first investigated. As these salts are soluble, immersion efflorescence was not possible and thus was not investigated. The crystal of Mg(ClO4)2 was chosen because it should act as a perfect crystal seed and be highly efficient at inducing heterogeneous nucleation. Past studies using the optical levitator have shown that seeded crystal growth will occur at RH values very near the DRH, effectively shutting down hysteresis and the potential for supersaturated brines (Davis et al., 2015a; Davis et al., 2015b). On Mars, the interaction between a brine and crystal of the same salt could occur if crystals from a dryer region were carried by wind to an area where its brine had formed. NaCl, which is also expected to be found on Mars (Hecht

77 et al., 2009; Osterloo et al., 2010), was chosen because it maintains its crystalline structure at higher RH values and has a different crystal lattice structure and shape. While (NH4)2SO4 is not thought to be a constituent of the Martian surface, it was chosen because it has a more similar crystalline structure to Mg(ClO4)2 than does NaCl. The effect of crystal structure on heterogeneous efflorescence will be analyzed based on the crystal lattice constants of the heterogeneous nuclei.

In addition to the soluble heterogeneous nuclei, Mg(ClO4)2 was exposed to montmorillonite and

MMS to observe the effect of the insoluble heterogeneous nuclei in both contact and immersion efflorescence. For MgCl2, only immersion efflorescence by montmorillonite was investigated as the contact ERH of MgCl2 by montmorillonite has already been examined and reported in a previous study using the optical levitator as seen in Chapter 2 (Ushijima et al., 2018). Finally, contact and immersion efflorescence of the two remaining brines of Ca(ClO4)2 and CaCl2 by montmorillonite were investigated.

4.2: Experimentation

4.2.1: Optical Levitation and Flow Cell Arrangement

The optical levitator used in this study has been described in detail previously (Davis et al.,

2015b) and thus will only be briefly discussed here. Droplets of aqueous salts are trapped by two vertical counterpropagating laser beams from a 532 nm Nd:YAG source that has been split in half.

The two beams enter a closed flow cell with multiple windows for observation. The beam entering the flow cell from below has a Gaussian profile, while the beam from above has a Bessel profile.

The laser beams exert a scattering force in the direction of the laser propagation. In addition, a gradient force toward higher intensity moves the particle toward the center of the beam. The scattering forces from the two counterpropagating lasers, the force of gravity, and the drag force

78 from air flowing up through the flow cell balance out in the vertical direction. The gradient forces of the lasers stabilize the trap by keeping the trapped particle on axis.

The air flowing through the cell consists of two streams of pure nitrogen gas controlled with mass flow controllers. The first stream flows through a bubbler containing HPLC-grade water to humidify the flow. The second stream flows through a diffusion drier and is then mixed with the humidified nitrogen prior to entering the cell. By adjusting the amounts of humidified and dry flows, the RH inside the flow cell is controlled. RH and temperature are measured by individual probes (Vaisala HMP 60) at the inlet and outlet of the cell. The RH inside the flow cell is calculated as the average of the two probes (±1 S.D.).

To create heterogeneous nuclei, the dry flow is diverted toward a nebulizer (Omron NE-

U22) filled with 3−4 mL of an aqueous solution or a suspension of mineral particles in water. The nebulizer creates a mist that is carried by the nitrogen gas into the diffusion drier to remove the water, forming heterogeneous nuclei of salt crystals or mineral particles. The gas carrying the heterogeneous nuclei mixes with the humidified flow and enters the flow cell where the nuclei interact with the trapped droplet. Forming heterogeneous nuclei of insoluble minerals from a mixture in water and drying have the potential to change the surface properties of the mineral particle due to a redistribution of the soluble portion of the mineral during the wetting (Garimella et al., 2014). However, the effect of any change in mineral surface properties on the heterogeneous efflorescence efficiency has not been studied. Further, minerals on Mars may have experienced such aging processes during their lifetime as they interact with brines or ice.

To manipulate the heterogeneous nuclei, the flow cell is attached to a slidable stage with mobility in both horizontal directions. The movable flow cell allows us to control where the heterogeneous nuclei flow relative to the position of the trapped droplet. The optics, such as

79 mirrors and lenses, that guide the laser into the flow cell are not moved and thus the trapped droplet does not move. However, by moving the flow cell, the particle stream of heterogeneous nuclei can be adjusted to be aligned or misaligned with the lasers. When the particle stream is misaligned, no collisions between heterogeneous nuclei and droplet occur and when aligned collisions occur.

4.2.2. Generating Droplets of Aqueous Solution

A droplet generator (Microfab MJ-APB-20) was used to produce droplets of aqueous solutions to be trapped. For droplet generation, solutions of 5 wt % Mg(ClO4)2 (Sigma-Aldrich),

MgCl2 (Mallinckrodt), Ca(ClO4)2 (Sigma-Aldrich), and CaCl2 (Fisher Scientific) were first filtered through a 0.22 μm pore nylon filter. The droplet generator was then filled with the solution and inserted into a side port near the top of the flow cell. The tip of the droplet generator is made of a glass capillary with a 20 μm orifice. An alternating positive and negative voltage on a ring made of a piezoelectric material near the tip of the glass capillary pushes droplets out into the flow cell to be caught by the lasers. After equilibration, the typical size of the trapped droplet is 10−15 μm in diameter.

4.2.3. Generating Contact Nuclei

For generation of soluble heterogeneous nuclei, aqueous solutions of 10 wt % Mg(ClO4)2,

NaCl (Mallinckrodt), and (NH4)2SO4 (Sigma-Aldrich) in HPLC-grade water were prepared. For nonsoluble mineral particles, montmorillonite (SWy-2b) was obtained from the Clay Mineral

Society and MMS was obtained from the Jet Propulsion Lab. A mixture of the mineral in HPLC grade water was stirred with a magnetic stirrer for 30 min to make a slurry of suspended minerals in water. Because the nebulizer utilizes vibrating mesh technology, to prevent the mesh from becoming clogged with mineral particles, the largest particles were settled out of solution before

80 use. Settling velocities (w (cm/s)) of mineral particles in water were calculated based on Stokes’ law

2 2(휌푝 − 휌푤)𝑔푟 푤 = (퐸푞. 4.1) 9휇

3 Here, ρp and ρw are the bulk densities (g/cm ) of the mineral particle and water, respectively, g is the acceleration of gravity (9.8 m/s2), r is the particle radius (cm), and μ is the dynamic viscosity of water (Pa·s). Densities of 2.2 and 2.9 g/cm3 were used for montmorillonite and MMS, respectively. Values of density and viscosity for water at 20 °C were 0.998 g/cm3 and 1.002 Pa·s, respectively. Particles larger than 1 μm in diameter settle out of a 3 cm column in 13 and 8 hrs for montmorillonite and MMS, respectively. The particle size distribution of the resulting stream of soluble and insoluble heterogeneous nuclei was measured using a scanning mobility particle sizer

(SMPS TSI model 3010). The average size distributions from three samples from each of the heterogeneous nuclei types are shown in Figure A1 in the Appendix. The mode diameters of the heterogeneous nuclei ranged from 200 to 400 nm.

4.2.4. Imaging Efflorescence and Collisions

To monitor collisions between the levitated droplet and heterogeneous nuclei and to detect efflorescence, two charge-coupled device (CCD) cameras are placed outside the windows of the flow cell, oriented perpendicular to the laser axis. Far-field scattering of the trapping laser as well as the bright-field image taken by adding an LED light source is captured by the cameras. The two

CCDs are controlled with a LabView program that records, stores, and replays video files. When observed in far field, the light scattering is used to determine the phase state of the levitated droplet.

As seen in Figure 4.1, when the levitated droplet is liquid, a Mie scattering pattern with alternating bright and dark bands occurs. The relative intensity and spacing of the bands depend on the size

Figure 4.1: Images of Mg(ClO4)2 liquid droplet and crystal in bright field and far field. Brightness and contrast of the bright field images have been increased to better distinguish the two images.

82 and refractive index of the droplet and change as RH changes. Larger particles have closer band spacings than smaller particles. Once the droplet has effloresced and become crystalline, the alternating bands disappear, and the scattering pattern becomes mosaic. As discussed in previous studies with the optical levitator, the loss of the linear Mie scattering pattern is consistent with crystallization (Davis et al., 2015a; Davis et al., 2015b). Further, as the trapped crystal constantly rotates in the trap, the scattering pattern changes with time as different crystal faces interact with the laser. Bright-field images of the crystal can also sometimes be used to determine efflorescence.

As seen in Figure 4.1, the liquid droplet appears round and homogeneous. In contrast, while the overall shape of this particular crystalline particle is spherical, inhomogeneities due to light passing through the crystal facets are observed.

