FORMALDEHYDE AS A PROBE OF RURAL

VOLATILE ORGANIC COMPOUND OXIDATION

by

Joshua P. DiGangi

A dissertation submitted in partial fulfillment of the

requirements for the degree of

Doctor of Philosophy

(Chemistry)

at the

University of Wisconsin–Madison

2012

i

To my wife, Easter, and my parents, Joseph and Cynthia

ii

Acknowledgements

Collaboration is an important part of any scientific work, but is absolutely crucial

for atmospheric chemistry field measurements. Entire teams of collaborators are vital to

measure the wide variety of species needed to draw any significant conclusions. For that,

I would like to thank all members of both the BEARPEX and BEACHON-ROCS science

teams, for their hard work, advice, and experience that made the work in this thesis possible.

Similarly, complex campaigns require a great deal of non-science logistics, such as housing,

transportation, materials, etc. The work of the Blodgett Forest Research Station Staff

and Rocky Mountain Research Station-Manitou Staff during these campaigns was much

appreciated, as it allowed the rest of us to focus on science.

I cannot imagine a place I would have rather spent my graduate career than in the

UW-Madison Chemistry department. It is full of wonderful people: supportive and available

faculty, knowledgeable and friendly staff, and enthusiastic students. Unfortunately, it is not

possible to list all of the people that have helped me along the way. I would like to thank all

of the members of my committee for their advice and guidance through my graduate process,

especially Ankur Desai, whose input was critical to my understanding of micrometeorology.

Thanks to the Crim and Wright groups, for helping us get on our feet by always lending that

crucial component for our experiments (hopefully, we gave it all back). I would like to thank

Rob McClain for lending us equipment, but also providing instrumental guidance. Thanks

to the members of the machine, electronics, and glass shops for their skill and patience. I

would like to, in particular, thank Jerry Stamn for his critical eye, good stories, and always

helping to make sure we were ready for campaigns. Finally, I would like to thank Mark

iii

Wendt and Samir Youssef, whose teaching examples I try to emulate with every student.

Were I half as good at instruction as either, I would consider myself a great educator.

I have also been fortunate to work with an amazing group of people in the Keutsch

group. John Hottle and Andy Huisman laid most of the ground work for much of the

instrumentation techniques I have used, and my work was truly performed by standing on

their shoulders. Melissa Galloway has been my compatriot since the beginning, and has

always been a wonderful sounding board, co-complainer, and friend. Sam Henry has worked

with me on most of my field campaigns, and has provided feedback and camaraderie, in

and out of the lab. Even though he is junior to me, I am sure I have learned at least as

much from Sam as he has from me. Glenn Wolfe’s experience and advice with fluxes and

atmospheric chemistry have been very helpful, and his patience has been greatly appreciated.

Sam Henry, Jen Knapp, and Kate Skog in particular were quite helpful in the preparation

of this manuscript. Finally, thanks to Ben Bratton, our unofficial group member, adept at

both MATLAB and the lifting of heavy objects.

The advisor-student relationship not only shapes one’s graduate experience, but can

shape the entire way one views science. I am lucky to have had Frank Keutsch as a mentor

for my graduate career. His support, critical eye, experience, and, most of all, patience were

central to the work in this thesis. As Frank often says, his graduate students do all the work,

but the best ideas are usually his.

I owe a great debt of gratitude to my parents, Joseph DiGangi and Cynthia Carney.

My thanks are not only for their hard work ensuring that I had every opportunity possible,

but for instilling the ethics and encouraging the questioning nature that have allowed me to

succeed.

Finally, thanks to my wife, Easter, for her cheerleading, her compassion, and being

my best friend. I love you, hon.

iv

Contents

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii

1 Introduction 1

1.1 Trace Species and Air Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1.1 Harmful effects of ozone and aerosol . . . . . . . . . . . . . . . . . . . 2

1.1.2 Quantifying poor air quality and mitigation . . . . . . . . . . . . . . . 3

1.2 Volatile Organic Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.2.1 Emission and Processing . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.2.2 Oxidized Volatile Organic Compounds . . . . . . . . . . . . . . . . . . 5

1.2.3 Models and VOC oxidation . . . . . . . . . . . . . . . . . . . . . . . . 6

1.3 Formaldehyde in the Atmosphere . . . . . . . . . . . . . . . . . . . . . . . . . 7

1.3.1 Production and destruction . . . . . . . . . . . . . . . . . . . . . . . . 7

1.3.2 Importance and Challenges . . . . . . . . . . . . . . . . . . . . . . . . 9

1.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

1.5 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2 Design and Characterization of the FILIF Technique for HCHO Detection 21

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

2.2 Instrument Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

2.2.1 Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

v

2.2.2 Fiber Laser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

2.2.3 Optical Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

2.2.4 Gas System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

2.2.5 Data Acquisition Principle . . . . . . . . . . . . . . . . . . . . . . . . 27

2.3 Instrument Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

2.3.1 Initial Studies - Prototype Laser . . . . . . . . . . . . . . . . . . . . . 28

2.3.2 Studies with Non-Prototype (Field Ready) Laser . . . . . . . . . . . . 30

2.4 Inlet Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

2.4.1 BEARPEX 2009 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

2.4.2 CalNex-SJV 2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

2.4.3 BEACHON-ROCS 2010 . . . . . . . . . . . . . . . . . . . . . . . . . . 34

2.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

2.6 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

2.7 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

3 First Direct Measurements of Formaldehyde Flux via Eddy Covariance 57

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

3.2 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

3.2.1 Field Campaign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

3.2.2 Fiber Laser-Induced Fluorescence (FILIF) of HCHO . . . . . . . . . . 60

3.2.3 Other Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

3.2.4 Eddy Covariance Measurements . . . . . . . . . . . . . . . . . . . . . 63

3.3 Data and Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

3.3.1 Gradient and Flux Profile . . . . . . . . . . . . . . . . . . . . . . . . . 67

3.3.2 Emission Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

3.4 Zero-Dimensional Box Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

vi

3.4.1 Chemical Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

3.4.2 Chemical Loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

3.4.3 Direct Emission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

3.4.4 Dry Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

3.5 Model Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

3.5.1 General Sensitivity Analyses . . . . . . . . . . . . . . . . . . . . . . . 77

3.5.2 PPine Emission Sensitivity (E350) . . . . . . . . . . . . . . . . . . . . 78

3.5.3 MBO Sensitivity (VOC-I) . . . . . . . . . . . . . . . . . . . . . . . . . 79

3.5.4 Monoterpene Sensitivity (VOC-II) . . . . . . . . . . . . . . . . . . . . 80

3.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

3.7 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

3.8 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

3a Supplementary materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

3a.1 HCHO Permeation Tube Calibration . . . . . . . . . . . . . . . . . . . 101

3a.2 Error in Flux Measurements . . . . . . . . . . . . . . . . . . . . . . . . 101

3a.3 HCHO Production via Methylperoxy Radical . . . . . . . . . . . . . . 102

3a.4 Aerodynamic and Laminar Sublayer Resistance . . . . . . . . . . . . . 103

3a.5 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

4 Observations of Glyoxal and Formaldehyde as Metrics for the Anthro-

pogenic Impact on Rural Photochemistry 114

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

4.2 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

4.2.1 Site Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

4.2.2 Gly and HCHO Measurements . . . . . . . . . . . . . . . . . . . . . . 117

4.2.3 Other Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

vii

4.3 Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

4.3.1 BEARPEX 2009 16-17 July, 2009: Mammoth Fire Incident (MFI) . . 119

4.3.2 BEACHON-ROCS 18 August 2010 (BN1) . . . . . . . . . . . . . . . . 121

4.3.3 BEACHON-ROCS 19 August 2010 (BN2) . . . . . . . . . . . . . . . . 123

4.3.4 BEACHON-ROCS 14 August 2010 (BN3) . . . . . . . . . . . . . . . . 123

4.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

4.4.1 Gly:HCHO Ratios from Anthropogenic and Biogenic VOC Oxidation:

Surface and Satellite Values . . . . . . . . . . . . . . . . . . . . . . . . 125

4.4.2 Anthropogenic Influence on BVOC Oxidation via NO . . . . . . . . . 128

4.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

4.6 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130

4.7 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

5 Conclusions 156

5.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156

5.2 Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

5.2.1 VOC oxidation chemistry in a plume . . . . . . . . . . . . . . . . . . . 159

5.2.2 Measurements of HCHO direct emission from trees and ground litter/soil160

5.2.3 Long-term investigations of Gly:HCHO ratios and alkylperoxy radical

fate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

5.3 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

viii

List of Figures

1.1 Correlations of aerosol concentrations with increased mortality rate in six U.S.

cities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

1.2 Formation of tropospheric ozone (O3) through the cycling of HOx and NOx. . 19

1.3 Schematic of methane (CH4) oxidation showing the production and destruc-

tion of HCHO. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2.1 Theoretical principle of LIF. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

2.2 Schematic of optical setup for FILIF system. . . . . . . . . . . . . . . . . . . 40

2.3 Schematic of gas handling system for HCHO FILIF instrument. . . . . . . . . 41

2.4 Field data example of FILIF data acquisition scheme. . . . . . . . . . . . . . 42

2.5 Overlay of the broad FILIF excitation spectrum using the prototype laser

with the 50 Torr absorption cross-sections reported by Co et al. (2005). . . . 43

2.6 Laser control voltage (proportional to laser wavelength) and fluorescence sig-

nal from the wavelength reference cell as laser is dithered in wavelength be-

tween the online and offline positions at 40 Hz. . . . . . . . . . . . . . . . . . 44

2.7 Photodiode voltage vs. laser power measured before the detection axis during

a field photodiode calibration. . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

2.8 Sensitivity analyses for the FILIF electronic gating parameters. . . . . . . . . 46

2.9 Sensitivity analysis of signal/noise of FILIF instrument vs. cell pressure. . . . 47

ix

2.10 Sensitivity analysis of the signal due to changing purge/main flow ratios. . . . 48

2.11 Humidity sensitivity of the HCHO FILIF instrument. . . . . . . . . . . . . . 49

2.12 PFA vs. PTFE ambient inlet comparison tests during BEARPEX 2009. . . . 50

2.13 Long vs. short PTFE inlet tests during BEARPEX 2009. . . . . . . . . . . . 51

2.14 Zeroing tests for the FILIF instrument during BEARPEX 2009. . . . . . . . . 52

2.15 Inlet comparison of four zeroing tests for the FILIF instrument during CalNex-

SJV. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

2.16 Long vs. short PFA inlet tests during BEACHON-ROCS 2010. . . . . . . . . 54

2.17 Heated vs. unheated PFA inlet tests during BEACHON-ROCS 2010. . . . . . 55

3.1 Lag time vs. correlation plot for vertical wind speed (w) with both HCHO

and virtual temperature (Tv). . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

3.2 Average cospectra of HCHO and virtual temperature with vertical wind speed. 96

3.3 Averaged ogives and weighted cospectra over entire campaign. . . . . . . . . . 97

3.4 Diurnal profiles of HCHO flux and concentrations over entire campaign. . . . 98

3.5 Diurnal medians of contributions to HCHO flux in the base case model. . . . 99

3.6 Comparison of model results with measured HCHO fluxes. . . . . . . . . . . . 100

3a.1 Average cospectra of HCHO and virtual temperature with vertical wind speed

during three half-hour periods. . . . . . . . . . . . . . . . . . . . . . . . . . . 108

3a.2 Time series of HCHO flux over entire flux measurement period. . . . . . . . . 109

3a.3 Temperature and PAR dependence of HCHO flux during BEACHON-ROCS. 110

4.1 Diurnal profiles of RGF, HCHO, Gly, and wind direction during BEACHON-

ROCS 2010 and BEARPEX 2009. . . . . . . . . . . . . . . . . . . . . . . . . 140

4.2 One hour bin averaged Gly, HCHO, and RGF during BEARPEX 2009. . . . . 141

4.3 One hour bin averaged Gly, HCHO, and RGF during BEACHON-ROCS 2010. 142

x

4.4 Median diurnal profiles of MBO+Isoprene, monoterpenes, and OH reactivity

during BEACHON-ROCS and BEARPEX 2009. . . . . . . . . . . . . . . . . 143

4.5 Diurnal median profiles of RGF, HCHO, Gly, and wind direction during

BEACHON-ROCS 2010. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144

4.6 Gly, HCHO, RGF, other tracer species, and meteorological data during the

two days of the MFI. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

4.7 Four hour forward HYSPLIT trajectories for 16 July, 2009 originating at the

MFI site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

4.8 Four hour forward HYSPLIT trajectories for 17 July, 2009 originating at the

MFI site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

4.9 Gly, HCHO, RGF, other tracer species, and meteorological data during event

BN1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148

4.10 Gly, HCHO, RGF, other tracer species, and meteorological data during event

BN2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

4.11 Gly, HCHO, RGF, other tracer species, and meteorological data during event

BN3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150

4.12 A closer view of sharp changes in Gly, HCHO, and RGF during BN3. . . . . . 151

4.13 Comparison of m/z 95 with Gly, HCHO, and RGF during BN1 and BN2. . . . 152

4.14 RGF ranges during campaigns presented in this work and the literature. . . . 153

4.15 Examination of RO2 fate and its relation to HCHO, Gly, and RGF on 24 Au-

gust during BEACHON-ROCS. . . . . . . . . . . . . . . . . . . . . . . . . . . 154

xi

List of Tables

2.1 Summary of laser gating parameter analyses. . . . . . . . . . . . . . . . . . . 56

3a.1 Comparison of detection limits and time resolution of HCHO measurement

techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

3a.2 Chemical production and loss rates and yields for zero-dimensional box model.

All rate constants have units of cm3 molec−1 s−1 unless otherwise specified. . 112

3a.3 Noon model case results in µg m−2 hr−1 by species. . . . . . . . . . . . . . . 113

4.1 Percent increases for Gly, HCHO, RGF, and other species for each transport

event. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

5.1 Optimum parameters determined for HCHO FILIF and fiber laser. . . . . . . 165

xii

FORMALDEHYDE AS A PROBE OF RURAL

VOLATILE ORGANIC COMPOUND OXIDATION

Joshua P. DiGangi

Under the supervision of Assistant Professor Frank N. Keutsch

At the University of Wisconsin–Madison

Formaldehyde (HCHO), one of the most common organic species present in the atmo-

sphere, is produced via the atmospheric oxidation of volatile organic compounds (VOCs).

Byproducts of VOC oxidation include tropospheric ozone and secondary organic aerosol,

both of which are correlated with increased incidences of cardiac/respiratory disease and

mortality. Accurate models are crucial to predicting these species’ behavior and require an

accurate understanding of the mechanisms of VOC oxidation. Oxidized VOCs (OVOCs),

produced as intermediates in VOC oxidation, can illustrate the chemical pathways involved.

Measurement/model comparisons of VOC oxidation in rural environments (e.g. forests) have

typically reported poor agreement, complicated by the advection of polluted anthropogenic

air. Forest photochemical measurements have also implied the emission of unmeasured

biogenic VOCs (BVOCs), specifically terpenes. As HCHO is an OVOC produced in the oxi-

dation of nearly all VOCs, comparisons of measured and modeled HCHO concentrations can

illustrate the accuracy with which models reproduce the overall VOC oxidation chemistry.

To address these issues, a new HCHO detection technique was developed, Fiber Laser-

Induced Fluorescence (FILIF). FILIF is capable of faster sampling rates (10 Hz) than previ-

ously reported with comparable limits of detection to literature methods (∼25 pptv in 1 s).

Using FILIF, HCHO concentration gradients and vertical fluxes were measured in a conif-

erous forest canopy during BEACHON-ROCS 2010. Ground litter/soil and tree emissions

xiii

were also quantified with chamber experiments. A zero-dimensional box model of HCHO

mass balance in the canopy largely underestimated (×6) the vertical flux, attributable to ei-

ther higher direct emissions than predicted by chamber measurements and/or missing VOCs

with solar-driven emission profiles (unlike most terpenes). Satellite measurements of RGF,

or the ratio of HCHO with glyoxal (another common OVOC), are used by global models to

estimate the VOC mixture in certain areas. Measurements of RGF during both BEARPEX

2009 and BEACHON-ROCS 2010 showed increases due to biomass burning and fresh/strong

anthropogenic influence while showing no change for old/weak anthropogenic influence. The

RGF trend from urban to rural environments based on ground measurements was opposite

of that observed by satellites. A solution to this discrepancy is vital to improving global

models.

1

Chapter 1

Introduction1

1.1 Trace Species and Air Quality

Our atmosphere consists of roughly 78% nitrogen gas, 20% oxygen gas, 1% argon gas,

as much as 1% water vapor, and less than 0.05% other gases and airborne particles. While

oxygen gas (O2) is incredibly important to human life on Earth, that small fraction of other

material often profoundly affects our everyday lives. In particular, two of these, ozone and

small particulate matter, can have a negative impact on the quality of air we breathe. Air

quality is a general concept which is related to how hazardous the air in a given area is to

animal and plant life. It is strongly dependent on the amount of these trace materials. If the

air quality is “poor”, this denotes that there is an increased health risk to people, animals,

and plants due to these materials. In order to attempt to control and analyze these effects,

we first must understand the effects that these species may have on animal and plant life

and second accurately predict the concentration of these species in the atmosphere.

1This introduction will be part of a series compiled by the Wisconsin Initiative for Science Literacy (WISL)to “promote literacy in science, mathematics and technology among the general public.” It is intended to“explain [this] scholarly research and its significance to a wider audience that includes family members,friends, civic groups, newspaper reporters, state legislators, and members of the U.S. Congress.”

2

1.1.1 Harmful effects of ozone and aerosol

Ozone is a trace gas naturally present in our atmosphere. Most people have heard of

ozone in the context of the “ozone layer” high in the atmosphere which protects us from solar

ultraviolet radiation. Ozone also exists in the troposphere, the layer of the atmosphere closest

to the surface, in which we live, though in much smaller concentrations. While beneficial

higher in the atmosphere, ozone near the surface can be quite harmful to plant and animal

life. Increased ozone concentrations have been associated with increased incidences (about

20%) of cardiac-related emergency room visits (Stieb et al., 2000). Additionally, as much as

4% of respiratory-related ER visits have been attributed to tropospheric ozone, comparable

to that of many common allergens (Stieb et al., 2000). A more recent study reported that

increased ozone concentrations can lead to a more than 300% increase respiratory-related

death in urban areas compared to rural/remote areas (Jerrett et al., 2009). Ozone has

also been shown to be detrimental to plant life, by damaging sensitive tissue and stimulating

stress reactions by the plants (Mauzerall and Wang, 2001). This can affect humans indirectly

by decreasing crop yields in areas with increased concentrations of ozone.

Airborne particulate matter, also known as aerosol, also can contribute to poor air

quality. Aerosols are small particles, either liquid or solid, that are suspended in air. They

are emitted and/or formed by various processes, such as volcanic eruption, wind sweeping

up dust, vehicle emissions, and chemical reactions which occur in the atmosphere. Many

studies have shown a high agreement between increased concentrations of aerosol and in-

creased mortality rates due to respiratory-related illnesses (Ostro, 1993; Laden et al., 2000).

Additionally, increased aerosol concentrations were found to correlate with increased ER

cases of cardiac dysrhythmia (irregular heart beat) (Stieb et al., 2000). Figure 1.1 shows the

results of a highly-cited case study of six small U.S. cities (Dockery et al., 1993). This study

reported a linear trend where increased aerosol concentrations corresponded to increased

3

mortality.

1.1.2 Quantifying poor air quality and mitigation

To raise awareness of current air quality conditions, the U.S. Environmental Protection

Agency reports a quantity called the Air Quality Index (AQI). The AQI acts similar to a

weather report, with a scale of 0 - 500, which denotes the level of risk that the air in

your area poses to your immediate health (http://www.airnow.gov/index.cfm?action=

aqibasics.aqi). Higher AQI indices indicate higher risks to your health. Also similar to

weather reports, the EPA forecasts future AQI for the following day. By predicting future

air quality, state and local government officials can dictate policy to try to avoid or minimize

the negative effects of the poor air quality. For example, when a high AQI is predicted for

a particular day, many cities will declare an “Ozone Awareness Day”, or something similar.

As a result, public transportation may operate at reduced or no cost to encourage its use,

people may be asked not to mow the grass, or relief shelters may be set up for groups at risk

without adequate shelter or protection.

While relief shelters help to reduce the consequences of the poor air quality, reducing

vehicle and lawnmower emissions focuses on lowering the intensity of the event by reducing

the sources. Computer models containing the meteorology and chemistry of the atmosphere

can help predict which approaches are both useful and cost-effective. The processes that

determine air quality are incredibly complex and interconnected. For example, ozone pro-

duction in the atmosphere is highly dependent on the particular trace species present in the

air. Under some conditions, decreasing emissions results in less ozone production while un-

der other conditions, decreasing emissions results in more ozone production (Finlayson-Pitts

and Pitts Jr., 2000). The effectiveness of these models depends on a thorough understanding

of the chemical and physical processes controlling the production of ozone and aerosols, such

4

as the oxidation of volatile organic compounds.

1.2 Volatile Organic Compounds

The chemical production in the atmosphere of both ozone and certain aerosols is tied

to a group of trace gases called volatile organic compounds, or VOCs. A VOC is any organic

(carbon-based) chemical that exists as a vapor in the atmosphere. The most common VOC in

the atmosphere is methane. Methane is emitted from various sources, ranging from livestock

emission to oil mining/refining to swamps and wetlands. Annually, roughly 600 Tg, or 650

million tons, of methane are emitted into the atmosphere globally (Seinfeld and Pandis,

1998). Isoprene is another common VOC and has larger emissions than any other non-

methane VOC, roughly 500 Tg, or 550 million tons, annually (Finlayson-Pitts and Pitts Jr.,

2000). Isoprene is a biogenic VOC, or BVOC, which specifies it is from natural sources. In

this case, isoprene is primarily emitted by leafy (deciduous) trees and plants. Other common

VOCs include aromatic species (benzene, toluene), small terpenes (plant resin), and other

hydrocarbons (propane, hexane) (Seinfeld and Pandis, 1998).

1.2.1 Emission and Processing

Emission of VOCs typically falls into one of two categories: anthropogenic, or the re-

sult of human activity, and biogenic, or the result of natural activity. Anthropogenic VOCs,

or AVOCs, are emitted by factories, refineries, vehicles, agriculture, livestock, or any other

forms directly attributable to humans. Biogenic VOCs, or BVOCs, are predominantly emit-

ted from wetlands and non-agricultural plant life (forests, jungles, natural plains, oceans).

Once in the atmosphere, all VOCs are oxidized in a light-driven process which cou-

ples two families of very reactive, or radical, atmospheric compounds called NOx and HOx

(Fig. 1.2). NOx consists of two chemical species: nitrogen oxide (NO) radical and nitrogen

dioxide (NO2) radical. Tropospheric ozone is produced by the conversion of NO2 to NO

5

by ultraviolet (UV) light from the sun. NO can then react with ozone to once again form

NO2. The presence of HOx, which consists of the chemical species hydroxyl (OH) radical

and hydroperoxyl (HO2) radical, results in another pathway to convert NO to NO2. HO2

can react with NO to form OH and NO2. OH is then converted back to HO2 by reaction

with VOCs. Since there is no net loss of HOx or NOx in this process, it can be described

simply as in Eqn. 1.1:

V OClight→ Products + Ozone (1.1)

Additionally, as NOx and HOx are not destroyed in this process, a small amount of these

radical species can create a large amount of ozone.

However, HOx and NOx can be destroyed through reactions between themselves.

Higher concentrations of HOx and NOx increases the rate by which they are destroyed.

For example, at very high concentrations of NOx, such as in heavily polluted environments,

the speed by which NO reacts with other radical species may be faster than the rate of reac-

tion of NO with HO2. If these very high NOx concentrations are decreased, ozone production

can actually increase as there is less destruction of these reactive species (Finlayson-Pitts

and Pitts Jr., 2000). Complexities, such as these, in the chemical pathways of the atmo-

sphere make regulation difficult, so it is important to characterize them properly in order to

accurately predict the effects of any mitigation strategy.

1.2.2 Oxidized Volatile Organic Compounds

The products created in Eqn. 1.1 are more specifically oxidized VOCs, or OVOCs.

These are essentially VOCs which have more oxygen attached to them. As each type of

VOC reacts differently, oxidation of each type of VOC can make different OVOCs. OVOCs

are usually relatively stable, meaning they can exist in the atmosphere for longer than a few

minutes and have higher boiling temperatures than VOCs. Once created in the atmosphere,

6

OVOCs can be lost in two ways. The first is to react just like a VOC to make different

OVOCs and additional ozone. This process can continue until carbon monoxide is formed.

Secondly, as OVOCs have higher boiling points, they may no longer be volatile enough

to stay in the atmosphere and can condense onto a surface. For example, if the OVOC

encounters a small particle in the atmosphere, like an aerosol, it can stick to it (Pankow,

1994a,b; Odum et al., 1996). This is how a particular form of aerosol, secondary organic

aerosol (SOA), is formed: by many gas molecules sticking together. SOA is usually a major

portion of the total amount of aerosol and can contribute 18-70% of the mass of all aerosols,

depending on the region (Zhang et al., 2007; Jimenez et al., 2009).

1.2.3 Models and VOC oxidation

Models of air quality contain mechanisms of VOC oxidation in order to account for

this production of ozone and SOA. VOC emissions for the model are estimated from satellite,

aircraft, and/or ground measurements. These models then calculate what they expect for

OVOC concentrations based on those VOC concentrations, the amount of solar radiation,

concentrations of HOx and NOx, as well as other factors. We can compare the OVOC

concentrations predicted by the models with measurements of these OVOCs. This allows us

to gauge how accurate the model is for those given conditions. By identifying which OVOCs

do not match with models, we can determine what parts of a model still need improvement.

A great deal of effort has been put into validating, or checking, the accuracy of these

models. This has been performed using both controlled conditions (Lee et al., 2006a,b;

Carrasco et al., 2007; Galloway et al., 2011b) and real world conditions present in ambient

air (Choi et al., 2010; Huisman et al., 2011). Overall, models and measurements of OVOC

concentrations can be fairly consistent when there are high concentrations of NOx, such as

in urban areas (since anthropogenic activity can emit large quantities of NOx). However, in

rural or remote areas where NOx is low, the consistency between models and measurements

7

breaks down, even with respect to concentrations of HOx (Hofzumahaus et al., 2009; Paulot

et al., 2009; Peeters et al., 2009). In order to improve these models, we must further develop

accurate ways to detect OVOCs in the atmosphere. This will grant additional points of

reference for model outputs.

1.3 Formaldehyde in the Atmosphere

The simplest aldehyde, formaldehyde (HCHO) is formed in the oxidation of nearly all

VOCs. Because of this, it is one of the most common OVOCs. As with most trace gases,

HCHO concentrations are quite small: 0.02-0.2 parts per billion by volume, or ppbv, high

in the troposphere (Fried et al., 2008a,b), 1-20 ppbv in rural areas near the surface (Apel

et al., 1998; Lee et al., 1998; Muller et al., 2006; Choi et al., 2010; Galloway et al., 2011a),

and as much as 45 ppbv in urban areas (Dasgupta et al., 2005). Due to its prevalence in the

atmosphere as a central OVOC, measurements of HCHO are of vital importance to validate

any model of atmospheric chemistry.

1.3.1 Production and destruction

Figure 1.3 shows a schematic of HCHO production and destruction in the atmosphere

from one of the simplest cases: the oxidation of methane. Tropospheric methane oxidation

starts with reaction with OH, which removes a hydrogen atom to make methyl radical (CH3)

and water (H2O). CH3 quickly reacts with (O2) in the air to form CH3O2, which is one of a

family of compounds called alkylperoxy radicals (generally represented as RO2). While RO2

molecules can react in various ways (see Chap. 4 for a brief discussion), one of the primary

ways involves converting NO to NO2. This results in the formation of CH3O, which reacts

quickly with O2 to form HO2 and HCHO. There are two additional points of interest about

this production method for HCHO. The first is that there is no net change in the amount

of HOx, only a conversion from the OH to HO2 form. Again, this results in increased ozone

8

production from the conversion of NO2, which was produced during this process, to NO.

The lower part of Fig. 1.3 shows three pathways through which HCHO can be destroyed

in the atmosphere, two initiated by sunlight and one by OH. The right pathway is the most

likely during the daytime (about 50%), and directly results in the production of carbon

monoxide (CO) and hydrogen gas (Seinfeld and Pandis, 1998). This pathway is largely

uninteresting in terms of atmospheric chemistry, as no HOx, NOx, or OVOCs are involved.

The left pathway is the second most likely in the daytime (about 39%). Light energizes the

HCHO molecule, after which it reacts with O2 to form HO2 and HCO, which also quickly

reacts with O2 to form CO and a second HO2 (Seinfeld and Pandis, 1998). As a result, this

pathway creates two HOx molecules, increasing the oxidizing capacity of the air. The third

pathway is the least likely during the daytime (about 11%). OH reacts with HCHO to make

water and HCO, which again quickly reacts with O2 to form CO and HO2 (Seinfeld and

Pandis, 1998).

There is an additional pathway of atmospheric HCHO loss that is not pictured in

Fig. 1.3, in which HCHO condenses onto a surface, called dry deposition (see Chap. 3.4

for more details). Contributions from dry deposition are highly dependent on the surface

conditions and aerosol concentrations, as it depends on the amount of surface area. At night,

this last pathway usually is the most significant near to the surface as no sunlight is present

to enable the other three.

Overall, atmospheric HCHO destruction results in a net increase in HOx, making

HCHO one of the largest sources of HOx. As HOx is needed to oxidize VOCs in the first

place in order to make HCHO, this creates positive feedback for the production of HOx

in the atmosphere. By monitoring concentrations of HCHO, we can measure the amount

of HOx added to the atmosphere by this pathway. As HOx is a major component in the

production of ozone due to VOC oxidation, HCHO concentrations are important to be able

9

to constrain HOx production in a model.

1.3.2 Importance and Challenges

As HCHO is formed in the oxidation of nearly all VOCs and OVOCs, concentrations

of HCHO can give us information about the total amount of VOCs present in a particular

parcel of air. If ambient measurements of HCHO concentrations do not match concentra-

tions predicted by computer models, then it shows that the models are still missing some

aspect of VOC oxidation. This makes measurements of HCHO critical to validate models of

atmospheric chemistry.

