Optical Measurements of Atmospheric Aerosols: Aeolian Dust, Secondary Organic Aerosols, and Laser-Induced Incandescence of Soot

Home , Kerosene lamp

Lulu Ma, B. S.

A Dissertation

Chemistry

Submitted to the Graduate Faculty of Texas Tech University in Partial Fulfillment of the Requirements for the Degree of

Doctor of philosophy

Approved

Jonathan E. Thompson Chair of Committee

Dimitri Pappas

Carol L. Korzeniewski

Dominick Casadonte Interim Dean of the Graduate School

August, 2013

Texas Tech University, Lulu Ma, August 2013

ACKNOWLEDGMENTS I would like to thank my advisor, Dr. Thompson for his guidance and encouragement throughout the past 4 years in Texas Tech University. His profound knowledge, diligence and patience encouraged me and has setup a great example for me in the future.

I also thank to Dr. Pappas and Dr. Korzeniewski for their help and for taking time as my committee members. I also would like to thank Dr. Ted Zobeck for his help on soil dust measurements and soil sample collection and also Dr. P. Buseck and his lab for TEM analysis.

My group members: Hao Tang, Fang Qian, Kathy Dial, Haley Redmond, Yiyi Wei, Tingting Cao, and Qing Zhang, also helped me a lot in both daily life, and research. At last, I would like to thank my parents and friends for their love and continuous encouragement and support.

Texas Tech University, Lulu Ma, August 2013

TABLE OF CONTENTS ACKNOWLEDGMENTS……………………………………………………………….ii

ABSTRACT ……………………………………………………………………………..vi

LIST OF TABLES …………………………………………………………………….viii

LIST OF FIGURES……………………………………………………..………………ix

LIST OF ABBREVIATIONS ………………………………………….………………xi

I. INTRODUCTION ...... 1

1.1 Introduction to Atmospheric Aerosols...... 1 1.1.1 Natural Sources ...... 1 1.1.2 Anthropogenic Sources ...... 2 1.1.3 Production of Atmospheric Aerosols ...... 2 1.2 Effects of Atmospheric Aerosols ...... 3 1.2.1 Climate Effects...... 3 1.2.2 Health effects ...... 5 1.3 Optical properties of Atmospheric Aerosols...... 6 1.3.1 Possible Phenomena of Light Incidence on Aerosols ...... 6 1.3.2 Extinction, Scattering and Absorption ...... 7 1.4 Optical Measurement of Aerosols ...... 8 1.4.1 Cavity Ring-down Spectroscopy (CRDS) ...... 8 1.4.2 Nephelometer ...... 9 1.4.3 Particle Soot Absorption Photometer (PSAP) ...... 10 II. OPTICAL PROPERTIES OF AEOLIAN DUSTS COMMON TO WEST TEXAS ...... 17

2.1 Introduction ...... 17 2.2 Experimental ...... 19 2.2.1. Dentition of Fundamental Optical Parameters ...... 19 2.2.2 Soil Types, Sampling Locations, and Physical Properties ...... 19 2.2.3 Soil Textures ...... 20 2.2.4 Organic Matter Content and Particle Density of Soil ...... 20 2.2.5. Generation of Soil Dust ...... 21 2.2.6. Measurement Instruments ...... 21 2.3. Results and Discussion ...... 22 2.3.1 Observed Particle Size Distributions for Pullman and Amarillo Soils ...... 22 iii

Texas Tech University, Lulu Ma, August 2013

2.3.2 Dust Optical Properties: the Spectral Dependence of Dust Extinction and Absorption...... 23 2.3.3 Angstrom Extinction Exponent (AEE) ...... 23 2.3.4 Angstrom Absorption Exponent (AAE) and Diffuse Reflectance ...... 24 2.3.5 Estimate of Single Scatter Albedo (SSA) of Pullman and Amarillo Soil Dust Aerosols ...... 26 2.3.6 Mass-Extinction Coefficient (MEC) of Pullman and Amarillo Soils ...... 26 2.4 Summary and conclusions ...... 27 III. OPTICAL PROPERTIES OF DISPERSED AEROSOLS IN THE NEAR ULTRAVIOLET (355NM): MEASURMENT, APPROACH AND INITIAL DATA 35

3.1 Introduction ...... 35 3.2 Experimental Methods ...... 37 3.2.1 Sample Flow Path ...... 37 3.2.2 Generation of Aerosol Particles...... 37 3.2.3 Optical measurements...... 39 3.3 Results and discussion ...... 40 3.3.1 Calibration Gas Experiments ...... 40 3.3.2 Size-Selected Ammonium Sulfate Aerosols and Secondary Organic Aerosol (SOA)...... 41 3.3.3 Soil Dusts and Biomass Burning Aerosols ...... 42 3.3.4 Ambient Measurements ...... 43 3.4 Conclusions ...... 46 IV. LIGHT SCATTERING AND EXTINCTION MEASUREMENTS CCOMBINED WITH LASER-INDUCED INCANDESCENCE WITHIN AN AEROSOL ALBEDOMETER: METHOD DEVELOPMENT AND EFFECT OF SOOT MORPHOLOGICAL CHANGES ...... 56

4.1 Introduction...... 56 4.2 Methods...... 57 4.2.1 Generation and Characteristics of Kerosene Soot ...... 57 4.2.2 Description of the Optical Measurements...... 58 4.2.3 Scanning Mobility Particle Sizing (SMPS) and Particle Soot Absorption Photometry (PSAP) ...... 59 4.2.4 Discrete Dipole Approximation (DDA) Modeling ...... 60 4.2.5 Transmission Electron Microscopy (TEM) Analysis ...... 60 4.3 RESULTS & DISCUSSION...... 60 4.3.1 Strategies to Measure the LII Signal ...... 60 4.3.2 Consequences of the LII Measurement: Soot Morphology Changes ...... 64 4.3.3 Suggested Measurement Sequence...... 65 iv

Texas Tech University, Lulu Ma, August 2013

V. CONCLUSION AND DISCUSSION ...... 76

APPENDIX ...... 79

A. OPERATING PROCEDURE FOR USING THE ADDA DISCRETE DIPOLE SOFTWARE FOR MODELING LIGHT SCATTERING AND ABSORPTION BY SMALL IRREGULAR PARTICLES...... 79

REFERENCES ...... 98

Texas Tech University, Lulu Ma, August 2013

ABSTRACT Aerosols suspended in earth’s atmosphere absorb and scatter sunlight. These optical effects can alter photochemistry or warm or cool the atmosphere and affect climate. This dissertation research contributes to a broader understanding of how aerosols affect our climate system by presenting advances in measurement devices used for optical measurements of aerosols. In addition, the dissertation provides salient optical property data for several aerosol types including Aeolian dusts, ammonium sulfate, secondary organic aerosols formed from pinene and toluene, biomass burning aerosols, black carbon (soot), and ambient aerosols collected at Lubbock, TX.

Chapter 2 outlines the optical properties of Amarillo and Pullman soil dusts that are widely dispersed in the Texas Panhandle. After collection of the soils, they were dispersed in air, and light absorption and extinction coefficients were measured at several visible wavelengths. Size distributions, mass concentration (µg particles / m3) and soil texture / color were also determined. Amarillo and Pullman soils had similar size distributions with generated particles having effective diameters in the 1-6 µm range. Both of the samples exhibited a wavelength dependent extinction and absorption. Extinction increased at larger wavelengths slightly with Angstrom extinction exponents of -0.11 and -0.17 for Pullman and Amarillo dust, respectively. Conversely, light absorption increased at shorter wavelengths with Angstrom absorption exponents of 1.73 and 2.17 for the Pullman and Amarillo dusts. Values for the mass-extinction coefficient ranged between 1.7 m2/g and 3.0 m2/g at 522 nm and single-scatter albedo (SSA) for both soil types ranged from 0.947 – 0.980 at visible wavelengths with SSA increasing at longer wavelengths.

Chapter 3 outlines the development of a measurement method that combines the advantages of cavity ring-down spectrometry (CRDS) and integrating sphere nephelometry to measure both extinction and scattering of dispersed aerosols at 355 nm simultaneously. Aerosols including ammonium sulfate, secondary organic aerosols (SOA) resulting from the ozonlysis of α-pinene and photo-oxidation of toluene, Aeolian dusts (Amarillo and Pullman), biomass burning aerosols, and ambient aerosols were

Texas Tech University, Lulu Ma, August 2013

monitored though this approach in the near ultraviolet range. Scattering and extinction cross-sections for size-selected (diameter = 300 nm) particles were between 1.65–2.60 × 10–9 cm2 with the largest value corresponding to ammonium sulfate and the lowest value for pinene SOA. Neither ammonium sulfate, nor α-pinene SOA absorbed light at 355 nm. Toluene photo-oxidation SOA showed slight absorption. Aeolian dust samples and biomass burning aerosols absorbed light significantly at 355 nm with single scatter albedo (SSA) between 0.74 and 0.84. The near UV optical measurements this system can provide are useful for assessing the impact of aerosols on tropospheric photochemistry.

The albedometer instrument described in chapter 3 was then modified to perform laser induced-incandescence (LII) of soot. The results of this study are presented in chapter 4. Instead of the 355 nm light beam, a co-linear beam containing both 532 and 1064 nm was applied to the aerosol. The 532 nm beam was meant to probe the particles scattering and extinction properties, while the 1064 nm beam heated the strongly absorbing soot resulting in broadband incandescence. The LII signal at 590 nm for fresh soot generated by a kerosene lamp was found to produce a linear response with soot mass concentration from the limit of detection of 0.5 to at least 100 µg / m3. Transmission electron microscopy (TEM) analysis and scanning mobility particle sizing (SMPS) data showed significant morphology changes and new particle production for lased soot samples compared with non-lased soot. These results suggest the LII process always irreversibly alters soot microphysical properties. Since these morphological changes were accompanied by a dramatic and significant drop in SSA after lasing with 1064 nm, we conclude that it is necessary for observations of aerosol optical properties and LII measurements to be made sequentially while allowing sufficient time for the measurement cell to be flushed with fresh aerosol sample.

vii

Texas Tech University, Lulu Ma, August 2013

LIST OF TABLES 1.1 Sources and Estimations of Global Emissions of Atmospheric Aerosols ...... 11 2.1 Observed single scatter albedo (SSA) values for Pullman and Amarillo soils types...... 29 3.1 Summary of Data and Extracted Refractive Indices (RI) for Dp = 300 nm Particles after Correction for Multiply Charged Particles ...... 47 3.2 Near UV Single Scatter Albedo (SSA) of Biomass Burning, Soil - Dust, and Organic Aerosols ...... 48 A. 1 Cartesian coordinates for points on the fundamental sphere. For this work we assumed a complex refractive index of 1.75 – 0.63i for soot...... 84

viii

Texas Tech University, Lulu Ma, August 2013

LIST OF FIGURES 1.1 Study of aerosol concentration and composition in 30 cities around world...... 12 1.2 A highly simplified mechanism of atmospheric oxidation of volatile organic compounds (VOCs)...... 13 1.3 A study of 52 - year haze data in Guangzhou, China...... 14 1.4 Possible phenomenon of light incidence on aerosols...... 15 1.5 An example of integrating nephelometer...... 16 2.1 Schematic of the dust generation and analysis system used in this study...... 30 2.2 Particle size distributions of Pullman (A) and Amarillo (B) soil types produced by the LDGASS...... 31 2.3 Plots of extinction vs. wavelength for lab generated (A) Pullman and (B) Amarillo soils...... 32 2.4 Plots of absorption coefficient vs. wavelength for lab generated (A) Pullman and (B) Amarillo soil dusts during a particular trial. (C) represents diffuse reflectance spectra on bulk Amarillo and Pullman soils (solid lines) and residual soil dust (dashed lines) collected from the main chamber of the LDGASS...... 33 2.5 Plot of observed extinction coefficient at λ= 522 nm vs. dust mass loading for Pullman and Amarillo soil samples...... 34 3.1 Schematic of the measurement apparatus...... 49 3.2 Observed size distributions for aerosolizd Pullman (A) and Amarillo (B) soils...... 50 3.3 Calibration of experiment using R - 134a...... 51 3.4 Plots of extinction and scattering coefficients at 355 nm vs. particle concentration for lab generated (A) ammonium sulfate aerosols and (B) α-pinene SOA...... 52 3.5 Plots of extinction and scattering coefficients (Mm-1) and mass concentrations (μg / m3) vs. time for 3 aerosol plumes observed in Lubbock, TX...... 53 3.6 Plot of extinction (black) and scattering (red) coefficient vs. mass concentration for ambient aerosol plumes in Lubbock, TX at 355 nm...... 54

Texas Tech University, Lulu Ma, August 2013

3.7 Plots of absorption coefficients (babs, Mm-1) at different wavelengths (355, 470, 552, and 660 nm) vs. time for three aerosol plumes observed at Lubbock, TX...... 55 4.1 Thermogravimetric analysis of kerosene soot...... 67 4.2 Experimental setup and transmission spectrum of high reflectivity mirror and near infrared (NIR) blocking filter...... 68 4.3 Three dimensional plot of the simulated soot aggregate used for DDA modeling...... 69 4.4 Oscilloscope trace for the scatter channel illustrating a typical signal response at 532 nm...... 70 4.5 Plots of laser incandescence signal vs. soot mass concentration at 20.8 mJ and normalized soot incandescence signal vs. laser pulse fluencies...... 71 4.6 Plots of soot LII signal at 590 nm and ambient measurement data using this instrumental setup...... 72 4.7 Morphology analysis and comparison of lased and non – lased soot. (A) Observed SMPS number concentration size distributions of fresh kerosene soot at various laser fluences...... 73 4.8 High resolution TEM images of fresh kerosene soot that has been lased vs. non-lased soot...... 74 4.9 Size distributions and single scatter of albedo (SSA) of fresh soot with and without 1064 nm illumination...... 75 A. 1 Primary sphere used for soot studies...... 83 A. 2 TEM images of fresh kerosene soot and (bottom) soot aggregate constructed from primary spheres used for soot modeling studies...... 84

Texas Tech University, Lulu Ma, August 2013

LIST OF ABBREVIATIONS AAE Angstrom absorption exponent

AEE Angstrom extinction exponent

AMS aerosol mass spectrometer

BC black carbon

CAPS cavity - attenuated phase - shift

CPC condensation particle counter

CRDS cavity ring-down spectroscopy

DDA discrete dipole approximation

DMA differential mobility analyzer

DTR diurnal temperature range

LED light-emitting diode

LII laser-induced incandescence

LOD limit of detection

LPM liter per minute

MAC mass absorption cross section

MEC mass-extinction coefficient

NIR near infrared

OC organic carbon

PAH polycyclic aromatic hydrocarbon

PSAP particle soot absorption photometer

PM particular matter

PMT photomultiplier tube

Texas Tech University, Lulu Ma, August 2013

R.I. refractive index r.o.c. radius of curvature

SMPS scanning mobility particle sizer

SCI stabilized Criegee intermediate

SOA secondary organic aerosol

SP2 single particle soot photometer

TEM transmission electron microscopy

TGA thermogravimetric analysis

UV ultraviolet

VOC volatile organic compound

xii

Texas Tech University, Lulu Ma, August 2013

CHAPTER I

INTRODUCTION

1.1 Introduction to Atmospheric Aerosols. An aerosol is a collection of fine solid or liquid particles suspended in a gas medium. Diameters of aerosol particles range from 10-9 to 10-4 m. 1 - 2 Thousands of substances can be defined as aerosols, for example, spray, smoke, dust, clouds and bioaerosols such as fungal spores. Aerosols are generated from both anthropogenic and natural sources.

1.1.1 Natural Sources Typical natural source aerosols include mineral dust, sea-salt, volcanic sulfate aerosols and secondary organic aerosols (SOA) formed from natural volatile organic compound (VOC) emissions. Table 1.1 lists a series of sources and estimates of global atmospheric aerosol emission.2 Though there is a huge uncertainty in mass flux, natural sources contribute a large portion of total atmospheric aerosols. A study considering 360 active volcanoes all around the world between 1991 and 2000 showed an annual sulfur emission of 9.5 - 10.5 Tg / year, which is about 25 - 30% anthropogenic sulfur emission (35 Tg / year). This emission of sulfur is even greater during major volcanic eruptions.3 Take the 1815 eruption of Tambora volcano (Sumbawa Island, Indonesia) as an example. During this eruption, about 60 Tg of sulfur was injected into atmosphere, forming a global sulfate aerosol veil in the stratosphere. The peak stratospheric loading was estimated to approach 200 Tg of sulfate aerosols.4 As another important source, biomass fires generate significant amount of trace gases and aerosols. Emissions of carbon in lower fire activity years such as 2000, 2001, 2008 and 2009 were about 1.6 Pg / year, while in higher fire activity years, for example, 1998, the emission reached 2.8 Pg / year.5 North Africa, the Earth's largest source of dust, has an annual dust emission ranging from 130 to 760 Tg/year, and estimates of global dust emissions ranges from 1000 to 3000 Tg/ year. 6 Atmospheric aerosols can be transported thousands of miles downwind from source regions. There are about 64 Tg of aerosols imported to North America annually, including 56 Tg of dust and 4 Tg of combustion aerosols transporting from the Pacific ocean and an additional 4 Tg of Saharan dust imported across the Atlantic coast. 7 The

Texas Tech University, Lulu Ma, August 2013

aerosols transported over the Pacific Ocean are from East Asia, Europe, and South Asia.7 Significant amounts of bacteria and fungi are also found to be transported with dust across the Atlantic Ocean.8 On a global scale, natural sources are distributed all around the world and contribute more than 50 % of global particulate emissions.

1.1.2 Anthropogenic Sources Compared with natural sources, anthropogenic sources have less contribution to the global particulate mass burden, but concentrations of anthropogenic aerosols can be much higher in localized regions of the world, for example, urban areas. Usually, sulfate, nitrate, organic carbon (OC) and black carbon (BC) generated by fossil fuel and biomass burning aerosols are considered significant components of anthropogenic aerosols.2 An investigation of primary anthropogenic aerosol in China showed emission of OC and BC at 1.51 Tg and 3.19 Tg in 2005.9 Anthropogenic aerosols are major components of urban aerosols with mass concentration ranging from 0.01-1 mg / m3.2,10 Aerosol chemical composition data obtained by the Aerosol Mass Spectrometer (AMS) in 30 cities around world is given in figure 1.1. 11 Total concentration and composition vary with locations. The analysis gave the lowest concentration of 1.5 μg / m3 in Mace Head, Ireland and the highest concentration of 71 μg / m3 in Beijing, China. Usually, urban areas have aerosol concentrations 2 - 5 times higher than that in rural areas. Because sources contributions vary with time, seasonal fluctuation of aerosol concentration is often observed. For example, a study in Guangzhou, China between 2002 and 2003 showed a 2 times-higher concentration of aerosols in winter than in spring.12 The largest contribution in winter came from n-alkanes, n-alkanoic acid and secondary oxidation products formed by the photo-oxidation of organic compounds.

1.1.3 Production of Atmospheric Aerosols Aerosols are generated by direct emission or chemical reactions in the atmosphere. Therefore, two terms, primary aerosols and secondary aerosols are used according to the two mechanisms. Primary aerosols contain different kinds of hydrocarbon, partially oxidized organics, combustion emissions, and a large number of suspended debris from anthropogenic and biology sources. Secondary aerosols, on the other hand, are formed by gas-phase reactions with atmospheric gases and gas - to - particle conversion or by

Texas Tech University, Lulu Ma, August 2013

chemical transformation of primary aerosols in the atmosphere.1,13 A study across the United States in 1999 found an equal contribution between primary organic and secondary organic carbon over the West and West Pacific areas. While in the Northeast, the dominant contribution was made by secondary organic aerosols.14 In MILAGRO 2006 campaign in Mexico City, secondary aerosols showed a larger influence on optical and aerosol mass measurements than primary aerosols. 15 Secondary organic aerosol (SOA) is a major contributor of secondary aerosols. The formation of SOA is considered to be dominated by oxidation of volatile organic compounds (VOCs). The formation of SOA is very complex. Figure 1.216 gives a highly simplified mechanism of atmospheric oxidation of VOC. A radical, for example, OH, NO3, or halogen atom initiated oxidation reaction is usually accomplished by capturing a hydrogen atom or adding to a C - C double bond. Oxidation of alkenes by ozone has a different mechanism, where ozone helps to cleave the double bond of alkenes and form a carbonyl and an energetically excited carbonyl oxide. The energetically excited carbonyl oxide then either decomposes to OH and an alkyl radical which may follow the mechanisms shown in figure 1.2 or form a stabilized Criegee intermediate (SCI), which can react with water or oxygenated organics in the atmosphere.

1.2 Effects of Atmospheric Aerosols

1.2.1 Climate Effects Atmospheric aerosols have great influence on global climate change. The term radiative forcing (ΔF), which means the net radiative change at the tropopause, is introduced to estimate climate effects. Net radiative change (forcing, ΔF), defined as the difference between incoming radiative energy and outgoing radiative energy in Watts per square meter (W / m2), is related to mean surface temperature change through equation (1):

∆푇푠푢푟푓푎푐푒 = 휆 × ∆F (1) where ΔTsurface and λ represent the change of surface temperature (K) and climate sensitivity (K / (W / m2)), respectively. In general, surface temperature increases with a positive radiative forcing and drops with a negative radiative forcing.