The cameras are also used to identify collisions between the trapped droplet and the incoming heterogeneous nucleus. While there are several ways to determine whether a collision occurs in the trap, here we show an example that only uses farfield imaging. Figure 4.2 shows an experiment where a droplet of Mg(ClO4)2 comes into contact with two different heterogeneous nuclei, crystalline Mg(ClO4)2 and NaCl, at a similar RH but with different outcomes. In panel A, crystalline Mg(ClO4)2 comes in from below and collides with the droplet inducing immediate efflorescence, as indicated by the change in the light scattering pattern. As the effloresced particle is stabilized in the trap, it rises in the trap at a higher position as a result of losing water and thus mass. The rise in the trapping position is also used as an additional indicator of efflorescence. In panel B, a crystal of NaCl collides with the droplet of Mg(ClO4)2, but unlike the previous experiment, it does not cause efflorescence. The crystal becomes immersed into the droplet and dissolves away without changing the light scattering or the trapping position. The diffused light coming from the bottom is the light scattered off the heterogeneous nucleus. The scattered light

Figure 4.2: Sequence of images from a contact efflorescence experiment of Mg(ClO4)2 by a crystal of A.) Mg(ClO4)2 and B.) NaCl. Both experiments were conducted at approximately 22%

RH. The red solid line is a reference for the vertical position of the droplet. The red arrows on the first frames point to the heterogeneous nucleus.

84 from the heterogeneous nucleus and levitated droplet appear to be the same size because we are observing in the far field and the scattered images are cut off by the windows on the flow cell. The circular windows cut off the scattered light; thus, all of the far-field images are circular and of the same size due to the window. However, the small heterogeneous nucleus has only one apparent band, while the large droplet has a scattering pattern with many bands.

4.2.5: Determining Contact and Immersion Heterogeneous ERH

Two separate experimental methods were utilized to determine contact and immersion

ERH. For contact efflorescence, droplets of the brine were trapped at an RH between the homogeneous ERH and DRH. Each trial consisted of exposing the trapped droplet to a stream of the dried heterogeneous nucleus for either 25 s for the soluble salts or 60 s for the insoluble mineral particles at a constant RH. The difference in exposure time was due to the difference in collision rates between insoluble and soluble heterogeneous nuclei, possibly arising from the different methods used to create nebulized particles. The soluble heterogeneous nuclei on average would collide with the trapped droplet every 4.2 ± 1.8 s, while the insoluble mineral particles would collide every 9.5 ± 1.6 s. Thus, the trapped droplet would be exposed to approximately six collisions totally per trial for both insoluble and soluble heterogeneous nuclei. However, the number of collisions may be an underestimate as heterogeneous nuclei scatter less light as their size decreases. Thus, a heterogeneous nucleus that is too small may not scatter ample light to be observed on the CCD camera. The droplet was monitored for whether it effloresced during the exposure period. If the droplet did not effloresce, then the droplet was ejected from the trap and a new droplet was trapped for the next trial. The process was repeated to calculate a probability of efflorescence (Peff) as the ratio of the observed efflorescence events to the number trials as a function of RH. Contact ERH was then reported as the RH, where Peff was equal to 0.5.

For immersion efflorescence, a droplet of the brine at an RH significantly higher than homogeneous ERH and the contact ERH was exposed to a stream of heterogeneous nuclei. The humid conditions ensured that the droplet did not undergo contact efflorescence during the process of immersing the mineral particle into the brine. The droplet was monitored until a collision with a mineral particle was visually observed and was then isolated from further collisions by adjusting the slidable mount to direct the particles away from the droplet. A video recording from the CCDs was reviewed to confirm a collision. If multiple collisions were observed, the number of collisions was kept to a maximum of three. Droplets with more than three collisions were expelled from the trap and a new droplet was caught. The RH inside the flow cell was then lowered at a rate of ≤1%

RH/min until efflorescence occurred. Immersion ERH was reported as the average RH (±1 S.D.) of all of the RH values where immersion efflorescence was observed.

4.3: Results

4.3.1: Homogeneous Efflorescence of Brines

Examples of the homogeneous phase transition of all four salts at room temperature are shown in Figure 4.3. In Figure 4.3A, we begin with a droplet of aqueous Mg(ClO4)2 at a humidity of 49.4% RH at the top right of the panel along the red curve. As the humidity is lowered, the droplet slowly loses water but remains a liquid as seen by the scattering pattern. The change in size due to loss of water is also seen through the changes to the Mie scattering peaks. At 12.1%

RH, the droplet effloresces, signified by the abrupt change in scattering pattern shown by two images taken immediately before and after efflorescence. The trapped crystal is subsequently humidified (blue line), but the scattering pattern does not change until 39.7% RH when the Mie scattering returns, signifying a deliquescence event. The average homogeneous ERH and DRH

86 values for Mg(ClO4)2 across all experiments conducted were 13.0 ± 0.5 and 39.7 ± 0.7% RH, respectively.

In contrast to Mg(ClO4)2, homogeneous efflorescence of Ca(ClO4)2 was not observed even when dried down to a humidity of 3.1% RH. This is shown in Figure 4.3B where a droplet of

Ca(ClO4)2 begins at a humidity of 28.8% RH and is dried along the red curve. As RH is lowered, the Ca(ClO4)2 droplet loses water as seen in the changes in Mie scattering, similar to Mg(ClO4)2.

However, as the humidity begins to reach the low single digits, the linear Mie scattering pattern remains indicating that the droplet has not effloresced. In all experiments conducted, dehumidification of Ca(ClO4)2 in our optical trap did not cause efflorescence. This study cannot rule out efflorescence occurring below 3.1% RH. However, a previous study that examined

Ca(ClO4)2 showed that the brine did not homogeneously effloresce at RH <1% at room temperature (Nuding et al., 2014).

The homogeneous phase transition of droplets of MgCl2 and CaCl2 is shown in Figure

4.3C,D, respectively. On average, MgCl2 undergoes efflorescence and deliquescence at 3.7 ± 0.4 and 13.7 ± 0.5% RH, respectively. Similar to Ca(ClO4)2, CaCl2 does not homogeneously effloresce when dried and thus the DRH could not be measured. However, at low RH, interferences in the scattering pattern appear, which could indicate changes to the morphology of the droplet or localized changes in the refractive index of the droplet. For our study, the homogeneous DRH value at room temperature was estimated to be 15% RH based on a study that utilizes the droplet on a plate technique coupled with Raman microscopy (Gough et al., 2016).

4.3.2. Heterogeneous Efflorescence of Brines

The probability of efflorescence for Mg(ClO4)2 in contact with the various heterogeneous nuclei is shown in Figure 4.4. The Peff for each heterogeneous nucleus, with the exception of

Figure 4.3: Homogeneous hysteresis of A.) Mg(ClO4)2, B.) Ca(ClO4)2, C.) MgCl2, and D.) CaCl2.

The red line indicates dehumidification and the blue line indicates humidification. Scattering images in the far field are shown for various humidities along hysteresis. The liquid water content and relative humidity are not to scale.

MMS, was fit with a sigmoidal curve that was bound between 0 and 1. From the fitted curves, the half-max value (Peff = 0.5) with a 90% confidence band was reported as the contact ERH.

Heterogeneous nuclei that are efficient at inducing efflorescence of Mg(ClO4)2 will have a higher contact ERH closer to the homogeneous DRH of 39.7% RH; thus, the contact efflorescence efficiency increases for the curves from left to right on the figure. From Figure 4.4, it can be seen that a crystal of Mg(ClO4)2 is the most effective heterogeneous nucleus. Because the crystal has an identical lattice structure to the nucleating material, crystalline Mg(ClO4)2 simply acts as a seed for further crystallization and thus it is not surprising that it is the most effective heterogeneous nucleus causing efflorescence at 37.7 ± 1.2% RH. Since the contact ERH is close to the homogeneous DRH, a crystal of Mg(ClO4)2 will prevent any supersaturation of Mg(ClO4)2 brine it contacts. The next two most effective heterogeneous nuclei were (NH4)2SO4 and NaCl with a contact ERH of 30.2 ± 0.6 and 19.3 ± 1.8% RH, respectively. The least effective heterogeneous nuclei were Na-montmorillonite and MMS. Na-montmorillonite induced contact efflorescence at

14.8 ± 1.0% RH, while MMS never induced efflorescence at the RH tested. Thus, an upper limit for contact ERH of Mg(ClO4)2 by MMS was chosen as the average RH (±1 S.D.) of the lowest set of RH tested at 15.1 ± 0.3% RH. When compared to the homogeneous ERH of Mg(ClO4)2 at 13.0%

RH, neither Na-montmorillonite nor MMS seems to significantly increase the ERH of Mg(ClO4)2.

A summary of the results for all heterogeneous efflorescence experiments for Mg(ClO4)2 is shown in Figure 4.5. In addition to the contact ERH values, the immersion efflorescence values are also included for the non-soluble heterogeneous nuclei. For Na-montmorillonite, the immersion ERH was 14.4 ± 1.5% RH, and for MMS, it was 12.7 ± 0.6% RH. As seen in the figure,

Na-montmorillonite seems to have minimal effect on the efflorescence of Mg(ClO4)2 with both

Figure 4.4: Contact efflorescence experiment results for Mg(ClO4)2 in contact with Mg(ClO4)2

(●), (NH4)2SO4 (▼), NaCl (▲), Na-montmorillonite (♦), and MMS(■). Solid lines are sigmoid curve fits.

Figure 4.5: Summary of results for heterogeneous efflorescence of Mg(ClO4)2 by Mg(ClO4)2,

NaCl, (NH4)2SO4, Na-montmorillonite (Swy-2b), and MMS. The solid boxes indicate contact

ERH, hatched boxes indicate immersion ERH. The two solid lines show the homogeneous ERH and DRH values for Mg(ClO4)2. The two dashed lines around the ERH line indicate the standard deviation in the homogeneous ERH as observed in this study. The *UL indicates that the value shown is the upper limit as contact efflorescence was never observed.