Measurements of HCHO do face significant challenges, however. Concentrations of

HCHO are quite small and are difficult to detect accurately and selectively (in other words,

detecting only HCHO). In order to detect concentrations of HCHO at typical atmospheric

levels, it is necessary to use sophisticated, usually expensive, equipment. In addition, the use

of HCHO to investigate VOC oxidation can be complicated by direct emissions of HCHO

from many sources. These include rural sources, such as trees (Kesselmeier et al., 1997;

Villanueva-Fierro et al., 2004) and soil or forest ground litter (see Chap. 3), as well as urban

sources, such as industrial processes and vehicles (Anderson et al., 1996; Kean et al., 2001;

Garcia et al., 2006; Reyes et al., 2006; Lei et al., 2009). While unimportant far above the

emission sources, these direct emissions can be significant near the surface. For HCHO to be

useful as a VOC oxidation tracer, it is necessary to know when these emissions are significant

and, if so, how to account for them.

1.4 Summary

The consequences of poor air quality to life on Earth are broad and include effects

such as increased respiratory/cardiac illness, premature death, and crop failure which can

result in famine. In order to understand and reduce poor air quality, it is vital to understand

10

the processes involved in its production. The major contributors to poor air quality include

the production of tropospheric ozone and aerosols. These are strongly tied to the oxidation

of volatile organic compounds, or VOCs. Oxidized VOCs, or OVOCs, are produced as inter-

mediates in this oxidation. They can either continue reacting in the atmosphere, resulting

in further production of ozone, or can stick to particles to form a particular form of aerosol,

secondary organic aerosol. HCHO is a key OVOC for two reasons. HCHO is created in the

oxidation of nearly all VOCs, making it an ideal marker of VOC oxidation. HCHO is also a

significant source of HOx, a family of compounds involved in VOC oxidation.

In the work presented in this thesis, HCHO is used as a way of quantifying the amount

of VOC oxidation occurring in a given volume of air. Chapter 2 discusses some of the specific

challenges with regard to HCHO measurement and the FILIF technique that was developed

to address these challenges. Chapter 3 discusses using the FILIF technique for HCHO

detection in combination with a technique called eddy covariance. The combination of these

techniques enabled the measurement of the amount of HCHO moving out of a forest canopy

which provided insight into the chemistry inside the forest canopy. Chapter 4 discusses the

combination of HCHO measurements with measurements of a different OVOC, glyoxal, to

be able to estimate changes in the types of VOCs reacting in a volume of air.

1.5 Bibliography

Anderson, L. G., Lanning, J. A., Barrell, R., Miyagishima, J., Jones, R. H., and Wolfe, P.:

Sources and sinks of formaldehyde and acetaldehyde: An analysis of Denver’s ambient

concentration data, Atmospheric Environment, 30, 2113–2123, 1996.

Apel, E. C., Calvert, J. G., Riemer, D., Pos, W., Zika, R., Kleindienst, T. E., Lonneman,

W. A., Fung, K., Fujita, E., Shepson, P. B., Starn, T. K., and Roberts, P. T.: Measure-

ments comparison of oxygenated volatile organic compounds at a rural site during the

11

1995 SOS Nashville Intensive, Journal of Geophysical Research, 103, 22 295–22 316, 1998.

Carrasco, N., Doussin, J. F., O’Connor, M., Wenger, J. C., Picquet-Varrault, B., Durand-

Jolibois, R., and Carlier, P.: Simulation chamber studies of the atmospheric oxidation

of 2-methyl-3-buten-2-ol: Reaction with hydroxyl radicals and ozone under a variety of

conditions, Journal of Atmospheric Chemistry, 56, 33–55, 2007.

Choi, W., Faloona, I. C., Bouvier-Brown, N. C., McKay, M., Goldstein, A. H., Mao, J.,

Brune, W. H., LaFranchi, B. W., Cohen, R. C., Wolfe, G. M., Thornton, J. A., Son-

nenfroh, D. M., and Millet, D. B.: Observations of elevated formaldehyde over a forest

canopy suggest missing sources from rapid oxidation of arboreal hydrocarbons, Atmo-

spheric Chemistry and Physics, 10, 8761–8781, 2010.

Dasgupta, P. K., Li, J., Zhang, G., Luke, W. T., McClenny, W. A., Stutz, J., and Fried,

A.: Summertime Ambient Formaldehyde in Five U.S. Metropolitan Areas: Nashville,

Atlanta, Houston, Philadelphia, and Tampa, Environmental Science and Technology, 39,

4767–4783, 2005.

Dockery, D. W., Pope, C. A., Xu, X., Spengler, J. D., Ware, J. H., Fay, M. E., Ferris, B. G.,

and Speizer, F. E.: An Association between Air Pollution and Mortality in Six U.S. Cities,

New England Journal of Medicine, 329, 1753–1759, 1993.

Finlayson-Pitts, B. and Pitts Jr., J.: Chemistry of the Upper and Lower Atmosphere: The-

ory, Experiments, and Applications, Academic Press, San Diego, 2000.

Fried, A., Olson, J. R., Walega, J. G., Crawford, J. H., Chen, G., Weibring, P., Richter, D.,

Roller, C., Tittel, F., Porter, M., Fuelberg, H., Halland, J., Bertram, T. H., Cohen, R. C.,

Pickering, K., Heikes, B. G., Snow, J. A., Shen, H., O’Sullivan, D. W., Brune, W. H., Ren,

X., Blake, D. R., Blake, N., Sachse, G., Diskin, G. S., Podolske, J., Vay, S. A., Shetter,

12

R. E., Hall, S. R., Anderson, B. E., Thornhill, L., Clarke, A. D., McNaughton, C. S.,

Singh, H. B., Avery, M. A., Huey, G., Kim, S., and Millet, D. B.: Role of convection in

redistributing formaldehyde to the upper troposphere over North America and the North

Atlantic during the summer 2004 INTEX campaign, Journal of Geophysical Research,

113, D17 306, 2008a.

Fried, A., Walega, J. G., Olson, J. R., Crawford, J. H., Chen, G., Weibring, P., Richter, D.,

Roller, C., Tittel, F. K., Heikes, B. G., Snow, J. A., Shen, H., O’Sullivan, D. W., Porter,

M., Fuelberg, H., Halland, J., and Millet, D. B.: Formaldehyde over North America

and the North Atlantic during the summer 2004 INTEX campaign: Methods, observed

distributions, and measurement-model comparisons, Journal of Geophysical Research, 113,

D10 302, 2008b.

Galloway, M. M., DiGangi, J. P., Hottle, J. R., Huisman, A. J., Mielke, L. H., Alaghmand,

M., Shepson, P. B., Weremijewicz, J., Klavon, H., McNeal, F. M., Carroll, M. A., Griffith,

S., Hansen, R. F., Stevens, P. S., Bertman, S. B., and Keutsch, F. N.: Observations

and modeling of formaldehyde at the PROPHET mixed hardwood forest site in 2008,

Atmospheric Environment, accepted, 2011a.

Galloway, M. M., Huisman, A. J., Yee, L. D., Chan, A. W. H., Loza, C. L., Seinfeld, J. H.,

and Keutsch, F. N.: Yields of oxidized volatile organic compounds during the OH radical

initiated oxidation of isoprene, methyl vinyl ketone, and methacrolein under highNOx

conditions, Atmospheric Chemistry and Physics Discussions, 11, 10 693–10 720, 2011b.

Garcia, A. R., Volkamer, R., Molina, L. T., Molina, M. J., Samuelson, J., Mellqvist, J.,

Galle, B., Herndon, S. C., and Kolb, C. E.: Separation of emitted and photochemical

formaldehyde in Mexico City using a statistical analysis and a new pair of gas-phase

tracers, Atmospheric Chemistry and Physics, 6, 4545–4557, 2006.

13

Hofzumahaus, A., Rohrer, F., Lu, K., Bohn, B., Brauers, T., Chang, C., Fuchs, H., Holland,

F., Kita, K., Kondo, Y., Li, X., Lou, S., Shao, M., Zeng, L., Wahner, A., and Zhang, Y.:

Amplified Trace Gas Removal in the Troposphere, Science, 324, 1702–1704, 2009.

Huisman, A. J., Hottle, J. R., Galloway, M. M., DiGangi, J. P., Coens, K. L., Choi, W.,

Faloona, I. C., Gilman, J. B., Kuster, W. C., de Gouw, J., Bouvier-Brown, N. C., Gold-

stein, A. H., LaFranchi, B. W., Cohen, R. C., Wolfe, G. M., Thornton, J. A., Docherty,

K. S., Farmer, D. K., Cubison, M. J., Jimenez, J. L., Mao, J., Brune, W. H., and Keutsch,

F. N.: Photochemical modeling of glyoxal at a rural site: observations and analysis from

BEARPEX 2007, Atmospheric Chemistry and Physics, 11, 8883–8897, 2011.

Jerrett, M., Burnett, R. T., Pope, C. A., Ito, K., Thurston, G., Krewski, D., Shi, Y., Calle,

E., and Thun, M.: Long-Term Ozone Exposure and Mortality, New England Journal of

Medicine, 360, 1085–1095, 2009.

Jimenez, J. L., Canagaratna, M. R., Donahue, N. M., Prevot, A. S. H., Zhang, Q., Kroll,

J. H., DeCarlo, P. F., Allan, J. D., Coe, H., Ng, N. L., Aiken, A. C., Docherty, K. S.,

Ulbrich, I. M., Grieshop, A. P., Robinson, A. L., Duplissy, J., Smith, J. D., Wilson, K. R.,

Lanz, V. A., Hueglin, C., Sun, Y. L., Tian, J., Laaksonen, A., Raatikainen, T., Rautiainen,

J., Vaattovaara, P., Ehn, M., Kulmala, M., Tomlinson, J. M., Collins, D. R., Cubison,

M. J., Dunlea, E. J., Huffman, J. A., Onasch, T. B., Alfarra, M. R., Williams, P. I., Bower,

K., Kondo, Y., Schneider, J., Drewnick, F., Borrmann, S., Weimer, S., Demerjian, K.,

Salcedo, D., Cottrell, L., Griffin, R., Takami, A., Miyoshi, T., Hatakeyama, S., Shimono,

A., Sun, J. Y., Zhang, Y. M., Dzepina, K., Kimmel, J. R., Sueper, D., Jayne, J. T.,

Herndon, S. C., Trimborn, A. M., Williams, L. R., Wood, E. C., Middlebrook, A. M.,

Kolb, C. E., Baltensperger, U., and Worsnop, D. R.: Evolution of Organic Aerosols in the

Atmosphere, Science, 326, 1525–1529, 2009.

14

Kean, A. J., Grosjean, E., Grosjean, D., and Harley, R. A.: On-Road Measurement of Car-

bonyls in California Light-Duty Vehicle Emissions, Environmental Science and Technology,

35, 4198–4204, 2001.

Kesselmeier, J., Bode, K., Hofmann, U., Muller, H., Schafer, L., Wolf, A., Ciccioli, P.,

Brancaleoni, E., Cecinato, A., Frattoni, M., Foster, P., Ferrari, C., Jacob, V., Fugit, J. L.,

Dutaur, L., Simon, V., and Torres, L.: Emission of short chained organic acids, aldehydes

and monoterpenes from Quercus ilex L. and Pinus pinea L. in relation to physiological

activities, carbon budget and emission algorithms, Atmospheric Environment, 31, 119–

133, 1997.

Laden, F., Neas, L. M., Dockery, D. W., and Schwartz, J.: Association of fine particulate

matter from different sources with daily mortality in six US cities, Environmental Health

Perspectives, 108, 941–947, 2000.

Lee, A., Goldstein, A. H., Keywood, M. D., Gao, S., Varutbangkul, V., Bahreini, R., Ng,

N. L., Flagan, R. C., and Seinfeld, J. H.: Gas-phase products and secondary aerosol

yields from the ozonolysis of ten different terpenes, Journal of Geophysical Research-

Atmospheres, 111, D07 302, 2006a.

Lee, A., Goldstein, A. H., Kroll, J. H., Ng, N. L., Varutbangkul, V., Flagan, R. C., and

Seinfeld, J. H.: Gas-phase products and secondary aerosol yields from the photooxidation

of 16 different terpenes, Journal of Geophysical Research-Atmospheres, 111, 2006b.

Lee, Y. N., Zhou, X., Kleinman, L. I., Nunnermacker, L. J., Springston, S. R., Daum, P. H.,

Newman, L., Keigley, W. G., Holdren, M. W., Spicer, C. W., Young, V., Fu, B., Parrish,

D. D., Holloway, J., Williams, J., Roberts, J. M., Ryerson, T. B., and Fehsenfeld, F. C.:

Atmospheric chemistry and distribution of formaldehyde and several multioxygenated car-

15

bonyl compounds during the 1995 Nashville Middle Tennessee Ozone Study, Journal of

Geophysical Research-Atmospheres, 103, 22 449–22 462, 1998.

Lei, W., Zavala, M., de Foy, B., Volkamer, R., Molina, M. J., and Molina, L. T.: Impact of

primary formaldehyde on air pollution in the Mexico City Metropolitan Area, Atmospheric

Chemistry and Physics, 9, 2607–2618, 2009.

Mauzerall, D. L. and Wang, X. P.: Protecting agricultural crops from the effects of tropo-

spheric ozone exposure: Reconciling science and standard setting in the United States,

Europe, and Asia, Annual Review of Energy and the Environment, 26, 237–268, 2001.

Muller, K., Haferkorn, S., Grabmer, W., Wisthaler, A., Hansel, A., Kreuzwieser, J., Co-

jocariu, C., Rennenberg, H., and Herrmann, H.: Biogenic carbonyl compounds within

and above a coniferous forest in Germany, Atmospheric Environment, 40, Supplement 1,

81–91, 2006.

Odum, J. R., Hoffmann, T., Bowman, F., Collins, D., Flagan, R. C., and Seinfeld, J. H.:

Gas/Particle Partitioning and Secondary Organic Aerosol Yields, Environmental Science

and Technology, 30, 2580–2585, 1996.

Ostro, B.: The Association of Air-Pollution and Mortality - Examining the Case for Infer-

ence, Archives of Environmental Health, 48, 336–342, 1993.

Pankow, J.: An absorption model of gas/particle partitioning of organic compounds in the

atmosphere, Atmospheric Environment, 28, 185–188, 1994a.

Pankow, J.: An absorption model of the gas/aerosol partitioning involved in the formation

of secondary organic aerosol, Atmospheric Environment, 28, 189–193, 1994b.

Paulot, F., Crounse, J. D., Kjaergaard, H. G., Krten, A., St. Clair, J. M., Seinfeld, J. H.,

16

and Wennberg, P. O.: Unexpected Epoxide Formation in the Gas-Phase Photooxidation

of Isoprene, Science, 325, 730–733, 2009.

Peeters, J., Nguyen, T. L., and Vereecken, L.: HOx radical regeneration in the oxidation of

isoprene, Physical Chemistry Chemical Physics, 11, 5935–5939, 2009.

Reyes, F., Grutter, M., Jazcilevich, A., and Gonzlez-Oropeza, R.: Tecnical Note: Analysis

of non-regulated vehicular emissions by extractive FTIR spectrometry: tests on a hybrid

car in Mexico City, Atmospheric Chemistry and Physics, 6, 5339–5346, 2006.

Seinfeld, J. and Pandis, S.: Atmospheric Chemistry and Physics: From Air Pollution to

Climate Change, Wiley, New York, 2nd edn., 1998.

Stieb, D. M., Beveridge, R. C., Brook, J. R., Smith-Doiron, M. A. R. C., Burnett, R. T.,

Dales, R. E., Beaulieu, S., Judek, S., and Mamedov, A.: Air pollution, aeroallergens and

cardiorespiratory emergency department visits in Saint John, Canada, Journal of Exposure

Analysis and Environmental Epidemiology, 10, 461–477, 2000.

Villanueva-Fierro, I., Popp, C. J., and Martin, R. S.: Biogenic emissions and ambient con-

centrations of hydrocarbons, carbonyl compounds and organic acids from ponderosa pine

and cottonwood trees at rural and forested sites in Central New Mexico, Atmospheric

Environment, 38, 249–260, 2004.

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., and Worsnop, D. R.: Ubiquity and domi-

17

nance of oxygenated species in organic aerosols in anthropogenically-influenced Northern

Hemisphere midlatitudes, Geophysical Research Letters, 34, L13 801, 2007.

18

Fig. 1.1.— Reproduced from Dockery et al. (1993). Correlations of aerosol concentrations

with increased mortality rate in six U.S. cities: Watertown, Massachusetts (W); Harriman,

Tennessee (H); St. Louis, Missouri (L); Steubenville, Ohio (S); Portage, Wisconsin (P); and

Topeka, Kansas (T). The y-axis represents the ratio of deaths in a given city normalized by

the deaths in Portage, WI.

19

Fig. 1.2.— Formation of tropospheric ozone (O3) through the cycling of HOx and NOx.

VOC oxidation results in an additional way to convert NO to NO2, which leads to more O3.

20

Fig. 1.3.— Schematic of methane (CH4) oxidation showing the production and destruction

of HCHO. The vertical pathway (reaction with NO) production pathway is dominant under

urban conditions. The reaction of the RO2 radical (CH3O2) with HO2 becomes more signif-

icant under remote conditions and leads to different oxidation products, such as peroxides

(ROOH). These peroxides can still form HCHO, but a different amount than formed by the

NO pathway.

21

Chapter 2

Design and Characterization of the Fiber

Laser-Induced Fluorescence Technique for HCHO

Detection

2.1 Introduction

Formaldehyde (HCHO) is an organic compound ubiquitous in the Earth’s atmosphere.

HCHO is formed by the oxidation of nearly all volatile organic compounds (VOCs). Un-

derstanding the processing of these VOCs is of vital importance to modern science due to

their contributions to the production of secondary organic aerosols and tropospheric ozone.

As HCHO is produced in the oxidation of nearly all VOCs, it is an ideal tracer for overall

VOC oxidation. Therefore, measurements of HCHO are crucial for constraining any model

of atmospheric chemistry.

Instrumentation for field detection of HCHO has both rigid and demanding require-

ments. The instrumentation must be rugged in order to operate in various environments.

Such environments may include remote, harsh conditions during ground measurements or

changes in motion and temperature during flight measurements. The instrument must have

minimal power, volume, and weight requirements, so as to facilitate field/aircraft installation

22

and operation. The sensitivity of the instrument must be sufficient in order to detect and

quantify the minute ambient concentrations of HCHO. The instrument must be selective so

that the measurement is reliable. Finally, the instrument must have high time-resolution,

so as to measure fast fluctuations in HCHO concentrations. For example, measuring HCHO

flux via eddy covariance requires 10 Hz measurements, while sufficient sampling of an airmass

from a fast aircraft requires accurate measurements at no less than 1 Hz.

Many contemporary methods have been reported for HCHO detection, summarized in

Table 3a.1. Direct infrared absorption spectroscopy (e.g. QCLAS, TDLAS, DFGLAS) is the

most common family of methods for research grade instrumentation (Weibring et al., 2007;

McManus et al., 2010). This involves some form of tunable laser which has been coupled

into a Herriot-type multipass cell. The laser, typically wavelength modulated, is swept over

an infrared absorption feature at a high rate (10 Hz-1 kHz). This technique is quite sensitive

(3σ limit of detection (LOD): ∼100 pptv in 1 s), while the use of rotational absorption

features result in excellent selectivity. However, errors in spectral background fitting can

lead to measurement interferences, while the cell volume inherent to a longpath Herriot-

type cell typically limits the time resolution of this technique (> 1 s). Proton Transfer

Reaction-Mass Spectrometry (PTR-MS) is another common method for HCHO detection,

but suffers from significant water interference (Vlasenko et al., 2010; Warneke et al., 2011).

Differential Optical Absorption Spectroscopy (DOAS) can detect a wide variety of molecules,

including HCHO, but suffers from a delocalized detection volume and spectral background

interferences (Wisthaler et al., 2008). Derivatization methods (e.g. DNPH or Hantzsch) are

common, inexpensive HCHO detection techniques commonly used in regulatory studies, but

suffer from low time resolution (∼1-15 min) and significant condensed-phase interferences

(Warneke et al., 2011).

Laser-induced fluorescence (LIF) is a sensitive, selective technique used to monitor

23

certain atmospheric species (Thornton et al., 1999; Perkins et al., 2001). Ambient detection

of HCHO using LIF was first reported by Mohlmann (1985), but suffered from bulky, unreli-

able laser systems that were impractical for field operation. Hottle et al. (2009) reported an

HCHO LIF instrument using a Ti-Sapphire laser with relatively low limit of detection (3σ

LOD: ∼50 pptv in 1 s). This instrument was also successfully deployed for ground HCHO

measurements during the PROPHET 2008 campaign (Galloway et al., 2011). However, while

the Ti-Sapphire laser represented a major improvement in HCHO LIF lasers, the Ti-Sapphire

LIF method still suffered from significant power, space, and weight requirements as well as

limited time resolution (>1 s).

In this chapter, I will summarize the design and characterization of a new HCHO LIF

detection technique called Fiber Laser-Induced Fluorescence (FILIF). I will also discuss the

general instrumental operation, sensitivity optimization experiments, interference tests, and

inlet characterization studies.

2.2 Instrument Description

2.2.1 Technique

FILIF uses the pervasive physical chemistry technique of laser-induced fluorescence

(LIF). In this case, a rotational feature in the A-X 401 absorption band of HCHO is excited

via a laser pulse at 353.15 nm (σHCHO: 6× 10−19 cm2 molec−1) (Co et al., 2005). An excited

HCHO molecule can then photolyze, collisionally-deexcite, or emit a photon (fluoresce) over

the range 390-510 nm (Fig. 2.1a). The probability of fluorescence is called the quantum

yield of fluorescence (Φf ):

Φf =kf

kf + kp + kq × PM(2.1)

where kf is the rate constant for the fluorescence (2× 105 s−1) (Yeung and Moore, 1973),

kp is the rate constant for photolysis (2.53× 106 s−1 at ∼353 nm) (Moortgat and Warneck,

24

1979), kq is the rate constant for collisional quenching (1.7× 104 Torr−1 s−1) (Moortgat and

Warneck, 1979), and PM is the pressure of the bath gas. The 401 vibronic absorption feature,

the lowest in energy of the A-X band, was chosen to minimize the contribution of kp (Φp =

27% at 353 nm) (Mohlmann, 1985). At 110 Torr and 353 nm, the quantum yield of HCHO

fluorescence is ∼4%.

Figure 2.1b shows the fluorescence process with respect to time. At t = 0, a pulse from

the excitation laser is introduced to the air sample. As fluorescence is a non-instantaneous

process (τf ' 150 ns), there is a time delay between the excitation photon and the emission

of a fluorescent photon. As a result, the fluorescence signal persists after the excitation

pulse has ended. By employing an empirically-determined electronic gate to isolate only

the fluorescence emitted after the end of the laser pulse, the fluorescence is observed with a

minimal background signal from laser scatter. This gating technique improves sensitivity by

minimizing background while improving selectivity by isolating the fluorescence signal from

other, faster processes (e.g. Raman scattering).

2.2.2 Fiber Laser

The key to the FILIF technique is a narrow-bandwidth UV pulsed fiber laser, recently

developed by NovaWave Technologies (TFL series). In contrast to Ti:Sapphire and other

forms of bulk lasers, a fiber laser uses a doped fiber optic cable as a laser gain medium. In

this case, a Yb-glass fiber is used to amplify a seed laser to ∼1 W of 1059 nm light, which

is then tripled with two sum frequency generation crystals (SHG and THG) to achieve 15-

20 mW of 353 nm light. The laser wavelength can be controlled by an analog voltage, which

controls the current through the seed diode. Fiber optic cable has a much higher surface

area to volume ratio than a bulk laser crystal, which allows a fiber laser to be convectively,

as opposed to liquid, cooled. This eliminates the size, weight, and power requirements of

a liquid chiller. Additionally, as a fiber laser is entirely fiber-coupled, there is no need to

25

realign the laser or clean optics, which results in a rugged, turn-key laser system. Finally,

the laser is seeded by a narrow-bandwidth distributed feedback diode laser, which results in

a narrow-bandwidth UV beam (see Sect. 2.3.1.1).

2.2.3 Optical Setup

The optical setup for the FILIF system is shown in Fig. 2.2. Using a telescope, the

laser is focused at the entrance of the detection axis, a White-type multipass cell. A beam-

splitter transmits ∼10% of the power, which is continuously monitored for changes using

a photodiode (UDT, 20-00-017). The front mirror of the cell (Spectrum Thin Films) has

separate cutouts for the ingoing and outgoing beams. Two identical mirrors (CVI Laser,

Y3-0537-0-0.25CC) on the opposite end of the cell align the beam to form a flat, White-type

multipass pattern (White, 1942). Laser scatter and stray light were minimized through the

use of slotted light baffles. Optimum HCHO sensitivity with this system was observed with

32 passes. Fluorescence was collected at a right angle to the laser axis with a collimating lens

(CVI Laser, BICX-38.1-100.0/30.9-UV-355-532), filtered using a 390 nm dielectric longpass

filter (Barr Assoc.), then focused onto a single-photon counting photomultiplier tube (Sen-

sTech, P25PC). A light trap opposing the photomultiplier tube minimizes background laser

scatter that makes it to the detector. The outgoing beam power was similarly monitored

using a beamsplitter and photodiode. As a change in alignment would result in a change

in power throughput through the cell, by comparing the two photodiode signals, alignment

changes in the multipass cell can be detected. The remainder of the beam was directed into

an LIF wavelength reference cell. This cell was prepared by being evacuated, then exposed

to the vapor from a 37 wt. % HCHO solution (Sigma-Aldrich, F1635-500ML). Fluorescence

in the reference cell was filtered with a 390 nm dielectric longpass filter, and then monitored

with a current-mode photomultiplier tube (Hamamatsu, H5783).

26

2.2.4 Gas System

Figure 2.3 shows the gas handling system for the FILIF instrument. Ambient air was

sampled through one of four separate inlets used for gradient studies, controlled electronically

by four three-way 1/2” PTFE Teflon solenoid valves (Teqcom Ind., M863W2DTS-HT). Gas

flow rate was controlled by an electronically-actuated (Hanbay, MPA02i) 1/2” PFA Teflon

needle valve (Swagelok, PFA-4RPS8). Air flow was produced by drawing vacuum in the

cell with a vacuum pump (BOC Edwards, XDS10). The flow of air through the cell was

perpendicular to both the detection and excitation axes. It was necessary only to overturn

the volume the beam pattern occupied (<1 cm thick along the air flow axes), which was

much less than the overall cell volume, resulting in fast time resolution. Dead volumes in the

cell (e.g. optics ports, light trap) were purged with Ultra Zero air (Airgas, Inc.) controlled

by a mass flow controller (MKS Instruments, M100B Series).

Zeroing and calibration experiments were performed using a bulk flow also of Ultra

Zero air (Airgas, Inc.), the flow of which was quantified using a mass flow meter (MKS

Instruments, 558A Series). For calibrations, known quantities of HCHO were supplied by a

HCHO permeation tube (VICI Metronics, 100-044-2300-U45). A HCHO permeation tube

consists of paraformaldehyde surrounded by a semi-permeable membrane. At a constant

temperature, this tube emits a constant mass of HCHO. To control temperature and flow, the

HCHO permeation tube was placed in a permeation oven (VICI Metronics, Dynacalibrator

Model 120), which held the permeation tube at a constant temperature of 85°C and a

constant flow of ∼1 standard liter per minute (SLM). This system was cross-calibrated with

FTIR absorption to verify its emission rate (see Chap 3a.1). The calibration flow was mixed

with the main bulk flow using a mass flow controller (MKS Instruments, 1179A Series).

Another mass flow controller monitored the excess flow in order to determine the calibrant

concentration. A solenoid shutoff valve isolated the calibrant flow from the main flow when

27

calibrations were not being performed.

2.2.5 Data Acquisition Principle

All data (including photodiode voltages, cell pressure, mass flow controller/meter

flows, etc.) was acquired using a National Instruments PCI-6229 data acquisition card

in conjunction with homebuilt routing electronics in a manner similar to that reported by

Huisman (2010). Data was read and stored using a single-board Mini-ITX computer (Jetway

JNC91-330-LF).

Figure 2.4 shows a typical data acquisition cycle for concentration and eddy covariance

measurements. At the start of the cycle, cell pressure was optimized using the actuated

needle valve and a PID algorithm. The laser wavelength was dithered between directly on

top of the absorption feature (online), then directly next to the absorption feature at a

wavelength of low absorption (offline). Periodically (every 5-30 min), the laser was scanned

over the absorption feature and then blocked to monitor dark signal. This cycle was repeated

continuously. After each scan, the online wavelength was automatically corrected for laser

wavelength drift based on the highest point in the reference cell scan. The offline wavelength

was always held to be at ∼0.0025 nm higher wavelength (35 mV lower laser control voltage)

than the online wavelength. HCHO concentrations were then calculated by the following

equation:

[HCHO] = kcal ×

(Conline

P onlinelaser

−Coffline

P offlinelaser

)× Rwave (2.2)

where kcal is the empirically-determined calibration factor, P onlinelaser and P offline

laser are the aver-

age laser power during respectively online and offline measurement, Conline and Coffline are

the number of detected fluorescence photons during respectively online and offline measure-

ment, and Rwave is the wavelength drift ratio correction. Rwave was calculated by taking

the ratio of the reference cell signal during a given online measurement and the interpolated

28

value of what that signal should be based on the previous and following scans. This corrected

for wavelength drift between scans, and was typically less than 10%.

2.3 Instrument Characterization

To characterize the instrument, it was necessary to optimize many empirical parame-

ters. Originally, only a prototype, table-top laser was available for a proof-of-concept anal-

ysis. The goal of the initial studies with the prototype laser was to characterize the laser so

as to determine the appropriate settings for the final laser. Due to equipment limitations,

the prototype laser was limited to wavelengths of ∼354 nm. While this limited excitation to

features with lower absorption cross-section, this did not affect the validity of the parameters

discussed here. Further studies were performed using the field-capable laser, which was ca-

pable of lasing at 353 nm. These determined the final parameters used in field measurements

and calibration.

2.3.1 Initial Studies - Prototype Laser

2.3.1.1 Laser Wavelength and Bandwidth

Figure 2.5 show the results of the initial experiments to determine the laser wavelength

and bandwidth. Low pressure (50 Torr) absorption cross-sections (Co et al., 2005) were

compared to the excitation spectrum to find similar line spacing and relative intensities; a

match was found at ∼354.26 nm. The laser bandwidth was determined from these spectra

by comparing the spectral linewidth of the excitation lines with the spectral linewidth of the

absorption cross-section lines, which were found to be indistinguishable. Therefore, using

the spectral resolution of the absorption cross-section data, an upper limit for the laser

bandwidth of 0.01 cm−1, or 300 MHz, was determined.