Texas Tech University, Lulu Ma, August 2013

The climate effects caused by atmospheric aerosols can be divided into direct and indirect effects. Both scattering and absorption of solar radiation are considered direct affects.17 Equation (2) describes the radiative forcing due the atmospheric aerosols’ direct affects:

1368 ∆퐹 = − 1 − 퐴 푇2 × (1 − 푅 )22훽푏 푧 − 4푅 푏 푧 (2) 4 퐶 푎푡푚 푠 푠푐푎푡 푠 푎푏푠

Where Tatm is atmospheric transmission; Ac and Rs are fraction of cloud cover and reflectivity of earth’s surface; β, bscat, babs and z represent up-scatter fraction, scattering coefficient, absorption coefficient, and path length, respectively. This equation demonstrates that the radiative forcing depends mostly on scattering coefficient, absorption coefficient and backscatter fraction. By combining equation (1) and (2), the negative sign in front of equation (2) demonstrates an expected decrease on surface temperature when scattering has a greater influence and an increase on surface temperature when light absorption is strong. Direct scattering of light back to space reduces the amount of light absorbed by earth-atmosphere system and therefore plays a cooling effect on global climate. The diameter of many atmospheric aerosols is similar to the wavelength of visible light. In this situation, Mie theory is often used to predictincident light scattered in different directions.

Major light scattering aerosol types include sulfate, nitrate, and mineral dust. Aerosols that absorb significant visible radiation are BC and OC. A study of aerosols during dust events over New Delhi, India in 2006 showed an increased abundance of coarse particles due to dust storms from March to June. Meanwhile, a stronger absorption of aerosols as well as larger radiative forcing was observed during the dust storm.18 Dust from different locations does not behave exactly the same. For example, dust transported from East Asia to the Pacific did not absorb as much light as those transported from South Asia or from the Sahara Desert. As one of the most important light absorbing aerosols, BC has a large effect on positive radiation forcing and global warming. Fossil fuel BC has a radiative forcing of 0.2 ± 0.15 W / m2. BC mixed with other aerosols may increase the radiative forcing. For example, internal mixing of BC with nonabsorbing aerosol compounds amplified absorption cross section up to a factor of 2 and more absorption of light was found as a result of multiple scattering effect,19,20

Texas Tech University, Lulu Ma, August 2013

To draw a conclusion, climate change caused by aerosols varies with chemical compositions, concentrations, morphology and seasonal changes. Radiative forcing, combining scattering and absorption, is used to estimate climate change. Backward scattering of radiation usually results in a cooling effect and negative radiative forcing; while absorption always has heating effect and therefore a positive radiative forcing.

1.2.2 Health effects The diameter of atmospheric aerosols is small, so aerosols can stay in the atmosphere, transport from location to another and finally into cells through the respiratory system. Figure 1.3 is a study of 52 - year (1954 - 2006) haze data in Guangzhou, China as reproduced from Tie et al 21. Figure 1.3. (A) plots aerosol extinction coefficient (AEC) and mortality of lung cancer (deaths / 100,000 people) versus time. AEC related to aerosol concentration; a larger aerosol concentration usually resulted in a higher extinction. Before the year of 1972, AEC was very low. From 1972, AEC started to increase rapidly and its value was almost doubled in 1980. AEC value stayed at a high concentration without significant increase from 1980 until 2006. The mortality of lung cancer increased as well and showed a similar trend as AEC with a time lag. Meanwhile, figure 1.3 (B) gives the rate of cigarette smoking (defined by total consumption of cigarettes per person per year) and mortality of other cancers versus time. Neither the rate of cigarette smoking nor the mortality of other cancers showed similar trend as AEC.

A term named particular matter (PM) is introduced to estimate air quality, where PM 10 refers to aerosols with diameter of 10 μm or less and PM 2.5 refers to aerosols with diameter of 2.5 μm or less. In the beginning of January, 2013, Beijing was sulfuring from strong and long-lasting air pollution with the peak concentration of PM2.5 at 886 μg/m3 on Jan 12th.22 As a result of the air pollution, the respiratory health department of Beijing Shijitan Hospital received about 20 % more patients than usual. 23 Epidemiological studies showed that air pollution resulting from atmospheric aerosols can cause severe health effects, including enhanced mortality, cardiovascular, respiratory, and allergic diseases. General symptoms of health damage contain cough, fatigue, muscle aches and discomfort in the neck. A linear relationship between PM (PM10 and PM2.5) and various health indicators (including mortality, hospital admissions, bronchodilators

Texas Tech University, Lulu Ma, August 2013

use, and symptom aggravation) has been found within the range from 0 to 200 µg/m-3. Ultraﬁne particles with diameter smaller than 100nm are thought to be particularly hazardous to human health, because they can penetrate sufficiently and deeply into lungs and enter the blood circulation. Toxicology studies showed that particles become more toxic per mass unit with decreasing size. 1,24,25

1.3 Optical properties of Atmospheric Aerosols

1.3.1 Possible Phenomena of Light Incidence on Aerosols Figure 1.4 gives possible phenomenon of light incidence on aerosols.26 Among these phenomena, reflection, refraction, and diffraction are elastic scattering. No energy different is observed during the change of direction of light beam, and wavelength remains the same. Absorption of light can also occur, resulting in reduction of light, fluorescence emission or thermal emission. Light emission of fluorescence always have larger wavelength compared with incident light because of the energy drop. Raman scatter may occur, but this phenomenon is much weaker than the elastic effects.

Velocity of light in (c, m / s) is shown in equation (3);

푐 = 휐휆 (3)

c is a constant (3.0 × 108 m / s in vacuum); ν and λ represent frequency (Hz) and wavelength of incident light (m), respectively. The ratio of velocity of light in a vacuum to the velocity of light in a particular matter (vp, m / s) is called absolute refractive index (m), which can be described by equation (4) for non-absorbing materials:

푐 푚 = (4) 푣푝

Because velocity of light in a particular matter is always lower than that in a vacuum, so absolute refractive index is greater than unity. The situation for light absorbing materials is more complex:

푚 = 푚′ − 푚′푎푖 (5)

Where m’ is the real refractive index; i equals to −1 and a is related to the absorption effects of material. The imaginary part of equation describes the extent of light absorption.

Texas Tech University, Lulu Ma, August 2013

Optical properties of aerosols are greatly influenced by size. The size effect is described by size parameter (α), the ratio of circumference to wavelength of light:

2휋푟 훼 = (6) 휆

Rayleigh scattering can be applied when α is smaller than 0.2. When α falls in the range of 0.2 - 10, Mie theory is used to calculate optical properties of aerosols instead of Rayleigh scattering. If diameter of aerosols continue increasing and α is larger than 10, classical laws of refraction, reflection and diffraction should be applied.26

1.3.2 Extinction, Scattering and Absorption 2 The relationship between incident light intensity (I0, W / m ) and the intensity of 2 light traversing though aerosols (I, W / m ) follow Lambert-Beer Law:

−(푏푒푥푡 푧) 퐼 = 퐼0푒 (7)

-1 Where bext is extinction coefficient aerosols (m ) and z (m) is path length of light travelling through aerosols. Extinction is the net loss of light. The ratio of I to I0 is called transmission of light (I / I0). A larger extinction coefficient with constant path length usually results in a lower transmission and a lower light intensity travelling through -1 aerosols. Extinction coefficient is the sum of absorption (babs, m ) and scattering -1 coefficient (bscat, m ).

푏푒푥푡 = 푏푎푏푠 + 푏푠푐푎푡 (8)

Both of absorption and scattering coefficients are influenced by aerosols sizes, compositions and concentrations. In the situation of monodisperse, absorption and scattering coefficients can be described by equation (9) and (10):

2 푏푠푐푎푡 = 휎푠푐푎푡 푁 = 휋푟 푄푠푐푎푡 푁 (9)

2 푏푎푏푠 = 휎푎푏푠 푁 = 휋푟 푄푎푏푠 푁 (10)

3 where N, r, σscat, σabs, Qscat, Qabs represent number concentration of aerosols (# / m ), radius of aerosols(m), scattering cross-section (m2), absorption cross-section (m2), single particle scattering efficiency (unitless) and single particle absorption efficiency (unitless), respectively. The ratio of scattering cross-section to extinction cross-section is called single scattering albedo (SSA).

Texas Tech University, Lulu Ma, August 2013

휎 푏 푏 −푏 푆푆퐴 = 푠푐푎푡 = 푠푐푎푡 = 푒푥푡 푎푏푠 (11) 휎푒푥푡 푏푒푥푡 푏푒푥푡

Therefore, an SSA value of 1 indicates a non-absorbing effect of aerosols. The more light that aerosols absorb, the smaller SSA it will be. SSA depends not only on composition and morphology of aerosols, but also on wavelength of light source. In order to evaluate spectral dependence of light absorption, Absorption Angstrom exponent (AAE) is introduced and described in equation (12):

−퐴퐴퐸 푏푎푏푠 = 훫휆 (12)

Where babs is absorption coefficient, K is a constant, and λ is wavelength of light. A larger AAE always shows greater dependence on wavelength. It was found that BC had AAE values at round 1.1.While aeolian dusts common to West Texas showed AAE at about 2, indicating a stronger wavelength dependence than BC.19,27,28

1.4 Optical Measurement of Aerosols

1.4.1 Cavity Ring-down Spectroscopy (CRDS) Cavity Ring-down Spectroscopy (CRDS) is a sensitive technique that determines extinction of samples. The history of CRDS can date back to 1980s. Herbelin et al. and Anderson et al. are regarded to be precursors of CRDS.29,30,31 Herbelin et al. developed a method named cavity - attenuated phase - shift (CAPS) combing laser and two aligned high reflectivity mirror together. Time constant of CAPS was determined through a phase shift in the light beam intensity that passed through the optical cell.30 Instead of measuring phase shift, Anderson found another method to determine time constant. In this setup, a light detector was used to measure intensity of light that transmitted through the second high reflectivity mirror directly. Laser source was a continuous - wave laser. Time constant was determined by turning off laser source when transmitted light reached predetermined level.31 This method reduced variability and noise compared with CAPS, but the problem of random mode coincidences still existed, which might cause unstable intensity and occasional nonexponential decays. O'Keefe and Deacon improved this method by using pulsed laser sources instead of continuous-wave laser sources. A pulsed laser source had much shorter pulses than the roundtrip time of mirror cavity, so

Texas Tech University, Lulu Ma, August 2013

fluctuation of signal intensity was reduced and sensitivity and data was improved.32 This design forms the embryo of frequently used CRDS setup.

Nowadays, a CRDS instrument is made up of two high reflectivity mirrors (R>0.999) and a pulsed laser source. Light beam resonates between the two high reflectivity mirrors. With each round trip, a fraction of light penetrates through mirrors; this round - trip loss then forms exponential decay of light signal. Time constant (τ, s), the time for signal to drop to 1 / e of is original intensity, where e is the base of natural logarithm (2.718), is shown in equation (13):

푡 휏 = 푟 (13) 2 1−푅 +푏푒푥푡푡표푎푙 퐿 where tr is the time of round trip for laser beam (s), R is reflectivity of mirror, L is the -1 path length of sample flow cell (m), and bexttotal is the total extinction coefficient (m ) of sample. bexttotal is related to scattering and absorption of both aerosol sample and gas -1 phase. Extinction coefficient of aerosol sample (bext, m ) can be obtained by measuring time constant of aerosol sample and filtered air.

1 1 1 푏푒푥푡 = ( − ) (14) 푐 휏푠푎푚 휏푎푖푟

Equation (14) shows the calculation of extinction coefficient of aerosol sample. c is the velocity of light in vacuum; τsam and τair represent time constant of aerosol sample and filtered air, respectively. Reflectivity of high reflectivity mirrors has a great influence on light signal. Usually, a higher reflectivity will result in a larger time constant and lower detection limit. Typical detection limit of CRDS can range from 0.2 to 12 Mm- 1.33,34,35,36

1.4.2 Nephelometer A nephelometer is an instrument measuring scattering coefficient of sample. One of the earliest nephelometer instruments was built by Beuttell and Breverin 1949. 37 Though there are many types of nephelometers available, both lab-made and commercial, the basic principles of nephelometers are similar.26,38,39,40 An example of a integrating nephelometer is shown in Figure 1.5. When flowing through the optical cell, aerosols are illuminated by light generated from a flash tube. The backscatter shutter placed in front

Texas Tech University, Lulu Ma, August 2013

of light source blocks the light that is reflected backward by aerosols. Light trap provides a dark background. Several apertures are used to eliminate light integration angle. Light detection is accomplished by photomultiplier tube (PMT). Light scattering by aerosols at the right side of optical cell is thought to be backward scattering, while light scattered by aerosols at the left side is considered as forward scattering. In a real nephelometer, it is impossible to collect scattered light over a full angle range. This limitation on measurement of scattering is called truncation angle. Determination of truncation angle depends on both device design and sizes of aerosols. The typical truncation angles of nephelometers are within the range of 7 to 15°. Usually, the larger the sizes of aerosols, the bigger truncation errors occur. Truncation errors can increase to about 50 % for aerosols at 5 μm. Truncation error for droplet measurement can also be significant due to heating of droplets under illumination.26,41,42,43,44

1.4.3 Particle Soot Absorption Photometer (PSAP) Particle Soot Absorption Photometer (PSAP) is a widely used filter - based methods measuring light absorption of aerosols. Multi - wavelength PSAP is now commercially available (467, 530, and 660 nm). PSAP has two flow cells: one for air samples containing aerosols and one for clean air which is used as blank. 45 Apparent -1 absorption coefficient (b’abs, m ) of sample is described in equation (15):

퐴 퐼 푏′ = ( 0) (15) 푎푏푠 푉 퐼

Where A is the area of sample spot (m2); V is the volume of air sample flowing 3 through the filter during the time period of sample (m ); and I0 and I are average filter transmittance in the prior time period and current time period, respectively. This apparent absorption coefficient does not account for the bias of absorption due to heavily loaded sample on filter which causes a decrease in actual path length during measurement. After -1 correction, actual absorption coefficient (babs, m ) is shown in equation (16):

푏′ 푏 = 푎푏푠 (16) 푎푏푠 2 0.5398휏+0.355

τ is filter transmission, which is always set to 1 for unloaded filters. The calibration is valid until the transmission of filter drops below 0.7. Therefore, once the transmission drops below 0.7, the PSAP filter must be replaced. 44,45,46,47

Texas Tech University, Lulu Ma, August 2013

Table 1.1 Sources and Estimations of Global Emissions of Atmospheric Aerosols2 Amount, Tg/yr [106metic tons/yt] Sources Range Best Estimate Natural Soil dust 1000-300 1500 Sea salt 1000-10000 1300 Botanical debris 26-80 50 Volcanic dust 4-10000 30 Forest fires 3-150 20 Gas-to-particle conversion 100-260 180 Photochemical 40-200 60 Total for natural sources 2200-24000 3100 Anthropogenic Direct emission 50-160 120 Gas-to-particle conversion 260-460 330 Photochemical 5-25 10 Total for anthropogenic sources 320-640 460

Texas Tech University, Lulu Ma, August 2013

Figure 1.1 Study of aerosol concentration and composition in 30 cities around world. Blue, black and pink indicate three types of sampling location: urban areas, <100 miles downwind of major cities and rural / remote areas >100 miles downwind. Average mass concentration and chemical composition of PM10 are shown in Pie charts; where green, red, blue, orange and purple represent organics, sulfate, nitrate, ammonium, and chloride, respectively.(Obtained copyright clearance from Zhang, Q.; Jimenez, J. L.; Canagaratna, M. R.; Allan, J. D.; Coe, H.; Ulbrich, I.; Alfarra, M. R.; Takami, A.; Middlebrook, A. M.; Sun, Y. L.; Dzepina, K.; Dunlea, E.; Docherty, K.; DeCarlo, P. F.; Salcedo, D.; Onasch, T.; Jayne, J. T.; Miyoshi, T.; Shimono, A.; Hatakeyama, S.; Takegawa, N.; Kondo, Y.; Schneider, J.; Drewnick, F.; Borrmann, S.; Weimer, S.; Demerjian, K.; Williams, P.; Bower, K.; Bahreini, R.; Cottrell, L.; Griffin, R. J.; Rautiainen, J.; Sun, J. Y.; Zhang, Y. M.; Worsnop, D. R. GEOPHYSICAL RESEARCH LETTERS 2007, 34, L13801.)11

Texas Tech University, Lulu Ma, August 2013

Figure 1.2 A highly simplified mechanism of atmospheric oxidation of volatile organic compounds (VOCs). (Reproduced from Kroll, J H.; Seinfeld, J. H. Atmospheric Environment 2008, 42, 3593-3642.)16

Texas Tech University, Lulu Ma, August 2013

( A)

( B)

Figure 1.3 A study of 52 - year haze data in Guangzhou, China. Year of measurement is represented in x-axis. Figure 1.3 (A) shows aerosol extinction coefficient (AEC, Mm-1) (1954 – 2006) in black dotted-line, its 10 - year mean values in black dots, and mortality of lung cancer (deaths / million people) (1964 – 2002) in red dots. A time lag between AEC and mortality of lung cancer is observed and shown by the double arrow. In comparison, figure 1.3 (B) gives mortality due to other cancers (deaths / million people) (1964 – 2002): nasopharyngeal in blue dots-line, leucocythemia in green dots - line, and cervical cancers in black dots-line. The red dots - line shows rate of cigarettes smoking (number / people / year) during 1954–2006. (Obtained copyright clearance from Tie, X.; Wu, D.; Brasseur, G. Atmospheric Environment 2009, 43, 2375-2377)21

Texas Tech University, Lulu Ma, August 2013

Figure 1.4 Possible phenomenon of light incidence on aerosols.(Reproduced from Redmond, H. E.; Dial, K. D.; Thompson, J. E. Aeolian Research2010, 2, 5-26)26

Texas Tech University, Lulu Ma, August 2013

Figure 1.5 An example of integrating nephelometer.

Texas Tech University, Lulu Ma, August 2013

CHAPTER II

OPTICAL PROPERTIES OF AEOLIAN DUSTS COMMON TO WEST TEXAS Source: Aeolian Research 2011, 3, 235–242

2.1 Introduction Aerosols, defined as suspensions of small liquid or solid particles in a gaseous medium, are significant sources of light extinction in the lower atmosphere.17 Particles between 0.2 and 10 µm in size are excellent scatterers of sunlight, so large concentrations can distort visibility and sky color from the ground. The particles can also modify the solar actinic flux at different levels of the atmosphere and at earth’s surface, thereby altering surface temperatures and climate. This behavior is known as the “direct climate effect” of aerosols because the airborne particles play an immediate role in affecting atmospheric temperature.10 These same particles can also affect clouds through so - called “indirect effects” by becoming seed nuclei for cloud formation, lengthening cloud lifetimes or changing the water droplet size.10, 48, 49 The extent of these effects has not yet been fully understood, but the mechanisms described explain how aerosols can affect climate change.

Several lines of experimental evidence support the view that aerosols influence climate. While greenhouse gases such as CO2 and methane warm the troposphere, aerosols can reverse this trend by scattering light away from the earth’s surface. Oppenheimer has documented the eruption of Mount Tambora in 1815, in which approximately 60 Mt of sulfur were expelled into the stratosphere.4 The immense amount of particulate matter in the atmosphere shielded enough sunlight to cause a global mean temperature drop the next year of 1-2℃ and famine due to widespread crop failure. According to Self et al.,50 the aerosols produced by the eruption of Mount Pinatubo in 1991 had a stronger cooling effect than the warming impact of either El Niño or greenhouse gases in that period. A study conducted by Travis et al.51 found that the lack of jet contrails in US airspace for three days after the 9 / 11 World Trade Center disaster may have produced a three-day increase in the US diurnal temperature range (DTR).

Texas Tech University, Lulu Ma, August 2013

In addition, Aeolian (wind - driven) processes play a key role in the effects of aerosols on climate. Wind disperses particles from earth’s surface into the atmosphere; a recent estimate suggests 1 - 3 Pg of dust aerosols are dispersed into the atmosphere every year.52 The wind often combines with loose soil and dust to form dust storms and sand storms. These clouds of particles often form in areas undergoing desertification and pose health risks to humans and animals downwind of source regions. For example, Beijing is plagued each spring by storms originating from the Gobi desert and Tibetan Plateau, almost a thousand miles away.53 Given the large quantities of dust aerosols entrained into the atmosphere, it is envisioned they can play a key role in affecting local and regional climate.

Since soil - originating dust varies with soil type, which varies with location, a number of studies have been conducted all over the world to investigate local soils and their characteristics when carried by the wind.53,54,55,56 However, research has yet to investigate the optical properties of West Texas soil dust aerosols. Such research merits interest because of the geographic and climatic nature of this region; flat plains, high temperatures and a dry environment promote topsoil erosion and dust storm formation. Additionally, a study conducted by Cook et al.57 using computer simulations found that wind - blown soil aerosols may have amplified the effects of the 1920s Dust Bowl. Soil already dry from drought, loosened by crops being removed from the ground, increased the atmospheric dust aerosol loading and reduced the formation of precipitation, creating feedback that accelerated into the Dust Bowl. Measuring the optical properties of regional mineral dust aerosols would aid in understanding how regional soils may have affected the Dust Bowl era and how dust storms might currently influence climate / radiative transport.

In this work, the optical properties of two types of soil dust aerosols common to West Texas were studied. A dust generator was used to simulate soil erosion and aerosolization. The resulting aerosol was then examined by several measurement instruments collecting data simultaneously. Four parameters were measured, including 3 particle size distribution, mass concentration (g / m ), optical absorption coefficient (babs, -1 -1 m ) and optical extinction coefficient (bext, m ). From the instrumental data we have extracted a mass - extinction coefficient for each soil type. This parameter relates mass 18

Texas Tech University, Lulu Ma, August 2013

concentration of dust aerosols and their optical effects. Additional intrinsic aerosol properties that have been determined include the aerosol Angstrom extinction exponent (AEE), Angstrom absorption exponent (AAE), and single scatter albedo (SSA) for three visible wavelengths. Experimentally measured values are compared with similar quantities from other soil studies both to relate soil characteristics and to explain differences. The results presented begin to allow a better understanding of how mineral dust/soil aerosols dispersed in West Texas may affect radiative transport in the atmosphere.