91 contact and immersion ERH not statistically different from the homogeneous ERH. Similarly, for

MMS, the contact ERH is shown as an upper limit that is already close to the homogeneous ERH.

Immersion efflorescence of Mg(ClO4)2 by MMS occurs at the homogeneous ERH within uncertainty. Thus, we conclude that at room temperature neither of these Mars soil analogues has a significant effect on the efflorescence of Mg(ClO4)2. The inactivity of the two insoluble Mars soil analogues as heterogeneous immersion nuclei for Mg(ClO4)2 was previously shown in a study performed on droplets on a plate at lower temperatures (Primm et al., 2018). The contact and immersion ERH for Na-montmorillonite with Mg(ClO4)2 were the same, within error, unlike other studies that have found contact efflorescence to be more effective than immersion (Ushijima et al.,

2018). Here, neither contact nor immersion nucleation was effective.

2018). Here, neither contact nor immersion nucleation was effective. The results for heterogeneous efflorescence of the remaining three brines MgCl2, Ca(ClO4)2, and CaCl2 by contact and immersion with Na-montmorillonite are shown in Figure 4.6. The contact ERH of MgCl2 at

10.9 ± 0.6% RH was reported in a previous study performed in our optical levitator (Ushijima et al., 2018), and the immersion ERH at 7.9 ± 2.7% RH is from the current study. Unlike Mg(ClO4)2, the efflorescence of MgCl2 is significantly influenced by the presence of Na-montmorillonite in

Figure 4.6: Summary of results for heterogeneous efflorescence experiments of A.) MgCl2, B.)

Ca(ClO4)2, and C.) CaCl2 by Na-montmorillonite. Solid boxes are for contact ERH values while hatched boxes are for immersion ERH values. Solid lines indicate homogeneous ERH and DRH values for each salt. DRH values for Ca(ClO4)2 (Nuding et al., 2014) and CaCl2 (Gough et al.,

2016) were taken from literature. *UL represents an upper limit as heterogeneous efflorescence was never observed in those experiments.

Figure 4.7: Contact efflorescence experiments for Ca(ClO4)2 and CaCl2 by Na-montmorillonite.

94 both contact and immersion modes. Additionally, contact efflorescence of MgCl2 by Na- montmorillonite occurs at a higher RH than immersion efflorescence. For the two calcium salts, since homogeneous deliquescence was not observed, the DRH value shown was taken from the literature (Gough et al., 2016; Nuding et al., 2014). Contact efflorescence experimental results for

Ca(ClO4)2 and CaCl2 are shown in Figure 4.7. As seen in the figure, contact efflorescence of

Ca(ClO4)2 and CaCl2 by Na-montmorillonite was not observed. The lowest average RH values achieved for Ca(ClO4)2 and CaCl2 were 5.2 ± 0.5 and 4.9 ± 0.6% RH, respectively. Thus, these values also represent upper limits for contact ERH. Similarly, immersion efflorescence was not observed for the two calcium salts by Na-montmorillonite. The average of the five lowest RH values reached in immersion experiments for Ca(ClO4)2 and CaCl2 were 3.1 ± 0.9 and 2.7 ± 0.3%

RH, respectively. Thus, these values represent the upper limits for immersion ERH. However, it is still possible for contact and immersion efflorescence to occur at RH values below the reported values.

4.4: Discussion

4.4.1: Heterogeneous Efflorescence on Mars

Comparisons of the homogeneous phase diagrams for the four magnesium and calcium salts to temperature and humidity conditions on Mars have shown that some salt brines on Mars may be stable for certain periods of time during a Martian sol (Gough et al., 2011, 2014; Nuding et al., 2015, 2014; Primm et al., 2017). When the homogeneous phase transitions of Mg(ClO4)2 are compared to modeled surface temperature and humidity conditions at the Viking Lander 2 site,

Mg(ClO4)2 does not seem to deliquesce for any significant amount of time (Gough et al., 2011).

However, when the same diurnal trajectory is compared to the stability phase diagram of Ca(ClO4)2, brine could be stable or metastable for 3−4 hrs per sol (Nuding et al., 2014). Additionally, the

95 temperature and humidity conditions of the shallow subsurface were modeled at the Phoenix landing site, and when that trajectory is applied to the stability diagram of Ca(ClO4)2 it was shown that the stability period could be as long as 17 hrs (Nuding et al., 2014). The modeled shallow subsurface was significantly warmer and more humid, never dropping below 8% RH, which allowed the brine to be stable for a longer time. Since the homogeneous ERH of Mg(ClO4)2 is above 8% RH, efflorescence is likely to occur even in the subsurface. In contrast, MgCl2,

Ca(ClO4)2, and CaCl2 have a lower homogeneous ERH, allowing them to possibly avoid efflorescence and persist as metastable, supersaturated brines. These studies did not, however, consider the possibility of heterogeneous nucleation of the briny droplets by contact or immersion of Martian regolith materials. Our studies show that Na-montmorillonite can act as a heterogeneous nucleus for the efflorescence of MgCl2, increasing the heterogeneous ERH to 10.9 and 7.9% for contact and immersion ERH, respectively. The presence of montmorillonite could thus cause MgCl2 to effloresce in the shallow subsurface, cutting into the stability period of MgCl2 brine. For the two calcium salts, the inactivity of montmorillonite as a heterogeneous nucleus confirms the current expectation that their brines are metastable in the subsurface of Mars as the reported upper limits for heterogeneous ERH were not above 8%.

While the current study was conducted at room temperature, the average temperature of

Mars is 210 K. During the warmer months on Mars, the temperatures can reach up to 300 K at equatorial regions,7 but those high temperatures are accompanied by low RH values. Thus, the question arises as to the relevance of these results for Mars. However, past work has shown that homogeneous efflorescence is less dependent on temperature than deliquescence for these salts

(Gough et al., 2011, 2014; Nuding et al., 2015, 2014; Primm et al., 2017). For example, Gough et al. (2011) examined the homogeneous efflorescence of Mg(ClO4)2 and found minimal temperature

96 dependence over the temperature range of 223−270 K. Further, a study by Primm et al. (2018) over a similar temperature range examined immersion efflorescence of Mg(ClO4)2 by montmorillonite and MMS and again saw that efflorescence was independent of temperature. Thus, we expect a minimal temperature dependence for contact efflorescence as well, although this has not been proven.

4.4.2. Crystal Lattice Match

The effectiveness of foreign nuclei on crystal nucleation is often discussed in terms of the crystal lattice match between the two solids. It is believed that when a heterogeneous nucleus has a similar crystal lattice structure to the nucleating crystal, it will be more effective than the one that lacks a crystal lattice match (Mithen & Sear, 2014; van Meel et al., 2010). For efflorescence, a better crystal match would induce efflorescence at a relative humidity closer to the homogeneous

DRH (Davis et al., 2015b; Ushijima et al., 2018), thus preventing brine supersaturation. To compare the various heterogeneous nuclei’s crystal lattice to that of the effloresced crystal from the brine, the crystal lattice mismatch (δ) between their crystal faces was calculated:

푎1,퐻푁 − 푎1,퐶푟 푎2,퐻푁 − 푎2,퐶푟 | 푎 | + | 푎 | 훿 = 1,퐶푟 2,퐶푟 (퐸푞. 4.2) 2

Here, a1 and a2 are the two lattice constants that define the specific crystal face for the heterogeneous nucleus (HN) and the nucleating crystal (Cr). A crystal lattice mismatch of zero means a perfect match, as the value for crystal lattice mismatch increases, the more dissimilar the two crystal systems are. A crystal lattice was simplified to the three faces: the 100 face is represented by the b and c lattice constants, the 010 by the c and a, and the 001 by the a and b lattice constants. Each of these faces of the brine crystal was compared to that of the heterogeneous nuclei, resulting in a total of nine calculated crystal lattice mismatches for each pair. Of the nine calculated values, the lowest value was determined as the crystal lattice mismatch. The crystal 97 lattice mismatch between Mg(ClO4)2 and four of the heterogeneous nuclei (Mg(ClO4)2,

(NH4)2SO4, NaCl, and Na-montmorillonite) was calculated and compared to their respective contact ERH in Figure 4.8. MMS was excluded from this analysis because it is a mixture of minerals. The data was fit with a linear fit, and the resulting R-squared value is shown. A list of the lattice constants and crystal systems of the species analyzed are shown in Table A1 of the

Appendix. There is a mild negative correlation between contact ERH and lattice mismatch, implying that a heterogeneous nucleus with a more similar crystal lattice to the brine tends to be more effective at inducing contact mode efflorescence. The result is as expected and is also consistent with previous studies on heterogeneous efflorescence in the optical trap (Davis et al.,

2015b; Ushijima et al., 2018). Those studies found that when the lattice mismatch value was above

0.12, the heterogeneous nuclei were typically not effective at inducing contact efflorescence of the nucleating salt.