29

2.3.1.2 Laser Repetition Rate

The prototype fiber laser was capable of operation at various repetition rates and pulse

widths, as driven by a function generator, while maintaining similar average laser power. To

maintain this consistent power, it was also necessary to maintain a ratio of 100 between the

pulse spacing and pulse width. Table 2.1 displays the results of testing to determine the

optimal repetition rate/pulse width pairing. As the electronic gating parameters changed

with the pulse width, it was necessary to optimize them for each repetition rate. Higher

repetition rates reduced the likelihood of detector saturation, while lower repetition rates

ensured that the fluorescence from each pulse had decayed before the arrival of the next

pulse. The optimum balance between these effects was found to be at 300 kHz repetition

rate and 30 ns pulse width.

2.3.1.3 Laser Wavelength Rise Time

In order to measure at 10 Hz, it was necessary to prove that the laser was capable

of tuning on a time scale which is small compared to the 100 ms measurements. This

was accomplished by monitoring both the wavelength control voltage for the laser and the

HCHO reference cell signals on an oscilloscope. The laser wavelength was then dithered at

40 Hz between the online and offline positions. Figure 2.6 shows the results of this analysis.

When switching from the offline to online position, the laser wavelength did not change

immediately, but equilibrated over the course of ∼10 ms. When switching from the online to

offline position, the laser wavelength dropped sharply with the control voltage, equilibrating

within 1-2 ms. This effect is likely explained by different difficulties in warming vs. cooling

the seed diode laser, which controls the laser wavelength. The result of this analysis was the

addition of a 10 ms delay after changing the laser wavelength before any data acquisition.

30

2.3.2 Studies with Non-Prototype (Field Ready) Laser

2.3.2.1 Photodiode Calibration

The sensitivity of the FILIF instrument is directly proportional to the amount of

laser power entering the detection axis. Therefore, it was necessary to accurately monitor

this laser power and normalize for any changes. This was measured by splitting a portion

of the ingoing beam with a beamsplitter and monitoring it with a photodiode. However,

the photodiode response was not linear with power. To determine the power dependence,

a thermopile power meter (Coherent, PS10Q) was placed between the beamsplitter and

the detection axis entrance, and multiple laser powers were measured by both photodiode

and power meter (Fig. 2.7). The photodiode was found to have a 2nd order polynomial

dependence on laser power, allowing for correction during data analysis. This photodiode

calibration was observed to deviate over time (on the order of weeks), and it was necessary

to periodically repeat this calibration.

2.3.2.2 Photon Gating Optimization

In order to optimize FILIF LOD, it was necessary to empirically determine the opti-

mum delay and gate width (Fig. 2.1b). Delays too short result in too high of a background,

leading to higher noise, while delays too long miss a large amount of the fluorescence signal,

resulting in low signal. Gate widths too short also miss a large amount of fluorescence sig-

nal, while gate widths too long detect too much dark noise interference. To determine the

optimum gate, the signal/noise ratio of a constant HCHO concentration was measured using

various delay and gate widths (Fig. 2.8). Optimum signal/noise was found with a delay of

325 ns, and a gate width of 212.5 ns.

31

2.3.2.3 Pressure Optimization

Detection cell pressure is another parameter affecting the sensitivity of the FILIF in-

strument. Higher pressures provide a higher number density to increase number of excited

HCHO but result in a lower Φf by increased quenching. To determine the optimum pres-

sure, the signal/noise ratio of a constant HCHO concentration was measured with varying

cell pressures (Fig. 2.9). The optimum pressure was found to be ∼110 Torr, with indistin-

guishable signal/noise over the range 90-130 Torr.

2.3.2.4 Purge Flow Sensitivity

The purge gas is necessary to isolate the optics from both ambient dust and chemicals

that may damage the dielectric coating, particularly when combined the UV light from the

laser. Additionally, the purge gas also ensures that dead volumes inside the cell do not retain

HCHO from previous air samples. Typically, it is easy to correct the small dilution that this

causes in the signal. However when purge flow rates begin to approach the bulk flow rates, it

is no longer a trivial correction. To characterize this effect, the HCHO signal was monitored

for a constant purge flow (0.5 SLM) and varying main flows (Fig. 2.10). The HCHO signal

deviated significantly from linearity at main flows below ∼4 SLM. As a result, calibration

factors for the instrument cannot be considered linear for main/purge flow ratios below ∼8.

If these lower ratios are used, a separate calibration must be performed and will only be

consistent for that particular flow ratio.

2.3.2.5 Humidity Studies

Many HCHO measurement techniques suffer from water interferences. HCHO LIF

could potentially suffer from a water interference, as water likely has a different quenching

rate than N2 or O2, affecting the fluorescence quantum yield. This could result in a low bias

to measurements at high humidity. To test for this, concentrations of HCHO were measured

32

at different relative humidities while at constant temperature (∼19° C). Other than the dry

air measurements, all measurements were corrected for background contamination from the

water used to humidify the air. Figure 2.11 shows the results of this analysis. There was

no observed significant deviation in HCHO concentrations at different humidities, which

indicates that there is no water interference for HCHO LIF under atmospherically-relevant

conditions.

2.4 Inlet Studies

As the FILIF system is intrinsically a closed-path system, it is necessary to bring the

air sample from the collection area to the detection axis. This was accomplished through

lengths of tubing, referred to as inlets. However, this tubing introduces the possibility

of artifacts, through HCHO deposition and emission inside the inlets. Wert et al. (2002)

reported that both stainless steel and PFA Teflon showed no HCHO emission/deposition

effects for short inlets (< 2 m). For tower gradient experiments, the inlets tend to be longer.

For this longer length of tubing, further studies to confirm the absence of artifacts were

required.

2.4.1 BEARPEX 2009

For long inlets in tower gradient experiments, stainless steel tubing was impractical

to both transport and install. Additionally, PFA Teflon tubing was financially demanding

for four 100 ft inlets. As a compromise, three inlets of less expensive PTFE Teflon and one

inlet of PFA Teflon were used for gradient measurements during the BEARPEX campaign.

During the first part of the campaign, all four inlets were collocated to analyze any artifacts.

Figure 2.12a shows a comparison between PTFE and PFA inlets, while Fig. 2.12b shows the

results of a comparison of two identical PTFE inlets. Both comparisons showed discrepancies

on the order of ∼10%, but mostly appeared randomly distributed. The PFA vs. PTFE

33

inlet suggested a slight enhancement of [HCHO] through the PFA inlet, but this was on

average not statistically significant. At ∼2: 00, there was an event where the PTFE [HCHO]

was significantly higher than the PFA [HCHO]. However, this corresponded to a sudden

discrepancy in the PTFE vs. PTFE [HCHO] as well, suggesting this is due to atmospheric

variability. Additional measurements were performed by collocating a ∼20 ft PTFE inlet

with a ∼100 ft PTFE inlet (Fig. 2.13). The goal of these measurements was to determine

the effect of scaling the inlet length. The results indicated no greater error between the

long and short PTFE inlets than existed for the identical long inlets, suggesting no length

dependence on inlet artifacts.

To test for potential HCHO inlet emission, zeroing tests were performed with high-

purity air (Ultra Zero air, Airgas, Inc.). A ground level inlet was set to constantly sample

ambient air. At the end of an inlet, a tube from a high-purity air tank was alternately added

and removed to the end of the inlet. Care was taken to not interfere with the detection

cell pressure or flow, which could have biased the results. Figure 2.14 shows the results

of this experiment. The measured HCHO concentration fell to statistically zero when the

high-purity air was sampled. However, a statistically insignificant positive bias of ∼0.3 ppbv

suggested that, should the instrumental sensitivity be higher, a significant inlet bias may be

observed.

2.4.2 CalNex-SJV 2010

During this campaign, a more comprehensive series of inlet experiments were per-

formed with a lower instrumental LOD. Four different inlet types were sampled: 1/8” ID

(1/4” OD) PTFE, 7/16” ID (1/2” OD) PTFE, 3/8” ID (1/2” OD) PTFE, and 3/8” ID

(1/2” OD) PFA. Zeroing experiments were performed with each inlet type in a manner sim-

ilar to that used during BEARPEX. In addition, these experiments were performed during

both the daytime and nighttime to observe any potential dependencies on solar radiation

34

or temperature. The results from this analysis are shown as Figure 2.15. All inlets were

observed to have statistically zero HCHO concentrations with a UZA flow during the night

measurements. The 1/8” ID PTFE and 3/8” ID PFA inlets also were observed to have

statistically zero HCHO concentrations during the day measurements. The 7/16” ID PTFE

measurements had a large offset of ∼2.2 ppbv HCHO during the day. The 3/8” ID PTFE

measurements showed a statistically insignificant offset of 0.13 ppbv HCHO during the day.

These results suggest that PTFE is not an acceptable inlet material for HCHO sampling

at high sensitivities, unless the residence time is short as with the 1/8” ID PTFE inlet.

Thin walled tubing (7/16” ID PTFE) contributed a large HCHO interference, though it is

possible that this particular product was defective. Finally, this study confirms that PFA is

an acceptable inlet material.

2.4.3 BEACHON-ROCS 2010

For the BEACHON-ROCS campaign, only 1/2” PFA tubing was used based on the

lessons of the CalNex-SJV inlet experiments. However, the BEACHON-ROCS site differed

from the previous sites in that it experienced significantly cooler nighttime temperatures

(¡10°C). This increased the likelihood of HCHO deposition inside of the inlets. In the first

test, a short inlet (∼50 ft) and long inlet (∼100 ft) were collocated near the ground for

∼2.5 days (Fig. 2.16). The differences between the inlets were randomly distributed around

zero and typically within 5%. This is insignificant in comparison to our calibration error

(∼30%), thus negligible, and likely attributable to atmospheric variability. Additionally, a

heating sleeve was added to a ∼100 ft inlet and collocated with an unheated ∼100 ft inlet

(Fig. 2.17). The differences between the heated and unheated inlets were also randomly dis-

tributed around zero, were typically within ∼10%, and showed no diurnal (i.e. temperature)

dependence.

35

2.5 Conclusions

In this chapter, I have summarized the design and characterization of a Fiber Laser-

Induced Fluorescence instrument for HCHO detection. The narrow-bandwidth fiber laser

was shown to be crucial to this instrument, as it provides a rugged, reliable laser source for

selective excitation of HCHO. The laser was found to have an upper limit of its bandwidth

of 0.01 cm−1, an optimum repetition rate/pulse width of 300 kHz and 30 ns for HCHO de-

tection, and a wavelength rise time of ∼10 ms. For optimum HCHO detection, an electronic

delay of 320 ns, an electronic gate width of 212.5 ns, and a cell pressure of 110±20 Torr

was used. Humidity was found to have no interference on measured HCHO concentrations.

PTFE tubing was found to contribute little to strong HCHO interference, depending on the

wall thickness and air residence time. PFA tubing was found to contribute no observable

HCHO interferences. There was no statistically significant difference between short and long

inlets of any material tested, and PFA tubing showed no statistically significant difference

between a heated or unheated inlet. These characterization studies allow the HCHO FILIF

instrument to be reliably deployed for field measurements. Its ruggedness, sensitivity, selec-

tivity, and high time-resolution make the FILIF instrument optimal for HCHO aircraft or

eddy covariance flux measurements.

2.6 Acknowledgements

The National Science Foundation (ATM 0852406) and NASA-SBIR Phase I and II

grants provided funding for the development of this instrument. I would also like to thank

John Hottle, Andrew Huisman, and Sam Henry for useful discussions and advice.

2.7 Bibliography

Co, D. T., Hanisco, T. F., Anderson, J. G., and Keutsch, F. N.: Rotationally resolved

absorption cross sections of formaldehyde in the 28100-28500 cm−1 (351-356 nm) spectral

36

region: Implications for in situ LIF measurements, Journal of Physical Chemistry A, 109,

10 675–10 682, 2005.

Galloway, M. M., DiGangi, J. P., Hottle, J. R., Huisman, A. J., Mielke, L. H., Alaghmand,

M., Shepson, P. B., Weremijewicz, J., Klavon, H., McNeal, F. M., Carroll, M. A., Griffith,

S., Hansen, R. F., Stevens, P. S., Bertman, S. B., and Keutsch, F. N.: Observations

and modeling of formaldehyde at the PROPHET mixed hardwood forest site in 2008,

Atmospheric Environment, accepted, 2011.

Hottle, J. R., Huisman, A. J., Digangi, J. P., Kammrath, A., Galloway, M. M., Coens,

K. L., and Keutsch, F. N.: A Laser Induced Fluorescence-Based Instrument for In-Situ

Measurements of Atmospheric Formaldehyde, Environmental Science and Technology, 43,

790–795, 2009.

Huisman, A.: Measurements and modeling of glyoxal: Insights into rural photochemistry

and secondary organic aerosol production, Ph.d., The University of Wisconsin - Madison,

2010.

McManus, J. B., Zahniser, M. S., Nelson, J. D. D., Shorter, J. H., Herndon, S., Wood, E.,

and Wehr, R.: Application of quantum cascade lasers to high-precision atmospheric trace

gas measurements, Optical Engineering, 49, 111 124–11, 2010.

Mohlmann, G. R.: Formaldehyde Detection in Air by Laser-Induced Fluorescence, Applied

Spectroscopy, 39, 98–101, 1985.

Moortgat, G. K. and Warneck, P.: CO and H2 quantum yields in the photodecomposition

of formaldehyde in air, The Journal of Chemical Physics, 70, 3639–3651, 1979.

Perkins, K. K., Hanisco, T. F., Cohen, R. C., Koch, L. C., Stimpfle, R. M., Voss, P. B.,

Bonne, G. P., Lanzendorf, E. J., Anderson, J. G., Wennberg, P. O., Gao, R. S., Del Negro,

37

L. A., Salawitch, R. J., McElroy, C. T., Hintsa, E. J., Loewenstein, M., and Bui, T. P.:

The NOx-HNO3 System in the Lower Stratosphere: Insights from In Situ Measurements

and Implications of the JHNO3-[OH] Relationship, Journal of Physical Chemistry A, 105,

1521–1534, 2001.

Thornton, J. A., Wooldridge, P. J., and Cohen, R. C.: Atmospheric NO2: In Situ Laser-

Induced Fluorescence Detection at Parts per Trillion Mixing Ratios, Analytical Chemistry,

72, 528–539, 1999.

Vlasenko, A., Macdonald, A. M., Sjostedt, S. J., and Abbatt, J. P. D.: Formaldehyde

measurements by Proton transfer reaction Mass Spectrometry (PTR-MS): correction for

humidity effects, Atmospheric Measurements and Technology, 3, 1055–1062, 2010.

Warneke, C., Veres, P., Holloway, J. S., Stutz, J., Tsai, C., Alvarez, S., Rappenglueck, B.,

Fehsenfeld, F. C., Graus, M., Gilman, J. B., and de Gouw, J. A.: Airborne formaldehyde

measurements using PTR-MS: calibration, humidity dependence, inter-comparison and

initial results, Atmospheric Measurements and Technology, 4, 2345–2358, 2011.

Weibring, P., Richter, D., Walega, J. G., and Fried, A.: First demonstration of a high

performance difference frequency spectrometer on airborne platforms, Optics Express, 15,

13 476–13 495, 2007.

Wert, B. P., Fried, A., Henry, B., and Cartier, S.: Evaluation of inlets used for the airborne

measurement of formaldehyde, Journal of Geophysical Research, 107, 4163, 2002.

White, J. U.: Long Optical Paths of Large Aperture, Journal of the Optical Society of

America, 32, 285–285, 1942.

Wisthaler, A., Apel, E. C., Bossmeyer, J., Hansel, A., Junkermann, W., Koppmann, R.,

Meier, R., Mller, K., Solomon, S. J., Steinbrecher, R., Tillmann, R., and Brauers, T.:

38

Technical Note: Intercomparison of formaldehyde measurements at the atmosphere simu-

lation chamber SAPHIR, Atmospheric Chemistry and Physics, 8, 2189–2200, 2008.

Yeung, E. S. and Moore, C. B.: Photochemistry of single vibronic levels of formaldehyde,

The Journal of Chemical Physics, 58, 3988–3998, 1973.

39

Fig. 2.1.— Theoretical principle of LIF. (a) Electronic structure schematic of LIF. (b)

Theoretical temporal plot of LIF, including electronic gate for signal/noise optimization.

40

Fig. 2.2.— Schematic of optical setup for FILIF system.

41

Fig. 2.3.— Schematic of gas handling system for HCHO FILIF instrument.

42

Fig. 2.4.— Field data example of FILIF data acquisition scheme. The raw WC counts denote

the signal from the photon-counting PMT on the detection axis. The laser power shown is

that measured by the photodiode measuring the throughput of the beamsplitter before the

detection axis. The laser position denotes the laser control voltage which is proportional to

the laser frequency.

43

Fig. 2.5.— Overlay of the broad FILIF excitation spectrum using the prototype laser with

the 50 Torr absorption cross-sections reported by Co et al. (2005).

44

Fig. 2.6.— Laser control voltage (proportional to laser wavelength) and fluorescence signal

from the wavelength reference cell as laser is dithered in wavelength between the online and

offline positions at 40 Hz.

45

Fig. 2.7.— Photodiode voltage vs. laser power measured before the detection axis during a

field photodiode calibration.

46

Fig. 2.8.— Sensitivity analyses for the FILIF electronic gating parameters. (a) Dependence

of signal/noise ratio on the gate delay with a constant gate width of 312.5 ns. (b) Dependence

of signal/noise ratio on the gate width with a constant gate delay of 325 ns.

47

Fig. 2.9.— Sensitivity analysis of signal/noise of FILIF instrument vs. cell pressure.

48

Fig. 2.10.— Sensitivity analysis of the signal due to changing purge/main flow ratios. Colors

denote the ratio of the main flow through the center of the detection axis to the purge flow,

while the purge flow remained constant at 0.5 SLM. The dotted line displays the expected

signal vs. [HCHO] line based on the calibration factor at high main flows.

49

Fig. 2.11.— Humidity sensitivity of the HCHO FILIF instrument. The measured [HCHO]

was background subtracted to account for the contribution of dissolved HCHO in the water

used to humidify the air. The dotted line represents a y=x line.

50

Fig. 2.12.— Ambient inlet comparison tests for (a) a 100’ 3/8” ID PFA inlet and a 100’

3/8” ID PTFE inlet and (b) two similar 100’ 3/8” ID PTFE inlets during BEARPEX 2009.

Y-axes denote the [HCHO] difference between the inlets normalized by the [HCHO]. Error

bars denote the 1σ measurement precision.

51

Fig. 2.13.— Ambient inlet comparison tests for a 20’ 3/8” ID PTFE inlet and a 100’ 3/8”

ID PTFE inlet during BEARPEX 2009. Y-axes denote the [HCHO] difference between the

inlets normalized by the [HCHO]. Error bars denote the 1σ measurement precision.

52

Fig. 2.14.— Zeroing tests for the FILIF instrument with a 100’ 3/8” PTFE inlet during

BEARPEX 2009. The inlet alternately sampled ambient and high-purity tank air. Error

bars denote the 1σ measurement precision.

53

Fig. 2.15.— Inlet comparison of zeroing tests for the FILIF instrument with four 100’

inlets of different materials during CalNex-SJV during (a) daytime and (b) nighttime. Each

inlet alternately sampled ambient and high-purity tank air. The HCHO signal is directly

proportional to [HCHO].

54

Fig. 2.16.— Ambient inlet comparison tests for a 50’ 3/8” ID PFA inlet and a 100’ 3/8” ID

PFA inlet during BEACHON-ROCS 2010. Y-axes denote the [HCHO] difference between

the inlets normalized by the [HCHO]. Error bars denote the 1σ measurement precision.

55

Fig. 2.17.— Ambient inlet comparison tests for a heated 100’ 3/8” ID PFA inlet and a

100’ 3/8” ID PFA inlet during BEACHON-ROCS 2010. Y-axes denote the [HCHO] differ-

ence between the inlets normalized by the [HCHO]. Error bars denote the 1σ measurement

precision.

56

Tab

le2.

1:S

um

mar

yof

lase

rga

tin

gp

aram

eter

anal

yse

s.

Pu

lse

Del

ay/G

ate

3σ

Exp

.R

epet

itio

nW

idth

Wid

thS

ign

alN

oise

S/N

[HC

HO

]L

OD

Rat

e(k

Hz)

(ns)

(ns)

( counts

mW×s

)(counts

mW×s

)(p

pbv)

(pp

bv

in1

s)

300

3090

0/31

2.5

104.

93.

232

.720

1.8

130

030

887.

5/31

2.5

134.

25.

026

.820

2.2

500

2588

7.5/

312.

597

.94.

123

.820

2.4

500

2587

5/31

2.5

146.

89.

914

.820

4.0

500

2588

7.5/

312.

511

4.0

3.9

29.2

313.2

500

2590

0/31

2.5

75.7

3.6

20.9

314.4

250

025

912.

5/31

2.5

52.5

3.4

15.6

315.9

300

3090

0/31

2.5

128.

33.

437

.431

2.5

300

3091

2.5/

312.

579

.03.

622

.131

4.2

200

5093

7.5/

312.

552

.62.

818

.727

4.3

330

030

912.

5/31

2.5

78.5

2.7

29.2

272.8

300

3092

5.5/

312.

569

.73.

023

.527

3.4

57

Chapter 3

First Direct Measurements of Formaldehyde Flux

via Eddy Covariance: Implications for Missing

In-Canopy Formaldehyde Sources1

3.1 Introduction

The oxidation of volatile organic compounds (VOCs) in the atmosphere occurs via the

HOx-NOx cycle, a photochemically-driven catalytic cycling of hydrogen oxide (OH + HO2)

and nitrogen oxide (NO + NO2) radicals. This process produces tropospheric ozone and

oxidized VOCs, the latter of which may condense to form secondary organic aerosol (SOA)

(Zhang et al., 2007; Jimenez et al., 2009). To accurately model both tropospheric ozone

and SOA, the processes involved in VOC oxidation must be characterized. Part of the

difficulty in understanding this cycle lies in the detection and quantification of all relevant

species of VOCs, particularly in forest environments. Multiple studies have reported a

significant discrepancy between measured and modeled OH concentrations and reactivities,

1Reprinted from: DiGangi, J. P., Boyle, E. S., Karl, T., Harley, P., Turnipseed, A., Kim, S., Cantrell,C., Maudlin III, R. L., Zheng, W., Flocke, F., Hall, S. R., Ullmann, K., Nakashima, Y., Paul, J. B., Wolfe,G. M., Desai, A. R., Kajii, Y., Guenther, A., Keutsch, F. N.: First direct measurements of formaldehydeflux via eddy covariance: implications for missing in-canopy formaldehyde sources, Atmos. Chem. Phys., 11,10565-10578, DOI: 10.5194/acp-11-10565-2011, 2011.

58

suggesting errors in our understanding of the emissions or processing of VOCs (Tan et al.,

2001; Di Carlo et al., 2004; Hofzumahaus et al., 2009; Lelieveld et al., 2008; Sinha et al., 2010;

Whalley et al., 2011). This discrepancy may be related to the fast in-canopy oxidation of

unmeasured biogenic VOCs (BVOCs), specifically terpenes (Di Carlo et al., 2004; Goldstein

et al., 2004; Holzinger et al., 2005). To confirm this, a method of determining the overall

VOC oxidation rate is needed.

Formaldehyde (HCHO) is both a significant participant in the cycling of HOx and

a major byproduct of the HOx-NOx cycle (Fried et al., 1997; Lee et al., 1998; Tan et al.,

2001). As a result, HCHO is an excellent tracer for overall VOC oxidation. Quantification

of HCHO production in forest environments could provide a valuable constraint for the

overall rate of VOC oxidation in this environment. There have been many reports of forest

HCHO mixing ratios (Munger et al., 1995; Slemr et al., 1996; Lee et al., 1998; Sumner

et al., 2001; Galloway et al., 2011), but a qualitative and quantitative understanding of in-

canopy HCHO production is still incomplete. One recent study (Choi et al., 2010) reported

a missing boundary layer HCHO production rate of as much as 1.6 ppbv h−1, nearly double

the calculated chemical production rate.

Measurements of HCHO vertical fluxes above and gradients throughout a forest canopy

may yield valuable insight into production and loss of HCHO inside the canopy. Gradient

measurements can give more detailed information about the sources and sinks in the canopy,

while vertical flux measurements are less influenced by advection, as the area sampled by

the flux is typically the area less than a kilometer upwind. HCHO fluxes have previously

been estimated based on flux-gradient calculations over polar icepack (Jacobi et al., 2002;

Hutterli et al., 2004), but there has been little work examining HCHO distribution in forest

canopies and no reported measurements of HCHO flux by eddy covariance (EC). Of the

many reported techniques to measure HCHO (Table 3a.1)(Weibring et al., 2007; Wisthaler

59

et al., 2008; Hottle et al., 2009; McManus et al., 2010), none have reported the capability

of performing the fast sampling needed for EC measurements with both the sensitivity

needed to quantify small perturbations in HCHO concentration and the selectivity inherent

to spectroscopic techniques.

In this work, we present HCHO gradients and EC flux observations using Fiber Laser-

Induced Fluorescence (FILIF), which has the high sensitivity and high time resolution

needed for EC measurements. Additionally, we discuss branch and soil enclosure experi-

ments performed to determine HCHO emission rates. To model HCHO flux, we present a

zero-dimensional box model used to apportion HCHO production and loss inside the canopy.

Finally, we discuss sensitivity studies with respect to both BVOC and direct HCHO emission

using the box model to ascertain their effect on measurement/model agreement.

3.2 Experimental

3.2.1 Field Campaign

All observations reported here were taken during the Bio-hydro-atmosphere interac-

tions of Energy, Aerosols, Carbon, H2O, Organics & Nitrogen - Rocky Mountain Organic

Carbon Study (BEACHON-ROCS) field campaign during 1 - 31 August, 2010 at Manitou

Experimental Forest (MEF, 39°06’02” N, 105°06’05” W, 2286 m), northwest of Colorado

Springs, CO. The site has been described in detail elsewhere (Kim et al., 2010). It is located

in a Central Rocky Mountains Ponderosa Pine (PPine) forest (canopy height: ∼18.5 m;

leaf area index (LAI) = 1.9) with minimal undergrowth, predominately clean air masses

transported from the southwest, and rare anthropogenic incursions.

60

3.2.2 Fiber Laser-Induced Fluorescence (FILIF) of HCHO

This technique is similar to that reported by Hottle et al. (2009), the primary differ-

ence being the laser, and will only be described briefly here. The 353 nm tunable, pulsed,

and narrow-bandwidth fiber laser (NovaWave Technologies, TFL Series) represents a signif-

icant improvement over previous field laser technology, as fiber lasers are inherently lighter,

smaller, and more stable than traditional lasers. The ∼10 mW laser was directed into a

32-pass White-type multipass cell, and the resulting HCHO fluorescence from 390 to 500 nm

was filtered using a 390 nm longpass filter then focused into a photomultiplier tube for detec-

tion. Laser power was monitored both before and after the multipass cell using photodiodes,

and a fraction (∼1 mW) of the outgoing beam was directed into a cell filled with concen-

trated gas-phase HCHO for wavelength reference. The separation between the multipass cell

mirrors was ∼25 cm. However, only ∼6 cm of each pass was through the 6 cm x 5 cm area

(cell depth: ∼6 cm) through which the ambient air was flowed perpendicular to the narrow

plane of the laser. The residence time of air in the cell was < 25 ms in the beam volume

(∼1 cm thick) at the ∼12 standard liters per minute (SLM) sampling flow rate. Remaining

volumes of the cell were purged using a zero air generator (AADCO 737-series) with a total

purge flow of 500 standard cubic centimeters per minute (SCCM) regulated by a mass flow

controller (MKS Instruments, M100B).

Measurements were performed by dithering the laser on and off a rovibronic absorption

line at 353.37 nm. The difference in fluorescence signal when the laser was centered on

these two positions was proportional to the HCHO concentration. Instrument calibrations

were performed weekly using a HCHO permeation tube (VICI Metronics, 100-044-2300-U45)

heated to 85 °C using a portable calibration gas generator (VICI Metronics, Model 120).

The output of the permeation tube device as characterized by Fourier Transform Infrared

(FTIR) spectroscopy was found to be 438±7 ng min−1; details on this calibration can be

61

found in the supplement. The calibration factor varied by less than 2.5% over the course of

the campaign. Field 3σ limits of detection were typically on the order of ∼300 pptv in 1 s,

with measurement accuracies of ∼20% limited by that of the permeation tube calibration.

Inlets for HCHO sampling with lengths of 30 to 45 m were located at heights of 25.1 m,

17.7 m, 8.5 m, and 1.6 m. Inlets consisted of ∼30 m 3/8” ID PFA Teflon tubing, short lengths

of which have been found to have a negligible effect on sampling (Wert et al., 2002). To test

for possible artifacts, both a 15 m and a 30 m inlet were collocated outside the instrument

trailer; resulting measurements agreed within 1.5%. Typically, ambient flow while sampling

through an inlet was ∼12 SLM. Inlets were continuously purged with ambient air at ∼3 SLM

when not in use. An additional scroll pump (Gast Manufacturing) with an average flow of

∼80 SLM was used to increase the flow rate of the 25.1 m inlet used for EC sampling to

reduce residence time and prevent laminar flow in the inlet. The 25.1 m inlet was placed

∼0.1 m below and ∼0.5 m upwind in the primary wind direction of the center of the sonic

anemometer (see Sect. 3.2.3).

Measurements were performed in an hourly cycle for 11 - 22 August. During this

period, HCHO was measured from the 25.1 m inlet for the first 35 min with online and

offline sampling times of 10 s and 1 s respectively at 10 Hz (for EC), following which was a

1.5 min diagnostic period. Then, each of the other three inlets was sampled sequentially with

online and offline sampling times of 20 s and 10 s respectively for 7 min, following each of

which was a 1.5 min diagnostic period. During 23 - 30 August, only EC measurements were

performed with 35 min collection periods and 1.5 min diagnostic periods. For this period,

as eddies with timescales on the order of 10 s contributed significantly to HCHO flux (see

Sect. 3.2.4.2), online and offline sampling times were changed to 290 s and 5 s respectively.

This change in sampling was to test for potential EC spectral interference, which was not

observed.