2.2 Experimental

2.2.1. Dentition of Fundamental Optical Parameters In optics, the Beer - Lambert Law describes the attenuation of a beam of monochromatic light passing through an optically dense medium, and is often written as:

I T = = b−bext Z (16) I0 where the transmittance (T, unitless) equals the ratio of the final intensity (I) to the initial intensity (I0 ); bext is the extinction coefficient and z is path length. Units can vary for bext and z; for this paper, extinction coefficients will be expressed in m-1 and path length in meters unless specified otherwise. The extinction coefficient can be expressed as the sum of the scattering (bscat) and absorption coefficients (babs):

bext = bscat + babs (17)

2.2.2 Soil Types, Sampling Locations, and Physical Properties Two types of West Texas soils were chosen for this research: Pullman clay loam and Amarillo loamy fine sand. These soils are common to the upper Panhandle of Texas (General Soil Map of Texas)58 and differ in the amount of fine suspendable particles. The Pullman clay loam collected from a site located at the Texas Tech University field laboratory in northeast Lubbock County, Texas (101.78°W; 33.75°N). The Amarillo loamy fine sand was collected at the USDA, Agricultural Research Service, Big Spring Field Station (101.49°W; 33.27°N). Both sites were planted to cotton (Gossypium hirsutum) production prior to the sample collection.

Texas Tech University, Lulu Ma, August 2013

The dry colors of the two test soils were brown as defined from Munsell Soil Color Charts but differed in dry soil color values and chromas. The hues for both soils were 7.5 YR. The Pullman had a higher color value (5) and chroma (4) than the Amarillo, with value of 4 and chroma of 2. Further optical characterization of the soils was completed by recording diffuse reflectance spectra via the dried-ground soil scanning method of Waiser et al.59 Spectra were recorded from 350 to 2450 nm using the ASD AgriSpec (Analytical Spectral Devices, Boulder, CO) on bulk Amarillo and Pullman soil samples (collected from field site) and on the dusts collected in the main metal chamber of the LDGASS system after aerosolization (see below). It should be noted that the diffuse reflectance experiment was conducted on bulk samples, not dispersed dusts.

2.2.3 Soil Textures Soil texture, classified based on the relative content of soil separates: clay (< 0.002 mm), silt (0.002 to 0.05 mm) and sand (0.05 to 2.0 mm), is a soil property strongly associated with soil chemical and physical properties useful in agricultural and engineering applications.60 The hydrometer method was employed to measure the soil texture.61 The Pullman soil had 37.4 % sand, 32.6 % clay and 30.0 % silt, and classified as a clay loam. The Amarillo soil is defined as loamy fine sand with 82.5 % sand, 7.4 % clay and 10.1 % silt.

2.2.4 Organic Matter Content and Particle Density of Soil The organic content of the soil was determined by loss on ignition.62 The soils were weighed (w1) after being heated at 110 ℃ for 24 h. During this process, water was evaporated. Then they were weighed again (w2) after being heated overnight at 400 ℃. Organic material is removed in such a high temperature environment. The organic matter content was 3.33 % and 1.15 % for Pullman and Amarillo, respectively. An AccuPyc 1330 Pycnometer (Micromeritics) was used to measure particle mass density (g / cm3) by a gas displacement technology. Air was filled in a sample chamber and then was discharged into a second empty chamber. By comparing different pressures using Ideal Gas Law, the volume of solid samples was found. Particle mass density was 2.46 g / cm3 for Pullman and 2.49 g / cm3 for Amarillo soil samples, respectively.

Texas Tech University, Lulu Ma, August 2013

2.2.5. Generation of Soil Dust A schematic of the Lubbock dust generation and analysis system (LDGASS) employed is shown in Figure 2.1. The purpose of the LDGASS was to model wind erosion of soil or dust in a laboratory setting.63 The generator works by mechanically agitating a sample of soil or other mixture inside a “dustfall tube”, producing a dust which is drawn into an approx. 1.5 cu. ft. metal chamber through use of a slight vacuum pull. The total flow from the dustfall tube was 6.5 LPM (litter per minute). From this housing, dust was sampled and analyzed by several instruments simultaneously. In the experiments described, a 25.00 g soil sample was added to the dustfall tube for each run.

2.2.6. Measurement Instruments Four instruments sampled from the dust generator main chamber (Figure 2.1). Particle size distribution was determined using a “Filter-check TM” SubMicron Aerosol Spectrometer / Filter Efficiency Monitor Model 1.108 (GRIMM Technologies, Inc.) that employs light scattering analysis at 655 and 780 nm wavelength to measure the number of particles in a sample within 15 size channels. The channels include > 0.3, > 0.4, > 0.5, > 0.65, > 0.8, > 1.0, > 1.6, > 2.0, > 3.0, > 4.0, > 5.0, > 7.5, > 10, > 15, and > 20 μm in diameter. Mass concentration of the dust was measured using a Thermo Anderson MIE DataRAM Aerosol Monitor DR - 2000 (Ashtead Technology) equipped with a PM10 inlet head. A PM10 inlet head prevented particles larger than 10 μm aerodynamic diameter from entering the instrument. The instrument estimated mass concentration based on the light scattering principle. Mass concentration of dust present in the measurement chamber was also reconstructed from the size distribution data by treating all particles within a bin as if they were spheres characterized by the bins midpoint diameter. Mass concentration was then determined through equation (18):

4 C = 20 πr2DP (18) r=0.2 3

C is mass concentration (g/m3), r represents the arithmetic average radius of particles (m), D is the density of dust, and P is the number of particles within r range (particles/m3).

The Radiance Research 3λ Particle Soot / Absorption Photometer (PSAP) measured the absorption coefficient (babs) of our dust samples at 470, 520, and 660 nm of wavelengths by measuring the transmittance of LED (Light-emitting diode) light through

Texas Tech University, Lulu Ma, August 2013

a particle laden filter as compared to a reference filter. Lastly, a transmissometer was constructed from a 2.5 cm diameter copper tube, apertures to eliminate forward scattered light, a fiber coupled UV - VIS light source (Hamamatsu UV - VIS tungsten - deuterium lamp), and a fiber optic photodiode array spectrometer (Compass X) (Figure 2.1). The copper tube itself served as a flow through sample cell and was approximately 1.5 m in length. Portions of the inside of the tube were painted black to minimize reflections. The optical path length was 0.90 m. Holes for sample entrance, air entrance, and air & sample exit were cut in the tube, and metal pipes were attached with cold weld glue. Small holes approximately 3 mm in size were drilled through rubber disks that were placed snugly inside the tube, in order to reduce internal reflections and prevent forward scattered light from reaching the detector. The PSAP instrument was placed directly after the transmission cell in an attempt to assure the measured aerosol sample is as similar as possible. No size selective inlet was used for either the GRIMM instrument or the optical measurement devices. Volumetric flow rate through the PSAP/transmission cell ranged from 0.65 to 1 LPM. Spectroscopic blanks (100 % T) were set using particle free room air provided through a filter and 3 - way ball valve plumbed into the system between the main mixing chamber and the optical instruments.

2.3. Results and Discussion

2.3.1 Observed Particle Size Distributions for Pullman and Amarillo Soils Figure 2.2 reports example particle size distributions measured by the GRIMM instrument for both soil types used in this study. As observed, most of the dust particles for both soils were found to be less than 5 μm in diameter with peaks in the distribution within 2 - 3 μm size channels. Of total particle counts recorded, 81 % were < 2.5 μm for the Amarillo soil, while only 50 % of counts were < 2.5 μm for the Pullman sample. This indicates the dispersed Amarillo soil features a higher fraction of fine - mode particles. It is difficult to compare the particle size distributions observed in this study with the case of authentic atmospheric dusts since coarse mode dust would be externally mixed with a large concentration of smaller particles produced from other sources in the latter case. Additionally, the size distribution would be expected to be heavily influenced by distance from source regions. Nonetheless, some available evidence from the literature suggests

Texas Tech University, Lulu Ma, August 2013

atmospheric dust also may consist of particles in similar size. For instance, Cheng et al. report a peak in the 2 - 3 μm size channel in the particle size distribution during heavy and moderate dust loadings at Hunshan Dake, China. 64 However, other reports for Saharan dusts suggest effective diameters in the 5 - 10 μm range with no simple parameterization possible for size distribution. 65 Also, Crabtree has documented differences in observed particle size distribution for sites on the Southern High Plains of differing distances from source regions. 66 Therefore, we report our observed size distributions simply as experimental documentation rather than suggesting they are representative of the authentic atmospheric case.

2.3.2 Dust Optical Properties: the Spectral Dependence of Dust Extinction and Absorption Measurement of the spectral dependence of light absorption and extinction by dusts is important for building models of how dust aerosols impact regional actinic flux and photochemistry. A simple model to quantitatively describe the spectral dependence of absorption and extinction by aerosols is to fit wavelength resolved extinction and absorption data to a power law such as:

τ = βλα (19)

In this equation, τ is the optical depth due to a particular effect (extinction, scattering, or absorption), β is a proportionality constant, λ is the wavelength, and α is a term known as the Angstrom exponent. The Angstrom exponent describes the spectral dependence of each optical effect and usually ranges between 0 and 4, but can even be negative. Generally, smaller particles produce a larger Angstrom extinction exponent (AEE), approaching 4 for very small particles. In this study, we report values for both Angstrom extinction exponent (AEE) and Angstrom absorption exponents (AAE) for lab generated Amarillo and Pullman soil dusts.

2.3.3 Angstrom Extinction Exponent (AEE) The wavelength resolved extinction coefficients of both soil types were computed from experimental transmittance results, normalized to extinction at 250 nm of wavelength, and graphed as a function of measurement wavelength. Figure 2.3 illustrates

Texas Tech University, Lulu Ma, August 2013

a plot of normalized extinction as a function of wavelength for both (A) Pullman and (B) Amarillo soil types. A best - fit power law distribution is also included in the plot.

The measured Angstrom extinction exponents for the soil types were 0.1112 and 0.1681 for Pullman and Amarillo, respectively. While the power law fit is a common approach to describe spectral dependence of optical effects, Figure 2.3 illustrates that the observed data often deviates significantly from the model. In fact, the R2 value for power law fits to the data were only 0.443 and 0.404 for the Pullman and Amarillo samples. This was particularly true for the Amarillo soil type. Nonetheless, the numerical value of AEE is strongly linked to particle size distribution of the aerosol itself and is therefore not an intrinsic property (at least not tied only to composition). Any differences between size distributions between our experiments and the dust suspended in the atmosphere would lead to differences in AEE. However, it is not uncommon to encounter AEE values near zero during dust events in the region. The AERONET site at Bushland, TX (190 km north of our location) reports AEE values approaching zero for several days dominated by dusts during a monitoring period in the spring / summer of 2008. We have attempted to model spectral extinction by the aerosolized dust samples using observed size distribution data and a Mie code, but attempts to replicate the complex spectrum observed has been largely unsuccessful. This is likely a consequence of the non - spherical nature of mineral dusts. Computer codes for non - spherical particles do exist, but they often are computationally expensive and complex. It is considered beyond the scope of this work to attempt to reconstruct optical data based on the more complex computational methods.

2.3.4 Angstrom Absorption Exponent (AAE) and Diffuse Reflectance The absorption of light at various wavelengths by aerosols is also commonly modeled by a power law, of which the exponent is known as the Angstrom absorption exponent (AAE). In this study, this quantity was calculated from absorption coefficients matched to the blue, green and red wavelengths (470, 520 and 660 nm, respectively) of the PSAP. A power law regression was performed on the three points and the AAE was determined. Figure 2.4 illustrates the absorption data for the Pullman and Amarillo soil. An AAE of 1.73 was derived for Pullman soil, while an AAE of 2.17 was found for Amarillo soil. It should be noted that the PSAP filter quickly became loaded with

Texas Tech University, Lulu Ma, August 2013

particulate matter and the filter transmittance dropped to unacceptable levels very rapidly. Only a small, early portion of the PSAP data was used in this analysis, because much of the later data was deemed potentially inaccurate. These values of AAE are considerably smaller than values of 3.75 - 6.4 previously reported for the AAE of Saharan mineral dusts.67 The exact reason for this difference is unclear, but likely due to the compositional differences between soil types. To the best of our knowledge, this is the first report of AAE for soil dust aerosol from the Southern High Plains of Texas. This parameter is of particular importance to the climate effects of dust aerosols as absorption drives atmospheric heating.

Diffuse reflectance spectra were also collected on both bulk soils and collected dusts generated with the LDGASS system. Diffuse reflectance spectra are shown in Figure 2.4 (C). It should be noted that single scatter albedo is not equivalent to diffuse reflectance due to the differences in photon transport within dispersed dusts versus optically dense media (such as the bulk samples). However, we might expect general trends in the diffuse reflectance spectra to also be observed in the absorption / albedo dataset. To a large extent, this similarity is observed for the visible region. Both the PSAP data and reflectance data suggests absorption increases with decreasing wavelength. Another interesting aspect of the dataset is the apparent similarity between the Pullman and Amarillo soil dusts collected after aerosolization (dashed line). Amarillo dust demonstrates slightly lower reflectance at higher visible frequencies and in the near IR when compared to Pullman. The visible data observed in Figure 2.4 (C) is consistent with the shape of the best fit curve applied to the PSAP absorption data reported in Figure 2.4 (A) and (B). A large difference exists in the reflectivity data for bulk soil samples compared to the dusts generated by the LDGASS system. This is likely a consequence of differences in the particle size distribution between the samples. Diffuse reflectance data is known to be affected by particle size. Lastly, an additional interesting feature of the diffuse reflectance data is the presence of distinct absorption bands in the near IR. Kaolinite, smectite, and muscovite exhibit several strong absorption bands between 1400 and 2600 nm that are associated with -OH and water.68 It is envisioned these spectral features could be used to develop optical dust concentration monitors more selective for

Texas Tech University, Lulu Ma, August 2013

soil dusts relative to other aerosol types. This could be of particular interest when coupled with the recently developed aerosol albedometer instrument.69,70

2.3.5 Estimate of Single Scatter Albedo (SSA) of Pullman and Amarillo Soil Dust Aerosols

In this experiment, extinction coefficient (bext) and absorption coefficient (babs) are measured in sequence. Use of both data sets allows estimation of dust single scatter albedo (SSA) at the three PSAP wavelengths (470 nm, 520 nm, 660 nm). SSA is simply the ratio between scattering and extinction coefficient:

푏 푏 −푏 푆푆퐴 = 푠푐푎푡 = 푒푥푡 푎푏푠 (20) 푏푒푥푡 푏푒푥푡

SSA describes the fraction of light lost from the beam which is removed through scattering. Table 2.1 reports observed SSA values for the soil dusts considered in this study. For the Pullman soil sample, the measured SSA ranged from 0.947 to 0.968 across the visible spectral range. The SSA for the Amarillo soil was slightly higher ranging from 0.972 to 0.986. The SSA of both soils increased at longer wavelengths. The SSA values observed in this study and their spectral dependence are within the range of values typically reported for mineral dust aerosols.26 The model of Claquin et al. is useful to predict SSA values for soils globally and in our region.71 These authors estimate SSA based on soil chemical composition and a mixing model to be 0.92–0.94 (550 nm) for dust aerosols typical of our location. This SSA value is slightly lower than what we have measured. Furthermore, recent measurements made on the ambient aerosol at Lubbock, TX during a period when substantial wind-blown dust was present indicated SSA ≈0.86 at 532 nm.70 However, it is unclear to what extent this ambient dust was influenced or modified by anthropogenic activities. Also, the particle size distribution in the two experiments would be expected to differ significantly.

2.3.6 Mass-Extinction Coefficient (MEC) of Pullman and Amarillo Soils A useful parameter linking aerosol optical effects and mass concentration (M, g / 3 2 m ) is the mass-extinction coefficient (σext, m / g).

푏푒푥푡 = 휎푒푥푡 × 푀 (21)

Texas Tech University, Lulu Ma, August 2013

The mass-extinction coefficient (σext) relates the optical extinction of an aerosol with its mass concentration. To calculate it, the average observed extinction coefficient (bext) at 522 nm wavelength was plotted against the measured mass concentration. The slope of a best - fit line through the data points is then reported as σext. The results of this analysis are illustrated in Figure 2.5. The observed MEC for the Pullman sample was roughly 2.75 m2 / g while the MEC for the Amarillo soil was 2.97 m2 / g when the DataRAM instrument was used to estimate mass concentration (shown in Figure 2.5 (A) and (B)). An alternate analysis was performed when mass concentration was estimated by reconstruction through use of the GRIMM size distribution data. This approach was pursued since the DataRAM instrument uses a PM 10 size-selective inlet while the optical cell and GRIMM instrument for size distribution data did not. As a result, the instruments are presented with slightly different samples. The MEC values were 1.77 and 1.74 m2 / g for Pullman and Amarillo soils when using the reconstruction approach, although linearity of the plots was not as good.

A review of recent literature concerning dust optical properties reveals that the mass-extinction coefficients found in this experiment are significantly higher than those reported elsewhere. Gerasopoulos et al.72 concluded a specific extinction cross-section of 0.64 ± 0.04 m2 / g at 500 nm for measurements in Athens. Malm and Day, in studying the composition of aerosols in the Grand Canyon assumed mass-scatter efficiency values of 0.4-0.6 m2 / g for soil-derived dust aerosols. 73 The IMPROVE program, which reconstructs estimates of aerosol light extinction based on the concentrations of different particle types, estimates a scatter efficiency of 1.0 m2 / g for soil dust particles in the air.74 The exact reason for the observed difference between our measurements and those referenced is not immediately clear.

2.4 Summary and conclusions This experiment examined laboratory generated Aeolian dust mimics from Pullman and Amarillo soil types to determine their optical properties. Values for the mass-extinction coefficient were measured as 2.75 and 2.97 m2 / g, respectively at 522 nm when using DataRAM data and 1.74-1.77 m2 / g when using reconstructed mass concentrations. The extinction coefficient (bext) was plotted over 200 - 800 nm and bext

Texas Tech University, Lulu Ma, August 2013

was found to slightly increase with wavelength for both soils. Angstrom extinction exponent ranging from 0.111 to 0.168 were observed for the soil types used. Angstrom absorption exponents modeling the dependence of babs on wavelength were measured as 1.73 for Pullman and 2.17 for Amarillo. Observed SSA values ranged from 0.947 to 0.980 depending on soil type and wavelength. SSA increased with increasing wavelength for both soil types. While observed SSA values were close to those previously reported for mineral dusts, the MEC we report are significantly higher than those previously reported and commonly used within the aerosol community for mineral dust aerosols.

Data reported in this manuscript may find use in estimating mass concentration of dust in the atmosphere or providing estimates of erosion rates from fields through use of optical sensors. Alternatively, the AEE, AAE and SSA values we report for soil types in our region may be useful to model current radiative transport in the atmosphere, and better understand the impact of dust on historical climatic effects like the Dust Bowl era.

Texas Tech University, Lulu Ma, August 2013

Table 2.1 Observed single scatter albedo (SSA) values for Pullman and Amarillo soils types. Soil type λ = 470 nm λ = 552 nm λ = 660 nm N (SSA ± 1 std. dev.) (SSA ± 1 std. dev.) (SSA ± 1 std. dev.) Pullman 0.947 ±0.012 0.958 ± 0.009 0.968 ± 0.006 5 Amarillo 0.964 ± 0.012 0.973 ± 0.01 0.980 ± 0.007 10

Texas Tech University, Lulu Ma, August 2013

Figure 2.1 Schematic of the dust generation and analysis system used in this study. Additional details of the generation system can be found in Gill et al.63 The DATARAM measures dust mass concentration, the GRIMM measures size distribution, and the transmissometer and PSAP measure extinction and absorption respectively.

Texas Tech University, Lulu Ma, August 2013

(A) (B)

Figure 2.2 Particle size distributions of Pullman (A) and Amarillo (B) soil types produced by the LDGASS. Both soil aerosols exhibited a similar size distribution dominated by super-micron particles.

Texas Tech University, Lulu Ma, August 2013

Figure 2.3 Plots of extinction vs. wavelength for lab generated (A) Pullman and (B) Amarillo soils. N refers to the number of individual spectra that were averaged. R2 values for the fits were 0.443 and 0.404 for Pullman and Amarillo, respectively.

Texas Tech University, Lulu Ma, August 2013

(A) (B)

(C)

Figure 2.4 Plots of absorption coefficient vs. wavelength for lab generated (A) Pullman and (B) Amarillo soil dusts during a particular trial. (C) represents diffuse reflectance spectra on bulk Amarillo and Pullman soils (solid lines) and residual soil dust (dashed lines) collected from the main chamber of the LDGASS. The exponent of the fit in plot (A) and (B) represents the Angstrom absorption exponent (AAE) of corresponding soil samples collected with PSAP on dispersed dust.

Texas Tech University, Lulu Ma, August 2013

(A) (B)

Figure 2.5 Plot of observed extinction coefficient at λ= 522 nm vs. dust mass loading for Pullman and Amarillo soil samples. (A) Pullman sample with mass concentration measured by the Data Ram instrument. B) Amarillo sample with mass concentration measured by the Data Ram instrument. (C) Pullman sample with mass concentration measured by reconstruction from particle size distribution. (D) Amarillo sample with mass concentration reconstructed from particle size distribution. The slope of the lines 2 represents the mass extinction coefficients (σext, m / g) as defined in this paper.