To predict the global effect of heterogeneous nuclei on brine stability, the crystal lattice mismatch between Mg(ClO4)2, MgCl2, Ca(ClO4)2, and CaCl2 and a set of minerals that have been detected either remotely or in situ were calculated and are tabulated in Table 4.1. The mismatch values were then designated a color based on the predicted effectiveness of the mineral in inducing efflorescence. Green (δ < 0.06) represents pairings where the heterogeneous nucleus is likely to impact brine stability, yellow (0.06 < δ < 0.12) represents minerals that may or may not impact the brine, and red (0.12 < δ) is unlikely to have an impact. As shown in the table, most of the brine and mineral pairings have a lattice mismatch above 0.12. Many of the clay minerals such as nontronite and montmorillonite have a lattice mismatch that is between 0.06 and 0.12 for the salts except for CaCl2. Additionally, of the four salts, MgCl2 has the lowest mismatch value with

Figure 4.8: Contact ERH of Mg(ClO4)2 compared to lattice mismatch (δ) with linear fit and R- squared values.

Table 4.1: Crystal lattice mismatch of select brines and minerals pairs.

Mg(ClO4)2 MgCl2 Ca(ClO4)2 CaCl2 ●6H2O ●4H2O ●4H2O ●6H2O Olivine 0.188 0.122 0.102 0.225 Orthopyroxene 0.175 0.237 0.14 0.226 Clinopyroxene 0.134 0.17 0.131 0.187 Feldspar 0.07 0.01 0.105 0.08 Hematite 0.032 0.277 0.136 0.318 Nontronite 0.088 0.099 0.097 0.201 Montmorillonite 0.084 0.08 0.099 0.199 Kaolinite 0.085 0.098 0.1 0.189 Chlorite 0.039 0.181 0.103 0.248 Illite 0.086 0.065 0.082 0.218 Prehnite 0.207 0.355 0.228 0.24 Quartz 0.186 0.336 0.195 0.279 Magnesite 0.117 0.361 0.227 0.292 Calcite 0.159 0.36 0.228 0.315 Kieserite 0.164 0.072 0.142 0.031 Gypsum 0.176 0.216 0.23 0.228

100 montmorillonite, which could explain why MgCl2 was the only brine whose stability was significantly decreased by montmorillonite. There are only four pairings that have a mismatch value below 0.06 with two of them being for contact with Mg(ClO4)2. At the surface of Mars,

Mg(ClO4)2 is already thought to not undergo deliquescence. However, in the shallow subsurface where Mg(ClO4)2 may transition into brine for short periods of time, the presence of hematite and chlorite may limit brine metastability. Kieserite (MgSO4), which has been detected by the

Opportunity rover and has been theorized to have an association with groundwater upwelling events,33 has a low mismatch value for CaCl2. Feldspar contacting MgCl2 is the final pair to have a mismatch value below 0.06 and is also the lowest mismatch value calculated from the selected list. Additionally, feldspar is the only mineral from the selected list that does not have a mismatch value above 0.12 with the four brines. As feldspar is believed to be globally distributed on Mars, its effect on brine stability may have the most impact.

4.5: Conclusion

On the surface of Mars, if droplets of a deliquesced salt were to form, they will likely interact with a heterogeneous nucleus such as other crystalline salts and mineral particles. Whether the heterogeneous nucleus will destabilize the brine and induce efflorescence at a significantly higher humidity than previously believed is dependent upon the combination of the aqueous salt and the nucleus. Some combinations, such as a crystalline Mg(ClO4)2 with a brine of itself or Na- montmorillonite with MgCl2 brine, can significantly lower the stability of the aqueous solution and will lower the probability or duration of brines on Mars. However, other combinations such as MMS with Mg(ClO4)2 and Na-montmorillonite with Mg(ClO4)2, Ca(ClO4)2, and CaCl2 did not impact the efflorescence when compared to homogeneous efflorescence. For the two calcium salts studied, the inactivity of montmorillonite as a heterogeneous nucleus is promising for liquid water

101 on Mars as these salts are theorized to be in the brine form for extended periods of time on Mars

(Gough et al., 2014; Nuding et al., 2014). However, to fully understand efflorescence and deliquescence on Mars’ surface and subsurface, a model for heterogeneous efflorescence is necessary as it is impractical to study every possible combination of brine or mineral in the laboratory. This work shows that crystal lattice is one factor that determines how effective a heterogeneous nucleus will be in inducing efflorescence, but crystal lattice is one of the many factors believed to drive heterogeneous nucleation. For example, the number of active sites is also believed to affect heterogeneous nucleation (Fletcher, 1969; Han et al., 2002; Martin et al., 2001).

Active sites are regions on a particle’s surface where nucleation is most likely to begin, believed to be formed from defects to the mineral’s structure. Studies that have examined the size effect of immersed metal oxide particles on efflorescence of (NH4)2SO4 have suggested that by increasing the surface area, the number of active sites were increased and thus a higher immersion ERH was observed (Han et al., 2002; Martin et al., 2001). For an accurate model of heterogeneous efflorescence, additional laboratory studies that probe and parameterize the many factors of heterogeneous nucleation will need to be conducted.

102

Chapter 5: Summary of Conclusions

Heterogeneous efflorescence of inorganic soluble salts relevant to both Earth and Mars were studied in the laboratory. The impact of mineral dust particles and organic compounds on the phase transition of salt aerosols in Earth’s atmosphere was examined. Additionally, the effect that Mars soil analogues have on the stability of salt brines believed to be on Mars were also studied. To investigate these systems of heterogeneous efflorescence, a room temperature optical levitation apparatus with humidity control was used. A single droplet of aqueous inorganic salt could be levitated in air and be exposed to heterogeneous nuclei of choice to investigate both contact and immersion efflorescence. The results and conclusions of the studies are summarized below.

In Chapters 2 and 3, the focus was on Earth’s atmosphere, with Chapter 2 investigating the effect that mineral particles of illite, Na-montmorillonite, and NX Illite have on the efflorescence of (NH4)2SO4 and NaCl. Both contact and immersion efflorescence of the two salts by the three mineral particles were examined. The study saw that mineral particles were able to heterogeneously effloresce the salts at higher RH than when compared to the homogeneous ERH.

For (NH4)2SO4 a particle of Illite increased the ERH the most and for NaCl particles of Na- montmorillonite increased ERH the most. For all pairs of salt and mineral particle, contact ERH was shown to be a higher value when compared to the immersion ERH of the same salt and mineral pair. The reported difference between contact and immersion efflorescence was the first of its kind for heterogeneous efflorescence by mineral particles to the extent of our knowledge. Analysis of the crystal lattice match between the salts and minerals showed that many of the pairs analyzed had a mismatch greater than 0.12 and does not seem to have a strong correlation between crystal lattice and heterogeneous ERH. Additionally, by increasing the number of immersed particles of

103 illite in (NH4)2SO4 the effect of active sites were investigated. The study showed that by increasing the total surface area of the immersed mineral particle, the immersion ERH could be increased.

However, neither crystal lattice nor active site could explain the observed difference between contact and immersion ERH. To address the difference, contact efflorescence of NH4Cl and

MgCl2 by Na-montmorillonite was also examined. Comparison of contact ERH for NH4Cl, NaCl,

2+ and MgCl2 showed that Mg , the cation with the highest hydration strength according to the

Hoffmeister series, was most affected by the negatively charged Na-montmorillonite particles.

The observed ion specific effect was proposed as an explanation for the difference in contact and immersion ERH. Finally, a model of collision lifetimes for various salt aerosols sizes and dust aerosol concentrations showed that collisions between salt aerosols and mineral particle were quite common especially for regions near a source such as the Saharan desert. Collisions were also shown to occur for regions further on days where dust emissions are higher than average.

In Chapter 3, Raman spectrometry was coupled to the optical levitation apparatus and the efflorescence behavior of (NH4)2SO4 mixed with equal weight of either PEG-400 or raffinose was studied. At room temperature, PEG-400 was known to LLPS from the aqueous (NH4)2SO4 and create a coating, whereas raffinose does not LLPS but was known to form a glass. The mixed

(NH4)2SO4:organic droplets were collided by crystalline (NH4)2SO4 to observe heterogeneous efflorescence. The organic coating of PEG-400 had minimal effect on effloresce, the observed homogeneous ERH was similar to that of pure (NH4)2SO4 and contact efflorescence of the mixed droplet by crystalline (NH4)2SO4 was observed for all RH tested between the ERH and DRH of pure (NH4)2SO4. The organic coating thickness as a function of RH and droplet size was estimated and based on the viscosities of PEG-400 at different RH a diffusion time for molecules of water and heterogeneous nuclei of 300 nm in diameter were calculated. The diffusion times for water

104 were fractions of a second and for the heterogeneous nuclei were tens of seconds. The calculated diffusion time also supported the conclusion that a non-viscous organic coating does not act as a physical barrier to either homogeneous or heterogeneous efflorescence. Raffinose on the other hand inhibited homogeneous efflorescence of the mixed droplet. Contact efflorescence was observed for the higher RH values but between 53.9 ± 1.2% RH and 45.5 ± 1.3% RH contact efflorescence no longer occurred. The viscosity of 1:1 (NH4)2SO4:raffinose at various RH showed that the viscosities at these RH values were 450 Pa∙s and 1.2∙104 Pa∙s respectively. The study suggests that the threshold viscosity for the aqueous fraction to effloresce from a mixed aerosol is somewhere between these viscosities.