62

3.2.3 Other Measurements

Unless otherwise noted, all other measurements used a valve switching system which

changed sampling lines every 5 min and cycled through six 1/4” OD Teflon inlets mounted

at 25.1 m, 17.7 m, 12.0 m, 8.5 m, 5.0 m, and 1.6 m over a 30 min period. Flow rates of

∼3.5 SLM through the sampling lines resulted in delay times between 8 to 12 s, measured

by spiking a VOC pulse at each sampling inlet.

A Proton-Transfer-Reaction Mass Spectrometer (PTR-MS, Ionicon Analytik GmbH)

was used for gradient measurements of selected VOCs. The instrument is based on soft

chemical ionization using protonated water ions (H3O+) (Hansel et al., 1998; Lindinger

et al., 1998), and was operated at 2.3 mbar drift pressure and 540 V drift voltage and

calibrated using two multi-component ppmv VOC standards (Karl et al., 2009).

OH, HO2, and RO2 were measured using chemical ionization mass spectrometry

(CIMS) as described by Tanner et al. (1997) and Hornbrook et al. (2011). The CIMS

acquired measurements at ∼10 m from the tower at a height of 2.7 m with the inlet facing

perpendicular to the primary wind direction. During periods with OH concentrations below

the detection limit (5× 105 molec cm−3), OH concentration was assumed to be equal to half

the detection limit (2.5× 105 molec cm−3).

Downwelling NO2 photolysis (JNO2) was measured from the top of the 30 m chem-

istry tower with commercially-available filter radiometers (Meteorologie Consult GmbH) as

described by Junkermann et al. (1989) and Volz-Thomas et al. (1996). The filtered mea-

surement was converted to a photolysis rate by comparison with spectrally-resolved actinic

flux measurements. Total JNO2 was estimated by measurement of the ratio of upwelling to

downwelling JNO2 as measured from the tower on 10 August, 2010.

OH reactivity was measured using a laser-induced pump and probe technique (Sadanaga

et al., 2004) at∼20 m from the tower and a height of∼4 m with a 2 min sampling rate. Perox-

63

yacetyl nitrate (PAN) was measured via Thermal Decomposition-Chemical Ionization Mass

Spectrometry, as described by Zheng et al. (2011). Ozone concentrations were measured

using a Model 205 Dual Beam Ozone Monitor (2B Technologies, Inc.). NO concentrations

were measured using an Ecophysics CLD-88Y analyzer. NO2 concentrations were measured

using a Droplet Measurement Technologies Blue Light Converter. A LI-COR LI-7000 mea-

sured CO2 and H2O concentrations at 25.1 m. LI-COR LI-190 quantum sensors measured

photosynthetically active radiation (PAR) at 27.8 m and 1.8 m. Vaisala HMP35C probes

measured temperature and relative humidity at 25.3 m and 7.0 m. A Vaisala PTB101B

barometer measured barometric pressure. A sonic anemometer (Campbell Scientific, CSAT-

3) at 25.1 m measured the three-dimensional wind vector, as well as virtual temperature, at

10 Hz.

3.2.4 Eddy Covariance Measurements

Eddy covariance (EC) is a widely-used micrometeorological technique for direct mea-

surement of surface-atmosphere exchange and will be discussed here briefly; further informa-

tion is available elsewhere (Baldocchi et al., 1988; Lee et al., 2004). EC uses the covariance

between vertical fluctuations in wind speed, caused by atmospheric eddies, and fast varia-

tions in tracer concentration to extract the mass transport through the plane of measurement.

Quantitatively, the turbulent flux of a species at a single height, assuming horizontal and

vertical advection is negligible, is defined as:

FEC ≡ w′ · c′ = w · c − w · c (3.1)

where w is the vertical wind speed, c is the tracer concentration, and x′ is the instantaneous

deviation of x from the ensemble mean value (i.e. x′ = x − x). For this study, a sonic

anemometer measured vertical wind speed, while the HCHO FILIF instrument measured

tracer concentration. As eddies occur on a wide range of timescales, the averaging time to

64

calculate the ensemble mean and fluctuating quantities can vary depending on measurement

height (Berger et al., 2001). For this study, a sampling period of ∼32 min was chosen, the

validity of which will be discussed in Sect. 3.2.4.2)

3.2.4.1 Data Reduction

Three-dimensional 10 Hz wind speeds from the sonic anemometer were rotated using

the natural wind coordinate (Lee et al., 2004) for each 35 min flux period. Sampling periods

with a friction velocity (u∗) less than 0.2 m s−1 were neglected as rotation has been shown

to result in poor data quality at low wind speeds (Lee et al., 2004). Vertical rotation

angles (e.g. tilt angles) were typically ∼2±4°. Additionally, a delay exists between HCHO

concentration and wind speed due to the residence time of the HCHO sample in the inlet

tubing. A correction was determined empirically by calculating w′HCHO′ at different time

delays, or lags, to find the maximum in covariance, as shown in Fig. 3.1, which should be

roughly equal to the residence time in the inlet tubing (Lee et al., 2004). In this study, an

additional variable lag was present as the computers recording the sonic anemometer and

HCHO data were not synchronized. This resulted in a lag time that varied considerably

over the campaign. Therefore, it was necessary to divide the dataset into 4 sections with

different linear trends, depending on computer resynchronization time. A sampling period

was considered to have a “good” lag when the covariance was greater than 20 µg m−2 hr−1

and the u∗ was greater than 0.3 m s−1, and these points were used to calculate the linear

trends. All sampling periods were then assumed to have a lag according to these trends.

Lag times over the course of the campaign ranged from -9.4 to 1.1 s. Finally, the EC data

was tested for stationarity (Foken and Wichura, 1996) by dividing each 30 min sampling

period into ∼5 min periods. The average of the 5 min flux measurements for each sampling

period was compared to the 30 min flux measurement for that period. The period was

considered stationary if the fluxes agreed within 30% (Foken and Wichura, 1996). Non-

65

stationary periods were rejected as invalid and not included in the analysis, which resulted

in the removal of 48% of daytime and 60% of nighttime data.

3.2.4.2 Spectral Analysis

To determine the validity of the remaining flux data, the cospectra of the HCHO fluxes,

which may be thought of as the frequency-dependent covariance between the species, were

investigated in further detail. As the cospectrum over a single period was typically quite

variable, cospectra were averaged over multiple periods. Figure 3.2 shows the average cospec-

trum for HCHO and virtual temperature fluxes over daily periods during the campaign. The

linear regions of each cospectrum indicative of the inertial sublayer (f > ∼0.04 Hz) exhibited

a lower slope than the expected value determined from a -7/3 power law (Lee et al., 2004).

A similar effect was observed at Blodgett Forest (Farmer et al., 2006; Wolfe et al., 2009)

and attributed to wake-generated turbulence present in forest canopies (Kaimal and Finni-

gan, 1994). Figure 3.2 also demonstrates that while the overall covariance for a given time

period is typically positive (upward flux), there are frequencies corresponding to different

sized eddies which can result in negative covariance (downward flux). The frequencies of

these eddies are highly variable between different sampling periods, which makes it difficult

to determine a cause. However, this variability also suggests that these negative covariance

events are not likely an artifact of data collection, as this would likely result in consistent

negative fluxes at a given frequency over multiple periods. As the field mission averaged

cospectrum (Fig. 3a.1) closely resembles that of the virtual temperature flux, the negative

covariance events are believed to have been due to atmospheric variability. These negative

values may also be responsible for the faster drop-off of the normalized cospectrum relative

to that for the temperature flux (Fig. 3.3a).

Spectral attenuation may be observed either when a sampling period is too short to

sample low-frequency eddies, or when the sample rate is too slow to sample high-frequency

66

eddies. The frequency-weighted cospectrum (Fig. 3.3a) peaked at the frequencies contribut-

ing most to total flux. For the HCHO cospectrum, three peaks were observed, corresponding

to characteristic eddy timescales of ∼0.5, 2.5, and 8 min. The 0.5 min peak corresponds to

the peak in the temperature flux cospectrum as well as in the momentum flux cospectrum,

which likely indicates that this is the integral time scale for turbulent transport. The 2.5 min

peak is similar in timing to one observed in PAN cospectra during other forest campaigns

(Turnipseed et al., 2006; Wolfe et al., 2009), which was on the same timescale of observed

canopy sweep events (Holzinger et al., 2005). The 8 min peak can likely be attributed to

a similar phenomenon. At frequencies greater than 0.04 Hz, the cospectrum appears to

decrease more quickly than the temperature flux cospectrum. The cause of this is not un-

derstood, but similar high frequency loss has been observed in PAN cospectra (Turnipseed

et al., 2006). If this loss is a result of spectral attenuation, it implies that fluxes are typically

underestimated. By comparing the difference between the integrated areas of the virtual

temperature and HCHO weighted cospectra, this underestimate would be on the order of

∼12%, which was included in the error analysis as a systematic low bias.

The cospectral cumulative distribution function, or ogive (Fig. 3.3b), is the cumulative

contribution to the flux as a function of frequency. The HCHO ogive is significantly shifted

towards lower frequencies compared to the temperature ogive indicating greater contribution

to the flux by lower frequency eddies than for temperature. The lack of an asymptote toward

the low frequency end of the ogive implies that the sampling period may not have been

sufficient to capture all of the low frequency eddies. However, analysis during the last half

of the campaign with longer sampling periods resulted in no significant gain in covariance

with periods greater than 30 min.

Other potential errors in the flux measurements are discussed in the supplement (see

Sect. 3a.2). By summing the systematic errors (response, sensor, dampening, attenuation),

67

then propagating with this the indeterminate errors (instrument noise, lag time, calibration),

we calculated the total error in the HCHO flux to be typically ∼38%.

3.3 Data and Observations

3.3.1 Gradient and Flux Profile

Daily HCHO fluxes typically showed a symmetric diurnal efflux centered at noon.

The median diurnal profile of HCHO flux is shown in Fig. 3.4a, while the full flux time

series is shown in Fig. 3a.2 (note that positive values denote an upward flux while neg-

ative values denote a downward flux). Median noontime fluxes were ∼80 µg m−2 hr−1

(∼24 pptv m s−1) with maxima as high as ∼170 µg m−2 hr−1 (∼50 pptv m s−1). For com-

parison, 2-methyl-3-buten-2-ol (MBO) fluxes have been observed in PPine forests on the

order of 8 to 9 mg m−2 hr−1 (Baker et al., 1999; Schade et al., 2000). HCHO fluxes were

also observed to have a significant dependence on both temperature and PAR (Fig. 3a.3).

Measured HCHO fluxes correspond to a median noontime net HCHO production rate of

∼3.2 ppbv hr−1 below the measurement height of 25.1 m. However, the net HCHO pro-

duction rate into the boundary layer from these fluxes, assuming a boundary layer height

of ∼1 km, is only ∼0.079 ppbv hr−1. This is small compared to the 2 to 3 ppbv hr−1 total

boundary layer production rates reported by the literature (Sumner et al., 2001; Choi et al.,

2010), implying that HCHO fluxes have only a small effect on boundary layer concentrations.

Figure 3.4b shows the median diurnal HCHO concentrations for each measurement

height. Nighttime hours show lower concentrations near ground level, suggesting dominance

of in-canopy sinks such as deposition. The peak in concentration around 8:00 corresponds

to increased wind speed and emission of precursors, followed by a sharp change in wind

direction. For most of the day, a negative gradient is present, with higher concentrations

near the ground. Daytime HCHO concentrations at the ground level (1.6 m) inlet were

68

typically 15 to 20% higher than concentrations in or above canopy. Qualitative testing

of campaign-related ground equipment (e.g. tarps) at the end of the campaign suggests

negligible emissions from these materials, and there was little ground level vegetation near

the site. This implies a significant direct and/or photochemical ground litter source of HCHO

or a significant difference in deposition loss between inside and below the canopy, with the

former supported by semi-quantitative testing of the ground litter at the end of the campaign

(Sect. 3.3.2). Canopy level enhancement of HCHO concentration was also observed in the

leafy part of the canopy, likely due to either fast oxidation of emitted BVOCs or direct

emission from the canopy.

The median diurnal profiles of the flux and concentration measurements do not ap-

pear to exhibit the same diurnal variation. During periods of changing wind speed and

direction occurring during early morning and mid-evening, the concentration profile changes

significantly while the flux profile does not. While these changes in airmass seem to affect

HCHO concentration, they have little effect on the flux, implying advection is a negligible

contributor to HCHO flux. This is supported by no significant correlation between tracers of

advection, such as SO2, CO2, or H2O, and HCHO flux. Nighttime deposition gradients are

not reflected as negative fluxes, as nighttime flux observations, even those with significant

turbulence (u∗ > 0.2 m s−1), are near zero. This is likely an effect of the low wind speeds in

the stable nighttime boundary layer, leading to less turbulence on which the EC technique

is dependent. In short, most of the expected drivers for HCHO fluxes (photochemistry,

emissions, stomatal uptake and turbulence), though not those for HCHO concentrations,

are linked to the solar cycle. However, we saw no evidence at this site of HCHO morning

entrainment from overnight oxidative production of HCHO above the canopy, as predicted

by Ganzeveld et al. (2008), in either the gradient or flux measurements.

69

3.3.2 Emission Studies

HCHO emission rates from canopy surfaces were measured via branch and soil/litter

enclosure experiments. Branch enclosures were performed using a ∼10 L Teflon chamber

on a branch located 2 m above the ground. Ultra-zero air enriched with CO2 to a final

concentration of ∼410 ppmv (Scott-Marin) was flowed through the chamber at ∼6 SLM and

was sampled using a 1/8” ID PTFE tube. While dry, this air was humidified by tree emission

to a typical relative humidity of 20-45%, comparable to the ambient humidity. Chamber

concentration was monitored for ∼4 hr. Blank experiments of the chamber without the

branch were performed before and after branch sampling. Average HCHO concentration

attributed to branch emission was 500±220 pptv with an average ambient temperature of

22.3±1.0 °C. Total dry needle mass was measured to be ∼14.37 g, yielding an average

emission rate of 15.4±6.9 ng (g dw)−1 hr−1 (dw = dry weight). This is significantly lower

than the 500 ng (g dw)−1 hr−1 reported by Villanueva-Fierro et al. (2004) for PPine but is

within the range of emissions reported for other conifers including Pinus pinea (Kesselmeier

et al., 1997) and Picea abies (Cojocariu et al., 2004) (see Sect. 3.5.2) for discussion). As

MEF had a measured specific leaf mass of 120±10 (g dw) m−2 and an LAI of 1.9 m2 m−2,

our measurement results in an average canopy emission rate of 3.5±1.6 µg m−2 hr−1.

Soil/litter enclosure experiments were performed using a ∼22 L steel chamber, sam-

pling at a flow rate of ∼2.5 SLM using a 1/8” ID PTFE tube. Blank experiments were

performed by holding the chamber in the air to measure ambient HCHO levels, then holding

the chamber firmly onto areas of ground with either undisturbed litter or soil with the surface

area of litter swept away and held until the HCHO concentration equilibrated. One experi-

ment each was performed using seemingly representative areas of ground litter and ground

soil. A blank experiment was also performed by placing a clean Teflon sheet on the ground

and pressing the chamber into the sheet as it was pressed into the soil, which resulted in no

70

significant difference from the ambient blanking method. The average HCHO concentration

attributed to litter & soil and bare soil were ∼900 pptv and ∼800 pptv respectively with

an average ambient temperature of 24.0±0.2 °C. Based on the ground area covered by the

chamber (∼800 cm2), the result is an average ground emission rate of 7.3±1.5 µg m−2 hr−1.

One known interference of the HCHO instrument with these measurements is due to the

significance of detection axis contamination at low flows (< 8 SLM) and changing humidity

conditions inside the chamber due to soil/litter moisture. Additionally, closed-chamber soil

measurements have been shown to affect pressure gradients in the soil, leading to enhanced

CO2 emission (Kanemasu et al., 1974; Rayment and Jarvis, 1997; Xu et al., 2006), which

may similarly affect HCHO emission. Finally, disturbance of the soil/litter and pressure

gradients in the chamber itself may have also resulted in increased HCHO emission. Each

of these interferences may have resulted in an overestimate of emission rate. Also, due to

the heterogeneity of the ground litter, it is quite likely that these two sites do not represent

the true soil/litter emission rate but provide simply a semi-qualitative estimate of soil/litter

emissions.

3.4 Zero-Dimensional Box Model

To quantify different contributions to HCHO flux, we have constructed a zero dimen-

sional box model to simulate HCHO flux above in the forest canopy similar to those that

have been reported in the literature (Sumner et al., 2001; Choi et al., 2010). The concept is

based on the need to maintain mass balance in a box vertically constrained by our HCHO

flux measurement. The contribution of vertical transport (flux) to this mass balance is de-

pendent on three other processes: horizontal transport of HCHO, sources/sinks of HCHO

inside the box, and changes in HCHO concentration inside the box (effectively ’storing’

source, sink, or transport effects). By account for each of these three terms, any remaining

HCHO production/loss in this box must correspond to the vertical flux. While gradient data

71

can yield vertically-resolved production/loss information, the goal is to integrate this over

the entire box to determine the overall estimated HCHO flux.

This mass balance is represented by the continuity equation:

δ[HCHO](z)

δt= P (z) − L(z) + E − D + A − δFHCHO(z)

δz(3.2)

P and L are respectively the height-dependent chemical production and loss, E is direct

emission, D is deposition, A is advection, and δFHCHO(z)/δz is the flux divergence. As

the area surrounding the site was remote and reasonably homogeneous, it was assumed that

horizontally-advecting airmasses were similar enough to neglect in this analysis. Solving for

FHCHO(h), the modeled flux at height h, yields the following equation:

FHCHO =

∫ h

0P (z) δz −

∫ h

0L(z) δz + E − VDep∗[HCHO] −

∫ h

0

δ[HCHO](z)

δtδz (3.3)

where VDep is the total deposition velocity of HCHO and the vertical dimension of the box

extends from 0 m to h, the EC measurement height (25.1 m). This assumes flux at z = 0

(i.e. ground level) is zero, as soil/litter contributions are treated as direct emission. The

end term corresponds to the time rate of change of the HCHO column density, referred

to as storage (S). To calculate S, vertically-resolved HCHO concentrations were linearly

extrapolated from the gradient data, with the concentration at heights between ground and

the bottom inlet assumed to be equal to the bottom inlet concentration. For clarity, we will

refer to each of the terms in Eq. 3.3 as the “flux contribution” for each respective process.

The methods used for the determination of these different processes are outlined below.

In addition to its simplicity, this model holds many advantages. Fluxes provide a con-

venient constraint on the vertical mixing at the measurement height, allowing this model to

be independent of boundary layer height. Measurements are also available for many heights

over the entire measurement volume, removing the need for concentration extrapolation.

The primary disadvantage is the absence of higher-order oxidative chemistry, which may

72

lead to significant in-canopy HCHO production from the further oxidation of VOCs formed

from the oxidation of BVOCs.

3.4.1 Chemical Production

HCHO chemical production is predicted from the first-order oxidation of different

VOCs by the following equation:

PHCHO(z) =∑i=0

αi,HCHO · kV OCi·Ox · [V OC]i(z) · [Ox](z) (3.4)

where αi,HCHO is the yield of HCHO and kV OCi·Ox is the rate constant for the respective

VOC and oxidant (Ox). Table 3a.2 shows a full list of modeled reactions with yields and rates

(Atkinson and Arey, 2003; Hasson et al., 2004; Atkinson et al., 2006; Lee et al., 2006; Carrasco

et al., 2007; Jenkin et al., 2007; Dillon and Crowley, 2008). Isoprene and its oxidation

products were neglected in this analysis, due to the low reported concentrations (0.1 to

0.3 ppbv) of isoprene at this site (Kim et al., 2010) and the short daytime lifetime of HCHO

(midday: 1 to 5 h), which likely limits the impact of HCHO advected from upwind production

sources. As a result, the total PTR-MS signal at m/z = 69 was considered to be MBO.

Monoterpene (MT) speciation was determined by previous observations at this site (Kim

et al., 2010), where α-pinene, β-pinene, and 3-carene were found to be 22%, 26%, and 21%

of total MT, respectively. The remaining MT (31%) were assumed to have a reaction rate

and HCHO yield equal to the average of the other three. HCHO production from the CH3O2

radical was calculated from methane and peroxyacetyl (PA) radical concentrations, where

PA concentrations were calculated using the steady state model presented by LaFranchi

et al. (2009) in a method similar to that used by Choi et al. (2010) (see Sect. 3a.3). The

oxidants used were OH and ozone. Nighttime oxidation by NO3 was neglected due to low

NOx concentrations at this site. Ozone gradients were available during the measurement

period while OH gradients were not. As a result, OH concentration was assumed constant

73

throughout the canopy. This assumption was validated by a series of vertical gradient studies

later in the campaign.

3.4.2 Chemical Loss

Chemical destruction of HCHO can proceed via reaction with OH or photolysis. Loss

due to OH was calculated with the rate constant described by Atkinson et al. (2006) and

assuming the OH concentrations were equal at all heights. Typical midday HCHO lifetime

with respect to OH was∼13 hr. Photolysis rates for HCHO were determined by weighting the

measured downwelling JNO2 values by the ratio of clear-sky HCHO and NO2 photolysis rates

estimated using the Tropospheric Ultraviolet and Visible (TUV) Radiation Model (http:

//cprm.acd.ucar.edu/Models/TUV/). To account for light extinction in the canopy, the

photolysis rates were weighted by the leaf area distribution function (LADF) using a modified

Weibull distribution (Teske and Thistle, 2004), for which parameters were determined by

destructive harvesting measurements of PPine at a similar PPine forest (Wolfe and Thornton,

2011). The extinction ratio was then calculated by:

Re(z) = e

−krad · LADF (z)cos(SZA) (3.5)

where SZA is solar zenith angle calculated from the TUV model and krad = 0.75, an empirical

parameter to scale the ground level extinction to be ∼25% at noon to match the measured

photosynthetically active radiation (PAR) profile. Integrating these photolysis rates over the

entire canopy, this yielded a typical noon lifetime of HCHO due to photolysis of∼3.5 hr above

the canopy and ∼14 hr near the ground. Actual loss of HCHO to photolysis was determined

by calculating the height-dependent loss using the HCHO gradient, then integrating to

calculate the overall HCHO photolysis loss.

74

3.4.3 Direct Emission

Emission flux contributions were extrapolated from the chamber experiments using a

simple exponential model (EHCHO = A · exp(βT )), where T is temperature in °C and β

= 0.07 °C−1 is an empirical constant found for HCHO by Villanueva-Fierro et al. (2004).

The PPine emissions were weighted by a factor of (0.85 ∗PAR/PAR0 + 15), where PAR0 is

the average, clear-sky, noontime measured PAR, thereby fixing nighttime emissions to 15%

of daytime emissions as observed by Villanueva-Fierro et al. (2004). The pre-exponential

factors (A) determined for both soil and branch emissions from the emission rates found in

the experiments (Sect. 3.3.2) were 1.52 and 0.74 µg m−2 hr−1 respectively.

3.4.4 Dry Deposition

Total dry deposition was estimated using a resistance model similar to that used

for PAN deposition in previous flux budget studies (Turnipseed et al., 2006; Wolfe et al.,

2009). The resistance model calculates the total deposition resistance (RDep) as the sum of

resistances from separate physical processes (Wesely, 1989; Wesely and Hicks, 2000):

Vdep =1

Rdep=

1

Ra + Rb + Rc(3.6)

Ra and Rb were calculated using standard literature methods (see Sect. 3a.4) (Monteith,

1965; Wesely, 1989; Jensen and Hummelshoj, 1995, 1997; Massman, 1998). Rc is the surface

resistance, or resistance to actual uptake or loss on the leaf, and consists of two parallel

terms, stomatal (RST ) and non-stomatal (RNS) resistance. As stomatal uptake is negligible

at night, RNS was estimated from the nighttime HCHO deposition velocity. At night,

the lack of thermal turbulence leads to very small fluxes. Therefore, we can estimate the

nighttime HCHO deposition rate by using Eq. 3.3, setting FHCHO to zero, and solving for

deposition:

VNS Dep ∗ [HCHO] = DNS =

∫ h

0P (z) δz −

∫ h

0L(z) δz + E −

∫ h

0

δ[HCHO](z)

δtδz (3.7)

75

Thus, our calculation of non-stomatal deposition is tied to the accuracy of our estimates

for nighttime chemical production/loss and emissions. Dividing the values of the non-

stomatal deposition flux contribution during the relatively constant nighttime hours (23:00

to 4:00) by the average canopy [HCHO] resulted in an average nighttime deposition velocity

of 0.18±0.08 cm s−1. RNS was then calculated by inverting the following equation:

Vdep,night =1

Ra,night + Rb,night + RNS(3.8)

where Ra,night and Rb,night are Ra and Rb averaged over the relatively constant nighttime

hours. It should be noted that this represents the total non-stomatal deposition velocity,

to which both cuticular and soil/ground uptake contribute, but are mathematically insepa-

rable by this method. Similarly, it was necessary to assume that the Rb,soil is equal to the

calculated Rb for a pine needle.

Literature values using the boundary layer budget method report HCHO nighttime

deposition velocity as ranging from 0.65 to 0.84 cm s−1 (Sumner et al., 2001; Choi et al.,

2010). The discrepancy between this work and the literature likely lies in the different

assumptions on which either model is based. The boundary layer method assumes similarity

between HCHO and ozone deposition and usually depends on literature estimates of ozone

deposition. This method also assumes that deposition is the only nighttime loss process

and there are no production processes. Finally, the boundary layer method is based on a

single measurement and assumes a continuous concentration throughout the boundary layer.

The gradient method used in this work makes no assumptions on the HCHO profile, as it

is measured directly, and does not depend on literature ozone deposition. The gradient

method also estimates nighttime production and loss via the model terms. However, the

gradient method still has limitations in that it is much more dependent on direct emission

measurements/estimates and assumes the canopy gradient is well represented by the available

measurements (in this case, four heights).

76

RST was calculated by the following equation (Wesely, 1989).

RST =DH2O

DHCHO·RST,H2O + Rm,HCHO (3.9)

Mesophyll resistance (Rm) is the resistance to absorption into the plant mesophyll once inside

the stomata, which is negligible for HCHO due to its large Henry’s law constant (Wesely,

1989; Zhang et al., 2002). RST,H2O was calculated using the Penman-Monteith equation

(Monteith, 1965; Monteith and Unsworth, 1990). The resulting average daily minimum Rc

was ∼180 s m−1. An alternative method used for estimating Rc was the parameterization

described by Wesely (1989) for an autumn coniferous forest, which yielded a comparable

daily minimum average of ∼226 s m−1. This latter method was not used in the final model,

as the measurement-based method was considered more accurate.

The daytime-maximum median VDep determined by this method was 0.39±0.11 cm s−1,

and had a diurnal profile peaking at 9:00, then gradually decreasing until a sharp decrease at

dusk. Similar to the nighttime deposition velocity, this daytime deposition velocity is con-

siderably smaller than the literature value of 1.5 cm s−1 (Krinke and Wahner, 1999). These

discrepancies may partly result from the lower LAI and less underbrush at the BEACHON

site compared to the literature sites. Additionally, the deposition term is highly dependent

on litter emission, which makes it very sensitive to the temperature-dependent method we

use to extrapolate litter emission rates. However, the method used in this work is also not

dependent on measured ozone deposition velocities, which may be influenced by chemistry

as well as deposition (Kurpius and Goldstein, 2003). Direct comparison of deposition using

the ozone similarity method described in the literature (e.g. the boundary-layer budget ap-

proach) (Sumner et al., 2001) was not possible for this dataset, as nighttime concentrations

did not exhibit clear first-order decay.

77

3.5 Model Results and Discussion

Modeled fluxes were calculated using data from 13 - 21 August. The major canopy-

integrated HCHO production and loss terms for the base (unaltered) version of the model

are shown in Fig. 3.5, while values for all terms are shown in Table 3a.3. The dominant

production terms are direct emission from both PPine and ground litter, OH oxidation of

MBO, CH4, and acetaldehyde, and chemical destruction of PA radicals. MT oxidation and

ozonolysis in general contribute minimally to the HCHO production. The total production

diurnal cycle is similar in form to the radiative diurnal cycle, reflecting the production

dependence on temperature and ambient radiation. HCHO loss was dominated by dry

deposition, as expected for an in-canopy airmass. The total loss diurnal cycle therefore

mostly reflects the diurnal cycle in stomatal uptake. As shown in Fig. 3.6, the base model

underpredicts the noontime HCHO fluxes by a factor of 6 during the day. Modeled nighttime

fluxes agree much better with observations, but this is expected as we have constrained

nighttime deposition via an assumption of no flux at night.

3.5.1 General Sensitivity Analyses

We performed a sensitivity analyses on a number of input parameters to determine

what model conditions resulted in the best agreement with measurements. Many of these

made little or no significant difference in model-measurement agreement. For example, we

assumed that OH mixing ratios below the CIMS detection limit were equal to half of the

detection limit value, or 2.5× 105 molec cm−3. To test this, we performed model calculations

with OH mixing ratios ranging from zero to the CIMS detection limit (5× 105 molec cm−3).

This effect was found to be negligible (< 5%) on the order of the missing HCHO flux.

We also tested the effect of separating HCHO PPine emission and stomatal deposition.

Strictly, PPine direct emission and stomatal deposition are not independent processes and

78

are related by the HCHO compensation point, the ambient HCHO concentration above

which stomatal deposition is expected and below which stomatal emission is expected. As

the compensation point can vary by tree species and environment (Seco et al., 2007, 2008), it

was not possible to treat this explicitly at this site. However, as an upper limit to the error

this assumption could contribute to the missing flux, we can neglect stomatal deposition

by assuming that we are strictly in an emission-only regime. This results in only a ∼10%

reduction in noontime missing HCHO flux. In an extreme case, we can also assume that our

measured soil/litter emission rate also represents the sum of both soil/litter emission and

deposition. This was simulated by also neglecting the non-stomatal deposition component

(therefore also neglected cuticular deposition) and resulted in only a ∼25% reduction in

noontime missing HCHO flux. As a result, this cannot explain the majority of the missing

HCHO flux.

3.5.2 PPine Emission Sensitivity (E350)

In an attempt to explain this missing flux, we scaled the modeled PPine emission rate

to reach the best match to the measured flux. We achieved the best match at a PPine

emission rate of 350 ng (g dw)−1 hr−1 (E350). The diurnal cycle of this case matches the

measured flux quite well, though the model we used for PPine emission was directly depen-

dent on temperature and PAR. The emission rate used in E350 is more than an order of

magnitude greater than the emission rate predicted by our branch enclosure studies. It is

comparable to the 500±400 ng (g dw)−1 hr−1 measured by Villanueva-Fierro et al. (2004).