Texas Tech University, Lulu Ma, August 2013

Chapter III

OPTICAL PROPERTIES OF DISPERSED AEROSOLS IN THE NEAR ULTRAVIOLET (355nm): MEASURMENT, APPROACH AND INITIAL DATA Source: Anal. Chem. 2012, 84, 5611−5617

3.1 Introduction Fine solid and liquid particles suspended in the atmosphere (aerosols) scatter and absorb solar radiation, and this causes a climate effect that is often quantitatively estimated by the aerosol’s “direct radiative forcing” (RF).10,75 Compared to well - mixed greenhouse gases that absorb long-wave radiation, aerosols have influence mainly on short - wave radiation with a total direct RF of approximately -0.5 W / m2.10,75,76 A positive RF indicates a heating influence, while a negative RF suggests a cooling influence.

-1 Light scattering, absorption, and extinction coefficients (bscat, babs, bext; m ) are the quantitative variables often used to describe these differing effects. The extinction coefficient (bext) is the sum of the effects of absorption and scattering (see equation (22)) and is defined through equation (23).

bext = babs + bscat (22)

−bext z I = I0e (23)

I and I0 are irradiances at point z meters and at z = 0 m, respectively. Light scattering leads to a cooling of climate, while strong absorption can lead to a warming. The interplay between scattering and absorption can be described through the aerosol single scatter albedo (SSA, unitless) defined as:

b SSA = scat (24) bext

For those aerosols that have no absorption, extinction of light is only composed of scattering, and SSA equals one. If light is absorbed, SSA will fall below one.

In this work, the optical properties of aerosols are studied by combining cavity ring-down spectroscopy (CRDS) and integrating sphere nephelometry.34,35,77,78,79 The

Texas Tech University, Lulu Ma, August 2013

innovative aspect of this project is the use of a pulsed Nd: YAG laser at λ = 355 nm as a light source to access the near ultraviolet (UV) region. The literature record consists of many papers that study the optical properties of aerosols in the visible spectral range; however, there are far fewer studies in the near UV. Of available data, many studies involve column integrated inversions of direct and diffuse sky radiances from radiometer and/or sun-photometer data. A recent example of this is found in Corr et al.80 These authors estimate near UV SSA in the vicinity of 0.8 for the Mexico City metropolitan area and also present a summary of additional inversions presented in the literature for a variety of global locations with resulting SSA values ranging from 0.65 to nearly 1.0 for the 300 - 400 nm spectral window. Unfortunately, extracting aerosol optical data through these inversion methods can be prone to difficulties due to variable light absorption by atmospheric gases aloft and poorly constrained surface reflectivity estimates in the UV. The latter alters actinic flux and consequently the observed radiances. In addition, the process is time - consuming and measurements can only be made during the day and in the absence of cloud cover.

UV light is known to play a very important role in tropospheric photochemistry, and significant recent evidence suggests that a class of material present in atmospheric particulate matter (commonly called “brown - carbon”) absorbs light strongly in the near UV spectral window.81,82,83,84,85 The potential impact of this material on atmospheric radiative transport and photochemistry makes the development of techniques for single - point measurement of light scattering and absorption between 290 and 400 nm a research priority for the atmospheric chemistry community. In addition, active, single point measurements of aerosol optics offer the ability to manipulate the aerosol sample chemically or physically prior to measurement. For instance, thermo - denuders can be used to remove volatile material, or the effect of relative humidity on optical signatures can be studied. This simply cannot be accomplished using columnar inversions. At present, it appears that only photoacoustic spectroscopy has been employed to make point measurements of aerosol absorption in the UV spectral window.86,87 This manuscript documents an alternative measurement instrument and reports initial optical data for several types of common aerosols for the near UV (355 nm). It is envisioned that the

Texas Tech University, Lulu Ma, August 2013

measurement method and resulting data will assist in improving models of radiative transport in atmosphere.

3.2 Experimental Methods Figure 3.1 is a schematic of the instrument setup. The apparatus used is very similar to that reported in Thompson et al. with the major exception of the different measurement wavelength.69,70 To summarize the experimental work, the instrument sensitivity and precision has been evaluated through a series of experiments using filtered air and R - 134a gas (tetra flouroethane). We find limits of detection for bscat, bext, and babs of 12, 11, and 16 Mm-1, respectively, for a 52 s integration time. In addition, optical measurements were made on size - selected 300 nm diameter (Dp) ammonium sulfate and α - pinene aerosols. Additional measurements were made on redispersed soil - dust proxies, toluene secondary organic aerosols (SOA), and biomass burning samples. Measurements on ambient aerosols have also been made with the device, and these were compared to measurements made simultaneously with commercial products.

3.2.1 Sample Flow Path Aerosol particles generated by different methods were pumped through an electrostatic classifier (TSI Inc., Model 3080) into the sphere. A pressure transducer (MKS Instruments, 722A13TCD2FA) and a humidity and temperature probe (Vaisala, HMP60) were installed within sphere to detect pressure, temperature, and relative humidity during experiments. The outlet of the sphere was connected to a condensation particle counter (TSI Inc., Model 3772) which measured particle concentrations (# / cm3). The electrostatic classifier selects a certain size of aerosol particles. In this work, 300 nm diameter (Dp) was chosen when size selection was employed. Sample flow (1 L / min) was driven by a vacuum pump located at the outlet of the particle counter, and flow rate controlled by a critical orifice. In order to minimize aerosol losses, conductive tubing (TSI Inc via VanGuard Products Corp.) was used to connect devices.

3.2.2 Generation of Aerosol Particles. Several kinds of aerosol particles were tested: ammonium sulfate aerosols, secondary organic aerosols (SOAs), soil dusts, biomass burning aerosols and ambient aerosols. Ammonium sulfate solution was prepared by dissolving ammonium sulfate

Texas Tech University, Lulu Ma, August 2013

(Mallinckrodt Inc.) in deionized water (18.2 MΩ / cm). Then small solution droplets were generated by an atomizer (TSI Inc., Model 9302). These droplets passed through a silica gel diffusion dryer (TSI Inc., Model 3062) before being size selected by the electrostatic classifier. Pinene SOAs were produced in a 67.5 ft3 Teflon chamber by mixing a vapor of (1s)-(-)-alpha-pinene (Acros Organics) and ozone at 10 ppmv with irradiation by UV light. The vapor of (1s)-(-)-alpha-pinene was produced by adding 10 - 15µL of (1s)-(-)- alpha-pinene to a heated flask. The flask was placed in a water bath with temperature of 50-80 ℃. The method for toluene SOA generation was similar to pinene except the reactants and their concentrations, which were ozone at 7-10 ppmv, NO2 at 30-120ppb, water vapor, and toluene. In this process, the 67.5 ft3 chamber was filled with ozone,

NO2 and water vapor first. Water vapor was made by running humidifier installed inside chamber for 5 min. Then 600-1600 µL of toluene was added into a heated flask placed in the same water bath system as described previously. Toluene vapor was introduced into the chamber and reacted with ozone, NO2 and water during irradiation by UV light. Toluene and pinene vapor phase concentrations were not measured.

For soil dust and biomass burning aerosols there was no size selection for optical measurements. Two kinds of soil dusts common to the West Texas Panhandle were chosen: Pullman and Amarillo soil dusts. Pullman soil was from the Texas Tech New Deal research farm and Amarillo soil was from Lloyd Field in Big Spring, Texas. Before measurements, the soil samples were dried in oven overnight. For an experiment, 1.0 - 1.2 g Pullman soil or 1.0 g Amarillo soil was dispersed inside a flask by mechanically stirring soil samples. The dispersed dust was then transferred directly to the sphere for measurement under an air purge. Figure 3.2 illustrates size distributions for the Amarillo and Pullman soil types as measured with a MetOne optical particle counter. These size distributions were determined in a separate experiment.

Combustion products of dry pear tree leaves and Afghanistan pine needles were used as biomass burning aerosols. A small quantity of biomass was placed in a dish within a fume hood and carefully ignited. Particles were sampled directly into the measurement sphere via a length of tubing with no further treatment. We were unable to control flaming vs. smoldering combustion, and our results likely represent an average of

Texas Tech University, Lulu Ma, August 2013

both cases. Any aerosol waste generated should be routed to fume hood waste or filtered to prevent exposure to laboratory workers.

In order to test the near UV albedometer for environmental monitoring, ambient aerosols were sampled through a laboratory window and measured without further treatment. Ambient aerosols often require drying prior to measurement, but this was not needed in this case since sample relative humidity (RH) was measured to be 10 – 12 % The optical results were compared with those from commercial monitoring products: a real - time aerosol mass concentration monitor (MIE INC. Model RAM-1) and particle / soot absorption photometer (PSAP, Radiance Research, Seattle WA). The real - time aerosol monitor recorded aerosol mass concentrations in micrograms per cubic meters, and the PSAP measured absorption coefficients of aerosols at three different wavelengths: 470nm, 552nm, and 660nm in real time. Both products were operated according to procedures specified in the user manual. Ambient results were averaged on an hourly time basis for presentation in this manuscript. The PSAP is a filter based device that is known to be subject to biases in measurement. The manufacturer of the device (Radiance Research) includes an empirical correction for non-linearity induced by effects of multiple scattering / optical absorption within the filter. This correction is based on observed filter transmittance as described in Bond et al.45 Bond et al. also comment that this correction is satisfactory if filter transmittance is > 0.7. For our measurements, we have used this correction (as computed internally within the PSAP) and filters were changed when T dropped to near 0.7. However, potential errors due to flow rate uncertainty and relative humidity effects remain uncertain. RH effects are not expected to be large for this work due to the very low humidity encountered during sampling (< 20%).

3.2.3 Optical measurements. Inner surfaces of the spherical chamber were coated with Duraflect (Labsphere Inc.), a diffuse reflective coating that has a high reflectance > 95% between 350 nm and 1200 nm. A 0.635 cm hole was drilled at each geometric pole of the sphere; and two high reflectivity mirrors (R ≈ 0.99963, 355 nm, ATFilms) were installed at the two holes outside sphere in mirror mounts. An approx. 5 ns beam of light shot by a 355nm Q- Switched Nd: YAG laser (Quantel, Brio) was introduced into sphere through the bottom

Texas Tech University, Lulu Ma, August 2013

mirror and then oscillated back and forth between the two high reflectivity mirrors. This oscillation produced cavity ring - down signal associated with extinction coefficient. The laser used in this experiment produces very powerful, short pulses of light. Care must be taken when operating the device and laser safety goggles used. A photomultiplier tube (PMT, Hamamatsu Photonics) located behind the second high reflectivity mirror collected light signals for measurement through a 355 nm band pass filter and neutral density filters were also used to attenuate this beam. There was a second PMT placed on the surface of the sphere. This was used to collect the scattered light signal. A light barrier (in shape of an arc) coated with Duraflect was placed in front of the PMT inside the sphere to block directly scattered light from reaching PMT. Again, a band pass filter was used to spectrally select the 355 nm beam. Additional details of the optical setup can be found in previous literature.70,88

3.3 Results and discussion

3.3.1 Calibration Gas Experiments Figure 3.3 (A) illustrates the results of an experiment in which the measurement cell was sequentially filled with filtered air, R - 134a, and then filtered air. First, when only filtered room air was present, both extinction and scattering coefficients were zero. Then, both measurements approached 370 Mm-1 when the sphere was filled by R - 134a. When the refrigerant gas was substituted by filtered room air again, the signals rapidly dropped back to zero. The measured absorption coefficient was zero during the entire experiment owing to the nonabsorptive nature of the gases. The spike in the data is an artifact that occurs when the refractive index (RI) of the gas within the cell changes very rapidly. It is believed this is due to a refractive effect at the interface between gases that are not yet well-mixed in the measurement cell. This experiment demonstrates monitoring changes in optical properties at λ = 355 nm is possible with the apparatus. Time resolution was limited by the washing out time of the 50 L sphere and varied between 20 and 40 min in this experiment depending on gas flow rate. The standard -1 -1 deviation of replicate filtered air measurements for bscat was 3.9 Mm and 3.5 Mm for -1 bext, and we are able to report 3σ limits of detection of 12 and 11 Mm accordingly (52 s integration time). This is roughly 1 order of magnitude poorer performance than what our

Texas Tech University, Lulu Ma, August 2013

previous instrument at λ = 532 nm was able to achieve.69 The cause of this is the poorer mirror reflectivity of UV optics compared to the visible. This decreases the effective sample path length for the experiment and reduces ring - down times, which increases relative error given a fixed imprecision in determining the ring - down constant, τ.

Absorption is measured by the difference between bext and bscat. As such, we can 89 propagate standard deviation to babs using the standard approach, and the result is 5.2 -1 Mm . This is in good agreement with the standard deviation of babs observations for zero air in Figure 3.3(A). The limit of detection for babs (3σ) for single point measurements is therefore 16 Mm-1 at 355 nm. Figure 3.3 (B) and (C) also illustrate how the variability in measured SSA scales with the magnitude of extinction / scatter coefficient. Figure 3.3 (B) is a plot of reported SSA in time as air is switched to R - 134a (same X axis as in Figure

3.3 (A)). As observed, when bext is near zero, SSA is highly variable but when bext increases, the precision of the SSA measurement improves significantly. This is a consequence of the SSA being a ratio of two measured values, each with its own imprecision. Propagation of relative uncertainty to SSA from the std. dev. of replicate bscat and bext measurements yields the plot shown in Figure 3.3 (C), which suggests bext ≥ 60 Mm-1 is required to achieve < 10% relative uncertainty for single point measurements of SSA.

3.3.2 Size-Selected Ammonium Sulfate Aerosols and Secondary Organic Aerosol (SOA). Figure 3.4 and Table 3.1 report results of experiments on size selected ammonium sulfate and secondary organic aerosol (SOA) produced from oxidation of α-pinene with ozone. A linear relationship between particle concentration and extinction coefficient is found in each measurement. Extinction cross sections (slope of the resulting lines) of Dp= 300 nm ammonium sulfate and pinene SOA were 2.4 (± 0.8) × 10-9 cm2 and 1.65 (± 0.14) × 10-9 cm2, respectively. Linear relationships were also found between particle concentration and scattering coefficient. The corresponding scattering cross sections are 2.6 (± 1) × 10-9 cm2 and 1.71 (± 0.09) × 10-9 cm2. Since the measured scattering cross sections for ammonium sulfate and α-pinene are statistically indistinguishable, no light absorption at λ = 355 was observed for these particle types.

Texas Tech University, Lulu Ma, August 2013

In these experiments, the highest particle concentration was about 103 cm-3 and particles with 300 nm diameter were selected with an electrostatic classifier, so aerosol particles are far away from each other inside the sphere and refractive index can be calculated from Mie theory via a method previously described.90 One known caveat to the size selection process is multiply charged particles of differing diameter making it through the size selection process and being measured. We have corrected the optical measurements for this effect when necessary (determined to be approximately 5 % of optical signal for ammonium sulfate and <1 % for α - pinene). Table 3.1 lists scattering and extinction cross sections observed and extracted refractive indices. Cross section uncertainties reported are 95 % confidence intervals. For refractive indices, uncertainty is governed by the approximate range of possible RI values that are bounded by the confidence intervals for the cross sections. As observed, best - fit refractive indices were 1.45 for α - pinene and 1.56 for ammonium sulfate. The refractive indices of ammonium sulfate and α - pinene SOA obtained from these measurements are very similar to values reported previously by Toon et al. and Nakayama et al. for 355 nm.91,92

3.3.3 Soil Dusts and Biomass Burning Aerosols Table 3.2 lists SSA values of soil dusts, biomass burning aerosol, and toluene SOA. The dust and smoke have SSA well below one, indicating that they absorb light at 355 nm. For soil dust and biomass burning aerosols, no size selection was used and the particles were likely not spherical so absolute cross sections cannot be determined. Pullman and Amarillo soil dusts are native to the West Texas Panhandle and regional dust storms commonly aerosolize these materials. Additionally, both soil dust and biomass burning aerosols have large global source strengths, and knowledge of typical SSA values is of scientific significance. Both biomass combustion aerosol samples had similar SSA of 0.83 - 0.84. This value is in reasonably good agreement with that expected from extrapolation of the data of Dubovik et al. for authentic atmospheric smoke aerosols.93 However, the combustion conditions (flaming vs. smoldering) are known to significantly affect soot production and therefore alter SSA. In this work, we took no controls to ensure either flaming or smoldering conditions, and both were present during the experiment. For the soil dusts, Amarillo soil appeared to be significantly more light absorbing compared to Pullman. This is consistent with previous results on these soil

Texas Tech University, Lulu Ma, August 2013

samples observed for the visible region.28 24 In addition, the SSA values reported here for the soil dusts at 355 nm are in good agreement with values expected from extrapolation of spectral trends previously presented in the literature.26,94 For the toluene SOA experiments, size selection was used, but we were unable to produce an aerosol of sufficient monodispersity to extract refractive index values. It is believed that 300 nm particles were by far the most abundant in the sample, but a sufficient number of approximately 500 nm diameter particles were present to increase the optical signals by an expected value of 10 - 20%. The mean SSA of the toluene aerosol mixture observed was 0.95 ± 0.03 (mean ± 95 % conf. int.). Nakayama et al. have recently reported that toluene SOA absorbs light at 355 nm (SSA of approximately 0.83 from their data).92 It is well - known that nitrated aromatics expected to be present in the toluene SOA at high 95,96 NOx absorb light in the near UV and visible region. The exact reason strong light absorption was not observed for toluene SOA in this experiment is unclear but is likely related to differences in reaction conditions between experiments.

3.3.4 Ambient Measurements Ambient samples were also measured during the course of our investigations. Aerosol samples were pulled into the sphere directly from outside a laboratory window. Absorption coefficients at 355 nm were obtained by subtracting scattering coefficient

(bscat) from extinction coefficient (bext). In addition, simultaneous measurements of aerosol absorption coefficient at visible wavelengths and aerosol mass concentrations were recorded using a particle - soot absorption photometer (PSAP) and real time aerosol mass concentration monitor. Figure 3.5 is a comparison of the optical results from the albedometer and the aerosol mass monitor. The left Y - axis of Figure 3.5 is extinction (black) and scattering (red) coefficients (Mm-1) of ambient aerosols. The Y-axis on the right plots the mass concentrations (green, μg / m3) measured simultaneously. For these ambient trials, the sample temperature was consistently ≈ 301 K and the aerosol sample was dry with a relative humidity of only 10 - 12%. The average SSA at 355 nm during the Dec 16 plume was 0.55; the Dec 20 plume yielded a mean of 0.57, and the Dec 27 plume had a mean of 0.61. This is much lower than the values of 0.83 − 0.87 reported by Gyawali et al. for winter-time ambient aerosol in Reno, NV.86 However, these authors did note the SSA tended to be lower during pollution plumes, so it is conceivable that our

Texas Tech University, Lulu Ma, August 2013

observations were made on discrete pollution plumes. In addition, measurements were made in winter during times of colder temperatures so smoke plumes from heating may influence the air sample. When comparing the mass concentration with extinction and scattering traces, we see both tend to track together in time. The magnitude of the measured values also suggest the atmospheric environment of Lubbock was clean as mass concentrations of ambient aerosols are often only a few μg / m3. The relationship between mass concentration and optical effects can be better observed in Figure 3.6, as plots of bscat and bext vs. particle mass concentration have been presented. An approximate linear relationship between the variables is observed, and best - fit lines for both data sets have been included. The size distribution of particles influences the slope of the line, so there is no fundamental reason or assurance this plot should be linear. In fact, it will only be linear if the ambient particle size distribution is relatively constant. Nonetheless, the data analysis is useful as it relates optical effects directly to mass concentration of aerosol. The slope of the lines for scattering and extinction data were 3.0 and 5.5 m2 / g, respectively, for this data set. This data is based on 3 plumes of aerosols studied over a short period of time, during low aerosol mass loading, and likely does not represent the typical case for Lubbock, TX.

Figure 3.7 plots absorption coefficients (Mm-1) at different wavelengths in time for the aerosol plumes previously mentioned. Absorption coefficients at 355 nm were obtained from the albedometer, and the visible wavelengths (470, 552, 660 nm) were obtained using the PSAP. The insets to the figure also show wind direction and speed along with aerosol Angstrom absorption exponent (AAE). AAE describes the wavelength dependence of light absorption by aerosol and was determined by fitting the PSAP data to a power law of the following form and extracting the exponent:

−퐴퐴퐸 푏푎푏푠 = 휅휆 (25)

Larger AAE values indicate strong enhancement in light absorption at higher frequencies or shorter wavelengths. A quick glance at Figure 3.7 suggests babs values measured at different wavelengths by different techniques tend to track one- another in time. More careful consideration of Figure 3.7 yields the insight that a change in babs is often accompanied by a shift in AAE, but AAE can shift to either higher or lower values during

Texas Tech University, Lulu Ma, August 2013

a plume. This is particularly evident for the trace of December 28 when AAE dipped from 1.2 to < 0.8 during the plume episode. However, on December 16 and December 20, larger AAE values were encountered during the plume. This is consistent with the magnitude of absorption at λ = 355 nm reflected in the data. For instance, in the plot for Dec 28, the b abs value for 355 nm is very close to that observed for 470 nm indicating there was not a strong wavelength dependence on absorption. However, during the Dec 16 and Dec 20 plumes, the observed near UV aerosol absorption was significantly higher than that observed by the PSAP at 470 nm. This is consistent with the increase in the AAE measured via the PSAP measurements during the plume. Inclusion of the albedometer babs data at 355 nm in the AAE calculation significantly changes the AAE value. For instance, for the plumes of Dec 16 and Dec 20, AAE during the plume increases to 2.03 and 2.01 if 355 nm data is included. For the Dec 28 aerosol plume, an AAE of 0.61 would be obtained which is similar to that obtained from the PSAP alone. The phenomenon of apparent increases in AAE relative to filter based measurement techniques has also recently been reported by Ajtai et al. when those authors compared photoacoustic absorption measurements to those made simultaneously with an aethalometer.97 A quick glance at Figure 3.7 is also enough to suggest the precision for the babs measurement afforded by the PSAP is superior to the albedometer difference -1 measurement. The PSAP LOD for babs at visible wavelengths is <1 Mm which is considerably better than what the difference method (albedometer) can offer in the UV.