In Chapter 4, the stability of brines believed to be on the surface and in the subsurface of

Mars was examined. Heterogeneous efflorescence of Mg(ClO4)2 by a crystal of itself, (NH4)2SO4,

NaCl, Na-montmorillonite and MMS was studied. As expected the crystal of Mg(ClO4)2 was most effective in inducing efflorescence of the Mg(ClO4)2. The two mineral particles, Na- montmorillonite and MMS, were shown to be ineffective at inducing efflorescence of Mg(ClO4)2 in either immersion or contact mode. Comparison of contact ERH to crystal lattice mismatch between Mg(ClO4)2 and the heterogeneous nuclei, excluding MMS, showed a mild correlation.

Heterogeneous efflorescence of brines of MgCl2, Ca(ClO4)2, and CaCl2 by Na-montmorillonite was then investigated. The only salt that was affected by the presence of Na-montmorillonite was

MgCl2 whose contact and immersion ERH was significantly higher than the homogenous ERH.

Both Ca(ClO4)2 and CaCl2 seemed to be unaffected by Na-montmorillonite, where at room temperature no efflorescence was observed. Comparison of crystal lattice mismatch between the four salts and a select set of minerals detected on Mars shows that MgCl2 may have been affected by Na-montmorillonite the most because it has the lowest mismatch value out of the four salts.

105

Additionally, a majority of salt and mineral pairs were shown to have a mismatch value greater than 0.12 suggesting that many of these minerals will have minimal effect on the efflorescence of the brines. However, feldspar was shown to have the best crystal lattice match with MgCl2 and moderately low mismatch values with the other three salts. Feldspar is believed to be globally distributed on Mars and thus may have a large impact on brine stability.

In conclusion, heterogeneous efflorescence was shown to be capable of affecting the phase stability of aqueous aerosols and brines. Depending on the salt and heterogeneous nuclei pair, the increase to ERH when compared to the homogeneous ERH varied greatly. Parameters such as active site, crystal lattice match, and ion-specific effects were all considered in attempting to explain the difference, but there is still no comprehensive model to explain the results of heterogeneous efflorescence. To fully understand heterogeneous efflorescence and to be able to model it, more laboratory studied will be needed. Heterogeneous efflorescence is expected to play a role in the Earth’s atmosphere where collisions between aerosol particles cause internal mixing.

The effect of mineral particles will have greater impacts for regions with higher mineral dust emissions. The effect on viscosity of salt aerosols due to mixing with organic aerosols is more important than whether the organic fraction LLPS from the droplet. Heterogeneous efflorescence and efflorescence in general will be more important in the lower troposphere where it is warm enough that the organic fraction has not caused the aerosols to form glasses. On Mars, the impact that the Martian regolith has on brines that form on the surface and in the subsurface is inevitable.

The time that a salt brine exists stably on Mars will depend on the mineralogy of the surrounding regolith, with some mineral particles being able to significantly shorten the time that brine exists.

If life on Mars exists on present day Mars, the lowered stability of brines may be detrimental for their survival.

106

Chapter 6: Bibliography

Abbatt, J. P. D., Benz, S., Cziczo, D. J., Kanji, Z., Lohmann, U., & Möhler, O. (2006). Solid ammonium sulfate aerosols as ice nuclei: a pathway for cirrus cloud formation. Science, 313(5794), 1770–1773. https://doi.org/10.1126/science.1129726 Andreae, M. O., Charlson, R. J., Bruynseels, F., Storms, H., VAN Grieken, R., & Maenhaut, W. (1986). Internal mixture of sea salt, silicates, and excess sulfate in marine aerosols. Science (New York, N.Y.), 232(4758), 1620–1623. https://doi.org/10.1126/science.232.4758.1620 Andreae, M. O., & Rosenfeld, D. (2008). Aerosol–cloud–precipitation interactions. Part 1. The nature and sources of cloud-active aerosols. Earth-Science Reviews, 89(1–2), 13–41. https://doi.org/10.1016/j.earscirev.2008.03.001 Angell, C. A. (1995). Formation of Glasses from Liquids and Biopolymers. Science, 267(5206), 1924 LP – 1935. https://doi.org/10.1126/science.267.5206.1924 Baldwin, R. L. (1996). How Hofmeister ion interactions affect protein stability. Biophysical Journal, 71(4), 2056–2063. https://doi.org/10.1016/S0006-3495(96)79404-3 Bibring, J.-P., Langevin, Y., Mustard, J. F., Arvidson, R., Poulet, F., Arvidson, R., … Neukum, G. (2006). Global Mineralogical and Aqueous Mars History Derived from OMEGA/Mars Express Data. Science, 312(5772), 400–404. https://doi.org/10.1126/science.1122659 Bibring, J., Langevin, Y., Gendrin, A., Gondet, B., Poulet, F., Berthé, M., … Drossart, P. (2005). Mars Surface Diversity as Revealed by the OMEGA/Mars Express Observations. Science, 307(5715), 1576 LP – 1581. https://doi.org/10.1126/science.1108806 Biskos, G., Malinowski, a., Russell, L. M., Buseck, P. R., & Martin, S. T. (2006). Nanosize Effect on the Deliquescence and the Efflorescence of Sodium Chloride Particles. Aerosol Science and Technology, 40(2), 97–106. https://doi.org/10.1080/02786820500484396 Blake, D., Vaniman, D., Achilles, C., Anderson, R., Bish, D., Bristow, T., … Yen, A. (2012). Characterization and Calibration of the CheMin Mineralogical Instrument on Mars Science Laboratory. Space Science Reviews, 170(1), 341–399. https://doi.org/10.1007/s11214-012- 9905-1 Bodsworth, A., Zobrist, B., & Bertram, A. K. (2010). Inhibition of efflorescence in mixed organic-inorganic particles at temperatures less than 250 K. Physical Chemistry Chemical Physics, 12(38), 12259–12266. https://doi.org/10.1039/c0cp00572j Boucher, O., Randall, D., Artaxo, P., Bretherton, C., Feingold, G., Forster, P., … Zhang, X.-Y. (2013). Clouds and Aerosols. In T. F. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S. K. Allen, J. Boschung, … P. M. Midgley (Eds.), Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climatte Change (pp. 571–657). Cambridge: Cambridge University Press. Broadley, S. L., Murray, B. J., Herbert, R. J., Atkinson, J. D., Dobbie, S., Malkin, T. L., … Neve, L. (2012). Immersion mode heterogeneous ice nucleation by an illite rich powder representative of atmospheric mineral dust. Atmospheric Chemistry and Physics, 12(1),

107

287–307. https://doi.org/10.5194/acp-12-287-2012 Buseck, P. R., & Pósfai, M. (1999). Airborne minerals and related aerosol particles: Effects on climate and the environment. Proceedings of the National Academy of Sciences of the United States of America, 96(7), 3372–3379. https://doi.org/10.1073/pnas.96.7.3372 Chou, C., Formenti, P., Maille, M., Ausset, P., Helas, G., Harrison, M., & Osborne, S. (2008). Size distribution, shape, and composition of mineral dust aerosols collected during the African Monsoon Multidisciplinary Analysis Special Observation Period 0: Dust and Biomass-Burning Experiment field campaign in Niger, January 2006. Journal of Geophysical Research, 113(23), D00C10. https://doi.org/10.1029/2008JD009897 Chu, Y., & Chan, C. K. (2017). Reactive Uptake of Dimethylamine by Ammonium Sulfate and Ammonium Sulfate–Sucrose Mixed Particles. The Journal of Physical Chemistry A, 121(1), 206–215. https://doi.org/10.1021/acs.jpca.6b10692 Ciobanu, V. G., Marcolli, C., Krieger, U. K., Zuend, A., & Peter, T. (2010). Efflorescence of ammonium sulfate and coated ammonium sulfate particles: Evidence for surface nucleation. Journal of Physical Chemistry A, 114(35), 9486–9495. https://doi.org/10.1021/jp103541w Clark, B. C., & Kounaves, S. P. (2016). Evidence for the distribution of perchlorates on Mars. International Journal of Astrobiology, 15(4), 311–318. https://doi.org/DOI: 10.1017/S1473550415000385 Clegg, S. L., Brimblecombe, P., & Wexler, A. S. (1998). Thermodynamic model of the system H+-NH4+-Na+-SO42-NO3--Cl--H2O at 298.15 K. Journal of Physical Chemistry A, 102(12), 2155–2171. https://doi.org/10.1021/jp973043j Cockell, C. S., Bush, T., Bryce, C., Direito, S., Fox-Powell, M., Harrison, J. P., … Zorzano, M. P. (2016). Habitability: A Review. Astrobiology, 16(1), 89–117. https://doi.org/10.1089/ast.2015.1295 Cziczo, D. J., & Abbatt, J. P. D. (1999). Deliquescence, efflorescence, and supercooling of ammonium sulfate aerosols at low temperature: Implications for cirrus cloud formation and aerosol phase in the atmosphere. Journal of Geophysical Research, 104(D11), 13781– 13790. https://doi.org/10.1029/1999JD900112 Cziczo, Daniel J., Froyd, K. D., Hoose, C., Jensen, E. J., Diao, M., Zondlo, M. a, … Murphy, D. M. (2013). Clarifying the dominant sources and mechanisms of cirrus cloud formation. Science (New York, N.Y.), 340(6138), 1320–1324. https://doi.org/10.1126/science.1234145 Davis, R.D., Lance, S., Gordon, J. a., & Tolbert, M. a. (2015). Long Working-Distance Optical Trap for in Situ Analysis of Contact-Induced Phase Transformations. Analytical Chemistry, 87(12), 6186–6194. https://doi.org/10.1021/acs.analchem.5b00809 Davis, Ryan D., Lance, S., Gordon, J. A., Ushijima, S. B., & Tolbert, M. A. (2015). Contact efflorescence as a pathway for crystallization of atmospherically relevant particles. Proceedings of the National Academy of Sciences, 112(52), 15815–15820. https://doi.org/10.1073/pnas.1522860113