However, the formaldehyde rates observed by Villanueva-Fierro et al. (2004) are consistently

an order of magnitude higher than those reported for similar tree species by other investi-

gators (Kesselmeier et al., 1997; Cojocariu et al., 2004). The cause of this discrepancy is

unclear, but the climate of the area studied by Villanueva-Fierro et al. (2004) was different,

and there may have been differences in other factors such as stress conditions. The formalde-

79

hyde quantification technique used for all of these studies (collection and storage on DNPH

cartridges followed by analysis with high pressure liquid chromatography) has potentially

large errors associated with background subtraction and differences due to analytical and

enclosure techniques may have contributed to the discrepancy in reported rates. It should

also be noted that emission rate is dependent on the β value used in the exponential model,

as described in Sect. 3.4.3. However, with no method of separating these quantities, we

continued to use the value found in Villanueva-Fierro et al. (2004) throughout this analysis.

3.5.3 MBO Sensitivity (VOC-I)

In another case, we simulated an increase in MBO concentrations, using it as a proxy

for a precursor with both a temperature and PAR dependent emission profile. The best

match to measured flux was an increase by a factor of 10 (VOC-I). This implies that HCHO

production could be significantly impacted by either contributions from higher-order oxida-

tion products of MBO or oxidation of an unmeasured BVOC/combination of BVOCs with

a similar temperature/PAR-dependent emission profile. As MBO emission is both a tem-

perature and PAR dependent process, the VOC-I and E350 model cases demonstrate that

the HCHO flux corresponds to a temperature/PAR dependent emission profile. However,

if these unmeasured BVOCs are assumed to have an OH reactivity similar to MBO, they

would contribute 9× the OH reactivity of MBO (median noontime MBO concentration:

∼1.1 ppbv; median noontime MBO contribution to OH reactivity: ∼1.3 s−1), which would

be on the order of ∼12 s−1. As the measured median noontime OH reactivity is on the order

of 6 to 7 s−1 in the canopy during the campaign, this suggests that the unmeasured BVOC

does not have a similar OH reactivity to MBO. Therefore, in order to form HCHO inside

the canopy faster than vertical transport out of the canopy, the primary oxidation pathway

of this unmeasured BVOC would need to be through a species other than OH (e.g. ozone).

80

3.5.4 Monoterpene Sensitivity (VOC-II)

As the missing BVOCs thought to cause the OH reactivity gap have been attributed

to terpenes, a final sensitivity analysis was simulating an increase in MT concentrations by

a factor of 10 (VOC-II). As MT concentrations are highest at night, but oxidation is highest

during the day, the result was an increase essentially independent of the time of day. This

does not match the observed HCHO flux diurnal cycle, suggesting that measurements of MT

are unlikely to be under-predicted. In Fig. 3.6, values are shown for VOC-II while using the

same dry deposition rates as the base case model. When dry deposition was calculated the

same as for the other cases, the net effect was an inverse diurnal cycle as HCHO production

from MT is greatest at night, a poor match to the measured fluxes. Additionally, the model

was no longer able to predict zero flux at night, as the nighttime deposition velocity reached

the aerodynamic limit due to a significantly decreased non-stomatal resistance. This further

supports that species with a temperature-dependent, PAR-independent emission profile, as

with MT at this site, are unlikely to be the source of the missing flux.

3.6 Conclusions

In this work, we demonstrate the first published measurements using the FILIF tech-

nique and the first published measurements of HCHO flux by eddy covariance. The ability

to use this emerging class of fiber laser technology now allows for more complex spectro-

scopic techniques to be used in field conditions, which was previously quite difficult due to

the sensitivities of traditional lasers. These advantages allow the FILIF technique to be one

of the fastest and most sensitive methods for HCHO detection, with laboratory limits of

detection (3σ) as low as ∼25 pptv in 1 s.

HCHO fluxes were found to have a median diurnal cycle quite similar to that of PAR,

with a median midday maximum of ∼80 µg m−2 hr−1 (∼24 pptv m s−1). Strong HCHO

81

gradients were observed at night implying deposition. Moderate inverted gradients were ob-

served during the day with higher concentrations near the ground during midday, implying

ground litter emission. These gradients were also observed in the canopy during mid/late af-

ternoon, implying PPine emission and/or fast, in-canopy, photochemical production. Branch

and soil chamber experiments confirmed HCHO emission of 3.5±1.6 µg m−2 hr−1 from PPine

and 7.3±1.5 µg m−2 hr−1 from soil and ground litter. While typical midday canopy HCHO

net production rates are ∼3.2 ppbv hr−1, this corresponds to only 0.079 ppbv hr−1 over the

entire boundary layer, insignificant with respect to the HCHO budget. Additionally, these

measured emissions, along with the gradient profiles, clarify the need to account for not only

HCHO emissions from the canopy and undergrowth in forests, but the soil and ground litter

as well.

A zero-dimensional box model of the forest canopy using first-order chemical produc-

tion of HCHO was shown to under-predict HCHO fluxes by a factor of 6. A sensitivity

analysis showed that the model would agree with measurements by increasing either the

PPine emission rate or the concentration of a species with an emission and reactivity profile

similar to MBO. This suggests that the missing HCHO flux is caused by a process that

is dependent on temperature and PAR. The disagreement of the measured flux with the

VOC-II case, with simulated increased MT, further supports this argument, as the domi-

nant MT at this site have a primarily temperature-dependent emission profile (Kim et al.,

2010). Potential explanations for this discrepancy are higher HCHO emission, production

by fast, higher-order chemistry of MBO oxidation products, or the processing of unmea-

sured BVOCs with emission profiles also dependent on temperature and PAR by oxidants

other than OH. A model including explicit chemistry of these oxidation products would dis-

tinguish between the latter two of these possibilities. The lack of agreement between both

non-stomatal HCHO deposition and emission rates between this work and the literature also

82

highlights a need to parameterize of HCHO compensation points, emission and deposition

rates for trees and soil as functions of temperature, radiation, and humidity.

These measurements provide a constraint on the oxidation in a forest canopy of unmea-

sured BVOCs, which have been attributed as a cause of the model/measurement mismatch in

OH reactivity and concentrations. To conclusively determine this effect, it will be necessary

to determine the amount of missing flux that is not due to either higher order chemistry or

direct emission. Calculations of OH reactivity compared to measurements have shown that

the missing flux cannot solely result from oxidation of missing VOC by OH. Additionally,

the minimal emissions of sesquiterpenes at this site (Kim et al., 2010) and the expected OH

reactivities suggest that VOC oxidation cannot explain the entire missing flux. As a result,

direct emission must be the cause of at least a portion of the missing flux, and this study does

not remove the possibility that it may be entirely due to this effect. Future investigations

into not only HCHO emission rates from the canopy, but also the soil and ground litter, will

be crucial to correctly apportioning HCHO flux.

3.7 Acknowledgements

The authors thank the National Science Foundation (ATM 0852406) and the NCAR

BEACHON project for support. We also thank the United States Forest Service, specifically

Richard Oakes, for logistical support, Melinda Beaver, John Crounse, and Paul Wennberg

for performing the permeation tube calibration, and Alan Fried for useful discussions on

calibration and zeroing techniques. NCAR is sponsored by the National Science Foundation.

3.8 Bibliography

Atkinson, R. and Arey, J.: Gas-phase tropospheric chemistry of biogenic volatile organic

compounds: a review, Atmospheric Environment, 37, S197–S219, 2003.

Atkinson, R., Baulch, D. L., Cox, R. A., Crowley, J. N., Hampson, R. F., Hynes, R. G.,

83

Jenkin, M. E., Rossi, M. J., and Troe, J.: Evaluated kinetic and photochemical data for

atmospheric chemistry: Volume II - gas phase reactions of organic species, Atmospheric

Chemistry and Physics, 6, 3625–4055, 2006.

Baker, B., Guenther, A., Greenberg, J., Goldstein, A., and Fall, R.: Canopy fluxes of 2-

methyl-3-buten-2-ol over a ponderosa pine forest by relaxed eddy accumulation: Field

data and model comparison, Journal of Geophysical Research-Atmospheres, 104, 26 107–

26 114, 1999.

Baldocchi, D. D., Hicks, B. B., and Meyers, T. P.: Measuring Biosphere-Atmosphere Ex-

changes of Biologically Related Gases with Micrometeorological Methods, Ecology, 69,

1331–1340, 1988.

Berger, B. W., Davis, K. J., Yi, C. X., Bakwin, P. S., and Zhao, C. L.: Long-term carbon

dioxide fluxes from a very tall tower in a northern forest: Flux measurement methodology,

Journal of Atmospheric and Oceanic Technology, 18, 529–542, 2001.

Carrasco, N., Doussin, J. F., O’Connor, M., Wenger, J. C., Picquet-Varrault, B., Durand-

Jolibois, R., and Carlier, P.: Simulation chamber studies of the atmospheric oxidation

of 2-methyl-3-buten-2-ol: Reaction with hydroxyl radicals and ozone under a variety of

conditions, Journal of Atmospheric Chemistry, 56, 33–55, 2007.

Choi, W., Faloona, I. C., Bouvier-Brown, N. C., McKay, M., Goldstein, A. H., Mao, J.,

Brune, W. H., LaFranchi, B. W., Cohen, R. C., Wolfe, G. M., Thornton, J. A., Son-

nenfroh, D. M., and Millet, D. B.: Observations of elevated formaldehyde over a forest

canopy suggest missing sources from rapid oxidation of arboreal hydrocarbons, Atmo-

spheric Chemistry and Physics, 10, 8761–8781, 2010.

Cojocariu, C., Kreuzwieser, J., and Rennenberg, H.: Correlation of short-chained carbonyls

84

emitted from Picea abies with physiological and environmental parameters, New Phytol-

ogist, 162, 717–727, 2004.

Di Carlo, P., Brune, W. H., Martinez, M., Harder, H., Lesher, R., Ren, X., Thornberry,

T., Carroll, M. A., Young, V., Shepson, P. B., Riemer, D., Apel, E., and Campbell,

C.: Missing OH Reactivity in a Forest: Evidence for Unknown Reactive Biogenic VOCs,

Science, 304, 722–725, 2004.

Dillon, T. J. and Crowley, J. N.: Direct detection of OH formation in the reactions of

HO2 with CH3C(O)O2 and other substituted peroxy radicals, Atmospheric Chemistry

and Physics, 8, 4877–4889, 2008.

Farmer, D. K., Wooldridge, P. J., and Cohen, R. C.: Application of thermal-dissociation

laser induced fluorescence (TD-LIF) to measurement of HNO3, Σalkyl nitrates, Σperoxy

nitrates, and NO2 fluxes using eddy covariance, Atmospheric Chemistry and Physics, 6,

3471–3486, 2006.

Foken, T. and Wichura, B.: Tools for quality assessment of surface-based flux measurements,

Agricultural and Forest Meteorology, 78, 83–105, 1996.

Fried, A., McKeen, S., Sewell, S., Harder, J., Henry, B., Goldan, P., Kuster, W., Williams,

E., Baumann, K., Shetter, R., and Cantrell, C.: Photochemistry of formaldehyde during

the 1993 Tropospheric OH Photochemistry Experiment, Journal of Geophysical Research-

Atmospheres, 102, 6283–6296, 1997.

Galloway, M. M., DiGangi, J. P., Hottle, J. R., Huisman, A. J., Mielke, L. H., Alaghmand,

M., Shepson, P. B., Weremijewicz, J., Klavon, H., McNeal, F. M., Carroll, M. A., Griffith,

S., Hansen, R. F., Stevens, P. S., Bertman, S. B., and Keutsch, F. N.: Observations

85

and modeling of formaldehyde at the PROPHET mixed hardwood forest site in 2008,

Atmospheric Environment, accepted, 2011.

Ganzeveld, L., Eerdekens, G., Feig, G., Fischer, H., Harder, H., Konigstedt, R., Kubistin,

D., Martinez, M., Meixner, F. X., Scheeren, H. A., Sinha, V., Taraborrelli, D., Williams,

J., de Arellano, J. V. G., and Lelieveld, J.: Surface and boundary layer exchanges of

volatile organic compounds, nitrogen oxides and ozone during the GABRIEL campaign,

Atmospheric Chemistry and Physics, 8, 6223–6243, 2008.

Goldstein, A. H., McKay, M., Kurpius, M. R., Schade, G. W., Lee, A., Holzinger, R., and

Rasmussen, R. A.: Forest thinning experiment confirms ozone deposition to forest canopy

is dominated by reaction with biogenic VOCs, Geophysical Research Letters, 31, L22 106,

2004.

Hansel, A., Jordan, A., Warneke, C., Holzinger, R., and Lindinger, W.: Improved detection

limit of the proton-transfer reaction mass spectrometer: on-line monitoring of volatile

organic compounds at mixing ratios of a few pptv, Rapid Communications in Mass Spec-

trometry, 12, 871–875, 1998.

Hasson, A. S., Tyndall, G. S., and Orlando, J. J.: A product yield study of the reaction

of HO2 radicals with ethyl peroxy (C2H5O2), acetyl peroxy (CH3C(O)O2), and acetonyl

peroxy (CH3C(O)CH2O2) radicals, Journal of Physical Chemistry A, 108, 5979–5989,

2004.

Hofzumahaus, A., Rohrer, F., Lu, K., Bohn, B., Brauers, T., Chang, C., Fuchs, H., Holland,

F., Kita, K., Kondo, Y., Li, X., Lou, S., Shao, M., Zeng, L., Wahner, A., and Zhang, Y.:

Amplified Trace Gas Removal in the Troposphere, Science, 324, 1702–1704, 2009.

Holzinger, R., Lee, A., Paw, K. T., and Goldstein, A. H.: Observations of oxidation prod-

86

ucts above a forest imply biogenic emissions of very reactive compounds, Atmospheric

Chemistry and Physics, 5, 67–75, 2005.

Hornbrook, R. S., Crawford, J. H., Edwards, G. D., Goyea, O., Mauldin Iii, R. L., Olson,

J. S., and Cantrell, C. A.: Measurements of tropospheric HO2 and RO2 by oxygen dilution

modulation and chemical ionization mass spectrometry, Atmospheric Measurements and

Technology, 4, 735–756, 2011.

Hottle, J. R., Huisman, A. J., Digangi, J. P., Kammrath, A., Galloway, M. M., Coens,

K. L., and Keutsch, F. N.: A Laser Induced Fluorescence-Based Instrument for In-Situ

Measurements of Atmospheric Formaldehyde, Environmental Science and Technology, 43,

790–795, 2009.

Hutterli, M. A., McConnell, J. R., Chen, G., Bales, R. C., Davis, D. D., and Lenschow,

D. H.: Formaldehyde and hydrogen peroxide in air, snow and interstitial air at South

Pole, Atmospheric Environment, 38, 5439–5450, 2004.

Jacobi, H. W., Frey, M. M., Hutterli, M. A., Bales, R. C., Schrems, O., Cullen, N. J.,

Steffen, K., and Koehler, C.: Measurements of hydrogen peroxide and formaldehyde ex-

change between the atmosphere and surface snow at Summit, Greenland, Atmospheric

Environment, 36, 2619–2628, 2002.

Jenkin, M. E., Hurley, M. D., and Wallington, T. J.: Investigation of the radical product

channel of the CH3C(O)O2 + HO2 reaction in the gas phase, Physical Chemistry Chemical

Physics, 9, 3149–3162, 2007.

Jensen, N. O. and Hummelshoj, P.: Derivation of Canopy Resistance for Water-Vapor Fluxes

over a Spruce Forest, Using a New Technique for the Viscous Sublayer Resistance, Agri-

cultural and Forest Meteorology, 73, 339–352, 1995.

87

Jensen, N. O. and Hummelshoj, P.: Erratum to ”Derivation of canopy resistance for water

vapor fluxes over a spruce forest, using a new technique for the viscous sublayer resistance”,

Agricultural and Forest Meteorology, 85, 289–289, 1997.

Jimenez, J. L., Canagaratna, M. R., Donahue, N. M., Prevot, A. S. H., Zhang, Q., Kroll,

J. H., DeCarlo, P. F., Allan, J. D., Coe, H., Ng, N. L., Aiken, A. C., Docherty, K. S.,

Ulbrich, I. M., Grieshop, A. P., Robinson, A. L., Duplissy, J., Smith, J. D., Wilson, K. R.,

Lanz, V. A., Hueglin, C., Sun, Y. L., Tian, J., Laaksonen, A., Raatikainen, T., Rautiainen,

J., Vaattovaara, P., Ehn, M., Kulmala, M., Tomlinson, J. M., Collins, D. R., Cubison,

M. J., Dunlea, E. J., Huffman, J. A., Onasch, T. B., Alfarra, M. R., Williams, P. I., Bower,

K., Kondo, Y., Schneider, J., Drewnick, F., Borrmann, S., Weimer, S., Demerjian, K.,

Salcedo, D., Cottrell, L., Griffin, R., Takami, A., Miyoshi, T., Hatakeyama, S., Shimono,

A., Sun, J. Y., Zhang, Y. M., Dzepina, K., Kimmel, J. R., Sueper, D., Jayne, J. T.,

Herndon, S. C., Trimborn, A. M., Williams, L. R., Wood, E. C., Middlebrook, A. M.,

Kolb, C. E., Baltensperger, U., and Worsnop, D. R.: Evolution of Organic Aerosols in the

Atmosphere, Science, 326, 1525–1529, 2009.

Junkermann, W., Platt, U., and Volzthomas, A.: A Photoelectric Detector for the Measure-

ment of Photolysis Frequencies of Ozone and Other Atmospheric Molecules, Journal of

Atmospheric Chemistry, 8, 203–227, 1989.

Kaimal, J. C. and Finnigan, J. J.: Atmospheric Boundary Layer Flows - Their Structure

and Measurement, Oxford University Press, New York, 1994.

Kanemasu, E. T., Powers, W. L., and Sij, J. W.: Field Chamber Measurements of CO2 Flux

from Soil Surface, Soil Science, 118, 233–237, 1974.

Karl, T., Guenther, A., Turnipseed, A., Tyndall, G., Artaxo, P., and Martin, S.: Rapid for-

88

mation of isoprene photo-oxidation products observed in Amazonia, Atmospheric Chem-

istry and Physics, 9, 7753–7767, 2009.

Kesselmeier, J., Bode, K., Hofmann, U., Muller, H., Schafer, L., Wolf, A., Ciccioli, P.,

Brancaleoni, E., Cecinato, A., Frattoni, M., Foster, P., Ferrari, C., Jacob, V., Fugit, J. L.,

Dutaur, L., Simon, V., and Torres, L.: Emission of short chained organic acids, aldehydes

and monoterpenes from Quercus ilex L. and Pinus pinea L. in relation to physiological

activities, carbon budget and emission algorithms, Atmospheric Environment, 31, 119–

133, 1997.

Kim, S., Karl, T., Guenther, A., Tyndall, G., Orlando, J., Harley, P., Rasmussen, R., and

Apel, E.: Emissions and ambient distributions of Biogenic Volatile Organic Compounds

(BVOC) in a ponderosa pine ecosystem: interpretation of PTR-MS mass spectra, Atmo-

spheric Chemistry and Physics, 10, 1759–1771, 2010.

Krinke, S. and Wahner, A.: Formaldehyde and ozone deposition velocities determined above

a deciduous forest during summer, Eos Trans AGU, 80(46), Fall Meeting Supplement,

1999.

Kurpius, M. R. and Goldstein, A. H.: Gas-phase chemistry dominates O3 loss to a forest, im-

plying a source of aerosols and hydroxyl radicals to the atmosphere, Geophysical Research

Letters, 30, Artn 1371, 2003.

LaFranchi, B. W., Wolfe, G. M., Thornton, J. A., Harrold, S. A., Browne, E. C., Min, K. E.,

Wooldridge, P. J., Gilman, J. B., Kuster, W. C., Goldan, P. D., de Gouw, J. A., Mckay, M.,

Goldstein, A. H., Ren, X., Mao, J., and Cohen, R. C.: Closing the peroxy acetyl nitrate

budget: observations of acyl peroxy nitrates (PAN, PPN, and MPAN) during BEARPEX

2007, Atmospheric Chemistry and Physics, 9, 7623–7641, 2009.

89

Lee, A., Goldstein, A. H., Kroll, J. H., Ng, N. L., Varutbangkul, V., Flagan, R. C., and

Seinfeld, J. H.: Gas-phase products and secondary aerosol yields from the photooxidation

of 16 different terpenes, Journal of Geophysical Research-Atmospheres, 111, 2006.

Lee, X., Massman, W., and Law., B.: Handbook of Micrometeorology - A Guide for Surface

Flux Measurement and Analysis, Kluwer Academic Publishers, Dordrecht, The Nether-

lands, 2004.

Lee, Y. N., Zhou, X., Kleinman, L. I., Nunnermacker, L. J., Springston, S. R., Daum, P. H.,

Newman, L., Keigley, W. G., Holdren, M. W., Spicer, C. W., Young, V., Fu, B., Parrish,

D. D., Holloway, J., Williams, J., Roberts, J. M., Ryerson, T. B., and Fehsenfeld, F. C.:

Atmospheric chemistry and distribution of formaldehyde and several multioxygenated car-

bonyl compounds during the 1995 Nashville Middle Tennessee Ozone Study, Journal of

Geophysical Research-Atmospheres, 103, 22 449–22 462, 1998.

Lelieveld, J., Butler, T. M., Crowley, J. N., Dillon, T. J., Fischer, H., Ganzeveld, L., Harder,

H., Lawrence, M. G., Martinez, M., Taraborrelli, D., and Williams, J.: Atmospheric

oxidation capacity sustained by a tropical forest, Nature, 452, 737–740, 2008.

Lindinger, W., Hansel, A., and Jordan, A.: Proton-transfer-reaction mass spectrometry

(PTR-MS): on-line monitoring of volatile organic compounds at pptv levels, Chemical

Society Reviews, 27, 347–375, 1998.

Massman, W. J.: A review of the molecular diffusivities of H2O, CO2, CH4, CO, O3, SO2,

NH3, N2O, NO3 and NO2 in air, O2 and N2 near STP, Atmospheric Environment, 32,

1111–1127, 1998.

McManus, J. B., Zahniser, M. S., Nelson, J. D. D., Shorter, J. H., Herndon, S., Wood, E.,

90

and Wehr, R.: Application of quantum cascade lasers to high-precision atmospheric trace

gas measurements, Optical Engineering, 49, 111 124–11, 2010.

Monteith, J.: Evaporation and Environment, Symposia of the Society for Experimental

Biology, 19, 205–234, 1965.

Monteith, J. and Unsworth, M.: Principles of Environmental Physics, Edward Arnold, Lon-

don, 2nd edn., 1990.

Munger, J. W., Jacob, D. J., Daube, B. C., Horowitz, L. W., Keene, W. C., and Heikes,

B. G.: Formaldehyde, Glyoxal, and Methylglyoxal in Air and Cloudwater at a Rural

Mountain Site in Central Virginia, Journal of Geophysical Research-Atmospheres, 100,

9325–9333, 1995.

Rayment, M. B. and Jarvis, P. G.: An improved open chamber system for measuring soil

CO2 effluxes in the field, Journal of Geophysical Research-Atmospheres, 102, 28 779–

28 784, 1997.

Sadanaga, Y., Yoshino, A., Watanabe, K., Yoshioka, A., Wakazono, Y., Kanaya, Y., and

Kajii, Y.: Development of a measurement system of OH reactivity in the atmosphere by

using a laser-induced pump and probe technique, Review of Scientific Instruments, 75,

2648–2655, 2004.

Schade, G. W., Goldstein, A. H., Gray, D. W., and Lerdau, M. T.: Canopy and leaf level

2-methyl-3-buten-2-ol fluxes from a ponderosa pine plantation, Atmospheric Environment,

34, 3535–3544, 2000.

Seco, R., Penuelas, J., and Filella, I.: Short-chain oxygenated VOCs: Emission and uptake

by plants and atmospheric sources, sinks, and concentrations, Atmospheric Environment,

41, 2477–2499, 2007.

91

Seco, R., Penuelas, J., and Filella, I.: Formaldehyde emission and uptake by Mediterranean

trees Quercus ilex and Pinus halepensis, Atmospheric Environment, 42, 7907–7914, 2008.

Sinha, V., Williams, J., Lelieveld, J., Ruuskanen, T. M., Kajos, M. K., Patokoski, J., Hellen,

H., Hakola, H., Mogensen, D., Boy, M., Rinne, J., and Kulmala, M.: OH Reactivity Mea-

surements within a Boreal Forest: Evidence for Unknown Reactive Emissions, Environ-

mental Science and Technology, 44, 6614–6620, 2010.

Slemr, J., Junkermann, W., and VolzThomas, A.: Temporal variations in formaldehyde,

acetaldehyde and acetone and budget of formaldehyde at a rural site in southern Germany,

Atmospheric Environment, 30, 3667–3676, 1996.

Sumner, A. L., Shepson, P. B., Couch, T. L., Thornberry, T., Carroll, M. A., Sillman, S.,

Pippin, M., Bertman, S., Tan, D., Faloona, I., Brune, W., Young, V., Cooper, O., Moody,

J., and Stockwell, W.: A study of formaldehyde chemistry above a forest canopy, Journal

of Geophysical Research-Atmospheres, 106, 24 387–24 405, 2001.

Tan, D., Faloona, I., Simpas, J. B., Brune, W., Shepson, P. B., Couch, T. L., Sumner, A. L.,

Carroll, M. A., Thornberry, T., Apel, E., Riemer, D., and Stockwell, W.: HOx budgets

in a deciduous forest: Results from the PROPHET summer 1998 campaign, Journal of

Geophysical Research-Atmospheres, 106, 24 407–24 427, 2001.

Tanner, D. J., Jefferson, A., and Eisele, F. L.: Selected ion chemical ionization mass spec-

trometric measurement of OH, Journal of Geophysical Research-Atmospheres, 102, 6415–

6425, 1997.

Teske, M. E. and Thistle, H. W.: A library of forest canopy structure for use in interception

modeling, Forest Ecology and Management, 198, 341–350, 2004.

Turnipseed, A. A., Huey, L. G., Nemitz, E., Stickel, R., Higgs, J., Tanner, D. J., Slusher,

92

D. L., Sparks, J. P., Flocke, F., and Guenther, A.: Eddy covariance fluxes of peroxy-

acetyl nitrates (PANs) and NOy to a coniferous forest, Journal of Geophysical Research-

Atmospheres, 111, 2006.

Villanueva-Fierro, I., Popp, C. J., and Martin, R. S.: Biogenic emissions and ambient con-

centrations of hydrocarbons, carbonyl compounds and organic acids from ponderosa pine

and cottonwood trees at rural and forested sites in Central New Mexico, Atmospheric

Environment, 38, 249–260, 2004.

Volz-Thomas, A., Lerner, A., Patz, H. W., Schultz, M., McKenna, D. S., Schmitt, R.,

Madronich, S., and Roth, E. P.: Airborne measurements of the photolysis frequency of

NO2, Journal of Geophysical Research-Atmospheres, 101, 18 613–18 627, 1996.

Weibring, P., Richter, D., Walega, J. G., and Fried, A.: First demonstration of a high

performance difference frequency spectrometer on airborne platforms, Optics Express, 15,

13 476–13 495, 2007.

Wert, B. P., Fried, A., Henry, B., and Cartier, S.: Evaluation of inlets used for the airborne

measurement of formaldehyde, Journal of Geophysical Research, 107, 4163, 2002.

Wesely, M. L.: Parameterization of Surface Resistances to Gaseous Dry Deposition in

Regional-Scale Numerical-Models, Atmospheric Environment, 23, 1293–1304, 1989.

Wesely, M. L. and Hicks, B. B.: A review of the current status of knowledge on dry deposi-

tion, Atmospheric Environment, 34, 2261–2282, 2000.

Whalley, L. K., Edwards, P. M., Furneaux, K. L., Goddard, A., Ingham, T., Evans, M. J.,

Stone, D., Hopkins, J. R., Jones, C. E., Karunaharan, A., Lee, J. D., Lewis, A. C., Monks,

P. S., Moller, S. J., and Heard, D. E.: Quantifying the magnitude of a missing hydroxyl

93

radical source in a tropical rainforest, Atmospheric Chemistry and Physics Discussions,

11, 5785–5809, 2011.

Wisthaler, A., Apel, E. C., Bossmeyer, J., Hansel, A., Junkermann, W., Koppmann, R.,

Meier, R., Mller, K., Solomon, S. J., Steinbrecher, R., Tillmann, R., and Brauers, T.:

Technical Note: Intercomparison of formaldehyde measurements at the atmosphere simu-

lation chamber SAPHIR, Atmospheric Chemistry and Physics, 8, 2189–2200, 2008.

Wolfe, G. M. and Thornton, J. A.: The Chemistry of Atmosphere-Forest Exchange (CAFE)

Model Part 1: Model description and characterization, Atmospheric Chemistry and

Physics, 11, 77–101, 2011.

Wolfe, G. M., Thornton, J. A., Yatavelli, R. L. N., Mckay, M., Goldstein, A. H., LaFranchi,

B., Min, K. E., and Cohen, R. C.: Eddy covariance fluxes of acyl peroxy nitrates (PAN,

PPN and MPAN) above a Ponderosa pine forest, Atmospheric Chemistry and Physics, 9,

615–634, 2009.

Xu, L. K., Furtaw, M. D., Madsen, R. A., Garcia, R. L., Anderson, D. J., and McDermitt,

D. K.: On maintaining pressure equilibrium between a soil CO2 flux chamber and the

ambient air, Journal of Geophysical Research-Atmospheres, 111, D08S10, 2006.

Zhang, L. M., Moran, M. D., Makar, P. A., Brook, J. R., and Gong, S. L.: Modelling gaseous

dry deposition in AURAMS: a unified regional air-quality modelling system, Atmospheric

Environment, 36, 537–560, 2002.

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.,

94

Weimer, S., Demerjian, K., Williams, P., Bower, K., Bahreini, R., Cottrell, L., Griffin,

R. J., Rautiainen, J., Sun, J. Y., Zhang, Y. M., and Worsnop, D. R.: Ubiquity and domi-

nance of oxygenated species in organic aerosols in anthropogenically-influenced Northern

Hemisphere midlatitudes, Geophysical Research Letters, 34, L13 801, 2007.