Literature reported limits of detection (LOD) for photoacoustic sensing of aerosol babs are also often < 1 Mm-1 in the visible region, but far less work has been reported in the UV region.98 However, there are some reports to provide a comparison. Ajtai et al. have recently reported a UV version of a photoacoustic instrument to measure aerosol absorption with LODs of 3.5 and 6.7 Mm-1 at 355 and 266 nm.97 Gyawali et al. have reported minimum detectable absorptions of < 0.5 Mm-1 at 355 nm.86 These data suggest the photoacoustic method offers improved sensitivity relative to the albedometer for absorption measurements in the UV. In addition, any small signal drift in the extinction or scatter channel can lead to a large relative error in babs. This was frequently encountered and necessitated regular calibration which disrupted monitoring efforts. Alternatively, the PSAP (and most photoacoustic instruments) currently do not make

Texas Tech University, Lulu Ma, August 2013

measurements in the UV, and these techniques do not measure scattering or extinction.

The albedometer babs measurement can provide some insight into ambient aerosol absorption, albeit not at the level of precision other techniques may offer. Therefore, the techniques are quite complementary, and the albedometer absorption measurement is best - suited for relatively polluted air masses.

3.4 Conclusions A method for measuring aerosol scattering, extinction, absorption coefficients, and single scatter albedo in the near UV spectral region (355 nm) has been described. The optical properties of atmospheric aerosols in the near UV have not been explored to a large extent, and measurement instruments like the albedometer presented here have utility for making point measurements of aerosols. Results suggest that ammonium sulfate and an organic aerosol produced through the ozonolysis of α - pinene do not absorb light at 355 nm (e.g., SSA ≈ 1). However, data supports the conclusion aerosol produced from toluene vapor at low concentrations of NO2 (< 150 ppbv) may weakly absorb light at 355 nm (SSA = 0.95). Single scatter albedo values for biomass burning samples (SSA = 0.83- 0.84) and windblown mineral dust (SSA = 0.75-0.84) have been reported for λ = 355 nm. These aerosol types are major constituents of authentic atmospheric aerosols, so data on their optical properties is relevant. In addition, the albedometer can be used for ambient aerosol measurements and study of ambient plumes at our location has allowed extraction of limited data for near UV mass scatter (3.0 m2 / g) and mass extinction efficiencies (5.5 m2 / g). Also, the albedometer derived absorption measurement tracked simultaneous measurements made in the visible spectrum via a filter based PSAP instrument.

Texas Tech University, Lulu Ma, August 2013

Table 3.1 Summary of Data and Extracted Refractive Indices (RI) for Dp = 300 nm Particles after Correction for Multiply Charged Particles

aerosol type σext ± 95% c.i. σscat ± 95% c.i. Measured RI Literature RI (× 10−9 cm2) (× 10−9 cm2) ammonium sulfate 2.4 ± 0.8 2.6 ± 1 1.56 ± 0.12 approx. 1.55 91 α-pinene SOA 1.65 ± 0.14 1.71 ± 0.09 1.45 ± 0.02 1.46 ± 0.02 92

Texas Tech University, Lulu Ma, August 2013

Table 3.2 Near UV Single Scatter Albedo (SSA) of Biomass Burning, Soil - Dust, and Organic Aerosols SSA ± 95% c.i. measurement range (Mm-1) SOA toluene SOA 0.95 ± 0.03 20−200 soil dusts pullman soil 0.84 ± 0.01 76−571 amarillo soil 0.75 ± 0.01 120−530 biomass burning aerosols pear tree leaves 0.84 ± 0.01 150−1390 Afghanistan pine needle 0.83 ± 0.01 43−1386

Texas Tech University, Lulu Ma, August 2013

Figure 3.1 Schematic of the measurement apparatus. The electrostatic classifier was used to select 300 nm diameter particles for ammonium sulfate and SOA experiments but removed for experiments on dust, ambient aerosol, and biomass burning aerosols. The 3rd harmonic of a Nd: YAG laser was used to produce 5 ns pulses at λ = 355 nm.

Texas Tech University, Lulu Ma, August 2013

(A)

(B)

Figure 3.2 Observed size distributions for aerosolizd Pullman (A) and Amarillo (B) soils. Distributions reconstructed from individual channels of a MetOne particle counter.

Texas Tech University, Lulu Ma, August 2013

Figure 3.3 Calibration of experiment using R - 134a. (A) Plots of observed scattering, extinction, and absorption coefficient in time for an experiment in which the cell was flushed sequentially with filtered air, R - 134a, and air again. Insets show CRDS traces for air and R - 134a demonstrating a shift in ring - down time constant (τ). In plot (B), the SSA is plotted for the data from (A). Point to point measurements of SSA are highly variable when bscat and bext are near zero, but much better precision is achieved for R - -1 134a with bext = bscat ≈ 370 Mm . This is consistent with the relative uncertainty in SSA propagated from observed standard deviations as shown in plot (C).

Texas Tech University, Lulu Ma, August 2013

) )

1 1

- -

Mm Mm

( (

Figure 3.4 Plots of extinction and scattering coefficients at 355 nm vs. particle concentration for lab generated (A) ammonium sulfate aerosols and (B) α-pinene SOA. Y-axis units are Mm-1.

Texas Tech University, Lulu Ma, August 2013

Figure 3.5 Plots of extinction and scattering coefficients (Mm-1) and mass concentrations (μg / m3) vs. time for 3 aerosol plumes observed in Lubbock, TX. Plume A occurred on 12 / 16 / 2011, plume B on 12 / 20 / 2011, and plume C on 12 / 28 / 2011. For all ambient trials, sample T ≈ 301 K, RH = 10 – 12 %, and P = 0.88 - 0.90 atm. The error bars are set to the observed precision of the scatter and extinction channel.

Texas Tech University, Lulu Ma, August 2013

Figure 3.6 Plot of extinction (black) and scattering (red) coefficient vs. mass concentration for ambient aerosol plumes in Lubbock, TX at 355 nm. The slopes of the lines are the mass scatter and mass extinction efficiencies, here 3.0 and 5.5 m2 / g.

Texas Tech University, Lulu Ma, August 2013

Figure 3.7 Plots of absorption coefficients (babs, Mm-1) at different wavelengths (355, 470, 552, and 660 nm) vs. time for three aerosol plumes observed at Lubbock, TX. The insets above the colored plots illustrate wind speed, wind direction, and Angstom absorption exponent (AAE) of the aerosol as measured from the PSAP. The constant error bars for the 355 nm trace are set to 5.2 Mm-1, the propagated precision of the albedometer babs channel. Missing points for the wind data plots correspond to “calm” conditions.

Texas Tech University, Lulu Ma, August 2013

CHAPTER IV

LIGHT SCATTERING AND EXTINCTION MEASUREMENTS CCOMBINED WITH LASER-INDUCED INCANDESCENCE WITHIN AN AEROSOL ALBEDOMETER: METHOD DEVELOPMENT AND EFFECT OF SOOT MORPHOLOGICAL CHANGES

4.1 Introduction. Bond et al. have recently suggested black carbon or soot exerts a radiative forcing in excess of +1 W / m2 on earth’s climate.99 If accurate, this warming effect on climate is second only to CO2. Soot exerts its warming influence through a variety of mechanisms including direct absorption of sunlight in the atmosphere, alteration of cloud optical properties and lifetimes, and deposition on snow and ice packs100 – causing increased melt in polar regions. In addition, soot often contains polycyclic aromatic hydrocarbons (PAHs) that are known to be mutagenic and carcinogenic.101,102,103

Given the importance of airborne soot to our climate and health, significant efforts have been made to develop schemes to quantify soot in air. These include a thermo - optical method that comprises the NIOSH 5040 standard,104,105 use of filter - based instruments such as aethalometers 106 , 107 , 108 and particle / soot absorption photometers45, or more recently through photoacoustic spectroscopy.98 Another technique to measure dispersed soot is laser-induced incandescence (LII).109,110 The LII technique employs a high – peak - power pulsed laser to rapidly heat the light absorbing soot to temperatures of 2500 - 4000 ℃. The soot reaches such high temperatures that it incandesces and produces broadband visible emission for 1 - 2  s after the laser pulse. The intensity of incandescence can be related to soot volume fraction or mass concentration through calibration. One such LII instrument that can measure incandescence of single particles has been commercialized by Droplet Measurement Technologies. 111 The so - called Single Particle Soot Photometer (SP2) probes incandescence of particles within a Nd: YAG laser resonator allowing laser fluences of approx. 3 MW / cm2. This affords the technique impressive detection limits on the order of 10 ng / m3 of air regardless of soot mixing state. The SP2 has been used in the recent

Texas Tech University, Lulu Ma, August 2013

HIPPO campaign to map atmospheric concentrations of soot over wide swaths of the globe.112,113

Because of the previous success of LII measurements for tracking soot mass concentrations, we became interested in combining the aerosol albedometer recently developed in our laboratory with LII measurements. The albedometer simultaneously measures light scattering, extinction, and absorption (by difference) on dispersed aerosols. In principle, combining the optical property measurements with an independent assessment of soot mass concentration through LII can provide direct determination of soot mass absorption cross section (MAC, m2 / g). Soot MAC is an important parameter that describes the degree of light absorption per unit mass of soot. Light scattering theory, and select measurements have suggested that when soot becomes coated with non- volatile or semi - volatile materials that MAC can increase by a factor of 1.5 - 2.20,114,115, Full development of a hybrid LII / albedometer instrument may allow direct measurement of soot MAC in nearly real-time. In this work, we report initial progress towards the ultimate goal of MAC measurement. We have explored the use of a 1064 nm laser line to promote incandescence of kerosene soot within the sphere albedometer. The analytical diagnostics of the measurement process has been studied and the consequences of the LII measurement understood. The net deliverable of the project is a framework within which a fully integrated, hybrid aerosol albedometer and LII-based soot mass concentration monitor can be constructed.

4.2 Methods

4.2.1 Generation and Characteristics of Kerosene Soot Kerosene soot was generated by combustion of kerosene (Coleman Fuel) in a fabric-wicked camping lantern (Walmart). The lamp was operated within a fume hood for proper venting of fumes. A ¼ inch diameter polyethylene tube approx. 30 m in length was used to pull the soot to the LII optical chamber. A 3 - way ball valve was placed immediately prior to the LII chamber. The third port of the ball valve was plumbed to an air filter to provide a source of clean air. The user could exert some control over the quantity of soot reaching the optical cell through carefully actuating the 3 - way valve. The kerosene soot generated is believed to be uncoated, and largely void of organics.

Texas Tech University, Lulu Ma, August 2013

This is supported by thermogravimetric analysis (TGA) data reported in Figure 4.1 which illustrates very little mass loss until approx. 550o C, a temperature at which elemental carbon / soot is known to oxidize or vaporize.116

4.2.2 Description of the Optical Measurements. A schematic of the optical setup used is shown in figure 4.2 (A). In the experiment, cavity ring-down spectroscopy (CRDS) is conducted on a dispersed aerosol sample placed within a transparent glass tube. The glass tube is essentially a flow - through optical cell with the laser beam propagating along the central axis of the tube.

The CRDS experiment provides a measurement of the aerosol extinction coefficient (bext, Mm-1) by measuring the time constant ( ) of the exponential decay of light intensity as it circulates in the optical cell. An integrating sphere is constructed around the CRDS optical cell. In this work, the sphere serves as both a nephelometer (for scattering measurement) and a means to collect and measure LII of dispersed soot aerosol. The -1 aerosol scattering coefficient (bscat, Mm ) is measured by ratioing the scatter channel signal to the CRDS channel signal and comparing this ratio to those observed for calibration gases of known optical loss. The device is very similar to what has been described in previous works (with the exception of the LII measurement) and we refer readers to the associated manuscripts for a more detailed description of the operation of the aerosol albedometer for light scattering and extinction measurements.69,70,85,88

Briefly, light from a frequency doubled Q - switched Nd:YAG laser (Quantel Brio, 5 ns pulses, 20 Hz maximum rep. rate, 4 mm diameter) is incident upon the optical cell used for spectroscopy. The laser produces the frequency - doubled line at 532 nm, but a significant amount of residual light at the fundamental wavelength (1064 nm) is also present in the co - linear beam. A near-infrared (NIR) blocking filter (Thor Labs, FM01) can be placed in the beam’s path when desired. The transmission spectrum of the NIR blocking filter is illustrated in figure 4.2 (B). The resulting probe beam was then passed into the optical cell formed from two 2.5 cm diameter, 6 m r.o.c. (radius of curvature) dielectric mirrors (Los Gatos Research / CRD Optics). Empty cavity ring - down time constants were generally between 20 - 40  s depending on the mirrors used and their state of cleanliness. Light was then scattered or incandesced by gases or aerosols present

Texas Tech University, Lulu Ma, August 2013

in the probe beam. Light collected by the integrating sphere was filtered through either a 532 nm or 590.4 nm band - pass filter and measured with a photomultiplier tube (PMT, H7732P-01). Current produced by the scattering channel PMT was passed through a 1000  load for measurement with either an 8 - bit, 100 MHz, 1 Gs / s oscilloscope (Tektronix 1012B) or a 12 - bit PCI oscilloscope card running at 100 Ms / s (Gage). CRDS signals measured at a separate PMT detector through a 532 nm laser line filter were passed through a 50  load and digitized using the same oscilloscope or DAQ (Data Acquisition) card.

In our previous experiments, the residual 1064 nm light has always been removed prior to the optical cell through use of a NIR blocking filter. In this work, the NIR blocking filter was intermittently removed to allow passage of the 1064 nm beam. As illustrated in figure 4.2 (C), the cavity ring-down mirrors feature very high reflectivity (consequently low % T) near 532 nm, but act like windows in the near IR at 1064 nm. This allows the aerosol sample within the optical cell to be illuminated with 1064 nm beam when the NIR blocking filter is removed. This is a key aspect of the experiment since it allows the test aerosol to interact with a high peak - power laser pulse in the MW range. Laser pulse energies of up to 40 mJ for the co - linear 532 and 1064 nm beams have been measured with a laser power meter with a calibration date in Feb. 2012 (Thor Labs PM1000 / 5310C). WARNING: Despite the high transmittance of the CRD mirrors at 1064 nm, a small amount of laser energy can be absorbed by the mirrors. If the laser fluence is too high, or optics dirty mirror damage will occur and the CRD mirrors will be permanently ruined. The damage threshold for the mirrors used here was specified to be approx. 500 mJ / cm2.

4.2.3 Scanning Mobility Particle Sizing (SMPS) and Particle Soot Absorption Photometry (PSAP) A SMPS system was used to obtain aerosol particle size distributions as measured through equivalent electrical mobility diameters. The SMPS was assembled from a TSI 3080 electrostatic classifier, a TSI 3081L differential mobility analyzer (DMA), and a TSI 3772 condensation particle counter (CPC). Butanol was used as the fill fluid in the CPC, and a sample – to - sheath flow rate ratio of 1:10 was maintained in the experiments. A sample flow rate of 0.6 LPM was commonly employed. The SMPS was run under

Texas Tech University, Lulu Ma, August 2013

computer control using software provided by the manufacturer. A Radiance Research (Seattle, WA) 3 particle / soot absorption photometer (PSAP) was used to make -1 measurements of soot absorption coefficient (babs, Mm ). Sample flow rates between 0.5 - 1.9 LPM were employed. The PSAP instrument collects particles onto a filter while simultaneously measuring the optical transmittance of the particle laden filter compared to a reference filter. For all experiments, only light absorption for the green channel (530 nm) was considered. Indicated light absorption coefficients were used to estimate soot mass concentration ( g / m3) using a mass absorption cross-section of 13.1 m2 / g advocated by Subramanian et al.117

4.2.4 Discrete Dipole Approximation (DDA) Modeling The discrete dipole code ADDA (http://code.google.com/p/a-dda/) was used. This code was developed by M.A. Yurkin and A.G. Hoekstra. 118 , 119 The soot aggregate consisted of 80 primary spheres, each with a diameter of 50 nm. A 3D model of the aggregate is illustrated in figure 4.3. Soot refractive index was assumed to be 1.75 + 0.63i. The ADDA simulation employed 19,344 total dipoles with 85.12 dipoles per  . The simulation was conducted with a 120x120x120 grid for  = 532 nm. For comparison, we also calculated the optical properties of a soot primary particle (e.g. one 50 nm sphere) using the same refractive index but in a 12x12x12 grid with 486 dipoles and 106.4 dipoles per  . The reported SSA values were the average over both polarization states.

4.2.5 Transmission Electron Microscopy (TEM) Analysis Fresh kerosene soot was collected on Lacey Formvar/Carbon 200 mesh copper TEM grids (Ted Pella, 01881-F) by passing a gentle flow of a soot - laden air stream through a glass Pasteur pipet containing the grid. Sampling times were ≤ 20 min. The TEM analysis was conducted by the research group of Peter Buseck at the LeRoy Eyring Center for Solid State Science at Arizona State University. Images were obtained on a CM200 (Philips Corp.) at an accelerating voltage of 200 kV.

4.3 RESULTS & DISCUSSION.

4.3.1 Strategies to Measure the LII Signal As mentioned previously, our intent was to develop a combined measurement of elastic light scattering, extinction, and laser induced incandescence for dispersed soot 60

Texas Tech University, Lulu Ma, August 2013

aerosols. This presents a measurement challenge since the elastic scatter and incandescence signals must be described separately. Incandescence occurs over a wide range of wavelengths, while elastic scatter will only occur at the laser line. If an additional detection channel is added to the integrating sphere for making incandescence measurements it would be simple to separate elastically scattered light from incandescence based on their unique spectral signatures. An alternate approach to separating the signals would be measuring at the same wavelength, using time - gated detection. Incandescence of soot is known to occur with the first 2 - 3 μs after the excitation laser pulse. 120 However, the resonant beam in a CRDS experiment may continue to circulate in the optical cell for 50 μs or longer in some cases. In principle, it should be possible to record LII data at / near the time of the laser pulse, and after delaying a few microseconds, extract ring - down transients to use for determining scattering and extinction coefficients. This latter method (time - gating) is appealing since a single detection channel could, in principle, make both scattering and LII measurements. Both time - gated and wavelength resolved approaches have been considered here.

4.3.1.1 Time-Gated Approach. The red data in figure 4.4 illustrates an oscilloscope trace observed for the integrating sphere channel after the laser pulse (both 532 and 1064 nm present) which occurs at approx. 50 μs. After a large initial signal peak, the scatter channel signal develops into a smooth monoexponential decay characteristic of CRDS experiments (solid black line is best fit and fit-residuals are shown at top of Figure 4.4). To determine whether the large signal observed within the first 3 μs is related to the LII effect we have performed several more experiments. First we have compared the initial signal pulse intensity when soot was in the optical cell compared to when it was filled with R-134a (tetrafluorothane) gas. As illustrated in Figure 4.4, the initial pulse was approximately double when soot was present in the optical cell (black trace) compared to the R - 134a signal (green trace). However, the presence of an initial peak in the green trace (for R - 134a) suggests a significant amount of scattered light reaches the detector during the laser pulse. This light results from both Rayleigh scatter from gas molecules and reflections of optical elements. To investigate whether the increase in signal observed in the presence of soot could be simply from additional light scattering, we compared the ratio of the

Texas Tech University, Lulu Ma, August 2013

scatter channel / CRDS channel for both experiments during the ring - down trace. The resultant data for soot and R-134a are plotted in the inset box in Figure 4.4. Despite the soot trace being significantly more noisy, we can see both traces produce a signal ratio in the vicinity of 200. This suggests both samples had a similar scattering coefficient (bscat) since this ratio is known to scale with bscat. As a result, it is not possible to explain the doubling of signal observed for the soot sample during the initial pulse by only elastic scattering effects. Therefore, we refer to the increase of the initial peak over the scattering background as the “Soot Incandescence Signal” as defined graphically in Figure 4.4.