108

Davis, Ryan D., & Tolbert, M. A. (2017). Crystal nucleation initiated by transient ion-surface interactions at aerosol interfaces. Science Advances, 3(July), e1700425. https://doi.org/10.1126/sciadv.1700425 Decesari, S., Mircea, M., Cavalli, F., Fuzzi, S., Moretti, F., Tagliavini, E., & Facchini, M. C. (2007). Source attribution of water-soluble organic aerosol by nuclear magnetic resonance spectroscopy. Environmental Science and Technology, 41(7), 2479–2484. https://doi.org/10.1021/es061711l Derieux, W.-S. W. W., Li, Y., Lin, P., Laskin, J., Laskin, A., Bertram, A. K., … Shiraiwa, M. (2018). Predicting the glass transition temperature and viscosity of secondary organic material using molecular composition. Atmos. Chem. Phys., 18(9), 6331–6351. https://doi.org/10.5194/acp-18-6331-2018 Downs, R. T., & Hall-Wallace, M. (2003). The American Mineralogist crystal structure database. American Mineralogist, 88(1), 247–250. Durant, A. J., & Shaw, R. A. (2005). Evaporation freezing by contact nucleation inside-out. Geophysical Research Letters, 32(20), L20814. https://doi.org/10.1029/2005GL024175 Ehlmann, B. L., Anderson, F. S., Andrews-Hanna, J., Catling, D. C., Christensen, P. R., Cohen, B. A., … Zahnle, K. J. (2016). The sustainability of habitability on terrestrial planets: Insights, questions, and needed measurements from Mars for understanding the evolution of Earth-like worlds. Journal of Geophysical Research: Planets, 121(10), 1927–1961. https://doi.org/10.1002/2016JE005134 Ehlmann, Bethany L., & Edwards, C. S. (2014). Mineralogy of the Martian Surface. Annual Review of Earth and Planetary Sciences, 42(1), 291–315. https://doi.org/10.1146/annurev- earth-060313-055024 Ehlmann, Bethany L., Mustard, J. F., Murchie, S. L., Bibring, J. P., Meunier, A., Fraeman, A. A., & Langevin, Y. (2011). Subsurface water and clay mineral formation during the early history of Mars. Nature, 479(7371), 53–60. https://doi.org/10.1038/nature10582 Fitzgerald, J. W. (1991). Marine aerosols: A review. Atmospheric Environment Part A, General Topics, 25(3–4), 533–545. https://doi.org/10.1016/0960-1686(91)90050-H Fletcher, N. H. (1969). Active Sites and Ice Crystal Nucleation. Journal of the Atmospheric Sciences. https://doi.org/10.1175/1520-0469(1969)026<1266:ASAICN>2.0.CO;2 Freedman, M. A. (2017). Phase separation in organic aerosol. Chemical Society Reviews, 46(24), 7694–7705. https://doi.org/10.1039/C6CS00783J Fukuta, N. (1975). A study of the mechanism of contact ice nucleation. Journal of the Atmospheric Sciences, 32, 1507–1603. Retrieved from http://journals.ametsoc.org/doi/abs/10.1175/1520- 0469(1975)032%3C1597:ASOTMO%3E2.0.CO%3B2 Garimella, S., Huang, Y.-W., Seewald, J. S., & Cziczo, D. J. (2014). Cloud condensation nucleus activity comparison of dry- and wet-generated mineral dust aerosol: the significance of soluble material. Atmospheric Chemistry and Physics, 14(12), 6003–6019. https://doi.org/10.5194/acp-14-6003-2014 109

Glaccum, R. A., & Prospero, J. M. (1980). Saharan aerosols over the tropical North Atlantic — Mineralogy. Marine Geology, 37(3–4), 295–321. https://doi.org/10.1016/0025- 3227(80)90107-3 Glavin, D. P., Freissinet, C., Miller, K. E., Eigenbrode, J. L., Brunner, A. E., Buch, A., … Mahaffy, P. R. (2013). Evidence for perchlorates and the origin of chlorinated hydrocarbons detected by SAM at the Rocknest aeolian deposit in Gale Crater. Journal of Geophysical Research: Planets, 118, 1955–1973. https://doi.org/10.1002/jgre.20144 Gough, R. V., Chevrier, V. F., Baustian, K. J., Wise, M. E., & Tolbert, M. A. (2011). Laboratory studies of perchlorate phase transitions: Support for metastable aqueous perchlorate solutions on Mars. Earth and Planetary Science Letters, 312(3–4), 371–377. https://doi.org/10.1016/j.epsl.2011.10.026 Gough, R. V., Chevrier, V. F., & Tolbert, M. A. (2014). Formation of aqueous solutions on Mars via deliquescence of chloride–perchlorate binary mixtures. Earth and Planetary Science Letters, 393, 73–82. https://doi.org/10.1016/j.epsl.2014.02.002 Gough, R. V., Chevrier, V. F., & Tolbert, M. A. (2016). Formation of liquid water at low temperatures via the deliquescence of calcium chloride: Implications for Antarctica and Mars. Planetary and Space Science, 131, 79–87. https://doi.org/10.1016/j.pss.2016.07.006 Grayson, J. W., Evoy, E., Song, M., Chu, Y., Maclean, A., Nguyen, A., … Bertram, A. K. (2017). The effect of hydroxyl functional groups and molar mass on the viscosity of non- crystalline organic and organic–water particles. Atmos. Chem. Phys., 17(13), 8509–8524. https://doi.org/10.5194/acp-17-8509-2017 Grayson, J. W., Zhang, Y., Mutzel, A., Renbaum-Wolff, L., Böge, O., Kamal, S., … Bertram, A. K. (2016). Effect of varying experimental conditions on the viscosity of α-pinene derived secondary organic material. Atmospheric Chemistry and Physics, 16(10), 6027–6040. https://doi.org/10.5194/acp-16-6027-2016 Han, F., Zhang, J., Chen, G., & Wei, X. (2008). Density, Viscosity, and Excess Properties for Aqueous Poly(ethylene glycol) Solutions from (298.15 to 323.15) K. Journal of Chemical & Engineering Data, 53(11), 2598–2601. https://doi.org/10.1021/je800464t Han, J.-H., Hung, H.-M., & Martin, S. T. (2002). Size effect of hematite and corundum inclusions on the efflorescence relative humidities of aqueous ammonium nitrate particles. Journal of Geophysical Research: Atmospheres, 107(D10), AAC 3-1-3-9. https://doi.org/10.1029/2001JD001054 Han, J.-H., & Martin, S. T. (1999). Heterogeneous nucleation of the efflorescence of (NH4)2SO4 particles internally mixed with Al2O3 , TiO2 , and ZrO2. Journal of Geophysical Research, 104(D3), 3543 3553. https://doi.org/10.1029/1998JD100072 Hecht, M. H., Kounaves, S. P., Quinn, R. C., West, S. J., Young, S. M. M., Ming, D. W., … Ugolini, F. C. (2009). Detection of Perchlorate and the Soluble Chemistry of Martian Soil at the Phoenix Lander Site. Science , 325(5936), 64–67. https://doi.org/10.1126/science.1172466