Zheng, W., Flocke, F. M., Tyndall, G. S., Swanson, A., Orlando, J. J., Roberts, J. M.,

Huey, L. G., and Tanner, D. J.: Characterization of a thermal decomposition chemical

ionization mass spectrometer for the measurement of peroxy acyl nitrates (PANs) in the

atmosphere, Atmospheric Chemistry and Physics Discussions, 11, 8461–8513, 2011.

95

Fig. 3.1.— Lag time vs. correlation plot for vertical wind speed (w) with both HCHO and

virtual temperature (Tv). Data shown is an average of all half-hour flux intervals from 6:00

to 18:00 on 30 August.

96

Fig. 3.2.— Average cospectra of HCHO and virtual temperature with vertical wind speed

during half-hour periods from 10:00 to 14:00 on 13 August (Day 225), 15 August (Day 227),

and 30 August (Day 242). Cospectra were binned into 50 bins spaced equally in logarithmic

frequency space, and each bin was averaged. The positive w′HCHO′ points (closed circles)

designate a positive covariance, whereas negative w′HCHO′ points (open circles) designate

negative covariance.

97

Fig. 3.3.— (a) Averaged, frequency-weighted, covariance-normalized cospectra for half-hour

periods from 10:00 to 14:00 over entire measurement period. Cospectra were binned into

200 bins spaced equally in logarithmic frequency space, and each bin was averaged. (b)

Averaged ogives for half-hour periods from 10:00 to 14:00 over entire measurement period.

98

Fig. 3.4.— (a) Hourly box-and-whisker plots of HCHO flux over entire measurement period.

Black and white targets denote the hourly medians, thick black lines denote the interquartile

range, and thin black lines denote the full range. (b) Diurnal medians of HCHO vertical

concentration profiles from 12 - 22 August.

99

Fig. 3.5.— Diurnal medians of contributions to HCHO flux in the base case model.

100

Fig. 3.6.— Comparison of model results with measured HCHO fluxes. Grey dots denote the

1 hr binned median of measured flux, while thick gray lines denote the interquartile range

of measured flux for each bin, and thin gray lines denote the entire range. Base refers to the

unaltered model result. VOC-I and VOC-II refer to the base model with tenfold simulated

increases in MBO and MT respectively. E350 refers to the base model with a direct HCHO

emission rate from PPine of 350 ng (g dw)−1.

101

3a Supplementary materials

3a.1 HCHO Permeation Tube Calibration

The FTIR gas cell pressure and temperature were held near ambient. Spectra were

acquired and integrated for 1 hr at 1 cm−1 resolution. The concentration of formaldehyde

(HCHO) in the calibration mixture was quantified using HITRAN absorption line lists (Roth-

man et al., 2005) and a multi-component least squares fitting algorithm (Griffith, 1996). The

C-H stretch region of 2620-2920 cm−1 was chosen as the fitting region. The permeation rate

determined via FTIR spectroscopy was found to be significantly lower (∼50%) than the rate

determined via mass loss over time.

3a.2 Error in Flux Measurements

Error resulting from instrument response time was estimated by:

∆w′HCHO′

w′HCHO′meas

= 2πfmτHCHO (3a.1)

where fm is the frequency maximum of the weighted cospectrum (Fig. 3.3a) and τHCHO

is the instrument response time, determined from the decay observed upon introducing a

sharp concentration change at the front of the inlet (Horst, 1997). The measured instrument

response time of ∼0.28 s resulted in an estimated error of ≤ 5%. Error from instrumental

noise as a result of the discrete method of detection (i.e. shot noise) was estimated by the

following equation:

∆w′HCHO′

w′HCHO′meas

=σ2wσ

2HCHO

fsT(3a.2)

where σ2x is the measurement variance in x, fs is the sampling frequency, and T is the length

of the sampling period (Lenschow and Kristensen, 1986; Ritter et al., 1990). This typically

102

resulted in an error of < 4%. Error due to the separation between the HCHO inlet and sonic

anemometer was determined by the following cospectral transfer function:

Ts(f) = e−9.9

(fsU

)1.5

(3a.3)

where f is the cospectral frequency, s is the sensor separation, and U is the wind speed

(Moore, 1986). During BEACHON-ROCS, the separation was ∼0.5 m and wind speeds

typically varied from 0.5 to 4.5 m s−1, leading to errors ranging from 0.84% to 6.6%. Error

resulting from dampening inside the inlet was predicted by the following cospectral transfer

function:

Ts(f) = e−(2πf)2ΛLa

u2 (3a.4)

where f is the cospectral frequency, Λ is the attenuation coefficient, L is the length of

tubing, a is the radius of the tubing inner diameter, and u is the flow rate through the

inlet (Massman, 1991). Dampening was considered for both the main inlet line (Λ = 1,

u = 18.7 m s−1, L = 38.5 m) and the internal instrument tubing (Λ = 20, u = 3.5 m s−1,

L = 1 m), resulting in a total error of 1.3%. Error resulting from the lag time calculation

was calculated using the error in the fitted linear trends. Fluxes were calculated for the lag

time range of the 1σ error in the trends, then the standard deviation over these fluxes were

taken to be the lag contribution to the error. Median daytime error due to lag time was

∼20%.

3a.3 HCHO Production via Methylperoxy Radical

PA concentrations were calculated with a steady-state model, based on observations

in a similar coniferous forest, which predicts the PA steady state concentration ([PA]ss) by

the steady-state equation:

[PA]SS =PMVK + PMACR + PCH3CHO + PMGLY + PBACE + PPAN

LNO2 + LNO + LHO2 + LRO2

(3a.5)

103

However, this equation simplifies significantly upon neglect of isoprene oxidation products,

as isoprene has been observed to be low at this site (Kim et al., 2010):

[PA]SS =kacetal·OH [CH3CHO][OH] + kd[PAN ]

kPA·NO2 [NO2] + kPA·NO[NO] + kPA·HO2 [HO2] + kPA·RO2 [RO2](3a.6)

Reactions of PA radical with NO and RO2 have a unity yield of methylperoxy radical

(CH3O2) (Atkinson et al., 2006), while reaction with HO2 has a 40% yield through methyl

hydrogen peroxide (Hasson et al., 2004; Jenkin et al., 2007; Dillon and Crowley, 2008).

CH3O2 has a net unity yield of HCHO via reactions with NO, RO2 (Tyndall et al., 2001;

Atkinson et al., 2006), and HO2 (Fried et al., 1997), which permits us to assume all CH3O2

radicals quickly react to form HCHO. This leads to a production rate of HCHO from PA

radicals of:

PPAHCHO = PPA

CH3O2= [PA]SS ·(kPA·NO[NO]+kPA·RO2 [RO2]+0.4·kPA·HO2 [HO2]) (3a.7)

Similarly, OH-initiated oxidation of methane produces CH3O2 radicals (and thus HCHO)

with unity yield. Methane concentrations were assumed to be constant at 1.7 ppmv.

3a.4 Aerodynamic and Laminar Sublayer Resistance

Ra is the aerodynamic resistance, the resistance to transfer between the measurement

height and the surface (Monteith, 1965).

Ra =u(z − d)

u2∗− ΨH(ξ)−ΨM (ξ)

k · u∗(3a.8)

where z is measurement height A.G.L., d is the displacement height (2/3 × h), u(x) is

the wind speed at height x, k is the von Karman constant (∼0.4), and ΨH and ΨM are

the sensible heat and momentum integrated stability corrections (Dyer, 1974), which are

a function of the stability parameter xi = (z − d)/L, where L is the Obukhov length.

Typical values of Ra range from 8 s m−1 at mid-day to 30 s m−1 at night. Rb is the laminar

104

sublayer resistance, the resistance to molecular diffusive transport through the viscous layer

surrounding leaf surfaces (Jensen and Hummelshoj, 1995, 1997).

Rb =ν

u∗ ·DHCHO·[

100 · l · u∗LAI2 · ν

]1/3(3a.9)

ν is the pressure-corrected kinematic viscosity of air (1.7 ×10−5 m2 s−1), DHCHO is the

pressure-corrected diffusion coefficient for HCHO (1.7× 10−5 m2 s−1) (Wesely, 1989; Mass-

man, 1998), and l is the “characteristic length scale”, or thickness, of a pine needle (1 mm).

Typical values of Rb range from 16 s m−1 at mid-day to 32 s m−1 at night.

3a.5 Bibliography

Atkinson, R. and Arey, J.: Gas-phase tropospheric chemistry of biogenic volatile organic

compounds: a review, Atmospheric Environment, 37, S197–S219, 2003.

Atkinson, R., Baulch, D. L., Cox, R. A., Crowley, J. N., Hampson, R. F., Hynes, R. G.,

Jenkin, M. E., Rossi, M. J., and Troe, J.: Evaluated kinetic and photochemical data for

atmospheric chemistry: Volume II - gas phase reactions of organic species, Atmospheric

Chemistry and Physics, 6, 3625–4055, 2006.

Carrasco, N., Doussin, J. F., O’Connor, M., Wenger, J. C., Picquet-Varrault, B., Durand-

Jolibois, R., and Carlier, P.: Simulation chamber studies of the atmospheric oxidation

of 2-methyl-3-buten-2-ol: Reaction with hydroxyl radicals and ozone under a variety of

conditions, Journal of Atmospheric Chemistry, 56, 33–55, 2007.

Dillon, T. J. and Crowley, J. N.: Direct detection of OH formation in the reactions of

HO2 with CH3C(O)O2 and other substituted peroxy radicals, Atmospheric Chemistry

and Physics, 8, 4877–4889, 2008.

Dyer, A.: A Review of Flux-Profile Relationships, Boundary-Layer Meteorology, 7, 363–372,

1974.

105

Fried, A., McKeen, S., Sewell, S., Harder, J., Henry, B., Goldan, P., Kuster, W., Williams,

E., Baumann, K., Shetter, R., and Cantrell, C.: Photochemistry of formaldehyde during

the 1993 Tropospheric OH Photochemistry Experiment, Journal of Geophysical Research-

Atmospheres, 102, 6283–6296, 1997.

Griffith, D. W. T.: Synthetic calibration and quantitative analysis of gas-phase FT-IR spec-

tra, Applied Spectroscopy, 50, 59–70, 1996.

Hasson, A. S., Tyndall, G. S., and Orlando, J. J.: A product yield study of the reaction

of HO2 radicals with ethyl peroxy (C2H5O2), acetyl peroxy (CH3C(O)O2), and acetonyl

peroxy (CH3C(O)CH2O2) radicals, Journal of Physical Chemistry A, 108, 5979–5989,

2004.

Horst, T. W.: A simple formula for attenuation of eddy fluxes measured with first-order-

response scalar sensors, Boundary-Layer Meteorology, 82, 219–233, 1997.

Hottle, J. R., Huisman, A. J., Digangi, J. P., Kammrath, A., Galloway, M. M., Coens,

K. L., and Keutsch, F. N.: A Laser Induced Fluorescence-Based Instrument for In-Situ

Measurements of Atmospheric Formaldehyde, Environmental Science and Technology, 43,

790–795, 2009.

Jenkin, M. E., Hurley, M. D., and Wallington, T. J.: Investigation of the radical product

channel of the CH3C(O)O2 + HO2 reaction in the gas phase, Physical Chemistry Chemical

Physics, 9, 3149–3162, 2007.

Jensen, N. O. and Hummelshoj, P.: Derivation of Canopy Resistance for Water-Vapor Fluxes

over a Spruce Forest, Using a New Technique for the Viscous Sublayer Resistance, Agri-

cultural and Forest Meteorology, 73, 339–352, 1995.

Jensen, N. O. and Hummelshoj, P.: Erratum to ”Derivation of canopy resistance for water

106

vapor fluxes over a spruce forest, using a new technique for the viscous sublayer resistance”,

Agricultural and Forest Meteorology, 85, 289–289, 1997.

Kim, S., Karl, T., Guenther, A., Tyndall, G., Orlando, J., Harley, P., Rasmussen, R., and

Apel, E.: Emissions and ambient distributions of Biogenic Volatile Organic Compounds

(BVOC) in a ponderosa pine ecosystem: interpretation of PTR-MS mass spectra, Atmo-

spheric Chemistry and Physics, 10, 1759–1771, 2010.

Lee, A., Goldstein, A. H., Kroll, J. H., Ng, N. L., Varutbangkul, V., Flagan, R. C., and

Seinfeld, J. H.: Gas-phase products and secondary aerosol yields from the photooxidation

of 16 different terpenes, Journal of Geophysical Research-Atmospheres, 111, 2006.

Lenschow, D. H. and Kristensen, L.: Sampling Errors in Flux Measurements of Slowly

Depositing Pollutants, Journal of Climate and Applied Meteorology, 25, 1785–1787, 1986.

Massman, W. J.: The Attenuation of Concentration Fluctuations in Turbulent-Flow through

a Tube, Journal of Geophysical Research-Atmospheres, 96, 15 269–15 273, 1991.

Massman, W. J.: A review of the molecular diffusivities of H2O, CO2, CH4, CO, O3, SO2,

NH3, N2O, NO3 and NO2 in air, O2 and N2 near STP, Atmospheric Environment, 32,

1111–1127, 1998.

McManus, J. B., Zahniser, M. S., Nelson, J. D. D., Shorter, J. H., Herndon, S., Wood, E.,

and Wehr, R.: Application of quantum cascade lasers to high-precision atmospheric trace

gas measurements, Optical Engineering, 49, 111 124–11, 2010.

Monteith, J.: Evaporation and Environment, Symposia of the Society for Experimental

Biology, 19, 205–234, 1965.

Moore, C. J.: Frequency-Response Corrections for Eddy-Correlation Systems, Boundary-

Layer Meteorology, 37, 17–35, 1986.

107

Ritter, J. A., Lenschow, D. H., Barrick, J. D. W., Gregory, G. L., Sachse, G. W., Hill, G. F.,

and Woerner, M. A.: Airborne Flux Measurements and Budget Estimates of Trace Species

over the Amazon Basin during the GTE/ABLE 2B Expedition, Journal of Geophysical

Research-Atmospheres, 95, 16 875–16 886, 1990.

Rothman, L. S., Jacquemart, D., Barbe, A., Benner, D. C., Birk, M., Brown, L. R., Carleer,

M. R., Chackerian, C., Chance, K., Coudert, L. H., Dana, V., Devi, V. M., Flaud, J. M.,

Gamache, R. R., Goldman, A., Hartmann, J. M., Jucks, K. W., Maki, A. G., Mandin,

J. Y., Massie, S. T., Orphal, J., Perrin, A., Rinsland, C. P., Smith, M. A. H., Tennyson,

J., Tolchenov, R. N., Toth, R. A., Vander Auwera, J., Varanasi, P., and Wagner, G.: The

HITRAN 2004 molecular spectroscopic database, Journal of Quantitative Spectroscopy &

Radiative Transfer, 96, 139–204, 2005.

Tyndall, G. S., Cox, R. A., Granier, C., Lesclaux, R., Moortgat, G. K., Pilling, M. J.,

Ravishankara, A. R., and Wallington, T. J.: Atmospheric chemistry of small organic

peroxy radicals, Journal of Geophysical Research-Atmospheres, 106, 12 157–12 182, 2001.

Weibring, P., Richter, D., Walega, J. G., and Fried, A.: First demonstration of a high

performance difference frequency spectrometer on airborne platforms, Optics Express, 15,

13 476–13 495, 2007.

Wesely, M. L.: Parameterization of Surface Resistances to Gaseous Dry Deposition in

Regional-Scale Numerical-Models, Atmospheric Environment, 23, 1293–1304, 1989.

Wisthaler, A., Apel, E. C., Bossmeyer, J., Hansel, A., Junkermann, W., Koppmann, R.,

Meier, R., Mller, K., Solomon, S. J., Steinbrecher, R., Tillmann, R., and Brauers, T.:

Technical Note: Intercomparison of formaldehyde measurements at the atmosphere simu-

lation chamber SAPHIR, Atmospheric Chemistry and Physics, 8, 2189–2200, 2008.

108

Fig. 3a.1.— Average cospectra of HCHO and virtual temperature with vertical wind speed

during half-hour periods from 10 AM to 2 PM over entire measurement period. To com-

pensate for noise, cospectra were binned into 200 bins spaced equally in frequency, and each

bin was averaged. The red dot-dashed region in w’HCHO’ denotes negative contributions

to flux. The positive w’HCHO’ points designate a positive covariance, whereas negative

w’HCHO’ points designate negative covariance.

109

Fig. 3a.2.— Time series of HCHO flux over entire flux measurement period (11 - 30 August).

Data has been corrected for unstationary conditions.

110

Fig. 3a.3.— Temperature and PAR dependence of HCHO flux during BEACHON-ROCS.

111

Tab

le3a

.1:

Com

par

ison

ofd

etec

tion

lim

its

and

tim

ere

solu

tion

ofH

CH

Om

easu

rem

ent

tech

niq

ues

.

Tec

hn

iqu

e3σ

Det

ecti

onL

imit

Ref

eren

ce

Qu

antu

mC

asca

de

Las

erS

pec

tros

copy

∼96

pp

t vin

1s

McM

anu

set

al.

(2010)

Tu

nab

leD

iod

eL

aser

Sp

ectr

osco

py

∼18

0p

pt v

in1

sW

eib

rin

get

al.

(2007)

Pro

ton

Tra

nsf

erR

eact

ion

-Mas

sS

pec

trom

etry

300

pp

t vin

2s

Wis

thal

eret

al.

(2008)

Han

tzsc

hD

eriv

itiz

atio

n75

pp

t vin

1m

inW

isth

aler

etal

.(2

008)

Mad

ison

Ti:

Sap

ph

ire

LIF

∼51

pp

t vin

1s

Hot

tle

etal

.(2

009)

Mad

ison

FIL

IF(fi

eld

)∼

300

pp

t vin

1s

this

wor

k

Mad

ison

FIL

IF(l

abor

ator

y)

∼25

pp

t vin

1s

this

wor

k

112

Table 3a.2: Chemical production and loss rates and yields for zero-dimensional box model.

All rate constants have units of cm3 molec−1 s−1 unless otherwise specified.

Reaction HCHO Yield Rate Constant Rate Constant

Yield Reference T = temperature (K) Reference

MBO + OH 0.33 a 8.2× 10−12 × e610/T d

α− pinene + OH 0.19 b 1.2× 10−11 × e440/T d

β − pinene + OH 0.51 c 7.89× 10−11 b

Methanol + OH 1.0 d 2.85× 10−12 × e−345/T d

3− carene + OH 0.28 c 8.68× 10−11 b

Acetaldehyde + OH 1.0 d 4.4× 10−12 × e365/T d

CH4 + OH 1.0 d 1.85× 10−12 × e−1690/T d

PAN → PA + NO2 - - ∗ k0 : 4.9× 10−3 × e−12100/T d∗ k∞ : 5.43× 1016 × e−13830/T

∗∗ Fc : 0.31

PA + NO2 - - ∗ k0 : 2.7× 10−28 × (T/300)7.1 d∗ k∞ : 1.2× 10−11 × (T/300)0.9

∗∗ Fc : 0.31

PA + NO 1.0 d 7.5× 10−12 × e290/T d

PA + HO2 ∼0.4 e,f,g 5.2× 10−13 × e980/T d

PA + RO2 1.0 d 2.0× 10−12 × e500/T d

MBO + O3 0.5 a 1.0× 10−17 d

α− pinene + O3 0.28 c 6.3× 10−16 × e−580/T d

β − pinene + O3 0.65 c 1.5× 10−17 b

3− carene + O3 0.25 c 3.61× 10−17 b

HCHO + OH - - 5.4× 10−12 × e135/T d

a. Carrasco et al. (2007) ∗ Units: s−1

b. Atkinson and Arey (2003) ∗∗ Unitless

c. Lee et al. (2006)

d. Atkinson et al. (2006)

e. Hasson et al. (2004)

f. Jenkin et al. (2007)

g. Dillon and Crowley (2008)

113

Table 3a.3: Noon model case results in µg m−2 hr−1 by species.

Species Base VOC-I E350 VOC-II

Production:

Litter Emission 8.43 (25%) 8.43 (7%) 8.43 (6%) 8.43 (15%)

MBO + OH 8.35 (24%) 83.5 (74%) 8.35 (6%) 8.35 (14%)

PPine Emission 4.24 (12%) 4.24 (4%) 105 (78%) 4.24 (7%)

PA 3.99 (12%) 3.99 (4%) 3.99 (3%) 3.99 (7%)

CH4 + OH 2.76 (8%) 2.76 (2%) 2.76 (2%) 2.76 (5%)

CH3CHO + OH 2.10 (6%) 2.10 (2%) 2.10 (2%) 2.10 (3%)

CH3OH + OH 1.16 (3%) 1.16 (1%) 1.16 (1%) 1.16 (2%)

β − pinene + OH 0.67 (2%) 0.67 (1%) 0.67 (<1%) 6.69 (12%)

α− pinene + OH 0.61 (2%) 0.61 (1%) 0.61 (<1%) 6.11 (11%)

Other MT + OH 0.49 (1%) 0.49 (<1%) 0.49 (<1%) 4.93 (9%)

MBO + O3 0.48 (1%) 4.80 (4%) 0.48 (<1%) 0.48 (1%)

3− carene + OH 0.33 (1%) 0.33 (<1%) 0.33 (<1%) 3.26 (6%)

Other MT + O3 0.32 (1%) 0.32 (<1%) 0.32 (<1%) 3.17 (6%)

α− pinene + OH 0.14 (<1%) 0.14 (<1%) 0.14 (<1%) 1.41 (2%)

β − pinene + O3 0.04 (<1%) 0.04 (<1%) 0.04 (<1%) 0.40 (1%)

3− carene + O3 0.03 (<1%) 0.03 (<1%) 0.03 (<1%) 0.30 (1%)

Loss:

Dry Deposition -19.30 (69%) -22.37 (72%) -27.19 (76%) -19.30 (69%)

Photolysis -4.90 (17%) -4.90 (16%) -4.90 (14%) -4.90 (17%)

OH -3.84 (14%) -3.84 (12%) -3.84 (11%) -3.84 (14%)

114

Chapter 4

Observations of Glyoxal and Formaldehyde as

Metrics for the Anthropogenic Impact on Rural

Photochemistry

4.1 Introduction

The oxidation of volatile organic compounds (VOCs) is directly coupled to the pro-

duction of tropospheric ozone, a major atmospheric pollutant correlated with increased inci-

dences of poor respiratory health and crop damage (Stieb et al., 2000; Mauzerall and Wang,

2001), and production of secondary organic aerosol (Dockery et al., 1993; Ostro, 1993; Laden

et al., 2000). Production of tropospheric ozone is dependent on two attributes: the reactive

mixture of VOCs in the atmosphere and the oxidation pathways & mechanisms of these

VOCs. Due to the wide variety of VOCs in the atmosphere, both explicit measurement of

the VOC mixture and a quantitative understanding of the oxidation pathways of the VOC

mixture are challenging. Observations of oxidation products are fundamental to testing and

improving our understanding of VOC oxidation. Measurements of oxidation products spe-

cific to only one VOC (e.g. MVK from isoprene oxidation) are valuable as are measurements

of species that are produced from oxidation of many VOCs (e.g. formaldehyde, a general

115

VOC oxidation tracer). By comparing general VOC oxidation tracers that are produced

from both anthropogenic and biogenic VOCs (AVOCs/BVOCs, respectively), but whose

relative yields vary between AVOCs and BVOCs, it is possible to determine which type of

reactive VOC is more important in a given air mass. Through observation of such oxidation

products, we can obtain a metric to identify changes in the overall reactive VOC mixture.

Formaldehyde (HCHO) and glyoxal (Gly) are ubiquitous oxidized VOCs (OVOCs)

formed as intermediates in the VOC-HOx-NOx cycle (HOx = HO + HO2, NOx = NO +

NO2), the catalytic photochemical cycle responsible for VOC oxidation in the atmosphere

(Fried et al., 1997; Lee et al., 1998; Tan et al., 2001). HCHO is produced in the oxidation

of nearly all VOCs and is often used as a tracer of overall VOC oxidation. It is also directly

emitted from various sources (Garcia et al., 2006)(see Chap. 3). Gly is similarly formed from

the oxidation of many VOCs, such as alkene and aromatic species. Additionally, Gly has

virtually no primary sources (Volkamer et al., 2005), except from biomass burning (McDon-

ald et al., 2000; Hays et al., 2002; Fu et al., 2008), which allows it to be used as a measure

of the rate of photochemical oxidation (Garcia et al., 2006). Both Gly and HCHO have

similar midday lifetimes on the order of a few hours (Atkinson, 2000). Due to these differ-

ences in sources and similarity in sinks, the Gly/HCHO ratio (RGF) has been proposed to

be indicative of changes in the atmospheric VOC mixture. Satellite retrievals and modeling

studies suggest higher values of RGF (4-6%) in biogenically-influenced regions and lower val-

ues (<4%) in anthropogenically-influenced regions (Wittrock et al., 2006; Myriokefalitakis

et al., 2008; Vrekoussis et al., 2009, 2010). However, there have been no reported inves-

tigations of using RGF as a tracer of local VOC oxidation or tracer of the reactive VOC

composition.

In this work, we present simultaneous, fast (<1 min), online, in situ observations of

both Gly and HCHO during two rural field intensives in Pinus Ponderosa forests, BEARPEX

116

2009 and BEACHON-ROCS. Additionally, we compare and discuss RGF values for these

campaigns, specifically the causes of variability of RGF within each dataset, with RGF re-

ported in the literature. Finally, we discuss multiple events during these campaigns in the

context of the fate of alkyl peroxy radicals (RO2) as well as the role of anthropogenic influ-

ence on rural regions via anthropogenic VOCs via influence via NOx.

4.2 Experimental

4.2.1 Site Information

The Biosphere Effects on AeRosols and Photochemistry EXperiment (BEARPEX)

2009 took place in a Sierra Pacific Industries Pinus Ponderosa plantation (canopy height

≈ 9 m; leaf area index (LAI) ≈ 3.7) in the Sierra Nevada Mountains (38°53’42.9” N,

120°37’59.7” W, 1315 m) near the Blodgett Forest Research Station from 15 June - 31 July

2009. This site has been described in detail elsewhere (Goldstein et al., 2000; Dillon et al.,

2002; Dreyfus et al., 2002) and exhibits a regular diurnal wind profile driven by the moun-

tain anabatic/katabatic winds. Local emissions are primarily 2-methyl-3-buten-2-ol (MBO)

and monoterpenes (MT) with consistent midday arrival of advective isoprene and oxidation

products (Dreyfus et al., 2002) and late afternoon/evening arrival of the Sacramento urban

plume (Dillon et al., 2002). Additionally, this site had a significant understory with an

estimated height of 2 m with an estimated LAI of ∼1.9 (Wolfe and Thornton, 2011).

The Bio-hydro-atmosphere interactions of Energy, Aerosols, Carbon, H2O, Organics

& Nitrogen - Rocky Mountain Organic Carbon Study (BEACHON-ROCS) 2010 took place

in the Manitou Experimental Forest (MEF, 39°06’02” N, 105°06’05” W, 2286 m) from 1-

31 August 2010. MEF is a Central Rocky Mountains Pinus Ponderosa forest (canopy height

≈ 18.5 m; LAI ≈ 1.9) located ∼40 km northwest of Colorado Springs, CO and ∼70 km south

of Denver, CO. This site has been described previously (Kim et al., 2010)(see Chap. 3) and

117

exhibited minimal undergrowth and predominantly biogenic-influenced air.

4.2.2 Gly and HCHO Measurements

Both Gly and HCHO measurements were obtained via laser induced photolumines-

cence by two different instruments: the Madison-Laser Induced Phosphorescence (Mad-LIP)

instrument and the Madison Fiber Laser-Induced Fluorescence (FILIF) instrument, respec-

tively. Details for these instruments can be found elsewhere (Huisman et al., 2008; Hottle

et al., 2009)(see Chap. 3) but will be briefly described here. The Mad-LIP instrument uses a

440 nm narrow-bandwidth (∼0.06 nm), Nd:YAG-pumped, doubled Ti:Sapphire laser (Pho-

tonics Industries, TU series), and FILIF uses a 353 nm narrow-bandwidth pulsed fiber laser

(NovaWave Technologies, TFL Series). The lasers are used to excite either Gly or HCHO

photoluminescence in the center of a 32-pass White-type multipass cell. Laser scatter was

minimized by the presence of carbon black-coated baffling as well as a light trap opposite

the detector. Photoluminescence from the analytes was collimated, passed through an op-

tical filter (HCHO: 390 nm longpass, Gly: 520 nm bandpass), and then focused onto the

entire active area of a photon-counting photomultiplier tube. The photoluminescence signal

was electronically gated to optimize detection of the photoluminescence signal and reduce

laser scatter, thus maximizing signal/noise. Since this electronic gating also preferentially

detected photons only in a specific time window after the laser pulse, only photons with

the lifetime of the luminescence were observed, increasing selectivity. Any remaining back-

ground was subtracted by dithering the laser periodically between near wavelengths of high

and low absorption cross-section. The difference between these two signals is proportional to

the analyte concentration. Weekly Mad LIP calibrations were performed using a gas stan-

dard quantified in the field via cavity ringdown spectroscopy as described by Huisman et al.

(Huisman et al., 2008), and weekly FILIF calibrations were performed using an FTIR cross-

calibrated permeation source as described in Chap. 3a.1. Gly and HCHO gradients were

118

measured by alternately sampling from four ∼30 m 3/8” ID PTFE (Gly) or PFA (HCHO)

inlets placed at heights of 25.1 m, 17.7 m, 8.5 m, and 1.6 m. Inlet comparison testing was

performed for both instruments and observed no detectable inlet artifacts (see Chap. 2.4).

4.2.3 Other Measurements

Details of the sensors for the meteorological measurements during BEARPEX 2009 can

be found elsewhere (Goldstein et al., 2000). Carbon monoxide (CO) was measured via a gas

correlation infrared spectrometer (Teledyne, Model 300E). Concentrations of benzene, sum

total of 2-methyl-3-buten-2-ol and isoprene (MBO+Isoprene: used as a BVOC tracer), and

BVOC first generation oxidation products (m/z 71) were measured via quadrupole proton

transfer reaction-mass spectrometry (PTR-MS).

Details of many of the sensors for measurements during BEACHON-ROCS can also

be found in Chap. 3. Benzene, MBO+Isoprene, m/z 71 and m/z 95 (C6H7O+, which likely

corresponds to phenol) concentrations were measured via proton transfer reaction-time of

flight-mass spectrometry (PTR-TOF-MS) (Jordan et al., 2009; Ruuskanen et al., 2011).