In addition, we have investigated how this incandescence signal scales with soot mass concentration and with laser fluence. The LII signal of this experiment is collected at 532 nm by placing a 532 nm optical filter in front of detector. Results of this analysis are illustrated in figure 4.5. Figure 4.5 (A) reports an approximately linear relationship between observed incandescence signals and soot mass concentration as inferred through use of PSAP and a mass absorption cross section of 13.1 m2 / g. Such a linear relationship would be expected if LII was the mechanism by which the signal was generated. The data plotted in black in figure 4.5 (B) illustrates the relationship between observed incandescence signals and laser fluence. As illustrated, the incandescence signal appears to saturate above 100 mJ / cm2. During this experiment, soot concentrations were observed to vary between 56-77 μg / m3 and all incandescence signals have been normalized to the signal expected at 77 μg / m3 for comparison of the effect of laser fluence. The saturation of LII signals has previously been reported for soot in flames at laser fluencies 50 - 100 MW / cm2.121 While the laser fluences at which we observe saturation are slightly lower (approx. 20 MW / cm2), the fact that this occurs further supports the claim the observed signal results from laser induced incandescence of soot since elastic scattering does not saturate in a similar manner. The grey data series in Fig. 4.5 (B) illustrates the ratio of soot incandescence signal-to-scattered-light-background (as defined in Figure 4.4). For this experiment, the scattered light background signal was taken for filtered air. It is demonstrated the incandescence signal-to-scattered-light- background ratio decreases at higher laser irradiances which supports the idea scattering continues to increase with no further increase in LII signal at high fluences. These results support the claim LII of soot occurred during our experiments resulting in an increase in

Texas Tech University, Lulu Ma, August 2013

observed signal at the scatter channel detector during the first 3  s after the laser pulse. The scatter channel signal after the initial 3  s was not affected and the CRDS / nephelometry measurement could be carried out during this period of time. Therefore, sequential LII and CRDS / nephelometry measurements are at least possible through the time - gated approach. However, the LII limit of detection for soot (based on a 3 criteria) at 532 nm was 12.4  g / m3. This level of performance is insufficient for most atmospheric monitoring applications in which typical soot concentrations are frequently < 2  g / m3.

4.3.1.2 Wavelength Resolved Approach. A significant limitation to the time-gated approach is that pulse energy variability in the laser source will cause significant noise in the LII signal since detection occurs on the laser line. To combat this, we decided to perform measurements using a different interference filter (590 + 1 nm) installed within the integrating sphere. Since incandescence is continuum emission, the resultant LII signal intensity can be measured at 590 nm, largely free from laser pulse energy noise. An experiment in which the concentration of kerosene soot was varied while LII signal recorded is presented in figure 4.6 (A). This figure illustrates a linear response between LII signal and soot concentrations < 10  g / m3. Compared with the LII measurement, negative influence of 532 nm was removed, and shape of the linear plot was improved significantly. The 3 LII limit of detection (LOD) for single point measurements at 590 nm was ≈ 0.6 μg / m3, however, this could be reduced to ≈ 0.2 μg / m3 for 5 minute running averages. Such performance can be useful for atmospheric monitoring applications. Indeed, figure 4.6 (B) reports results from such an experiment in which ambient air was sampled through a laboratory window, swept through the LII optical cell, and finally measured with the PSAP. As illustrated, between approx. 6:45 – 8:30 AM an increase in soot was observed on both the LII and PSAP detectors. The wavelength resolved measurement clearly offers much better performance when compared to measurements at 532 nm. As such, addition of a dedicated detection channel solely to LII is recommended.

Texas Tech University, Lulu Ma, August 2013

4.3.2 Consequences of the LII Measurement: Soot Morphology Changes Either the time - gated or spectral selection approach allows separation of the LII signal from the measurement of scattering and extinction coefficient. However, the LII measurement should not alter or affect the particles under study if scattering and extinction is to be measured separately. The LII experiment involves rapidly heating the soot particles to temperatures of 2500 – 4000 K within 100 ns of the laser pulse. Therefore, it is conceivable significant changes to the particles may occur during the LII experiment. We have investigated the effect of 532 and 1064 nm laser light on the observed size distributions and morphology of soot. For these experiments, size distributions of the soot aerosols were obtained using an electrostatic classifier with a condensation particle counter, and soot morphology was studied with transmission electron microscopy (TEM). Figure 4.7 (A) presents several SMPS size distributions for soot aerosols at various laser fluences. When soot was illuminated with < 14 mJ / cm2, the most probable electrical mobility diameter of the soot aggregate particles was between 100 – 200 nm. As laser power increased (532 + 1064 nm), a large shift in number concentration distribution towards particles with very small diameters (Dp < 50 nm) was observed. This effect was not observed when only 532 nm light was used with CRDS mirrors in the beam path. A large number of very small particles were produced due to lasing with the 1064 nm beam. This result was confirmed by the TEM analysis presented in figure 4.7 (B) which illustrates a large number of new soot fragment particles present on the TEM grid (some of which are circled in the figure). This phenomenon of new particle formation has previously been reported and characterized extensively by Michelson et al.122 Since the resultant fragment particles are similar in size to primary soot spheres, it is tempting to consider disaggregation of the fractal soot chains as a mechanism through which the fragment particles are produced. However, high-resolution TEM analysis of the fragment particles as shown in Figure 4.8, suggests to us this mechanism is inconsistent with observations. The small particles produced via lasing with 1064 nm light exhibit large differences in microstructure. In some cases, the nanoparticle fragments exhibit hollow cores with onion - like layers surrounding. In other cases, ribbon-like structures of carbon form. This structural reorganization has also been observed in previous work by Michelson et al.11 and Vander Wal et al.123 and suggests

Texas Tech University, Lulu Ma, August 2013

the soot aggregate chains do not simply disassemble into primary particles. Regardless of the mechanism of such transformation, the SMPS results of figure 4.7(A) suggest this process begins to occur for the kerosene soot sample when the laser fluence was approximately 30 - 40 mJ / cm2. This fluence is significantly lower than what other investigators have reported for the threshold. For instance, Michelson et al. (ref 11) report fluences of 120 mJ / cm2 and 220 mJ / cm2 are required for new particle generation using 532 and 1064 nm, respectively. Kruger et al. have reported observing a change in soot size distribution by virtue of laser irradiation at 532 nm when irradiance was 120 mJ / cm2.124 However Stipe et al. reported particle production at far lower fluence (70 mJ / cm2) when using UV photons of 193 nm. 125 The exact cause of the discrepancy in required laser fluence is not clear. We measured laser power prior to the optical cell and not within the measurement cell. It is possible multiple reflections (etaloning) off of the CRDS mirrors raises the irradiance in a way we are not accounting for. In practice, an analyst would wish to carry out LII measurements in the signal saturation regime, where fluctuations in laser pulse energy does not influence LII signal appreciably. Figure 4.5 (B) suggests for our experiments this occurred at > 100 mJ / cm2 laser fluence. Since this value is considerably larger than the energy density at which we observed the onset of soot morphological changes, our results support the idea that it is simply not possible to perform LII measurements without inducing morphological changes in the soot sample.

4.3.3 Suggested Measurement Sequence. Because of the change in soot morphology, light scattering / extinction measurements and LII measurements made with the albedometer should occur sequentially in time. The optimum sequence would involve first making a scatter / extinction measurement with only the 532 nm beam present, and then rapidly adding the 1064 nm beam to complete the LII measurement. Switching the beam condition could easily be achieved in only a few seconds with a motorized optical filter mount. This would help assure measurements were made on roughly the same aerosol sample. Between subsequent measurements, a sufficient period of time should elapse to allow the sample cell or probed volume to completely flush with a fresh sample of aerosol. The exact amount of time required would depend on the volume of the sample chamber and

Texas Tech University, Lulu Ma, August 2013

purge flow rate. Nonetheless, this sample flush step would likely limit the duty cycle of the device, but is required for accurate results.

We demonstrate such a sequence in figure 4.9. In this experiment fresh kerosene soot was passed into the albedometer optical cell and illuminated with 532 nm laser light. The size distribution was simultaneously measured at the outlet of the cell using the SMPS. Then, the near IR blocking filter was removed from the beam path and the soot present in the cell was illuminated with the 1064 nm light. Finally, the near IR blocking filter was re-inserted into the beam path. Because IR block filter changes light intensity passing through CRDS mirrors, two calibrations, one with IR blocking filter and the other one without IR blocking filter were carried out. The results of the experiment illustrate a dramatic shift in soot size distribution when illuminated with 1064 nm. This shift was accompanied by a significant change in aerosol single scatter albedo (SSA) at 532 nm. Prior to 1064 nm illumination, the soot SSA was measured to be 0.158 + 0.005. Upon interaction with the 1064 nm light, the SSA dropped to 0.09 + 0.01 and smaller particles were formed. After the near IR was again blocked, and the measurement cell flushed, the SSA and size distribution were indistinguishable from the original case (SSA = 0.155 + 0.004). The two SSA values of soot without LII match with each other, showing nice reproducibility of the setup. The drop in SSA observed is consistent with predictions from light scattering theory for a model in which smaller fragment particles are present as compared to only chain aggregates. For instance, if the discrete dipole approximation code ADDA is used to predict the SSA of an 80 primary particle soot aggregate of m = 1.75+0.63i and volume equivalent size parameter of 1.23, a resulting SSA of ≈ 0.20 is obtained for the aggregate. However, the SSA of individual 50 nm primary particles was modeled to be < 0.02. So a shift towards smaller SSA values when large quantities of very small soot fragment particles are produced would be expected. Closure (e.g. exact match) between measured SSA and modeled SSA was not pursued or expected because the fractal soot aggregates are not completely fragmented, and the resulting soot fragments are often not perfectly spherical (see Fig. 4.8). Nonetheless, this experiment illustrates sequential scatter / extinction and LII measurements are possible, and provides further evidence for LII producing a large number of new, very small particles.

Texas Tech University, Lulu Ma, August 2013

Figure 4.1 Thermogravimetric analysis of kerosene soot. The purge gas was air for this experiment. As observed, significant mass loss did not occur until approx. 550o C. This suggests a large mass of volatile organic molecules was not present in the sample.

Texas Tech University, Lulu Ma, August 2013

Figure 4.2 Experimental setup and transmission spectrum of high reflectivity mirror and near infrared (NIR) blocking filter. (A) Experimental apparatus. A NIR blocking filter can be placed in the beam path or removed as desired. When the filter is removed, the 1064 nm beam passes through the cavity ring-down mirrors and illuminates the test soot aerosol causing laser-induced incandescence (LII). (B) Transmission spectrum of the NIR blocking filter. (C) Transmission spectrum of the high reflectivity cavity ring-down (CRD) mirrors. The mirrors feature very high reflectivity near 532 nm, but act like windows in the near IR at 1064 nm.

Texas Tech University, Lulu Ma, August 2013

Figure 4.3 Three dimensional plot of the simulated soot aggregate used for DDA modeling. Each black point represents a dipole within the particle.

Texas Tech University, Lulu Ma, August 2013

Figure 4.4 Oscilloscope trace for the scatter channel illustrating a typical signal response at 532 nm. The signal observed at the laser pulse consists of light elastically scattered by the sample or optical elements combined with soot incandescence. This large initial pulse decays rapidly (approx. 3 μs), and the response curve develops into a monoexponential decay consistent with CRDS. In the initial peak in the inset, the soot incandescence signal nearly doubles the size of the initial peak relative to the case in which R-134a filled the measurement cell. Analysis of the subsequent CRDS/ nephelometry data suggests the scattering coefficient of the soot and R-134a samples are very similar since observed scatter / CRDS ratio are nearly identical (here a ratio of approx. 200). Therefore, the signal increase observed during the initial 3 μs did not result from additional light scattering caused by soot.

Texas Tech University, Lulu Ma, August 2013

Figure 4.5 Plots of laser incandescence signal vs. soot mass concentration at 20.8 mJ and normalized soot incandescence signal vs. laser pulse fluencies. (A) Plot of observed soot incandescence signal a 532 nm vs. soot mass concentration. Soot mass concentration estimated through use of PSAP and mass absorption cross section of 13.1 m2/g. (B) Plot of soot incandescence signals (black solid line) vs. laser fluence. Incandescence signals have been normalized to equivalent soot concentration for direct comparison. The second y-axis plots the observed incandescence signal-to-scatter background ratio. The incandescence signal appears to saturate at high laser fluence. Error bars are + 1 std. dev.

Texas Tech University, Lulu Ma, August 2013

Figure 4.6 Plots of soot LII signal at 590 nm and ambient measurement data using this instrumental setup. (A) Plot of peak soot incandescence signal at 590 nm vs. the aerosol absorption coefficient (babs) indicated by the PSAP for lab generated kerosene soot. The top x-axis reports the soot mass concentration as estimated using 13.1 m2 / g for the mass absorption cross section. (B) Monitoring soot through LII and a particle soot absorption photometer on the morning of January 29 2013 at Lubbock, TX. An increase in soot was observed between approx. 6:30 – 8:30 AM local time.

Texas Tech University, Lulu Ma, August 2013

Figure 4.7 Morphology analysis and comparison of lased and non – lased soot. (A) Observed SMPS number concentration size distributions of fresh kerosene soot at various laser fluences. At laser fluences > 28 mJ / cm2, a significant shift in particle size distribution was observed indicating small particle fragments dominate the size distribution. (B) TEM images of lased and not-lased kerosene soot. A large number of very small carbonaceous fragment particles (some circled) are observed in the lased samples. This is consistent with the SMPS results. Notice, the soot agglomerate structure is not completely disintegrated via lasing. The TEM grid appears like a spider web in the image.

Texas Tech University, Lulu Ma, August 2013

Figure 4.8 High resolution TEM images of fresh kerosene soot that has been lased vs. non-lased soot. Lased soot exhibits vastly different morphology on the nanoscale, including formation of graphitic layers with hollow particle cores.

Texas Tech University, Lulu Ma, August 2013

Figure 4.9 Size distributions and single scatter of albedo (SSA) of fresh soot with and without 1064 nm illumination. The data on the left illustrates a size distribution and SSA of fresh soot that has not been illuminated with 1064 nm. The middle plots illustrate the SSA and size distribution for soot illuminated with 1064 nm in the CRDS cell. We refer to this soot as being “re-ordered.” The data on the right illustrate observations after the NIR blocking filter was replaced and the measurement cell flushed with a fresh soot sample. When soot was illuminated with 1064 nm, a significant drop in SSA was observed. However, the soot properties in the right-most traces were indistinguishable from the initial measurements. This suggests sequential LII and optical property measurements are possible only if sufficient time is provided for the optical cell to fill with fresh sample.

Texas Tech University, Lulu Ma, August 2013

CHAPTER V

CONCLUSION AND DISCUSSION Aerosols suspended in atmosphere play an important role in climate change. The direct climate effect of aerosols is related to the optical properties, which is a major project in this dissertation. Optical properties were measured through different methods in the dissertation.

Very year, Aeolian processes blow a great amount of soil dust aerosols into atmosphere. Two types of soil dust common to the Texas panhandle were studied (Amarillo and Pullman soils). A system named the Lubbock Dust Generation, Sampling and Analysis System (LDGASS) was applied to generate dust aerosol particles in the laboratory. Absorption and extinction of soil dust were measured at different wavelengths. Pullman soil had Angstrom absorption exponents (AAE) of 1.73 and Angstrom extinction exponents (AEE) of 0.111. Amarillo showed larger wavelength dependence with AAE and AEE of 2.17 and 0.168. Mass concentrations of samples were determined through two different methods. The DataRAM instrument gave concentration of PM10 directly, and mass extinction coefficients were 2.97 m2 / g and 2.75 m2 / g for Amarillo soil and Pullman soil. The second method calibrated mass concentration from size distribution by treating all particles as they were sphere characterized. This approach resulted in mass extinction coefficients of 1.74 m2 / g for Amarillo soil and 1.77 m2 / g for Pullman soil. Wavelength of 522nm was chosen for all the mass extinction coefficient measurements. Single scatter albedo (SSA) was obtained at 3 different wavelengths (470, 520 and 660 nm). SSA of Pullman ranged from 0.947 to 0.968. Amarillo had larger SSA ranging from 0.964 to 0.980. Both of the soil samples showed smaller SSA at lower wavelength.

An aerosol albedometer combining cavity ring-down spectroscopy (CRDS) with integrating sphere nephelometer was developed at a near ultraviolet (UV, λ=355nm) wavelength with a pulsed Nd: YAG laser. CRDS measured extinction and nephelometer measured scattering of aerosols Therefore, this setup collect both extinction and scattering signals simultaneously, overcoming the deviation occurred when measuring extinction and scattering in two individual processes. Several samples were analyzed, including ammonium sulfate, secondary organic aerosols (SOA) resulting from the

Texas Tech University, Lulu Ma, August 2013

ozonolysis of α-pinene and toluene, soil dust samples, biomass burning aerosols, and ambient aerosols. Particle size of 300nm diameter was chosen during ammonium sulfate and α-pinene SOA measurement. A linear relationship was found between particle number concentration and extinction and scattering coefficients. Extinction cross sections for the two samples were 2.4 (±0.8) × 10−9 cm2 (ammonium sulfate) and 1.65 (±0.14) × 10−9 cm2 (α-pinene SOA). Refractive indices (RI) were extracted through Lorentz-Mie Theory. The near UV RI of ammonium sulfate was 1.56 ± 0.12 and the RI of pinene SOA was 1.45 ± 0.02, showing high similarity compared with results from other groups. No light absorption was observed for the two types of aerosols and SSA values of them are 1. SSA of toluene SOA was slightly smaller than 1. Soil dusts (Amarillo and Pullman soil) and biomass burning aerosols had significant absorption at 355nm with absorption coefficients in the range of 0.74 and 0.84. Results for ambient aerosols using the albedometer were compared with commercial product named particle-soot absorption photometer (PSAP), which measured absorption at visible wavelengths. Both of the results gave similar trend of absorption changing with time and correlated with each other very well.

The aerosol albedometer is then combined with laser induced incandescence (LII) measurement. A co - linear laser beam containing both 532 and 1064nm was applied, where 532nm was used to collect extinction and scattering signal at 532nm and aerosol samples were incandesced at 1064 nm. Difference between aerosols with and without LII was achieved by moving an IR blocking filter into or out of the beam’s path. New particle formation was observed by scanning mobility particle sizing (SMPS) and transmission electron microscopy (TEM) as a result of lasing. SMPS data showed a dramatic shift of particle diameter changing from about 150nm to less than 30nm. TEM analysis confirmed new particle formation with appearance of smaller particles and significant difference in nanoscale morphologies compared to non-lased soot. SSA of fresh kerosene soot dropped from 0.158 ± 0.005 to 0.09 ± 0.001 after lasing, which was in agreement with light scattering theory. LII usually occurs within the first 3  s after laser illumination and the signal was measured by oscilloscope in volts. LII of fresh kerosene soot were measured at both 532 and 590. A linear relationship was found between soot concentration and LII voltage signal. Measurement taking at 590 showed

Texas Tech University, Lulu Ma, August 2013

better linear relationship than that taken at 532nm. The detection limit at the 590 nm measurement was 0.5  g/m3. The reason was that laser source also emitted light at 532nm and had a bad influence to the 532nm detection channel. This influence was removed by the 590nm band - pass optical filter placed in front of detector and showed a better result at 590nm.

In atmosphere, aerosols can stay in the air for a long time, travelling from one location to another. New aerosols, which are usually named secondary aerosols, are then formed through reactions with gases and other aerosols. In many places, secondary aerosols constitute the major component compared with directly emitted aerosols, which are called primary aerosols. Since the albedometer showed successful long term ambient measurements, it is a good opportunity to apply the instruments on experiment of aging of aerosols. A Teflon chamber acts as reaction vessel, and chemical reactions can be monitored by purging gases and aerosols into the chamber. Formation of new aerosols can be tracked by pumping aerosols into detector continuously. CRDS, integrating nephelometer, and SMPS measure extinction coefficient, scattering coefficient and size distribution of aerosols. Structures of aerosols, for example, coatings and morphology, can be obtained through the use of TGA and TEM. Detailed component analysis can be made by mass spectrometry measurement. The study of aging of aerosols will help setup a model of aerosols analysis and give a better understanding of generation, growing and changing of atmospheric aerosols.

Texas Tech University, Lulu Ma, August 2013

APPENDIX A

OPERATING PROCEDURE FOR USING THE ADDA DISCRETE DIPOLE SOFTWARE FOR MODELING LIGHT SCATTERING AND ABSORPTION BY SMALL IRREGULAR PARTICLES.

The ADDA software can be used to compute Q – terms for light scattering, extinction, and absorption for irregular particles of varying composition (e.g. different refractive indices). ADDA was developed by M. A. Yurkin and A. G. Hoekstra. The reader can learn more about the specifics of the discrete dipole technique by reading ref. .

STEPS:

1. Visit the ADDA website at http://code.google.com/p/a-dda/ and download the software into a directory named “ADDA.” The software will be downloaded as a “.zip” file that will need to be extracted into a folder on your computers hard drive. A version operating in either 32-bit or 64-bit Windows is available. 2. Double click on the “win32” or “win64” folder present within the ADDA directory and find the file “adda.exe.” This is the executable file. 3. The adda.exe file needs certain input parameters specified when run. Therefore, we must run the executable under the MS-DOS command prompt. Click on the windows icon at bottom of screen and search for “command prompt.” Then click on the icon for command prompt to run it. 4. A black screen appears with some typing. You should type in “cd\” and hit enter. Then change the directory to the one ADDA.exe is present in. Use the “cd” (change directory command). For example, if you installed this file in C:\\adda\win32 one would type:

“cd ADDA\win32” and hit enter

5. The directory should change. Verify ADDA.exe is preset in this directory by typing the “dir” command and hitting enter. 6. Navigate your web browser to http://code.google.com/p/a- dda/wiki/Tutorial for a quick tutorial on the software. Its important to familiarize yourself with the command lines. The command line for ADDA contains important information about the simulation. If you don’t specify certain conditions, the software uses a default value and this may not reflect your experiment. A typical command line would read:

adda –size 6 –shape coated 0.5 0.2 0 0 –m 1.05 0 1.2 0.25 –lambda 0.532

Text written in black are ADDA commands while text written in red are values specified by the user. For this simulation, the size of the modeled domain is 6 units

Texas Tech University, Lulu Ma, August 2013

(microns usually used as units). This simulation is for a coated particle. The particle would be assumed spherical by the software since that is the default and a different shape is not specified in the command line. The 0.5 after “coated” specifies that the inclusion has a diameter half the size of the particle. The “0.2 0 0” commands specify the inclusion is not centered within the coating, but offset 20% along the x-axis (but 0 and 0% offset on y and z axis.. The refractive index is set by the “-m” command. Here, the first material has an index of 1.05 – 0i and the second material (coating) 1.2 - 0.25i. The wavelength of light considered is fixed by the –lambda command. Here, we specify 532 nm (0.532 microns).