110

Hennings, E., Schmidt, H., & Voigt, W. (2014). Crystal structures of Ca(ClO4)2·4H2O and Ca(ClO4)2·6H2O. Acta Crystallographica Section E: Structure Reports Online, 70(12), 489–493. https://doi.org/10.1107/S1600536814024532 Hoffmann, N., Duft, D., Kiselev, A., & Leisner, T. (2013). Contact freezing efficiency of mineral dust aerosols studied in an electrodynamic balance: quantitative size and temperature dependence for illite particles. Faraday Discussions, 165, 383–390. https://doi.org/10.1039/c3fd00033h Jerome, F. S., Tseng, J. T., & Fan, L. T. (1968). Viscosities of aqueous glycol solutions. Journal of Chemical & Engineering Data, 13(4), 496. https://doi.org/10.1021/je60039a010 Jimenez, J. L., Canagaratna, M. R., Donahue, N. M., Prevot, a. S. H., Zhang, Q., Kroll, J. H., … Worsnop, D. R. (2009). Evolution of Organic Aerosols in the Atmosphere. Science, 326(5959), 1525–1529. https://doi.org/10.1126/science.1180353 Kiselev, A., Bachmann, F., Pedevilla, P., Cox, S. J., Michaelides, A., Gerthsen, D., & Leisner, T. (2017). Active sites in heterogeneous ice nucleation—the example of K-rich feldspars. Science, 355(6323), 367–371. https://doi.org/10.1126/science.aai8034 Koop, T., Bookhold, J., Shiraiwa, M., & Pöschl, U. (2011). Glass transition and phase state of organic compounds: dependency on molecular properties and implications for secondary organic aerosols in the atmosphere. Physical Chemistry Chemical Physics, 13(43), 19238. https://doi.org/10.1039/c1cp22617g Kroll, J. H., & Seinfeld, J. H. (2008). Chemistry of secondary organic aerosol: Formation and evolution of low-volatility organics in the atmosphere. Atmospheric Environment, 42(16), 3593–3624. https://doi.org/10.1016/j.atmosenv.2008.01.003 Ladino Moreno, L. A., Stetzer, O., & Lohmann, U. (2013). Contact freezing: a review of experimental studies. Atmospheric Chemistry and Physics, 13(19), 9745–9769. https://doi.org/10.5194/acp-13-9745-2013 Li, J., Anderson, J. R., & Buseck, P. R. (2003). TEM study of aerosol particles from clean and polluted marine boundary layers over the North Atlantic. Journal of Geophysical Research, 108(D6), 4189. https://doi.org/10.1029/2002JD002106 Mangold, N., Quantin, C., Ansan, V., Delacourt, C., & Allemand, P. (2004). Evidence for Precipitation on Mars from Dendritic Valleys in the Valles Marineris Area. Science, 305(5680), 78–81. https://doi.org/10.1126/science.1097549 Marcolli, C., & Krieger, U. K. (2006). Phase Changes during Hygroscopic Cycles of Mixed Organic/Inorganic Model Systems of Tropospheric Aerosols. The Journal of Physical Chemistry A, 110(5), 1881–1893. https://doi.org/10.1021/jp0556759 Marcolli, C., Luo, B., & Peter, T. (2004). Mixing of the Organic Aerosol Fractions: Liquids as the Thermodynamically Stable Phases. Journal of Physical Chemistry A, 108(12), 2216– 2224. https://doi.org/10.1021/jp036080l Marcus, Y. (1988). Ionic Radii in Aqueous Solutions, 1475–1498. https://doi.org/10.1021/cr00090a003

111

Martín-Torres, F. J., Zorzano, M.-P., Valentín-Serrano, P., Harri, A.-M., Genzer, M., Kemppinen, O., … Vaniman, D. (2015). Transient liquid water and water activity at Gale crater on Mars. Nature Geoscience, 8(5), 357–361. https://doi.org/10.1038/ngeo2412 Martin, S. T. (2000). Phase Transitions of Aqueous Atmospheric Particles. Chemical Reviews, 100(9), 3403–3454. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11777428 Martin, S. T., Han, J.-H., & Hung, H.-M. (2001). The size effect of hematite and corundum inclusions on the efflorescence relative humidities of aqueous ammonium sulfate particles. Geophysical Research Letters, 28(13), 2601–2604. https://doi.org/10.1029/2001GL013120 Martin, S. T., Schlenker, J., Chelf, J. H., & Duckworth, O. W. (2001). Structure−Activity Relationships of Mineral Dusts as Heterogeneous Nuclei for Ammonium Sulfate Crystallization from Supersaturated Aqueous Solutions. Environmental Science & Technology, 35(8), 1624–1629. https://doi.org/10.1021/es001535v Martínez, G. M., & Renno, N. O. (2013). Water and brines on mars: Current evidence and implications for MSL. Space Science Reviews, 175(1–4), 29–51. https://doi.org/10.1007/s11214-012-9956-3 McEwen, A. S., Ojha, L., Dundas, C. M., Mattson, S. S., Byrne, S., Wray, J. J., … Gulick, V. C. (2011). Seasonal Flows on Warm Martian Slopes. Science, 333(6043), 740–743. https://doi.org/10.1126/science.1204816 Ming, D. W., Jr, P. D. A., Glavin, D. P., Eigenbrode, J. L., Franz, H. B., Sutter, B., … Yingst, R. A. (2014). Volatile and Organic Compositions of Sedimentary Rocks in Yellowknife Bay, Gale Crater, Mars. Science, 343(6169), 1245267. https://doi.org/10.1126/science.1245267 Mithen, J. P., & Sear, R. P. (2014). Computer simulation of epitaxial nucleation of a crystal on a crystalline surface. The Journal of Chemical Physics, 140(8), 084504. https://doi.org/10.1063/1.4866035 Murray, B. J., O’Sullivan, D., Atkinson, J. D., & Webb, M. E. (2012). Ice nucleation by particles immersed in supercooled cloud droplets. Chemical Society Reviews, 41(19), 6519–6554. https://doi.org/10.1039/c2cs35200a Niehaus, J., Bunker, K. W., China, S., Kostinski, A., Mazzoleni, C., & Cantrell, W. (2014). A Technique to Measure Ice Nuclei in the Contact Mode. Journal of Atmospheric and Oceanic Technology, 31(4), 913–922. https://doi.org/10.1175/JTECH-D-13-00156.1 Niehaus, J., & Cantrell, W. (2015). Contact Freezing of Water by Salts. The Journal of Physical Chemistry Letters, 6(17), 3490–3495. https://doi.org/10.1021/acs.jpclett.5b01531 Nuding, D. L., Davis, R. D., Gough, R. V, & Tolbert, M. A. (2015). The aqueous stability of a Mars salt analog: Instant Mars. Journal of Geophysical Research: Planets, 120(3), 588– 598. https://doi.org/10.1002/2014JE004722 Nuding, D. L., Rivera-Valentin, E. G., Davis, R. D., Gough, R. V., Chevrier, V. F., & Tolbert, M. a. (2014). Deliquescence and efflorescence of calcium perchlorate: An investigation of stable aqueous solutions relevant to Mars. Icarus, 243, 420–428. https://doi.org/10.1016/j.icarus.2014.08.036

112

Ojha, L., Wilhelm, M. B., Murchie, S. L., McEwen, A. S., Wray, J. J., Hanley, J., … Chojnacki, M. (2015). Spectral evidence for hydrated salts in recurring slope lineae on Mars. Nature Geoscience, 8(11), 829–832. https://doi.org/10.1038/ngeo2546 Orosei, R., Lauro, S. E., Pettinelli, E., Cicchetti, A., Coradini, M., Cosciotti, B., … Seu, R. (2018). Radar evidence of subglacial liquid water on Mars. Science, 7268(July), 1–8. https://doi.org/10.1029/2002GC000367","ISSN":"15252027","author":[{"family":"Hodell", "given":"David Osterloo, M. M., Anderson, F. S., Hamilton, V. E., & Hynek, B. M. (2010). Geologic context of proposed chloride-bearing materials on Mars. Journal of Geophysical Research E: Planets, 115(10), 1–29. https://doi.org/10.1029/2010JE003613 Pant, A., Parsons, M. T., & Bertram, A. K. (2006). Crystallization of aqueous ammonium sulfate particles internally mixed with soot and kaolinite: crystallization relative humidities and nucleation rates. The Journal of Physical Chemistry. A, 110(28), 8701–8709. https://doi.org/10.1021/jp060985s Peters, G. H., Abbey, W., Bearman, G. H., Mungas, G. S., Smith, J. A., Anderson, R. C., … Beegle, L. W. (2008). Mojave Mars simulant-Characterization of a new geologic Mars analog. Icarus, 197(2), 470–479. https://doi.org/10.1016/j.icarus.2008.05.004 Petters, S. S., Kreidenweis, S. M., Grieshop, A. P., Ziemann, P. J., & Petters, M. D. (2019). Temperature‐ and Humidity‐Dependent Phase States of Secondary Organic Aerosols. Geophysical Research Letters, 46(2), 1005–1013. https://doi.org/10.1029/2018GL080563 Poulet, F., Bibring, J. P., Mustard, J. F., Gendrin, A., Mangold, N., Langevin, Y., … Forget, F. (2005). Phyllosilicates on Mars and implications for early martian climate. Nature, 438(7068), 623–627. https://doi.org/10.1038/nature04274 Price, H. C., Mattsson, J., & Murray, B. J. (2016). Sucrose diffusion in aqueous solution. Physical Chemistry Chemical Physics, 18(28), 19207–19216. https://doi.org/10.1039/c6cp03238a Primm, K. M., Gough, R. V., Chevrier, V. F., & Tolbert, M. A. (2017). Freezing of perchlorate and chloride brines under Mars-relevant conditions. Geochimica et Cosmochimica Acta, 212, 211–220. https://doi.org/10.1016/j.gca.2017.06.012 Primm, K. M., Gough, R. V., Wong, J., Rivera-Valentin, E. G., Martinez, G. M., Hogancamp, J. V., … Tolbert, M. A. (2018). The Effect of Mars-Relevant Soil Analogs on the Water Uptake of Magnesium Perchlorate and Implications for the Near-Surface of Mars. Journal of Geophysical Research: Planets, 123(8), 2076–2088. https://doi.org/10.1029/2018JE005540 Prospero, J. M. (1999). Long-term measurements of the transport of African mineral dust to the southeastern United States: Implications for regional air quality. Journal of Geophysical Research: Atmospheres, 104(D13), 15917–15927. https://doi.org/10.1029/1999JD900072 Ramanathan, V., Crutzen, P. J., Kiehl, J. T., & Rosenfeld, D. (2001). Aerosols, Climate, and the Hydrological Cycle. Science, 294(5549), 2119–2124. https://doi.org/10.1126/science.1064034