Additional VOCs, including isoprene, toluene, benzene, and the toluene:benzene ratio (RTB),

were measured with a Total Organic Carbon Analyzer (Apel et al., 2003, 2010; Hornbrook

et al., 2011). Nitrogen oxide (NO) was measured via chemiluminescence (Eco Physics AG,

Model CLD-88Y).

4.3 Observations

Figure 4.1 shows the diurnal profiles of Gly and HCHO concentrations, Gly/HCHO

ratios (RGF), and wind direction for both sites, while Fig. 4.2- 4.3 shows Gly and HCHO

concentrations for each campaign. Gly and HCHO concentrations at both sites (Fig. 4.1c-f)

exhibited a moderately variable diurnal profile, with typical concentration maxima in the

early evening and minima around sunrise. Average Gly and HCHO concentrations were

119

significantly higher during BEARPEX 2009 than during BEACHON-ROCS, which concurs

with higher OH reactivity during the BEARPEX 2009 campaign (Fig. 4.4). RGF during each

campaign (Fig. 4.1a-b) showed a remarkably consistent diurnal profile peaking at midday

that is virtually independent of Gly or HCHO concentration changes.

The diurnal wind direction profile during BEARPEX 2009 (Fig. 4.1h) was very con-

sistent, with wind from the east overnight and the west during the day. During BEACHON-

ROCS, the diurnal wind direction profile (Fig. 4.1g) was best represented by two regimes.

Winds consistently originated from the south at night, while during the day the dominant

wind direction was either southwesterly (∼60%) or northeasterly (∼40%). Diurnal Gly and

HCHO concentrations were higher in the afternoon for the northeasterly regime, while RGF

were similar between the regimes (Fig. 4.5). Days in the northeast regime were more likely

to include transport events, indicated by a sudden transition from low to high concentrations

of many species. These events during BEACHON-ROCS were observed as intermittent but

regular occurrences, while only a two-day series of events with higher concentrations oc-

curred during BEARPEX 2009. We will use a selection of representative events to probe the

chemistry driving the behavior of RGF.

4.3.1 BEARPEX 2009 16-17 July, 2009: Mammoth Fire Incident (MFI)

During BEARPEX 2009, only one event, consisting of two consecutive days (16-17

July, 2009), was observed that showed deviation from the regular diurnal trends in HCHO,

Gly and RGF. To demonstrate the irregularity of this event, the days of the event are shown

as blue (16 July) and red (17 July) traces in the right panels of Fig. 4.1(b, d, f). The event

corresponds to the Mammoth Fire Incident (MFI) which occurred in the American River

Canyon east of Auburn, CA (38.93° N, 120.99° W, ∼400 m). The MFI began on 16 July,

2009 at 14:33 and was reported contained on 18 July, 2009 at 19:30 (http://bof.fire.ca.

gov/incidents/incidents_details_info?incident_id=340).

120

Figure 4.6 shows concentrations of Gly, HCHO, RGF, and other relevant species and

parameters during this event. The morning and afternoon of 16 July had Gly and HCHO

concentrations and RGF typical for BEARPEX 2009. At ∼19:00, there was a sharp increase

in RGF, as well as in concentrations of Gly, HCHO, benzene, acetonitrile, and CO (see Ta-

ble 4.1). The timing of these concentration increases is consistent with the arrival of the

plume from the MFI site modeled by HYSPLIT forward trajectories (Fig. 4.7)(Draxler and

Rolph, 2011; Rolph, 2011), which predict arrival of the plume at the BEARPEX 2009 site be-

tween ∼18:30 and ∼19:30. RGF, Gly, HCHO, benzene, acetonitrile, and CO concentrations

all exhibited similar decays as the evening progressed. No sharp rise was observed in oxida-

tion products specific to BVOCs, as judged by m/z 71, while MBO+Isoprene (representative

of BVOC emissions) followed the campaign-averaged evening decrease in concentration. The

coincidence of the rises in acetonitrile, a tracer of biomass burning (Holzinger et al., 1999,

2005), and CO, a tracer of combustion (Khalil and Rasmussen, 1988), with the rise in Gly,

HCHO and RGF is consistent with influence from the MFI. This rise in RGF is noteworthy,

as it was the only time during the 20 days of Gly and HCHO measurements at BEARPEX

2009 that showed a pronounced and rapid change in RGF. This fact demonstrates that RGF

was clearly enhanced due to a biomass-burning plume. This effect is consistent with satellite

retrievals (Vrekoussis et al., 2010) and can likely be attributed to differing primary emission

rates of Gly and HCHO from the burning event.

On 17 July, there was a second sharp increase in Gly, HCHO, benzene, acetonitrile,

and CO concentrations at ∼13:20, consistent with a HSYPLIT predicted arrival time of the

MFI plume (Fig. 4.8)(Draxler and Rolph, 2011; Rolph, 2011) of 13:15-13:45. Although both

HCHO and Gly strongly increased compared to the rest of campaign, the small increase in

RGF of 18% was not atypical for BEARPEX 2009 (see Fig. 4.1). This appears inconsistent

with the previous day as the increases in acetonitrile were very similar between the two

121

plume arrivals (Table 4.1), suggesting RGF should have risen as well due to biomass burning

influence. Additionally, HYSPLIT trajectory models (Fig. 4.8) predict that the MFI plumes

on the two days should have roughly the same age (∼3 h), which is comparable to the

lifetime of Gly and HCHO at that time. One fundamental difference between the events is

that MBO+Isoprene and m/z 71 also exhibited significant increases during the latter, which

should decrease the relative contribution from the biomass burning plume. In contrast to the

previous evening, the arrival of the MFI plume coincided closely with the arrival of isoprene

at the measurement site from a band of oak trees to the west, which makes it difficult to

discern the influences of the isoprene vs. the biomass burning plume. However, since no

other isoprene plume yielded Gly and HCHO concentrations as high as during this day,

combined with the large concentrations of acetonitrile, it is clear that biomass burning is a

significant influence.

Overall, the MFI event shows that biomass burning can influence RGF, but not always

noticeably. Acetonitrile concentrations from the MFI plume for the two days are quite

similar, but RGF is very different. This may point to differences in the emission from a

biomass burning event as it evolves, as the plume on the first day was from the freshly-started

fire while the plume the second day was when the fire was more than a day old. Additionally,

on the first day, the biomass burning plume arrived later during the day increasing the

lifetime of glyoxal and formaldehyde compared to the second day. It is possible that this

resulted in a stronger influence of glyoxal and formaldehyde from the biomass burning event

at the site.

4.3.2 BEACHON-ROCS 18 August 2010 (BN1)

Figure 4.9 shows an example of one of the events during BEACHON-ROCS, designated

BN1. At sunrise, concentrations of Gly, HCHO, and particularly m/z 71 began to rise.

This was coincident with the onset of photochemistry, BVOC emissions (indicated by the

122

increase in concentration of MBO+Isoprene), and vertical mixing estimated by u∗. At

∼09:30, a sudden shift in wind direction from south to northeast occurred, accompanied by

a fast drop in OVOC concentrations (Gly, HCHO, and m/z 71) but only a small decrease in

MBO+Isoprene and benzene concentrations. Acetonitrile concentration remained constant,

implying that they were already at/near the regional background. Constant MBO+Isoprene

concentrations at the same time as a dramatic decrease in glyoxal and no significant change

in OH concentration imply a shift from a more photochemically-aged airmass to one that is

less aged, at least in the presence of NO. RGF shows no discontinuity at this time, which

demonstrates that RGF is insensitive to the extent of airmass processing.

In the afternoon, RGF continued its regular diurnal increase until the wind direction

shifted to the south at ∼15:00, after which the ratio began to slowly decrease following

the average diurnal RGF pattern. At ∼17:00, both Gly and HCHO concentrations, which

had been nearly level all afternoon, roughly doubled within a few minutes. This fast rise

in Gly and HCHO was accompanied by a fast rise in benzene concentrations, a mild rise

in m/z 71, and no significant change in MBO+Isoprene or acetonitrile (see Table 4.1 for

values). The trend in many of these tracers was quite similar to those during the MFI

at BEARPEX 2009, with the exceptions of acetonitrile and RGF. RGF did not only lack

an increase, but it in fact continued decreasing on its normal diurnal trend. The lack of

change in acetonitrile concentrations with rising benzene concentrations implies that this

event arose from anthropogenic influence, rather than biomass burning as during the MFI.

Despite this noticeable increase in anthropogenic influence and sizable increases in both Gly

and HCHO concentrations, RGF was unaffected. This likely arises from the fact that the

same BVOCs still dominate the site’s reactive VOC mixture, and the rise in benzene likely

indicates increased anthropogenic influence via NO rather than via anthropogenic VOCs.

Combined with the lack of discontinutity in RGF in the morning, this suggests that RGF is

123

independent of NO concentrations.

4.3.3 BEACHON-ROCS 19 August 2010 (BN2)

Figure 4.10 shows another event, referred to as BN2. The morning had a similar rise

in MBO+Isoprene and OVOC concentrations as the morning of BN1, while Gly and HCHO

increased only slightly. The enhanced morning concentrations slowly decreased as vertical

mixing increased. RGF exhibited a morning profile very similar to that in BN1, despite

these differing conditions. At ∼14:30, Gly, HCHO and benzene once again rose very sharply,

whereas there was no discernable change in MBO+Isoprene and only a slight rise in m/z 71

and acetonitrile (see Table 4.1). Similar to BN1, RGF was unaffected by the sudden change

in concentrations. This reinforces that RGF did not exhibit responses to sudden changes in

airmass at the BEACHON-ROCS site. Finally, a rain event at ∼19:00 caused a fast decrease

in all VOC concentrations, including Gly, HCHO, and m/z 71. RGF remained surprisingly

constant during this transition, given the large differences in gas/liquid partitioning between

Gly and HCHO (Staudinger and Roberts, 1996; Ip et al., 2009).

4.3.4 BEACHON-ROCS 14 August 2010 (BN3)

During both BN1 and BN2, we observed no change in RGF despite sharp increases

in Gly, HCHO, and benzene (an anthropogenic tracer). However, benzene concentrations

remained fairly low, especially compared to the MFI event. During the BN3 event (Fig. 4.11),

the morning and afternoon had similar behavior as during BN1 and BN2, but the evening

showed different behavior as it was the only time substantially increased RGF values were

observed. Unfortunately, Gly measurements did not start until early evening of this day, but

the various other measured species indicate that the site was undergoing a similar change in

airmass as during events BN1 and BN2.

As with BN1, the morning exhibited a strong decrease in OVOC concentrations, sug-

124

gesting a transition to a less NO photochemically-processed airmass. VOC concentrations

remained mostly level until a wind direction change at∼13:15, when HCHO, MBO+Isoprene,

m/z 71 and benzene concentrations rose quickly, whereas acetonitrile stayed constant. Both

benzene and toluene rose significantly, and so did RTB, a measure of the processing of an-

thropogenic air . This confirms that anthropogenic VOCs in this new airmass were less

processed than the previous airmass. Isoprene also shows a marked increase at this time.

As this rise was quite similar to the rise in BN1, it is likely that there was a similar spike in

isoprene during BN1 as well.

Of particular interest is the region from ∼19:00 to ∼23:00 (see Fig. 4.12). Both HCHO

and Gly exhibited a series of brief (∼2-5 min) spikes in concentration. Even more significant

was that for the only time during either of the campaigns discussed in this work, RGF also

increased significantly over these short timescales, yielding the largest values of RGF during

both campaigns and approaching values observed in urban areas. At the same time, we

observed very high benzene and toluene concentrations, including the highest benzene con-

centrations during BEACHON-ROCS, a slight rise in both MBO+Isoprene concentrations

and m/z 71 with no observable change in acetonitrile concentrations. The highly elevated

benzene with constant acetonitrile shows that these spikes were caused by relatively fresh

and strong anthropogenic influence. This is further supported by measurements of m/z 95

(C6H7O+), which is attributed to phenol. Phenol has a much shorter lifetime than toluene

and BN3 was the only event for which we observed deviation of C6H7O+ from background

levels (Fig. 4.13), which demonstrates the very fresh anthropogenic influence. At the same

time, we observed the only deviation and fast change in RGF for the entire campaign. Regard-

less of the reasoning, the spikes during BN3 are distinct evidence that fresh anthropogenic air

(i.e. with reactive anthropogenic VOCs) has higher RGF, a trend opposite of that predicted

by satellites (see Chap. 4.1).

125

4.4 Discussion

The observation of changes in reactive VOC composition between BVOCs and biomass

burning/AVOCs on a very short timescale afforded the opportunity to evaluate trends in

RGF without potential instrumental changes, such as changes in calibration factors. The

results from BEARPEX 2009 and BEACHON-ROCS demonstrate that RGF is a tracer

sensitive to the reactive VOC composition of the measured air mass. RGF was observed to

be distinctly elevated for the oxidation of anthropogenic VOCs and conditionally elevated

for biomass-burning plumes, compared to the lower RGF values for oxidation of BVOCs.

Therefore, RGF represents a useful metric for the degree of anthropogenic influence on rural

areas via transport of anthropogenic VOCs. As RGF did not even vary for large and rapid

changes in absolute concentrations of glyoxal and formaldehyde, RGF is insensitive to the

extent of oxidative processing of air masses with similar reactive VOC composition. In the

following sections, we discuss the degree of agreement between surface and satellite retrievals

of RGF and the origin of the observed large changes reported in this study in absolute Gly

and HCHO concentrations at constant RGF (i.e. constant reactive VOC composition).

4.4.1 Gly:HCHO Ratios from Anthropogenic and Biogenic VOC Oxidation:

Surface and Satellite Values

Figure 4.14 shows a comparison between RGF in this work and the literature. Our mea-

surements during BEACHON-ROCS and BEARPEX 2009 typically had low values of RGF

(<2%). Sites in urban areas such as the Mexico City Metropolitan Area (MCMA)(Garcia

et al., 2006), Pasadena, CA (Washenfelder et al., 2011), and Bakersfield, CA (Henry et al.,

2011) typically had higher values of RGF (2.5 - 3.5%). This trend of increased RGF in

air masses with anthropogenic influence matches with observations for the transport events

during BEACHON-ROCS. Specifically, the BN3 event reached RGF values up to 4%. In

126

contrast to these surface based measurements from multiple locations, RGF values based

on GOME and SCIAMACHY satellite retrievals have been observed to be lower (<4%) in

urban/polluted areas as compared to rural areas (4-6%) with vegetated land cover (Vrek-

oussis et al., 2010). Satellite-driven global models tend to agree with these satellite re-

trievals (Myriokefalitakis et al., 2008). Median daytime RGF during BEACHON-ROCS and

BEARPEX as measured on the ground were one third or less than that retrieved by satellites

in rural regions. We believe this is not an LIF/LIP instrumental artifact, as RGF measured

using the same instrumentation in Bakersfield, CA were consistent with other urban field

sites (Henry et al., 2011). Furthermore, the trend of higher RGF values for anthropogenic

VOC mixtures over the very short timescales of individual events during these campaigns

is independent of instrumental changes, such as calibration factors, and even the absolute

value of RGF. The cause of the disagreement between surface measurements and satellite

retrievals is unclear but may be partially explained by the inherent limitations of comparing

a near-surface point measurement with column-averaged satellite retrieval. In rural forests,

there is evidence that direct emission is a major source of HCHO within the forest canopy

(Chap. 3) and may result in significantly lower RGF near the canopy. However, these emis-

sions are too small to be significant on the scale of the boundary layer and would only

result in decreased RGF at night when turbulence is low. Another possibility is that bound-

ary layer ratios are overall significantly lower than free tropospheric ratios. However, the

majority of formaldehyde and glyoxal are expected to be in the boundary layer, hence the

satellite retrievals should be strongly influenced by boundary layer values. The validity of

this hypothesis is difficult to ascertain, as there are have been no published simultaneous

measurements of Gly and HCHO in the free troposphere.

In addition to differences in absolute values in RGF between urban and rural sites,

there are significant differences in variability. As mentioned in Chap. 3, BEACHON-ROCS

127

and BEARPEX overall exhibited very consistent RGF. However, urban ground sites report

considerably higher variance. One potential explanation for these differences is the variability

in reactive VOC mixture resulting from influence between different emission sources among

these sites. Gly and HCHO concentrations at the BEACHON-ROCS and BEARPEX sites

are determined by local BVOC emissions (MBO, MT, isoprene). In urban areas (e.g. Mexico

City, Pasadena, and Bakersfield), the reactive VOC mixture is much more diverse, being a

mix of advected BVOC and emitted anthropogenic VOCs. As different VOCs have different

rates and yields of Gly and HCHO production, a fast-changing mix of VOCs, as would be

expected in an urban setting, could result in a widely variable RGF. Traditionally, primary

emissions of HCHO have been considered anthropogenic in nature (Garcia et al., 2006) and

variability in these primary emissions certainly contributes to the observed urban variability

in RGF. Recent evidence suggests significant primary biogenic HCHO sources may also exist

(Chap. 3) which would lower RGF in rural settings. Interestingly, sites at rural locations

such as George Smith State Park, GA (Lee et al., 1995) and Shenandoah National Park,

VA (Munger et al., 1995) have been reported to have RGF similar in variability and typical

values to urban areas. This is potentially due to the greater anthropogenic influence, near

these Southeastern US field sites compared to sparser population nearer the Western US

sites. The variable RGF values for the George Smith State Park and Shenandoah National

Park sites imply that these sites experience a greater deal of anthropogenic VOC influence

than the more remote BEARPEX and BEACHON-ROCS sites. However, it should be noted

that the DNPH measurement technique used by these investigators has been shown to be

prone to interferences and averages over a long time (Arnts and Tejada, 1989; Kleindienst

et al., 1998).

128

4.4.2 Anthropogenic Influence on BVOC Oxidation via NO

One of the most striking features in the BEACHON-ROCS RGF is the lack of a change

during quite significant changes in Gly and HCHO except during fresh anthropogenic VOC

influence. Additionally, overall BVOC (MBO+Isoprene) concentrations typically do not

significantly change during these events. As BVOC concentrations do not change while

oxidation products do, this suggests a difference in the oxidation pathways. For example,

the primary radical species controlling the oxidation of isoprene to MVK are OH and NO.

Lower NO concentrations in rural areas could result both in the RO2+NO products becoming

less significant and in reduced OH/HO2 ratios. Most models predict lower yields of Gly

and HCHO with decreasing NO (i.e. RO2 self reaction and especially RO2+HO2 become

increasingly dominant) (Galloway et al., 2011). The morning decreases of Gly and HCHO

observed during BEACHON-ROCS are consistent with a transition from a higher NO regime

to a lower NO regime, and the afternoon/evening increases are the opposite. RGF does not

change significantly during the fast increases and decreases of Gly and HCHO concentrations,

which is consistent with the similarity of the Gly and HCHO dependence on NO during

MBO and Isoprene oxidation based on common chemical mechanisms (e.g. University of

Leeds Master Chemical Mechanism)(Bloss et al., 2005). The available NO data, although

limited temporally and in sensitivity, generally agrees with this hypothesis (campaign midday

median: <100 pptv).

Figure 4.15 shows one such event during BEACHON-ROCS on 24 August for which

the low morning HCHO and Gly concentrations correlate with low NOx and in particular

low NO values between 40-70 pptv. Due to limitations of the data set the other events could

not be compared nor could the HO2/OH ratio be further analyzed. Therefore, it is not clear

how consistent this correlation was or whether the low HCHO and Gly concentrations are

the result of low NO concentrations and not the result of air masses that have accumulated

129

more oxidation products. The similar lifetime of MBO, isoprene and glyoxal make this

less likely, but from the available dataset we can only conclude that the observed behavior

is consistent with transitions between high and low NO conditions. To further characterize

these events, measurements of the RO2+HO2 oxidation products (i.e. hydroperoxides) would

be of significant importance, as these would be expected to increase relative to the RO2+NO

products. For example, in Fig. 4.15c, the RO2+HO2 channel changes from >70% to <10 %

during the low NO to high NO transition, as calculated from NO and HO2 concentrations

and rate constants from the University of Leeds Master Chemical Mechanism.

4.5 Conclusions

In this work, we present the first simultaneous forest online measurements of Gly

and HCHO during the BEARPEX 2009 and BEACHON-ROCS field intensives. Gly and

HCHO concentrations at both sites showed significant variability, while RGF diurnal values

were typically remarkably consistent and peak at midday. A fast change of airmass during

BEARPEX 2009 due to the Mammoth Fire Incident resulted in a sharp 79% increase in

RGF, the only large and/or rapid change or deviation in RGF from the diurnal cycle during

that campaign. This demonstrates that biomass burning influence can result in higher RGF.

Similarly, fast, strong increases in particular of very short-lived anthropogenic tracers during

BEACHON-ROCS coincided with fast increases in RGF, suggesting that fresh anthropogenic

air mass influence also results in higher RGF.

An increase in RGF was not observed during multiple events of weaker anthropogenic

influence, because the majority of the reactive VOCs in that airmass were still biogenic in

nature, although there is evidence that the NOx levels were substantially elevated. RGF was

not observed to change during these events, despite very rapid increases in both Gly and

HCHO. BVOC concentrations during these events were usually constant, suggesting that

the rise in Gly and HCHO, and other oxidation products, was more likely due to a shift

130

in the fate of the RO2 radical from a low NO regime to a higher one. Measurements of

low-NO oxidation products, such as hydroperoxides, would confirm this effect. This likely

illustrates a way in which anthropogenic influence can affect rural photochemistry by simply

altering the pathway by which local VOC emissions are processed. Ozone production is

sensitive to NOx, and large differences in SOA yields have been reported between high and

low NOx regimes (Ng et al., 2007; Lane et al., 2008; Chan et al., 2010). We propose that

RGF together with the absolute concentrations could be an important and useful metric

of the biogenic versus anthropogenic origin of a reactive VOC mixture as well as the NO

regime. Coincident measurements of RGF with a low NOx tracer, such as hydroperoxides

are necessary to confirm this.

RGF at these rural sites were observed to be typically lower than at any ground sites re-

ported in the literature. Possible anthropogenic VOC influence at other rural sites may have

contributed to higher ratios there, whereas other reports have been from urban campaigns.

While ground-based urban campaigns are consistent with satellite column retrievals of urban

areas, satellites retrievals show higher RGF in areas with greater BVOC influence. The trend

of increased RGF from anthropogenic reactive VOC mixtures and biomass burning compared

to biogenic reactive VOC mixtures from our work is robust due to the short timescales over

which the observed changes in RGF occurred. Similarly, observations in Bakersfield directly

preceding the BEACHON-ROCS campaign with the same instrumentation gave higher RGF

values (Henry et al., 2011). The cause of this discrepancy between the ground and satellite

retrievals is unclear. Considering the importance of satellite retrievals for global models, it

is important to resolve this discrepancy.

4.6 Acknowledgements

I would like to thank the National Science Foundation (ATM 0852406), the NCAR

BEACHON project, NASA SBIR Phase I and II, and the Camille and Henry Dreyfus Foun-

131

dation for support. I would like to thank Rainer Volkamer, Jose Garcıa, Joachim Stutz,

Catalina Tsai, Rebecca Washenfelder, and Cora Young for sharing data and for useful dis-

cussions. I would like to thank Erin Boyle for her assistance in the construction and operation

of the FILIF instrument during BEACHON-ROCS. I would also like to thank the staff of

Blodgett Forest Research Station and the U.S. Forest Service, particularly Richard Oakes.

Finally, I would also like to thank the following people for their data contributions:

Sam Henry and Aster Kammrath for glyoxal concentrations during both campaigns; Lisa

Kaiser, Ralf Schnitzhofer, Thomas Karl, and Armin Hansel for the PTR-TOF-MS VOC data

during BEACHON-ROCS; Jeung-Hoo Park and Allen Goldstein for the PTR-MS VOC data

from BEARPEX 2009; Robin Weber and Andrew Turnipseed for meterological and trace gas

data during BEARPEX 2009 and BEACHON-ROCS, respectively; Rebecca Hornbrook and

Eric Apel for the TOGA VOC data; and Saewung Kim, Chris Cantrell, and Lee Maudlin

for the CIMS OH data.

4.7 Bibliography

Apel, E. C., Hills, A. J., Lueb, R., Zindel, S., Eisele, S., and Riemer, D. D.: A fast-

GC/MS system to measure C2 to C4 carbonyls and methanol aboard aircraft, Journal of

Geophysical Research, 108, 8794, 2003.

Apel, E. C., Emmons, L. K., Karl, T., Flocke, F., Hills, A. J., Madronich, S., Lee-Taylor,

J., Fried, A., Weibring, P., Walega, J., Richter, D., Tie, X., Mauldin, L., Campos, T.,

Weinheimer, A., Knapp, D., Sive, B., Kleinman, L., Springston, S., Zaveri, R., Ortega, J.,

Voss, P., Blake, D., Baker, A., Warneke, C., Welsh-Bon, D., de Gouw, J., Zheng, J., Zhang,

R., Rudolph, J., Junkermann, W., and Riemer, D. D.: Chemical evolution of volatile

organic compounds in the outflow of the Mexico City Metropolitan area, Atmospheric

Chemistry and Physics, 10, 2353–2375, 2010.

132

Arnts, R. R. and Tejada, S. B.: 2,4-Dinitrophenylhydrazine-coated silica gel cartridge

method for determination of formaldehyde in air: identification of an ozone interference,

Environmental Science and Technology, 23, 1428–1430, 1989.

Atkinson, R.: Atmospheric chemistry of VOCs and NOx, Atmospheric Environment, 34,

2063–2101, 2000.

Bloss, C., Wagner, V., Bonzanini, A., Jenkin, M. E., Wirtz, K., Martin-Reviejo, M., and

Pilling, M. J.: Evaluation of detailed aromatic mechanisms (MCMv3 and MCMv3.1)

against environmental chamber data, Atmospheric Chemistry and Physics, 5, 623–639,

2005.

Chan, A. W. H., Chan, M. N., Surratt, J. D., Chhabra, P. S., Loza, C. L., Crounse, J. D.,

Yee, L. D., Flagan, R. C., Wennberg, P. O., and Seinfeld, J. H.: Role of aldehyde chemistry

and NOx concentrations in secondary organic aerosol formation, Atmospheric Chemistry

and Physics, 10, 7169–7188, 2010.

Dillon, M. B., Lamanna, M. S., Schade, G. W., Goldstein, A. H., and Cohen, R. C.: Chemical

evolution of the Sacramento urban plume: Transport and oxidation, Journal of Geophys-

ical Research-Atmospheres, 107, Artn 4045, 2002.

Dockery, D. W., Pope, C. A., Xu, X., Spengler, J. D., Ware, J. H., Fay, M. E., Ferris, B. G.,

and Speizer, F. E.: An Association between Air Pollution and Mortality in Six U.S. Cities,

New England Journal of Medicine, 329, 1753–1759, 1993.

Draxler, R. and Rolph, G.: HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Tra-

jectory) Model access via NOAA ARL READY Website, 2011.

Dreyfus, G. B., Schade, G. W., and Goldstein, A. H.: Observational constraints on the

133

contribution of isoprene oxidation to ozone production on the western slope of the Sierra

Nevada, California, Journal of Geophysical Research-Atmospheres, 107, Artn 4365, 2002.

Fried, A., McKeen, S., Sewell, S., Harder, J., Henry, B., Goldan, P., Kuster, W., Williams,

E., Baumann, K., Shetter, R., and Cantrell, C.: Photochemistry of formaldehyde during

the 1993 Tropospheric OH Photochemistry Experiment, Journal of Geophysical Research-

Atmospheres, 102, 6283–6296, 1997.

Fu, T.-M., Jacob, D. J., Wittrock, F., Burrows, J. P., Vrekoussis, M., and Henze, D. K.:

Global budgets of atmospheric glyoxal and methylglyoxal, and implications for formation

of secondary organic aerosols, Journal of Geophysical Research, 113, D15 303, 2008.

Galloway, M. M., DiGangi, J. P., Hottle, J. R., Huisman, A. J., Mielke, L. H., Alaghmand,

M., Shepson, P. B., Weremijewicz, J., Klavon, H., McNeal, F. M., Carroll, M. A., Griffith,

S., Hansen, R. F., Stevens, P. S., Bertman, S. B., and Keutsch, F. N.: Observations

and modeling of formaldehyde at the PROPHET mixed hardwood forest site in 2008,

Atmospheric Environment, accepted, 2011.

Garcia, A. R., Volkamer, R., Molina, L. T., Molina, M. J., Samuelson, J., Mellqvist, J.,

Galle, B., Herndon, S. C., and Kolb, C. E.: Separation of emitted and photochemical

formaldehyde in Mexico City using a statistical analysis and a new pair of gas-phase

tracers, Atmospheric Chemistry and Physics, 6, 4545–4557, 2006.

Goldstein, A. H., Hultman, N. E., Fracheboud, J. M., Bauer, M. R., Panek, J. A., Xu, M.,

Qi, Y., Guenther, A. B., and Baugh, W.: Effects of climate variability on the carbon

dioxide, water, and sensible heat fluxes above a ponderosa pine plantation in the Sierra

Nevada (CA), Agricultural and Forest Meteorology, 101, 113–129, 2000.

Hays, M. D., Geron, C. D., Linna, K. J., Smith, N. D., and Schauer, J. J.: Speciation

134

of Gas-Phase and Fine Particle Emissions from Burning of Foliar Fuels, Environmental

Science and Technology, 36, 2281–2295, 2002.

Henry, S. B., DiGangi, J. P., Weber, R. J., Gentner, D. R., Goldstein, A. H., and Keutsch,

F. N.: Formaldehyde influences from combustion and photochemistry at a California urban

site, in preparation, 2011.

Holzinger, R., Warneke, C., Hansel, A., Jordan, A., Lindinger, W., Scharffe, D. H., Schade,

G., and Crutzen, P. J.: Biomass burning as a source of formaldehyde, acetaldehyde,

methanol, acetone, acetonitrile, and hydrogen cyanide, Geophysical Research Letters, 26,

1161–1164, 1999.

Holzinger, R., Williams, J., Salisbury, G., Klupfel, T., de Reus, M., Traub, M., Crutzen,

P. J., and Lelieveld, J.: Oxygenated compounds in aged biomass burning plumes over

the Eastern Mediterranean: evidence for strong secondary production of methanol and

acetone, Atmospheric Chemistry and Physics, 5, 39–46, 2005.