The user can specify as many commands (starting with -) as needed to describe his / her situation. After running “command prompt” and entering the ADDA directory, type “ADDA –h” and press enter to view all of the options available to you. The “–h” command is to obtain help and can be used to obtain general assistance or specific information on particular options available. For instance, try entering “adda –h shape” and hit enter. The information that appears on the screen will describe all of the particle shape options available.

7. Define the particle shape. One of the most powerful features of ADDA is the ability to define any particle shape / morphology one can imagine. The user achieves this by using a command like: “-shape read filename.txt”. The specified tab delimited text file contains x y z coordinates that will function as locations of dipoles within the particles of interest. Therefore, by assigning coordinates one can describe any particle of interest. Such coordinates would appear like:

x y z material 0 0 1 0 0 0 2 0 0 1 3 0 0 0 3 0 0 0 4 1 0 1 4 1 0 1 4 1 0 2 0 1 0 1 0 1 0 1 1 1

In addition, by adding a fourth value to the line of information (0 or 1), a user can specify whether the site is made up of one of two materials of differing refractive index. More information about the contents of such a text file can be found in the user manual and appendices for ADDA. Developing a good shape file is often the most difficult aspect of modeling using ADDA.

For this work, we have been interested in soot aerosols. The optical and physical properties of soot (including values for refractive index) has been recently reviewed. For modeling purposes, it is necessary to first develop a soot particle model.

Texas Tech University, Lulu Ma, August 2013

We have done this by assigning 720 coordinates to build a sphere as shown in figure A-1. The actual x, y, z coordinates used for the primary sphere are listed in table A-1 at end of this appendix. However, soot is not simply a spherical particle but rather composed of fractal aggregate chains of very small primary particles that are 40 – 50 nm in diameter. The chain aggregates can often exceed 1 – 2 microns in length. Such primary spheres and aggregates are illustrated in the TEM images of kerosene burner soot presented in figure A-2. To model such a particle in ADDA we first constructed the primary sphere of figure A1. After a primary sphere was built, the sphere can be “copied and pasted” within the 3D coordinate system by simply adding or subtracting integers such that the added primary sphere is translated in space to touch along the edges of the spheres. The resulting particle (also shown in figure A-2) mimics soot aerosols.

8. After a particle shape file is complete, run ADDA typing the correct command line. For this work, the command lines were:

adda -m 1.75 0.63 -size 0.75 -lambda 0.532 -shape read eightysoot.txt –grid 120 120 120

adda -m 1.75 0.63 -lambda 0.532 -shape read part.txt -size 0.060 –grid 12 12 12 (for primary soot sphere)

In the first example, grid size was specified in the command line to be a box 120x120x120. Paying attention to grid size is very important because the software scales the dimensions of the box to what is indicated by the “-size” command. The grid must be large enough to include all points specified in the shape file, but excessively large grids will lead to very long computations. In this case the coordinate system was 0.75 microns in length on edge and broken into 120 coordinates spaced by 0.00625 micrometers. Since the primary soot spheres are modeled in the shape file as being 8 dipoles in width, the resultant primary soot particle is ≈ 50 nm in diameter.

9. Once the run is complete, it’s a good idea to check the “log.txt” file present in the directory created by the software for the run (the directories are numbered automatically). A good way to verify the software interpreted your desire correctly (in terms of defining the particle shape) is to look at the “volume equivalent size parameter” listed in the log file. We can compare the listed volume equivalent size parameter with what we expect it to be based on the model. To start this process, the “volume equivalent diameter” (Dvolume) of a particle is the diameter of a sphere with the same volume as the particle in question (Vparticle). The relationship between these variables is: 1  6  3 Dvolume   Vparticle  (A-1)   Since the volume of any sphere is 4/3 r3 we can easily compute the volume of any soot aggregate particle by simply adding up the volume of n primary spheres. Then, use the total particle volume (Vparticle) to compute Dvolume. With this value in hand we can compute volume equivalent size parameter (VESP) by:

2r VESP  (A-2) 

Texas Tech University, Lulu Ma, August 2013

Where r is the particle radius (e.g. r = Dvolume / 2), and  is the wavelength of light considered in the simulation. A user should get good agreement between the VESP value reported in the log file and the computed value. If good agreement is not found, the particle model /size used is not correct. The grid size, or shape file, or number of grid points likely needs to be adjusted.

Texas Tech University, Lulu Ma, August 2013

Figure A. 1 Primary sphere used for soot studies. The sphere is composed of 720 x, y, z coordinates that are listed in table A-1.

Texas Tech University, Lulu Ma, August 2013

Figure A. 2 TEM images of fresh kerosene soot and (bottom) soot aggregate constructed from primary spheres used for soot modeling studies. This example sphere is composed of 80 primary spheres translated in x,y,z space to touch along edges. TEM images are courtesy of P. Buseck, Arizona State University.

Table A. 1 Cartesian coordinates for points on the fundamental sphere. For this work we assumed a complex refractive index of 1.75 – 0.63i for soot.

x y z 0 0 1 0 0 2 0 1 3 0 0 3 0 0 4 0 1 4 0 1 4 0 2 0 0 1 0 0 1 1 0 1 2 84

Texas Tech University, Lulu Ma, August 2013

0 2 0 0 2 1 0 2 2 0 2 3 0 3 1 0 3 0 0 2 4 0 3 2 0 3 3 0 4 0 0 4 1 0 4 2 1 0 0 1 0 1 1 0 2 1 0 3 1 0 4 1 1 0 1 1 1 1 1 2 1 1 3 1 1 4 1 2 0 1 2 1 1 2 2 1 2 3 1 2 4 1 3 0 1 3 1 1 3 2 1 3 3 1 4 0 1 4 1 1 4 2 2 0 0 2 0 1 2 0 2 2 0 3 2 0 4 2 1 0 2 1 1 2 1 2 2 1 3 2 1 4 2 2 0 2 2 1 2 2 2 2 2 3 2 2 4 2 3 0 2 3 1 2 3 2 2 3 3 2 4 0 2 4 1 2 4 2

Texas Tech University, Lulu Ma, August 2013

3 0 0 3 0 1 3 0 2 3 0 3 3 1 0 3 1 1 3 1 2 3 1 3 3 2 1 3 2 2 3 2 3 3 3 0 3 3 1 3 3 2 4 0 0 4 0 1 4 0 2 4 1 0 4 1 2 4 2 0 4 2 1 4 2 2 4 1 1 0 0 -1 0 0 -2 0 -1 -3 0 0 -3 0 0 -4 0 -1 -4 0 -1 -4 0 -2 0 0 -1 0 0 -1 -1 0 -1 -2 0 -2 0 0 -2 -1 0 -2 -2 0 -2 -3 0 -3 -1 0 -3 0 0 -2 -4 0 -3 -2 0 -3 -3 0 -4 0 0 -4 -1 0 -4 -2 -1 0 0 -1 0 -1 -1 0 -2 -1 0 -3 -1 0 -4 -1 -1 0 -1 -1 -1 -1 -1 -2 -1 -1 -3 -1 -1 -4

Texas Tech University, Lulu Ma, August 2013

-1 -2 0 -1 -2 -1 -1 -2 -2 -1 -2 -3 -1 -2 -4 -1 -3 0 -1 -3 -1 -1 -3 -2 -1 -3 -3 -1 -4 0 -1 -4 -1 -1 -4 -2 -2 0 0 -2 0 -1 -2 0 -2 -2 0 -3 -2 0 -4 -2 -1 0 -2 -1 -1 -2 -1 -2 -2 -1 -3 -2 -1 -4 -2 -2 0 -2 -2 -1 -2 -2 -2 -2 -2 -3 -2 -2 -4 -2 -3 0 -2 -3 -1 -2 -3 -2 -2 -3 -3 -2 -4 0 -2 -4 -1 -2 -4 -2 -3 0 0 -3 0 -1 -3 0 -2 -3 0 -3 -3 -1 0 -3 -1 -1 -3 -1 -2 -3 -1 -3 -3 -2 -1 -3 -2 -2 -3 -2 -3 -3 -3 0 -3 -3 -1 -3 -3 -2 -4 0 0 -4 0 -1 -4 0 -2 -4 -1 0 -4 -1 -2 -4 -2 0 -4 -2 -1 -4 -2 -2

Texas Tech University, Lulu Ma, August 2013

-4 -1 -1 0 0 1 0 0 2 0 1 3 0 0 3 0 0 4 0 1 4 0 1 4 0 2 0 0 1 0 0 1 1 0 1 2 0 2 0 0 2 1 0 2 2 0 2 3 0 3 1 0 3 0 0 2 4 0 3 2 0 3 3 0 4 0 0 4 1 0 4 2 -1 0 0 -1 0 1 -1 0 2 -1 0 3 -1 0 4 -1 1 0 -1 1 1 -1 1 2 -1 1 3 -1 1 4 -1 2 0 -1 2 1 -1 2 2 -1 2 3 -1 2 4 -1 3 0 -1 3 1 -1 3 2 -1 3 3 -1 4 0 -1 4 1 -1 4 2 -2 0 0 -2 0 1 -2 0 2 -2 0 3 -2 0 4 -2 1 0 -2 1 1 -2 1 2 -2 1 3 -2 1 4

Texas Tech University, Lulu Ma, August 2013

-2 2 0 -2 2 1 -2 2 2 -2 2 3 -2 2 4 -2 3 0 -2 3 1 -2 3 2 -2 3 3 -2 4 0 -2 4 1 -2 4 2 -3 0 0 -3 0 1 -3 0 2 -3 0 3 -3 1 0 -3 1 1 -3 1 2 -3 1 3 -3 2 1 -3 2 2 -3 2 3 -3 3 0 -3 3 1 -3 3 2 -4 0 0 -4 0 1 -4 0 2 -4 1 0 -4 1 2 -4 2 0 -4 2 1 -4 2 2 -4 1 1 0 0 1 0 0 2 0 -1 3 0 0 3 0 0 4 0 -1 4 0 -1 4 0 -2 0 0 -1 0 0 -1 1 0 -1 2 0 -2 0 0 -2 1 0 -2 2 0 -2 3 0 -3 1 0 -3 0 0 -2 4 0 -3 2 0 -3 3 0 -4 0

Texas Tech University, Lulu Ma, August 2013

0 -4 1 0 -4 2 1 0 0 1 0 1 1 0 2 1 0 3 1 0 4 1 -1 0 1 -1 1 1 -1 2 1 -1 3 1 -1 4 1 -2 0 1 -2 1 1 -2 2 1 -2 3 1 -2 4 1 -3 0 1 -3 1 1 -3 2 1 -3 3 1 -4 0 1 -4 1 1 -4 2 2 0 0 2 0 1 2 0 2 2 0 3 2 0 4 2 -1 0 2 -1 1 2 -1 2 2 -1 3 2 -1 4 2 -2 0 2 -2 1 2 -2 2 2 -2 3 2 -2 4 2 -3 0 2 -3 1 2 -3 2 2 -3 3 2 -4 0 2 -4 1 2 -4 2 3 0 0 3 0 1 3 0 2 3 0 3 3 -1 0 3 -1 1 3 -1 2 3 -1 3 3 -2 1 3 -2 2

Texas Tech University, Lulu Ma, August 2013

3 -2 3 3 -3 0 3 -3 1 3 -3 2 4 0 0 4 0 1 4 0 2 4 -1 0 4 -1 2 4 -2 0 4 -2 1 4 -2 2 4 -1 1 0 0 -1 0 0 -2 0 1 -3 0 0 -3 0 0 -4 0 1 -4 0 1 -4 0 2 0 0 1 0 0 1 -1 0 1 -2 0 2 0 0 2 -1 0 2 -2 0 2 -3 0 3 -1 0 3 0 0 2 -4 0 3 -2 0 3 -3 0 4 0 0 4 -1 0 4 -2 1 0 0 1 0 -1 1 0 -2 1 0 -3 1 0 -4 1 1 0 1 1 -1 1 1 -2 1 1 -3 1 1 -4 1 2 0 1 2 -1 1 2 -2 1 2 -3 1 2 -4 1 3 0 1 3 -1 1 3 -2 1 3 -3 1 4 0

Texas Tech University, Lulu Ma, August 2013

1 4 -1 1 4 -2 2 0 0 2 0 -1 2 0 -2 2 0 -3 2 0 -4 2 1 0 2 1 -1 2 1 -2 2 1 -3 2 1 -4 2 2 0 2 2 -1 2 2 -2 2 2 -3 2 2 -4 2 3 0 2 3 -1 2 3 -2 2 3 -3 2 4 0 2 4 -1 2 4 -2 3 0 0 3 0 -1 3 0 -2 3 0 -3 3 1 0 3 1 -1 3 1 -2 3 1 -3 3 2 -1 3 2 -2 3 2 -3 3 3 0 3 3 -1 3 3 -2 4 0 0 4 0 -1 4 0 -2 4 1 0 4 1 -2 4 2 0 4 2 -1 4 2 -2 4 1 -1 0 0 1 0 0 2 0 -1 3 0 0 3 0 0 4 0 -1 4 0 -1 4 0 -2 0 0 -1 0

Texas Tech University, Lulu Ma, August 2013

0 -1 1 0 -1 2 0 -2 0 0 -2 1 0 -2 2 0 -2 3 0 -3 1 0 -3 0 0 -2 4 0 -3 2 0 -3 3 0 -4 0 0 -4 1 0 -4 2 -1 0 0 -1 0 1 -1 0 2 -1 0 3 -1 0 4 -1 -1 0 -1 -1 1 -1 -1 2 -1 -1 3 -1 -1 4 -1 -2 0 -1 -2 1 -1 -2 2 -1 -2 3 -1 -2 4 -1 -3 0 -1 -3 1 -1 -3 2 -1 -3 3 -1 -4 0 -1 -4 1 -1 -4 2 -2 0 0 -2 0 1 -2 0 2 -2 0 3 -2 0 4 -2 -1 0 -2 -1 1 -2 -1 2 -2 -1 3 -2 -1 4 -2 -2 0 -2 -2 1 -2 -2 2 -2 -2 3 -2 -2 4 -2 -3 0 -2 -3 1 -2 -3 2 -2 -3 3 -2 -4 0

Texas Tech University, Lulu Ma, August 2013

-2 -4 1 -2 -4 2 -3 0 0 -3 0 1 -3 0 2 -3 0 3 -3 -1 0 -3 -1 1 -3 -1 2 -3 -1 3 -3 -2 1 -3 -2 2 -3 -2 3 -3 -3 0 -3 -3 1 -3 -3 2 -4 0 0 -4 0 1 -4 0 2 -4 -1 0 -4 -1 2 -4 -2 0 -4 -2 1 -4 -2 2 -4 -1 1 0 0 -1 0 0 -2 0 1 -3 0 0 -3 0 0 -4 0 1 -4 0 1 -4 0 2 0 0 1 0 0 1 -1 0 1 -2 0 2 0 0 2 -1 0 2 -2 0 2 -3 0 3 -1 0 3 0 0 2 -4 0 3 -2 0 3 -3 0 4 0 0 4 -1 0 4 -2 -1 0 0 -1 0 -1 -1 0 -2 -1 0 -3 -1 0 -4 -1 1 0 -1 1 -1 -1 1 -2

Texas Tech University, Lulu Ma, August 2013

-1 1 -3 -1 1 -4 -1 2 0 -1 2 -1 -1 2 -2 -1 2 -3 -1 2 -4 -1 3 0 -1 3 -1 -1 3 -2 -1 3 -3 -1 4 0 -1 4 -1 -1 4 -2 -2 0 0 -2 0 -1 -2 0 -2 -2 0 -3 -2 0 -4 -2 1 0 -2 1 -1 -2 1 -2 -2 1 -3 -2 1 -4 -2 2 0 -2 2 -1 -2 2 -2 -2 2 -3 -2 2 -4 -2 3 0 -2 3 -1 -2 3 -2 -2 3 -3 -2 4 0 -2 4 -1 -2 4 -2 -3 0 0 -3 0 -1 -3 0 -2 -3 0 -3 -3 1 0 -3 1 -1 -3 1 -2 -3 1 -3 -3 2 -1 -3 2 -2 -3 2 -3 -3 3 0 -3 3 -1 -3 3 -2 -4 0 0 -4 0 -1 -4 0 -2 -4 1 0 -4 1 -2 -4 2 0

Texas Tech University, Lulu Ma, August 2013

-4 2 -1 -4 2 -2 -4 1 -1 0 0 -1 0 0 -2 0 -1 -3 0 0 -3 0 0 -4 0 -1 -4 0 -1 -4 0 -2 0 0 -1 0 0 -1 -1 0 -1 -2 0 -2 0 0 -2 -1 0 -2 -2 0 -2 -3 0 -3 -1 0 -3 0 0 -2 -4 0 -3 -2 0 -3 -3 0 -4 0 0 -4 -1 0 -4 -2 1 0 0 1 0 -1 1 0 -2 1 0 -3 1 0 -4 1 -1 0 1 -1 -1 1 -1 -2 1 -1 -3 1 -1 -4 1 -2 0 1 -2 -1 1 -2 -2 1 -2 -3 1 -2 -4 1 -3 0 1 -3 -1 1 -3 -2 1 -3 -3 1 -4 0 1 -4 -1 1 -4 -2 2 0 0 2 0 -1 2 0 -2 2 0 -3 2 0 -4 2 -1 0 2 -1 -1 2 -1 -2

Texas Tech University, Lulu Ma, August 2013

2 -1 -3 2 -1 -4 2 -2 0 2 -2 -1 2 -2 -2 2 -2 -3 2 -2 -4 2 -3 0 2 -3 -1 2 -3 -2 2 -3 -3 2 -4 0 2 -4 -1 2 -4 -2 3 0 0 3 0 -1 3 0 -2 3 0 -3 3 -1 0 3 -1 -1 3 -1 -2 3 -1 -3 3 -2 -1 3 -2 -2 3 -2 -3 3 -3 0 3 -3 -1 3 -3 -2 4 0 0 4 0 -1 4 0 -2 4 -1 0 4 -1 -2 4 -2 0 4 -2 -1 4 -2 -2 4 -1 -1

Texas Tech University, Lulu Ma, August 2013

REFERENCES

1. Pöschl, U. Angew.Chem. Int. Ed.2005, 44, 7520-7540.

2. Hinds, W. C. Aerosol Technology: Properties, Behavior, and Measurement of Airborne Particles, 2nd Edition, Wiley Interscience, New York, 1999.

3. Halmer, M. M.; Schmincke, H.-U.; Graf, H.-F. Journal of Volcanology and Geothermal Research 2002, 115, 511-528.

4. Oppenheimer, C. Progress in Physical Geography 2003, 27, 230-259.

5. van der Werf, G. R.; Randerson, J. T.; Giglio, L.; Collatz, G. J.; Mu, M.; Kasibhatla, P. S.; Morton, D. C.; DeFries, R. S.; Jin, Y.; van Leeuwen, T. T. Atmos. Chem. Phys.2010, 10, 11707-11735.

6. Engelstaedter, S.; Tegen, I.; Washington, R. Earth-Science Reviews 2006, 79, 73-100.

7. Yu, H.; Remer, L. A.; Chin, M.; Bian, H.; Tan, Q.; Yuan, T.; Zhang, Y. Science 2012, 337, 566-569.

8. Prospero, J. M.; Blades, E.; Mathison, G.; Naidu, R. Aerobiologia 2005, 21, 1-19.

11. Lei, Y.; Zhang, Q.; He, K. B.; Streets, D. G. Atmos. Chem. Phys.2011, 11, 931-954.

10. Forster, P.; Ramaswamy, V.; Artaxo, P.; Berntsen, T.; Betts, R.; Fahey, D. W.; Haywood, J.; Lean, J.; Lowe, D. C.; Myhre, W.; Nganga, J.; Prinn, R.; Raga, G.; Schulz, M.; Van Dorland, R. Changes in atmospheric constituents and in radiative forcing. In Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University press: Cambridge, U.K. and New York, U.S., 2007.

11. Zhang, Q.; Jimenez, J. L.; Canagaratna, M. R.; Allan, J. D.; Coe, H.; Ulbrich, I.; Alfarra, M. R.; Takami, A.; Middlebrook, A. M.; Sun, Y. L.; Dzepina, K.; Dunlea, E.; Docherty, K.; DeCarlo, P. F.; Salcedo, D.; Onasch, T.; Jayne, J. T.; Miyoshi, T.; Shimono, A.; Hatakeyama, S.; Takegawa, N.; Kondo, Y.; Schneider, J.; Drewnick, F.; Borrmann, S.; Weimer, S.; Demerjian, K.; Williams, P.; Bower, K.; Bahreini, R.; Cottrell, L.; Griffin, R. J.; Rautiainen, J.; Sun, J. Y.; Zhang, Y. M.; Worsnop, D. R. Geophysical Research Letters 2007, 34, L13801.

12. Bi, X.; Simoneit, B. R. T.; Sheng, G.; Ma, S.; Fu, J. Atmospheric Research 2008, 88, 256- 265.

13. Kolb, C. E.; Worsnop, D. R. Annu.Rev. Phys. Chem. 2012, 63, 471-91.

Texas Tech University, Lulu Ma, August 2013

14. Yu, S.; Dennis, R. L.; Bhave, P. V.; Eder, B. K.; Atmospheric Environment 2004, 38, 5257-5268.

15. Paredes-Miranda, G.; Arnott, W. P.; Jimenez, J. L.; Aiken, A. C.; Gaffney, J. S.; Marley, N. A.; Atmos. Chem. Phys.2009, 9, 3721-3730.