113

Richards, D. S., Trobaugh, K. L., Hajek-Herrera, J., & Davis, R. D. (2020). Dual-Balance Electrodynamic Trap as a Microanalytical Tool for Identifying Gel Transitions and Viscous Properties of Levitated Aerosol Particles. Analytical Chemistry, 92(4), 3086–3094. https://doi.org/10.1021/acs.analchem.9b04487 Richardson, M. S., DeMott, P. J., Kreidenweis, S. M., Cziczo, D. J., Dunlea, E. J., Jimenez, J. L., … Lersch, T. L. (2007). Measurements of heterogeneous ice nuclei in the western United States in springtime and their relation to aerosol characteristics. Journal of Geophysical Research Atmospheres, 112(2), 1–16. https://doi.org/10.1029/2006JD007500 Robertson, K., & Bish, D. (2011). Stability of phases in the Mg(ClO 4 ) 2 · n H 2 O system and implications for perchlorate occurrences on Mars. Journal of Geophysical Research, 116(E7), E07006. https://doi.org/10.1029/2010JE003754 Robinson, C. B., Schill, G. P., & Tolbert, M. A. (2014). Optical growth of highly viscous organic/sulfate particles. Journal of Atmospheric Chemistry, 71(2), 145–156. https://doi.org/10.1007/s10874-014-9287-8 Schmidt, H., Hennings, E., & Voigt, W. (2012). Magnesium chloride tetrahydrate, MgCl 2 ·4H 2 O. Acta Crystallographica Section C Crystal Structure Communications, 68(1), i4–i6. https://doi.org/10.1107/S0108270111054709 Seinfeld, J. H., & Pandis, S. N. (2006). Atmospheric Chemistry and Physics (2nd Ed.). Hoboken, NJ: Wiley-Interscience. Shaw, R., Durant, A., & Mi, Y. (2005). Heterogeneous surface crystallization observed in undercooled water. The Journal of Physical Chemistry B, 109, 9865–9868. Retrieved from http://pubs.acs.org/doi/abs/10.1021/jp0506336 Shiraiwa, M., Li, Y., Tsimpidi, A. P., Karydis, V. A., Berkemeier, T., Pandis, S. N., … Po, U. (2017). Global distribution of particle phase state in atmospheric secondary organic aerosols. Nature Communications, 8, 15002. https://doi.org/10.1038/ncomms15002 Sidey, V. (2016). On the effective ionic radii for ammonium. Acta Crystallographica Section B: Structural Science, Crystal Engineering and Materials, 72(4), 626–633. https://doi.org/10.1107/S2052520616008064 Song, Y. C., Haddrell, A. E., Bzdek, B. R., Reid, J. P., Bannan, T., Topping, D. O., … Cai, C. (2016). Measurements and Predictions of Binary Component Aerosol Particle Viscosity. Journal of Physical Chemistry A, 120(41), 8123–8137. https://doi.org/10.1021/acs.jpca.6b07835 Tegen, I., & Fung, I. (1994). Modeling of mineral dust in the atmosphere’ transport, and optical thickness. Journal of Geophysical Research, 99(94), 22897–22914. https://doi.org/10.1029/94JD01928 Tian, R., Yang, G., Li, H., Gao, X., Liu, X., Zhu, H., & Tang, Y. (2014). Activation energies of colloidal particle aggregation: towards a quantitative characterization of specific ion effects. Physical Chemistry Chemical Physics : PCCP, 16(19), 8828–8836. https://doi.org/10.1039/c3cp54813a

114

Toner, J. D., Catling, D. C., & Light, B. (2014). The formation of supercooled brines, viscous liquids, and low-temperature perchlorate glasses in aqueous solutions relevant to Mars. Icarus, 233, 36–47. https://doi.org/10.1016/j.icarus.2014.01.018 Usher, C. R., Michel, A. E., & Grassian, V. H. (2003). Reactions on mineral dust. Chemical Reviews, 103(12), 4883–4940. https://doi.org/10.1021/cr020657y Ushijima, S. B., Davis, R. D., & Tolbert, M. A. (2018). Immersion and Contact Efflorescence Induced by Mineral Dust Particles. The Journal of Physical Chemistry A, 122(5), 1303– 1311. https://doi.org/10.1021/acs.jpca.7b12075 van Meel, J. A., Sear, R. P., & Frenkel, D. (2010). Design Principles for Broad-Spectrum Protein-Crystal Nucleants with Nanoscale Pits. Physical Review Letters, 105(20), 205501. https://doi.org/10.1103/PhysRevLett.105.205501 Vaniman, D. T., Bish, D. L., Ming, D. W., Bristow, T. F., Morris, R. V, Blake, D. F., … Zorzano Mier, M.-P. (2014). Mineralogy of a Mudstone at Yellowknife Bay, Gale Crater, Mars. Science, 343(6169), 1243480–1243480. https://doi.org/10.1126/science.1243480 Wan, E. C. H., & Jian, Z. Y. (2007). Analysis of sugars and sugar polyols in atmospheric aerosols by chloride attachment in liquid chromatography/negative ion electrospray mass spectrometry. Environmental Science and Technology, 41(7), 2459–2466. https://doi.org/10.1021/es062390g Wehner, B., Wiedensohler, A., Tuch, T. M., Wu, Z. J., Hu, M., Slanina, J., & Kiang, C. S. (2004). Variability of the aerosol number size distribution in Beijing, China: New particle formation, dust storms, and high continental background. Geophysical Research Letters, 31(22), 1–4. https://doi.org/10.1029/2004GL021596 Zhang, Q., Jimenez, J. L., Canagaratna, M. R., Allan, J. D., Coe, H., Ulbrich, I., … Worsnop, D. R. (2007). Ubiquity and dominance of oxygenated species in organic aerosols in anthropogenically-influenced Northern Hemisphere midlatitudes. Geophysical Research Letters, 34(13), 1–6. https://doi.org/10.1029/2007GL029979 Zobrist, B., Marcolli, C., Pedernera, D. A., & Koop, T. (2008). Do atmospheric aerosols form glasses? Atmos. Chem. Phys., 8(17), 5221–5244. https://doi.org/10.5194/acp-8-5221-2008 Zorzano, M. P., Mateo-Martí, E., Prieto-Ballesteros, O., Osuna, S., & Renno, N. (2009). Stability of liquid saline water on present day Mars. Geophysical Research Letters, 36(20), 2–5. https://doi.org/10.1029/2009GL040315 Zuberi, B., Bertram, A. K., Cassa, C. A., Molina, L. T., & Molina, M. J. (2002). Heterogeneous nucleation of ice in (NH4)2SO4-H2O particles with mineral dust immersions. Geophysical Research Letters, 29(10), 142–144. https://doi.org/10.1029/2001GL014289

115

Chapter 7: Appendix

Supporting Figures

Figure A1: Particle size distribution of Mg(ClO4)2, NaCl, (NH4)2SO4, Na-montmorillonite, and MMS particles made with the nebulizer (Omron NE-U22). Distributions are the average of 3 samples taken in an SMPS (TSI model 3010). The high-end cutoff of particle size is due to the upper limit of the instrument and not due to the actual distribution.

116

Figure A2: A plot of the log of viscosity vs RH for raffinose (Grayson et al., 2017), and

(NH4)2SO4:Raffinose measured in the electrical dynamic balance. The data were linearly fit and are shown as dotted lines. The intercept (a) and slope (b) values are shown next to the data. The lowest RH data from Grayson et al. (2017) was not included in the fit as it was a lower limit.

117

Table A1: Lattice Constants and Crystal Systems.

Lattice Constants a b c Crystal System Reference

(NH4)2SO4 7.73 10.56 5.95 Orthorhombic (Downs & Hall-Wallace, 2003) NaCl 5.64 5.64 5.64 Cubic (Robertson & Bish, Mg(ClO4)2●6H2O 7.775 13.494 5.27 Orthorhombic 2011) (Schmidt, Hennings, & MgCl2●4H2O 7.26 8.43 11.04 Orthorhombic Voigt, 2012) (Hennings, Schmidt, & Ca(ClO4)2●4H2O 5.49 7.85 11.57 Triclinic Voigt, 2014)

CaCl2●6H2O 7.88 7.88 3.95 Hexagonal Olivine 4.779 10.277 5.995 Orthorhombic Orthopyroxene 18.316 8.907 5.218 Orthorhombic Clinopyroxene 9.794 8.906 5.319 Monoclinic Feldspar 8.56 12.96 7.299 Monoclinic Hematite 5.038 5.038 13.772 Hexagonal Nontronite 5.277 9.14 9.78 Monoclinic Na- Montmorillonite 5.17 8.94 9.95 Monoclinic (Downs & Hall-Wallace, Kaolinite 5.13 8.89 7.25 Triclinic 2003) Chlorite 5.3363 9.24 14.37 Monoclinic Illite 5.18 8.98 10.32 Monoclinic Prehnite 4.646 5.483 18.486 Orthorhombic Quartz 4.9133 4.9133 5.4053 Hexagonal Magnesite 4.633 4.633 15.016 Rhombohedral 17.061 Calcite 4.99 4.99 5 Rhombohedral Kieserite 6.891 7.624 7.645 Monoclinic 15.104 Gypsum 5.674 9 6.4909 Monoclinic

118