Hornbrook, R. S., Blake, D. R., Diskin, G. S., Fried, A., Fuelberg, H. E., Meinardi, S.,

Mikoviny, T., Richter, D., Sachse, G. W., Vay, S. A., Walega, J., Weibring, P., Wein-

heimer, A. J., Wiedinmyer, C., Wisthaler, A., Hills, A., Riemer, D. D., and Apel, E. C.:

Observations of nonmethane organic compounds during ARCTAS—Part 1: Biomass burn-

ing emissions and plume enhancements, Atmospheric Chemistry and Physics, 11, 11 103–

11 130, 2011.

Hottle, J. R., Huisman, A. J., Digangi, J. P., Kammrath, A., Galloway, M. M., Coens,

K. L., and Keutsch, F. N.: A Laser Induced Fluorescence-Based Instrument for In-Situ

Measurements of Atmospheric Formaldehyde, Environmental Science and Technology, 43,

790–795, 2009.

135

Huisman, A. J., Hottle, J. R., Coens, K. L., DiGangi, J. P., Galloway, M. M., Kammrath,

A., and Keutsch, F. N.: Laser-induced phosphorescence for the in situ detection of glyoxal

at part per trillion mixing ratios, Analytical Chemistry, 80, 5884–5891, 2008.

Ip, H. S. S., Huang, X. H. H., and Yu, J. Z.: Effective Henry’s law constants of glyoxal,

glyoxylic acid, and glycolic acid, Geophysical Research Letters, 36, Artn L01 802, 2009.

Jordan, A., Haidacher, S., Hanel, G., Hartungen, E., Mrk, L., Seehauser, H., Schottkowsky,

R., Sulzer, P., and Mrk, T. D.: A high resolution and high sensitivity proton-transfer-

reaction time-of-flight mass spectrometer (PTR-TOF-MS), International Journal of Mass

Spectrometry, 286, 122–128, 2009.

Khalil, M. A. K. and Rasmussen, R. A.: Carbon-Monoxide in the Earths Atmosphere -

Indications of a Global Increase, Nature, 332, 242–245, 1988.

Kim, S., Karl, T., Guenther, A., Tyndall, G., Orlando, J., Harley, P., Rasmussen, R., and

Apel, E.: Emissions and ambient distributions of Biogenic Volatile Organic Compounds

(BVOC) in a ponderosa pine ecosystem: interpretation of PTR-MS mass spectra, Atmo-

spheric Chemistry and Physics, 10, 1759–1771, 2010.

Kleindienst, T. E., Corse, E. W., Blanchard, F. T., and Lonneman, W. A.: Evaluation of

the Performance of DNPH-Coated Silica Gel and C18 Cartridges in the Measurement of

Formaldehyde in the Presence and Absence of Ozone, Environmental Science and Tech-

nology, 32, 124–130, 1998.

Laden, F., Neas, L. M., Dockery, D. W., and Schwartz, J.: Association of fine particulate

matter from different sources with daily mortality in six US cities, Environmental Health

Perspectives, 108, 941–947, 2000.

136

Lane, T. E., Donahue, N. M., and Pandis, S. N.: Effect of NOx on secondary organic aerosol

concentrations, Environmental Science and Technology, 42, 6022–6027, 2008.

Lee, Y. N., Zhou, X. L., and Hallock, K.: Atmospheric carbonyl compounds at a rural south-

eastern United States site, Journal of Geophysical Research-Atmospheres, 100, 25 933–

25 944, 1995.

Lee, Y. N., Zhou, X., Kleinman, L. I., Nunnermacker, L. J., Springston, S. R., Daum, P. H.,

Newman, L., Keigley, W. G., Holdren, M. W., Spicer, C. W., Young, V., Fu, B., Parrish,

D. D., Holloway, J., Williams, J., Roberts, J. M., Ryerson, T. B., and Fehsenfeld, F. C.:

Atmospheric chemistry and distribution of formaldehyde and several multioxygenated car-

bonyl compounds during the 1995 Nashville Middle Tennessee Ozone Study, Journal of

Geophysical Research-Atmospheres, 103, 22 449–22 462, 1998.

Mauzerall, D. L. and Wang, X. P.: Protecting agricultural crops from the effects of tropo-

spheric ozone exposure: Reconciling science and standard setting in the United States,

Europe, and Asia, Annual Review of Energy and the Environment, 26, 237–268, 2001.

McDonald, J. D., Zielinska, B., Fujita, E. M., Sagebiel, J. C., Chow, J. C., and Watson,

J. G.: Fine Particle and Gaseous Emission Rates from Residential Wood Combustion,

Environmental Science and Technology, 34, 2080–2091, 2000.

Munger, J. W., Jacob, D. J., Daube, B. C., Horowitz, L. W., Keene, W. C., and Heikes,

B. G.: Formaldehyde, Glyoxal, and Methylglyoxal in Air and Cloudwater at a Rural

Mountain Site in Central Virginia, Journal of Geophysical Research-Atmospheres, 100,

9325–9333, 1995.

Myriokefalitakis, S., Vrekoussis, M., Tsigaridis, K., Wittrock, F., Richter, A., Bruhl, C.,

Volkamer, R., Burrows, J. P., and Kanakidou, M.: The influence of natural and anthro-

137

pogenic secondary sources on the glyoxal global distribution, Atmospheric Chemistry and

Physics, 8, 4965–4981, 2008.

Ng, N. L., Chhabra, P. S., Chan, A. W. H., Surratt, J. D., Kroll, J. H., Kwan, A. J., McCabe,

D. C., Wennberg, P. O., Sorooshian, A., Murphy, S. M., Dalleska, N. F., Flagan, R. C.,

and Seinfeld, J. H.: Effect of NOx level on secondary organic aerosol (SOA) formation

from the photooxidation of terpenes, Atmospheric Chemistry and Physics, 7, 5159–5174,

2007.

Ostro, B.: The Association of Air-Pollution and Mortality - Examining the Case for Infer-

ence, Archives of Environmental Health, 48, 336–342, 1993.

Rolph, G.: Real-time Environmental Applications and Display sYstem (READY) Website,

2011.

Ruuskanen, T. M., Mller, M., Schnitzhofer, R., Karl, T., Graus, M., Bamberger, I., Hrtnagl,

L., Brilli, F., Wohlfahrt, G., and Hansel, A.: Eddy covariance VOC emission and depo-

sition fluxes above grassland using PTR-TOF, Atmospheric Chemistry and Physics, 11,

611–625, 2011.

Staudinger, J. and Roberts, P. V.: A critical review of Henry’s law constants for environmen-

tal applications, Critical Reviews in Environmental Science and Technology, 26, 205–297,

1996.

Stieb, D. M., Beveridge, R. C., Brook, J. R., Smith-Doiron, M. A. R. C., Burnett, R. T.,

Dales, R. E., Beaulieu, S., Judek, S., and Mamedov, A.: Air pollution, aeroallergens and

cardiorespiratory emergency department visits in Saint John, Canada, Journal of Exposure

Analysis and Environmental Epidemiology, 10, 461–477, 2000.

Tan, D., Faloona, I., Simpas, J. B., Brune, W., Shepson, P. B., Couch, T. L., Sumner, A. L.,

138

Carroll, M. A., Thornberry, T., Apel, E., Riemer, D., and Stockwell, W.: HOx budgets

in a deciduous forest: Results from the PROPHET summer 1998 campaign, Journal of

Geophysical Research-Atmospheres, 106, 24 407–24 427, 2001.

Volkamer, R., Molina, L. T., Molina, M. J., Shirley, T., and Brune, W. H.: DOAS measure-

ment of glyoxal as an indicator for fast VOC chemistry in urban air, Geophysical Research

Letters, 32, Artn L08 806, 2005.

Vrekoussis, M., Wittrock, F., Richter, A., and Burrows, J. P.: Temporal and spatial variabil-

ity of glyoxal as observed from space, Atmospheric Chemistry and Physics, 9, 4485–4504,

2009.

Vrekoussis, M., Wittrock, F., Richter, A., and Burrows, J. P.: GOME-2 observations of

oxygenated VOCs: what can we learn from the ratio glyoxal to formaldehyde on a global

scale?, Atmospheric Chemistry and Physics, 10, 10 145–10 160, 2010.

Washenfelder, R. A., Young, C. J., Brown, S. S., Angevine, W. M., Atlas, E. L., Blake,

D. R., Bon, D. M., Cubison, M. J., de Gouw, J. A., Dusanter, S., Flynn, J., Gilman, J. B.,

Graus, M., Griffith, S., Grossberg, N., Hayes, P. L., Jimenez, J. L., Kuster, W. C., Lefer,

B. L., Pollack, I. B., Ryerson, T. B., Stark, H., Stevens, P. S., and Trainer, M. K.: The

glyoxal budget and its contribution to organic aerosol for Los Angeles, California, during

CalNex 2010, Journal of Geophysical Research-Atmospheres, 116, Artn D00v02, 2011.

Wittrock, F., Richter, A., Oetjen, H., Burrows, J. P., Kanakidou, M., Myriokefalitakis, S.,

Volkamer, R., Beirle, S., Platt, U., and Wagner, T.: Simultaneous global observations of

glyoxal and formaldehyde from space, Geophysical Research Letters, 33, Artn L16 804,

2006.

Wolfe, G. M. and Thornton, J. A.: The Chemistry of Atmosphere-Forest Exchange (CAFE)

139

Model Part 1: Model description and characterization, Atmospheric Chemistry and

Physics, 11, 77–101, 2011.

140

Fig. 4.1.— Diurnal profiles of RGF, HCHO, Gly, and wind direction during BEACHON-

ROCS 2010 and BEARPEX 2009. Grey dots represent individual data points and black

dots represent the 1 h binned medians. Note that the Gly and HCHO y-axes are not the

same between the two campaigns. The different symbols in the bottom left panel denote the

two dominant diurnal wind profiles for that campaign, with triangles indicating the more

dominant (∼60%) southwesterly wind direction and circles indicating the less dominant

(∼40%) northeasterly wind direction. The blue points on the right panels of the figure

denote the first day of the MFI, while the red points denote the second day of the MFI (see

Chap. 4.3.1).

141

Fig. 4.2.— One hour bin averaged Gly, HCHO, and RGF during BEARPEX 2009.

142

Fig. 4.3.— One hour bin averaged Gly, HCHO, and RGF during BEACHON-ROCS 2010.

143

Fig. 4.4.— Thirty minute binned median diurnal profiles of MBO+Isoprene, monoterpenes,

and OH reactivity during BEACHON-ROCS and BEARPEX 2009.

144

Fig. 4.5.— Diurnal median profiles of RGF, HCHO, Gly, and wind direction for the two

different wind regimes during BEACHON-ROCS 2010.

145

Fig. 4.6.— Gly, HCHO, RGF, other tracer species, and meteorological data during the two

days of the MFI. Gly, HCHO, and RGF are shown as binned averages, whereas other species

are shown at full resolution. Data for MBO+Isoprene, m/z 71, benzene, and acetonitrile

were measured via PTR-MS.

146

Fig. 4.7.— Four hour forward HYSPLIT trajectories for 16 July, 2009 originating at the MFI

site (red symbol) at 50 m above ground level. Lines denote initial times of 14:00 (red), 15:00

(blue), 16:00 (green), 17:00 (grey), and 18:00 (purple). Filled circles denote plume position

after each hour of travel time. The BEARPEX site is denoted by the green triangle.

147

Fig. 4.8.— Four hour forward HYSPLIT trajectories for 17 July, 2009 originating at the MFI

site (red symbol) at 50 m above ground level. Lines denote initial times of 10:00 (red), 11:00

(blue), 13:00 (green), 14:00 (grey), and 16:00 (purple). Filled circles denote plume position

after each hour of travel time. The BEARPEX site is denoted by the green triangle.

148

Fig. 4.9.— Gly, HCHO, RGF, other tracer species, and meteorological data during event

BN1. Gly, HCHO, and RGF are shown as binned averages, whereas other species are shown

at full resolution. Data for MBO+Isoprene, m/z 71, benzene, and acetonitrile were measured

via PTR-TOF-MS.

149

Fig. 4.10.— Gly, HCHO, RGF, other tracer species, and meteorological data during event

BN2. Gly, HCHO, and RGF are shown as binned averages, whereas other species are shown

at full resolution. Data for MBO+Isoprene, m/z 71, benzene, and acetonitrile were measured

via PTR-TOF-MS.

150

Fig. 4.11.— Gly, HCHO, RGF, other tracer species, and meteorological data during event

BN3. Gly, HCHO, and RGF are shown as binned averages, whereas other species are shown

at full resolution. Data for MBO+Isoprene, m/z 71, benzene, and acetonitrile were measured

via PTR-TOF-MS. Data shown for toluene was measured via TOGA, and RTB is based on

TOGA measurements of benzene and toluene.

151

Fig. 4.12.— A closer view of sharp changes in Gly, HCHO, and RGF during BN3 on 14

August during BEACHON-ROCS 2010.

152

Fig. 4.13.— Comparison of m/z 95 with Gly, HCHO, and RGF during (a) BN1 and (b) BN2.

Note the lack of change in m/z 95 during the fast rises/falls in Gly and HCHO.

153

Fig. 4.14.— RGF ranges during campaigns presented in this work and the literature. Circlesdenote the campaign medians, the squares denote the campaign means, and the lines denotethe interquartile range (middle 50%) of the datasets.1 Myriokefalitakis et al. (2008)

2 Vrekoussis et al. (2010)

3 Volkamer et al. (2005)

4 Gly: Washenfelder et al. (2011); HCHO: Personal communication from J. Stutz

5 Henry et al. (2011)

6 Munger et al. (1995)

7 Lee et al. (1995), estimated from data in Figure 8

8 this work

154

Fig. 4.15.— Examination of RO2 fate and its relation to HCHO, Gly, and RGF on 24 August

during BEACHON-ROCS. (a & b) Thirty minute binned medians of RGF and concentrations

of Gly, HCHO, NO, NO2, HO2, and HO2+RO2 over the course of the day. (c) Thirty

minute binned medians of the percent of RO2 loss from reaction with NO or HO2 based

on concentrations in (b) and rate constants from the University of Leeds Master Chemical

Mechanism (Bloss et al., 2005).

155

Table 4.1: Percent increases for Gly, HCHO, RGF, and other species for each transport event.

Event MFI MFI BN1 BN2 BN3 BN3

(Day 1) (Day 2) (afternoon) (evening)

Event Time 18:00 12:30 16:30 14:00 13:00 20:00

-19:30 -14:30 -18:00 -15:30 -14:00 -21:15

Gly 280% 140% 94% 120% - 380%

HCHO 110% 120% 160% 54% 250% 67%

RGF 79% 18% -18% 3% - 190%

MBO+Isoprene -58% -18% -2% 7% -12% 84%

m/z 71 6% 77% 42% 200% 320% 120%

Benzene 280% 180% 150% 320% 260% 230%

Acetonitrile 150% 100% 4% 20% 12% 3%

156

Chapter 5

Conclusions

5.1 Summary

Tropospheric ozone and aerosols are two atmospheric species that have a profound

effect on air quality. High ozone concentrations are correlated with increased incidences

of both cardiovascular and respiratory illnesses, while aerosol concentrations are positively

correlated with increased mortality. To improve and mitigate the effects of these trace

atmospheric species on the quality of air, we must understand the fundamental processes

governing their production and destruction. This greater understanding would allow us

to more accurately predict air quality as well as determine the most effective mitigation

strategies.

The oxidation of volatile organic compounds (VOCs) in the atmosphere is strongly

tied to the production of both tropospheric ozone and secondary organic aerosol, which

contributes a significant percentage of total aerosol mass at most sites (Zhang et al., 2007;

Jimenez et al., 2009). This oxidation occurs via a solar-driven catalytic cycle of the radical

families HOx (OH+HO2) and NOx (NO+NO2). By studying the intermediates of VOC

oxidation, called oxidized volatile organic compounds (OVOCs), we can understand and

quantify the individual pathways by which these VOCs are oxidized. Formaldehyde (HCHO)

157

is one of the most common OVOCs, as some HCHO is formed in the oxidation of nearly all

VOCs. As a result of this and its short atmospheric lifetime of a few hours (Atkinson, 2000),

HCHO can be used as a tracer of the amount of overall VOC oxidation.

As previous techniques for HCHO detection exhibited insufficient detection limits and

sensitivities for fast aircraft sampling and eddy covariance, a new technique was developed

called Fiber Laser-Induced Fluorescence (FILIF, Chap. 2). The combination of a novel,

low-power, and rugged fiber laser with the highly-sensitive technique of LIF yielded an in-

situ technique with superior time resolution, high sensitivity, high selectivity, and excellent

reliability. Optimum conditions were established for FILIF of HCHO, as summarized in

Table 5.1. Additionally, inlet tests were performed over three separate campaigns to ascertain

any potential sampling artifacts. No significant sampling artifacts were discovered using PFA

Teflon tubing of any length, while larger inner diameter PTFE Teflon tubing did exhibit

artifacts, in particular sampling using the thin 1/32” walled tubing. However, the 1/8”

PTFE tubing did not exhibit any significant artifacts, suggesting that shorter residence

times minimize any inlet effects.

Measurements at many forest sites have reported a discrepancy between measured and

modeled OH reactivity, or the lifetime of OH in ambient air, which indicates a fundamental

problem with current understanding of forest VOC oxidation. Di Carlo et al. (2004) reported

an exponential increase in this discrepancy with respect to temperature, attributed to the

emission profile of a class of VOC called terpenes. If these terpenes are emitted and quickly

oxidized inside the forest canopy, as would be necessary to cause this reactivity, a discrep-

ancy in HCHO measurements should be observed as well. The FILIF HCHO instrument was

employed during BEACHON-ROCS 2010 to investigate the possibility of missing BVOCs

inside of forest canopies (Chap. 3). Gradients and vertical fluxes of HCHO were used to con-

strain HCHO inside the canopy. This study represented the first measurements of HCHO

158

flux by eddy covariance. HCHO fluxes exhibited a diurnal profile similar to the solar cycle,

peaking at noon with a median flux of ∼80 µg m−2 h−1. Enclosure experiments were per-

formed to determine the HCHO branch (3.5 µg m−2 h−1) and soil (7.3 µg m−2 h−1) direct

emission rates in the canopy, the latter of which represented the first reported measurements

of HCHO soil emission. A zero-dimensional canopy box model, used to determine the appor-

tionment of HCHO source and sink contributions to the flux, underpredicted the observed

HCHO flux by a factor of six. Simulated increases in concentrations of species similar to

monoterpenes resulted in poor agreement with measurements, while simulated increases in

direct HCHO emissions and/or concentrations of species similar to 2-methyl-3-buten-2-ol

(MBO) best improved model/measurement agreement. Given the typical diurnal variability

of these BVOC emissions and direct HCHO emissions, this suggests that the source of the

missing flux is a process with both a strong temperature and radiation dependence.

Finally, the production of ozone in the troposphere is highly dependent on the types

of VOCs that are oxidized. Global models use satellite measurements of the ratio of glyoxal

(Gly), another common OVOC, to HCHO (RGF) as a measure of the type of VOCs over a

certain area. In order to validate these ratios at the ground level, the area most relevant

to air quality, measurements of HCHO by FILIF and Gly were performed during two field

campaigns: BEARPEX 2009 and BEACHON-ROCS 2010 (Chap. 4). RGF was found to have

a very consistent diurnal profile during both campaigns, despite considerable variability in

both HCHO and Gly concentrations. Multiple events during which both HCHO and Gly

exhibited large (∼100%) increases in concentration resulted in changes in RGF of less than

20%, even when showing mild or aged anthropogenic influence. During these events, as

neither RGF nor BVOC measurements changed appreciably, the rises in Gly and HCHO also

suggest an anthropogenically-influence change in the fate of the RO2 radical, which requires

further study. Strong and/or fresh anthropogenic influence resulted in a marked increase

159

in RGF (∼200%), as did influence from a fresh biomass burning event (∼80%). However,

in comparing these RGF values with those bserved during ground-based urban campaigns

and from satellites, opposite trends were observed. While urban RGF measurements agree

between ground and satellite-based measurements, satellites measure greater RGF in BVOC

dominated areas compared to these urban areas, while ground measurements observe smaller

RGF than in urban areas. Resolving this discrepancy is of the utmost importance as it may

indicate that the satellite measurements are not indicative of ground concentrations of these

trace species.

5.2 Future Directions

5.2.1 VOC oxidation chemistry in a plume

Ground-based measurements are limited as they can sample an airmass only at a

given point in its processing. Two of the larger challenges in modeling an airmass are the

determination of initial conditions and accounting for the effects of advection. Aircraft

measurements have the advantage of sampling the same airmass at multiple points in time

by flying across its path at different points downwind. However, typical aircraft, such as jets

and airplanes, require high velocities which limit the averaging time, and so the resolution,

of the data from the plume. Helicopters can hover in place, but the rotor wash can disturb

and affect measurements. dirigibles do not suffer either of these problems, as they move at a

slow rate of speed without the disruptive rotor wash of a helicopter. As a result, instruments

mounted on dirigibles can begin measuring at a plume source, such as an urban/industrial

area or forest, and move along with it, watching the VOC oxidation over a long period

of time. This can eliminate much of the speculation that goes into model/measurement

comparison.

As with other aircraft, instrumentation power, size, and weight requirements for di-

160

rigibles are quite strict due to limited availability. The FILIF instrument has minimal

requirements in all three of these categories, making it ideally suited for aircraft measure-

ments. As a result, the FILIF HCHO instrument will be participating in the Pan-European

Gas-AeroSOls-climate interaction Study (PEGASOS) in Italy and Finland in the summer

of 2012 on a Zeppelin NT dirigible. The goals of this campaign that HCHO will help to

address is to identify the main processes currently missing in models predicting both air

quality and climate and to determine the potential feedback effects of climate change on the

global HOx budget. A retrofit of the FILIF system and test flights have already successfully

been performed by Glenn Wolfe and Jen Knapp, during which the instrument performed

over the flights entirely unmanned.

5.2.2 Measurements of HCHO direct emission from trees and ground litter/soil

Though measurements of HCHO flux during BEACHON-ROCS 2010 predicted a large

missing in-canopy production of HCHO, they were unable to conclusively establish the ex-

istence of the oxidation of considerable amounts of missing BVOCs in the canopy. This was

due to the incomplete characterization of the direct emissions of HCHO from the trees. In

order for HCHO to be useful in addressing VOC oxidation in forest canopies, it is necessary

to properly characterize these direct emission sources. This can be accomplished by a suite

of measurements to characterize the two primary emission sources. The first involves mea-

suring the HCHO compensation point, which is the ambient concentration at which emission

and deposition by the tree are equal (Kesselmeier, 2001). The second involves measuring

the direct emissions of HCHO from various types of soil, each under various temperature

and light conditions.

The model used to estimate HCHO flux contained an implicit approximation in regard

to tree emission and stomatal deposition in which it treats them as independent processes. In

fact, these processes are coupled, and their net effect is described by the compensation point.

161

Therefore, the actual production or deposition of HCHO by trees is more accurately described

as proportional to the difference between the ambient concentration and the compensation

point. At concentrations below the compensation point, there will be net emission, while at

concentrations above the compensation point, there will be net deposition. However, this

compensation point varies greatly depending on ambient temperature, relative humidity,

tree species and age, and other environmental factors (Kesselmeier et al., 1997; Kesselmeier,

2001; Seco et al., 2007, 2008). Thus, accurate characterization of this phenomenon would

require an extensive series of measurements at a given measurement site. A representative

sample of each species of tree would be sampled by measuring multiple examples of branches

from each species of tree at varying heights and varying tree ages (when possible), as well

as during different times of day.

Unlike direct tree HCHO emission, the processes underlying emission from either the

soil or ground litter are unknown. Initial characterization of these processes could begin with

field samples of soil and ground litter under controlled conditions. At a measurement site,

another comprehensive series of measurements of HCHO direct emission, similar to that for

tree emission, would be necessary for both soil and ground litter. To more accurately estimate

the soil/litter emission rate, ground chamber experiments would be performed at varying

ground locations and ground types (i.e. heavy vs. light litter cover) over the estimated

fetch area of the flux measurements. Similarly, we will perform enclosure experiments of the

various forms of undergrowth vegetation, if applicable.

Additionally, all of these emission measurements require a full tree, ground, and un-

dergrowth survey of the estimated fetch area, or area upwind that contributes to the HCHO

flux, to properly parameterize the emission contribution. Finally, a model similar to that

described in Chap. 3, with properly characterized HCHO emissions, would then be able to

isolate the effect of any unknown BVOC on in-canopy HCHO production.

162

5.2.3 Long-term investigations of Gly:HCHO ratios and alkylperoxy radical

fate

Satellite measurements of RGF are increasingly used as a metric for reactive VOCs in

global models. However, Chapter 4 showed that the trends in rural-urban RGF are oppo-

site for ground-based and satellite measurements. In order to resolve this satellite/ground

discrepancy, long-term ground measurements of RGF at multiple sites would increase the

reliability and applicability of the ground RGF trend. In other words, further measurements

would decrease the likelihood of a given site or given time acting as artifacts. Additionally,

more characterization is needed to apply RGF as a tracer of reactive VOC mix on a larger

scale.

The Environmental Protection Agency (EPA) has approved funding for a study of Gly

and HCHO similar to that described in Chap. 4, but for long-term (1 year) measurements

at the Horicon National Core Monitoring Station in Wisconsin. Part of the goal of these

measurements is to further characterize RGF as a tracer of the reactive VOC mixture present

in an airmass. An additional tracer for such will also be explored: the ratio of Gly to

photochemical HCHO, where photochemical HCHO is separated using a multidimensional

linear fit (Garcia et al., 2006; Henry et al., 2011). This new ratio is potentially more useful

for studying reactive VOC mixtures, as it removes the highly-variable influence of direct

HCHO emission that is insignificant on the scale of the entire surface (boundary) layer.

Finally, this study will further investigate the effects of anthropogenic influence on BVOC

oxidation, as current data is insufficient to draw broader conclusions. A long-term dataset

will enable this effect on RO2 fate to be further characterized.

163

5.3 Bibliography

Atkinson, R.: Atmospheric chemistry of VOCs and NOx, Atmospheric Environment, 34,

2063–2101, 2000.

Di Carlo, P., Brune, W. H., Martinez, M., Harder, H., Lesher, R., Ren, X., Thornberry,

T., Carroll, M. A., Young, V., Shepson, P. B., Riemer, D., Apel, E., and Campbell,

C.: Missing OH Reactivity in a Forest: Evidence for Unknown Reactive Biogenic VOCs,

Science, 304, 722–725, 2004.

Garcia, A. R., Volkamer, R., Molina, L. T., Molina, M. J., Samuelson, J., Mellqvist, J.,

Galle, B., Herndon, S. C., and Kolb, C. E.: Separation of emitted and photochemical

formaldehyde in Mexico City using a statistical analysis and a new pair of gas-phase

tracers, Atmospheric Chemistry and Physics, 6, 4545–4557, 2006.

Henry, S. B., DiGangi, J. P., Weber, R. J., Gentner, D. R., Goldstein, A. H., and Keutsch,

F. N.: Formaldehyde influences from combustion and photochemistry at a California urban

site, in preparation, 2011.

Jimenez, J. L., Canagaratna, M. R., Donahue, N. M., Prevot, A. S. H., Zhang, Q., Kroll,

J. H., DeCarlo, P. F., Allan, J. D., Coe, H., Ng, N. L., Aiken, A. C., Docherty, K. S.,

Ulbrich, I. M., Grieshop, A. P., Robinson, A. L., Duplissy, J., Smith, J. D., Wilson, K. R.,

Lanz, V. A., Hueglin, C., Sun, Y. L., Tian, J., Laaksonen, A., Raatikainen, T., Rautiainen,

J., Vaattovaara, P., Ehn, M., Kulmala, M., Tomlinson, J. M., Collins, D. R., Cubison,

M. J., Dunlea, E. J., Huffman, J. A., Onasch, T. B., Alfarra, M. R., Williams, P. I., Bower,

K., Kondo, Y., Schneider, J., Drewnick, F., Borrmann, S., Weimer, S., Demerjian, K.,

Salcedo, D., Cottrell, L., Griffin, R., Takami, A., Miyoshi, T., Hatakeyama, S., Shimono,

A., Sun, J. Y., Zhang, Y. M., Dzepina, K., Kimmel, J. R., Sueper, D., Jayne, J. T.,

Herndon, S. C., Trimborn, A. M., Williams, L. R., Wood, E. C., Middlebrook, A. M.,

164

Kolb, C. E., Baltensperger, U., and Worsnop, D. R.: Evolution of Organic Aerosols in the

Atmosphere, Science, 326, 1525–1529, 2009.

Kesselmeier, J.: Exchange of Short-Chain Oxygenated Volatile Organic Compounds (VOCs)

between Plants and the Atmosphere: A Compilation of Field and Laboratory Studies,

Journal of Atmospheric Chemistry, 39, 219–233, 2001.

Kesselmeier, J., Bode, K., Hofmann, U., Muller, H., Schafer, L., Wolf, A., Ciccioli, P.,

Brancaleoni, E., Cecinato, A., Frattoni, M., Foster, P., Ferrari, C., Jacob, V., Fugit, J. L.,

Dutaur, L., Simon, V., and Torres, L.: Emission of short chained organic acids, aldehydes

and monoterpenes from Quercus ilex L. and Pinus pinea L. in relation to physiological

activities, carbon budget and emission algorithms, Atmospheric Environment, 31, 119–

133, 1997.

Seco, R., Penuelas, J., and Filella, I.: Short-chain oxygenated VOCs: Emission and uptake

by plants and atmospheric sources, sinks, and concentrations, Atmospheric Environment,

41, 2477–2499, 2007.

Seco, R., Penuelas, J., and Filella, I.: Formaldehyde emission and uptake by Mediterranean

trees Quercus ilex and Pinus halepensis, Atmospheric Environment, 42, 7907–7914, 2008.

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., and Worsnop, D. R.: Ubiquity and domi-

nance of oxygenated species in organic aerosols in anthropogenically-influenced Northern

Hemisphere midlatitudes, Geophysical Research Letters, 34, L13 801, 2007.

165

Table 5.1: Optimum parameters determined for HCHO FILIF and fiber laser (see Chap. 2).

Parameter Property/Optimum Setting

Laser Bandwidth < 0.01 cm−1

Laser Repetition Rate 300 kHz

Laser Pulse Width 30 ns

Laser Rise Time ∼10 ms

Photon Gate Delay from Laser Pulse 325 ns

Photon Gate Width 212.5 ns

Detection Axis Cell Pressure 110±20 Torr

Purge Flow Rate ≥ 8 × Main Flow Rate

Humidity Sensitivity None observed