16. Kroll, J H.; Seinfeld, J. H. Atmospheric Environment 2008, 42, 3593-3642.

17. Finlayson-Pitts, B. J.; Pitts, Jr. J. N. Chemistry of the Upper and Lower Atmosphere: Theory, Experiments, and Applications, Academic Press, San Diego CA, 2000.

18. Pandithurai, G.; Dipu, S.; Dani, K. K.; Tiwari, S.; Bisht, D. S.; Devara, P. C. S.; Pinker, R. T. Journal of Geophysical Research 2008, 113, D13209.

19. Seinfeld, J. H.; Carmichael, G. R.; Arimoto, R.; Conant, W. C.; Brechtel, F. J.; Bates, T. S.; Cahill, T. A.; Clarke, A. D.; Doherty, S. J.; Flatau, P. J.; Huebert, B. J.; Kim, J.; Markowicz, K. M.; Quinn, P. K.; Russell, L. M.; Russell, P. B.; Shimizu, A.; Shinozuka, Y.; Song, C. H.; Tang, Y.; Uno, I.; Vogelmann, A. M.; Weber,R. J.; Woo, J.; Z HANG, X. Y. Bull. Am. Meteorol. Soc.2004, 85, 367-380.

20. Schnaiter, M.; Linke, C.; MÖhler, O.; Naumann, K.-H.; Saathoff, H.; Wagner, R.; Schurath, U.; Wehner, B. Journal of Geophysical Research 2005, 110, D19204.

21. Tie, X.; Wu, D.; Brasseur, G. Atmospheric Environment 2009, 43, 2375-2377.

22. http://www.bloomberg.com/news/2013-01-30/beijing-air-akin-to-living-in-smoking- lounge-chart-of-the-day.html

23. http://www.cbsnews.com/8301-205_162-57564051/nasa-releases-images-of-beijing- air-pollution/

24. Politis, M.; Pilinis, C.; Lekkas, T. D.; Global NEST Journal 2008, 10(3), 439-452.

25. Kennedy, I. M. Proceedings of the Combustion Institute 2007, 31, 2757-2770.

26. Redmond, H. E.; Dial, K. D.; Thompson, J. E. Aeolian Research 2010, 2, 5-26.

27. Lan, Z.; Huang, X.; Yu, K.; Sun, T.; Zeng, L.; Hu, M.; Atmospheric Environment 2013, 69, 118-123.

28.Ma, L.;,Zobeck, T. M.; Hsieh, D. H.; Holder, D.; Morgan, C. L.S.; Thompson, J. E. Aeolian Research 2011, 3, 235-242.

29.Berden, G.; Peeters, R.; Meijer, G.;In t. Reviews in Physical Chemistry 2000,19(4), 565-607.

30.Herbelin, J. M.; McKay, J. A. Applied Optics 1981, 20(19), 3341-3344.

Texas Tech University, Lulu Ma, August 2013

31.Anderson, D. Z.; Frisch,J. C.;Masser, C. S. Applied Optics 1984, 23(8), 1238-1245.

32.O'Keefe, A.; Deacon, D. A.G. Review of Scientific Instruments 1988, 59, 2544.

33.Baynard, T.; Lovejoy, E. R.; Pettersson, A.; Brown, S. S.; Lack,D.; Osthoff, H.; Massoli, P.; Ciciora, S.; Dube, W. P.; Ravishankara, A. R. Aerosol Science and Technology 2007,41, 447–462.

34.Thompson, J. E.; Smith, B. W.;Winefordner, J. D. Anal.Chem. 2002, 74, 1962-1967.

35.Thompson, J. E.; Nasajpour, H. D.; Smith, B. W.; Winefordner, J. D. Aerosol Science and Technology 2003, 37, 221-230.

36.Ma, L.; Thompson, J.E. Anal.Chem. 2012, 84, 5611−5617.

37.Beuttell,R. G.; Brewer, A. W.Journal of Scientific Instruments 1949, 26(11),357-359.

38.Heintzenberg, J.; Charlson, R. J. Journal of Atmospheric and Oceanic Technology 1996, 13, 987-1000.

39.Johnson, R. W. Atmospheric Environment 1981, 15(10), 1835-1845.

40.Anderson, T. L.; Covert, D. S.; Marshall, S. F.; Laucks, M. L.; Charlson, R. J.; Waggoner, A. P.; Ogren, J. A.; Caldow, R.; Holm, R. L.; Quant, F. R.; Sen, G. J.; Wiedensohler, A.; Ahlquist, N. A.; Bates, T. S.Journal of Atmospheric and Oceanic Technology 1996, 13, 967-986.

41.McMurry, P. H. Atmospheric Environment 2000, 34, 1959-1999.

42. Moosmüller, H.;Arnott, W. P. Rev. Sci. Instrum. 2003, 74, 3492-3501.

43.Mulholland, G. W.;Bryner, N. P. Atmospheric Environment 1994,28(5), 873-887.

44.Moosmüller, H.; Chakrabarty, R. K.; Arnott, W. P. Journal of Quantitative Spectroscopy &Radiative Transfer 2009, 110, 844–878.

45.Bond, T. C.; Anderson,T. L.; Campbell, D. Aerosol Science and Technology 1999, 30, 582-600.

46.Schmid, O.; Artaxo, P.; Arnott, W. P.; Chand, D.; Gatti, L. V.; Frank, G. P.; Hoﬀ er, A.; Schnaiter, M.; Andreae, M. O.;Atmos. Chem. Phys. Discuss.2005, 5, 9355– 9404.

47.Sheridan, P. J.; Arnott, W. P.; Ogren, J. A.; Andrews, E.; Atkinson, D. B.; Covert, D. S.;Moosmüller, H.; Petzold, A.; Schmid,B.; Strawa, A. W.; Varma, R.;Virkkula, A. Aerosol Science and Technology 2005, 39, 1-16.

48. Twomey, S.A. J. Atmos. Sci. 1977 34, 1149-1152.

100

Texas Tech University, Lulu Ma, August 2013

49. Albrecht, B.A. Science 1989, 245, 1227-1230.

50. Self, S.; Zhao, J. X.; Holasek, R. E.; Torres, R.C.; King, A.J. The Atmospheric Impact of the 1991 Mount Pinatubo Eruption FIRE and MUD: Eruptions and Lahars of Mount Pinatubo, Philippines , Philippine Institute of Volcanology and Seismology, Quezon City and University of Washington Press, 1996.

51 Travis, D.J.; Carleton, A.M.; Lauritsen, R.G. J. Climate 2004, 17, 1123–1134.

52. Textor, C.; Schulz, M.; Guibert, S.; Kinne, S.; Balkanski, Y.; Bauer, S.; Berntsen, T.; Berglen, T.; Boucher, O.; Chin, M.; Dentener, F.; Diehl, T.; Easter, R.; Feichter, H.; Fillmore, D.; Ghan, S.; Ginoux, P.; Gong, S.; Grini, A.; Hendricks, J.; Horowitz, L.; Huang, P.; Isaksen, I.; Iversen, T.; Kloster, S.; Koch, D.; Kirkevåg, A.; Kristjansson, J. E.; Krol, M.; Lauer, A.; Lamarque, J. F.; Liu, X.; Montanaro, V.; Myhre, G.; Penner, J.; Pitari, G.; Reddy, S.; Seland, Ø.; Stier, P.; Takemura, T.; Tie, X. Atmos. Chem. Phys. Discuss. 2005, 5, 8331-8420.

53. Guo, J.; Rahn, K. A.; Zhuang, G. Atmos. Environ. 2004. 38, 855-862.

54. Malm, W. C; Day, D. E. Atmos. Environ. 2000, 34, 3373-3391.

55. Chow, J. C.; Watson, J. G.; Park, K.J. J. Air Waste Man. Assoc. 2006, 54(6), 398-410.

56. Gerasopoulos, E.; Kokkalis, P.; Amiridis, V.; Liakakou, E.; Perez, C.; Haustein, K.; Eleftheratos, K.; Andreae, M. O.; Andreae, T. W.; Zerefos, C. S. Ann. Geophys. 2009, 27, 2903-2912.

57. Cook, B. I.; Miller, R. L.; Seager, R. Proc. Natl. Acad. Sci. USA 2009, 106(13), 4997- 5001.

58. USDA Natural Resources Conservation Service. ftp://ftp- fc.sc.egov.usda.gov/TX/soils/tx_gsm_map.pdf

59. Waiser, T. H.; Morgan, C. L. S.; Brown, D. J., Hallmark, C.T., soil sci. Soc. Am. J. 2007, 71 (2), 389-396.

60. Soil Survey Division Staff. Soil survey manual. Soil Conservation Service. U.S. Department of Agriculture Handbook 18. U.S. Govt. Printing Off., Washington, DC. 1993.

61. Day, P. R. Particle fractionation and particle size analysis. In Methods of Soil Analysis. Part 1. Physical and Mineralogical Properties, Including Statistics of Measurement and Sampling, Agronomy Monograph 9.1 Black, C. A. (Ed.) American Society of Agronomy, Soil Science Society of America, Madison, WI, pp. 545-567, 1965.

101

Texas Tech University, Lulu Ma, August 2013

62. Burt, R. Soil survey laboratory methods manual. Soil Survey Investigation Report No. 42, Ver. 4.0 USDA, Natural Resources Conservation Service, Sol Survey Laboratory. Lincoln, NE. 2004.

63. Gill, T. E.; Zobeck, T. M.; Stout, J. E. J. Hazard. Mater. 2006, 132 (1), 1-13.

64. Cheng, T.; Lu, D.; Chen, H.; Xu, Y. Atmos. Environ. 2005, 39, 1237-1243.

65. Weinaierl, B.; Petzold, A.; Esselborn, M.; Wirth, M.; Rasp, K.; Kandler, K.; Schütz, L.; Koepke, P.; Fiebig, M. Tellus 2009, 61B, 96-117.

66. Crabtree, G.W. Dustfall on the Southern High Plains of Texas. M.S. Thesis, Texas Tech University, Lubbock, TX, 2005.

67. Müller, T.; Petzold, A.; Rasp, K.; Schladitz, A.; Vragel, M.; Schnaiter, M.; Linke, C. European Aerosol Conference 2009, Karlsruhe, Abstract.

68. Waiser, T. H.; Morgan, C. L. S.; Brown, D. J.; Hallmark, C. T. Soil Sci. Soc. Am. J. 2007, 71, 389-396.

69. Thompson, J.E.; Barta, N.; Policarpio, D.; DuVall, R. Opt. Exp. 2008, 16(3), 2191- 2205.

70. Dial, K.D.; Hiemstra, S.; Thompson, J.E. Anal. Chem. 2010, 82, 7885-7896.

71. Claquin, T.; Schulz, M.; Balkanski, Y.J., J. Geophys. Res. 1999, 104(18), 22243- 22256.

72. Gerasopoulos, E.; Kokkalis, P.; Amiridis, V.; Liakakou, E.; Perez, C.; Haustein, K.; Eleftheratos, K.; Andreae, M.O.; Andreae, T.W.; Zerefos, C.S. Ann. Geophys. 2009, 27, 2903–2912.

73. Malm, W.C.; Day, D.E. Atmos. Environ. 2000, 34, 3373-3391.

74. DeBell, L.J.; Gebhart, K.A.; Hand, J.L.; Malm, W.C.; Pitchford, M.L.; Schichtel, B.A.; White, W.H. Chapter 2: Spatial Distributions of Reconstructed Mass and Mass Budgets and Reconstructed Light Extinction and Light-Extinction Budgets In Spatial and Seasonal Patterns and Temporal Variability of Haze and its Constituents in the United States Report IV 2006.

75. Ramaswamy, V.; Boucher, O.; Haigh, J.; Hauglustaine, D.; Haywood, J.; Myhre, G.; Nakajima, T.; Shi, G.Y.; Solomonet, S. Radiative forcing of climate change, In Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change; Cambridge University Press: Cambridge, UK, 2001.

102

Texas Tech University, Lulu Ma, August 2013

76. Ramanathan, V.; Crutzen, P.J.; Kiehl, J.T.; Rosenfeld, D. Science 2001, 294, 2119−2124.

77. Bulatov, V.; Fisher, M.; Schechter, I. Anal. Chim. Acta 2002, 466, 1-9.

78. Lang-Yona, N.; Rudich, Y.; Segre, E.; Dinar, E.; Abo-Riziq, A. Anal. Chem. 2009, 81, 1762-1769.

79. Brown, S.S.; Stark, H.; Chicora, S.J.; McLaughlin, R.J.; Ravishankara, A.R. Rev. Sci. Instrum. 2002, 73, 3291-3301.

80. Corr, C. A.; Krotkov, N.; Madronich, S.; Slusser, J. R.; Holben, B.; Gao, W.; Flynn, J.; Lefer, B.; Kreidenweis, S. M. Atmos. Chem. Phys. 2009, 9, 5813−5827.

81. Alexander, D. T. L.; Crozier, P. A.; Anderson, J. R. Science 2008, 321, 833-836.

82. Moosmüller, H.; Chakrabarty, R. K.; Ehlers, K. M.; Arnott, W. P. Atmos. Chem. Phys. 2011, 11, 1217-1225.

83. Hecobian, A.; Zhang, X.; Zheng, M.; Frank, N.; Edgerton, E. S.; Weber, R. J. Atmos. Chem. Phys. Discuss. 2010, 10, 7601−7639.

84. Chang, J.; Thompson, J. E. Atmos. Environ. 2010, 44 (4), 541-551.

85. Thompson, J. E.; Hayes, P. L.; Jimenez, J. L.; Adachi, K.; Zhang, X.; Liu, J.; Weber, R. J.; Buseck, P. R. Atmos. Environ. 2012, 55, 190-200.

86. Gyawali, M.; Arnott, W. P.; Zaveri, R. A.; Song, C.; Moosm üller, H.; Liu, L.; Mishchenko, M. I.; Chen, L.-W. A.; Green, M. C.; Watson, J. G.; Chow, J. C. Atmos. Chem. Phys. 2012, 12, 2587−2601.

87. Ajtai, T.; Filep, A.; Schnaiter, M.; Linke, C.; Vragel, M.; Bozoki, Z.; Szabo, G.; Leisner, T. J. Aerosol Sci. 2010, 41, 1020−1029.

88. Thompson, J.E.; G.I.T. Laboratory Journal 2010, 9-10.

89. Bevington, P. R.; Robinson, D.K. Data Reduction for the Physical Sciences; McGraw-Hill: New York, 1992; 41−50.

90. Redmond, H.; Thompson, J. E. Phys. Chem. Chem. Phys. 2011, 13, 6872−6882.

91. Toon, O. B.; Pollack, J. B.; Khare, B. N. J. Geophys. Res. 1976, 81, 5733−5748.

92. Nakayama, T.; Matsumi, Y.; Sato, K.; Imamura, T; Yamazaki, A.; Uchiyama, A. J. Geophys. Res. 2010, 115, D24204.

93. Dubovik, O.; Holben, B. N.; Kaufman, J.; Yamasoe, M.; Smimov, A.; Tan, D.; Slutsker, I. J. Geophys. Res. 1998, 103 (D24), 31903−31923.

103

Texas Tech University, Lulu Ma, August 2013

94. Linke, C.; Móhler, O.; Veres, A.; Mohácsi, A.; Bozóki, Z.; Szabó, G.; Schnaiter, M. Atmos. Chem. Phys. 2006, 6, 3315−3323.

95. Zhang, X.; Lin, Y.-H.; Surratt, J. D.; Zotter, P.; Prevot, A. S. H.; Weber, R. J. Geophys. Res. Lett. 2011, 38, L21810.

96. Jacobson, M. Z. J. Geophys. Res. 1999, 104 (D3), 3527−3542.

97. Ajtai, T.; Filep, A.; Utry, N.; Schnaiter, M.; Linke, C.; Bozoki, Z.; Szabo, G.; Leisner, T. J. Atmos. Sci. 2011, 42, 859−866.

98. Arnott, W. P.; Moosmüller, H.; Rogers, C. F.; Jin, T.; Bruch, R. Atmos. Environ. 1999, 33 (17), 2845−2852.

99. Bond, T. C.; Doherty, S. J.; Fahey, D. W.; Forster, P. M.; Berntsen, T.; DeAngelo, B. J.; Flanner, M. G.; Ghan, S.; Kärcher, B.; Koch, D.; Kinne, S.; Kondo, Y.; Quinn, P. K.; Sarofim, M. C.; Schultz, M. G.; Schulz, M.; Venkataraman, C.; Zhang, H.; Zhang, S.; Bellouin, N.; Guttikunda, S. K.; Hopke, P. K.; Jacobson, M. Z.; Kaiser, J. W.; Klimont, Z.; Lohmann, U.; Schwarz, J. P.; Shindell, D.; Storelvmo, T.; Warren, S. G.; Zender, C. S. J. Geophys. Res. Atmos. 2013, DOI: 10.1002/jgrd.50171.

100. McConnell, J., R. Edwards, G. Kok, M. Flanner, C. Zender, E. Saltzman, J. Banta, D. Pasteris, M. Carter, J. Kahl, Science 2007, 317, 1381-1384.

101. Kaden, D.A.; Hites, R.A.; Thilly, W.G. Cancer Res. 1979, 39, 4152-4159.

102. Götze, H.J.; Schneider, J.; Herzog, H.G. Fresenius J Anal Chem 1991, 340, 27-30.

103. Olson, K. L.; Harris, S.J.; Weiner, A.M. Combustion Science & Technol. 1987, 51(1-3), 97-102.

104. Birch, M. E.; Cary, R. A. Aerosol Science and Technology 1996, 25, 221-241.

105. Sunlab.com accessed Feb 1 2013.

106. Borakab, J.; Siriannib, G.; Cohenc, H.J., Chemerynskiab, S.; Wheelerd, R. AIHA Journal, 2003, 64(2), 260-268.

107. Mageesci.com accessed Feb 1 2013.

108. Miyazaki, Y.; Kondo, Y; Sahu, L. K; Imaru, J.; Fukushima, N.; Kano, M.; J. Environ Monit. 2008, 10(10), 1195-1201.

109. Stephens, M.; Turner, N.; Sandberg, J. Applied Optics, 2003, 42(19), 3726-3736.

110. Thomson, K.; Reimann, J.; Kuhlmann, S.-A.; Will, S.; Cignoli, F.; Bellomunno, C.; Maffi, S.; Zizak, G.; Vander Wal, R. L.; Goulay, F.; Schrader, P. E.; Michelsen, H. 104

Texas Tech University, Lulu Ma, August 2013

A.; Liu, F.; Snelling, D. R.; Thomson, K. A.; Smallwood, G. J.; Migliorini, F.; Iuliis, S. D; Maffi, S.; Cignoli, F.; Zizak, G.; Bladh, H.; Johnsson, J.; Bengtsson, P.-E.; Snelling, D. R.; Zerbs, J.; Geigle, K. P.; Lammel, O.; Hader, J.; Stirn, R.; Hadef, R.; Meier, W.; Ochoterena, R. Applied Physics B, 2009, 96(4), 581-709.

111. http://www.dropletmeasurement.com/products/carbon-sensing-instruments/sp2.html accessed Feb 1 2013.

112. Schwarz, J. P.; Spackman, J. R.; Gao, R. S.; Watts, L. A.; Stier, P.; Schulz, M.; Davis, S. M.; Wofsy, S. C.; Fahey, D. W. Geophys. Res. Lett. 2010, 37, L18812.

113. Wofsy, S. C. Phil. Trans. R. Soc. A 2011, 369, 2073–2086.

114. Bond, T.C.; Bergstrom, R.W. Aerosol Sci. & Technol. 2006, 40, 27-67.

115. Bond, T.C.; Habib, G.; Bergstrom, R.W. J. Geophys. Res. 2006, 111, D20211.

116. Lu, T.; Cheung, C.S.; Huang, Z. Aerosol Sci. & Technol. 2012, 46, 1349-1358.

117. Subramanian, R.; Kok, G. L.; Baumgardner, D.; Clarke, A.; Shinozuka, Y.; Campos, T. L.; Heizer, C. G.; Stephens, B. B.; de Foy, B.; Voss, P. B.; and Zaveri, R. A. Atmos. Chem. Phys. 2010, 10, 219-237.

118. Yurkin, M.A.; Hoekstra, A.G. J.Quant.Spectrosc.Radiat.Transf. 2007, 106, 558-589.

119. Yurkin, M.A.; Hoekstra, A.G. J.Quant.Spectrosc.Radiat.Transf. 2011, 112, 2234- 2247.

120. Yoder, G.D.; Diwakar, P.K.; Hahn, D.W. Applied Optics 2005, 44(20), 4211-4219.

121. Vander Wal, R. L.; Weiland, K. J. App. Phys. B 1994, 59, 445-452.

122. Michelson, H.A.; Tivanski, A.V.; Gilles, M.K.; van Poppel, L.H.; Dansson, M.A.; Buseck, P.R. Applied Optics 2007, 46(6), 959-977.

123. Vander Wal, R. L.; Ticich, T.M.; Stephens, A. B. Appl. Phys. B 1998, 67, 15-123.

124. Krüger, V.; Wahl, C.; Hadef, R.; Geigle, K.P.; Stricker, W.; Aigner, M. Meas. Sci. & Technol. 2005, 16, 1477-1486.

125. Stipe, C.B.; Choi, J.H.; Lucas, D.; Koshland, C.P.; Sawyer, R.F. Journal of Nanoparticle Research 2004, 6, 467–477.

105