UC Berkeley

The vast majority of earth’s atmosphere is outside, yet humans spend ~90% of their time indoors so the air they breathe is dominantly indoor air. Indoor air differs substantially from outdoor air in terms of its organic chemical composition and transformation processes, driven by key features including lower oxidant levels, less intense sunlight, higher surface area to volume environment, and more direct human influence While this dissertation looks at indoor air at a variety of scales, a consistent focus is that the human occupant is a defining feature of indoor air. This dissertation investigates how human activity affects indoor oxidant levels, how the human body directly influences the composition of indoor volatile organic compounds (VOCs), and how human activities control the emission of VOCs into the indoor environment.

Chapter 1 introduces the motivation for this work, provides a brief overview of key issues in atmospheric chemistry associated with the study of indoor air, reviews key prior knowledge of indoor air VOCs and oxidants that set the groundwork for the following chapters. A brief description of the instrumentation used to measure indoor VOCs is discussed, and a roadmap to this dissertation is provided.

In Chapter 2 we present direct indoor measurements of the nitrate radical (NO3¬) and dinitrogen pentoxide (N2O5) produced from combustion cooking emissions in a residential kitchen. The presence and importance of NO3 indoors had been hypothesized as early as 1986, however this chapter presents the first ever direct measurement made indoors. When indoor ozone (O3) concentration was low (~4 ppbv), nitric oxide (NO) emitted from gas-stove combustion suppressed NO3 formation. However, at moderate O3 levels (~40 ppbv), measured NO3 concentrations reached 3 to 4 pptv, and the indoor NO3 reactivity loss rate coefficient reached 0.8 s-1. A box model of known chemistry agrees with the reactivity estimate and shows that moderate O3 levels led to a nitrate radical production rate of 7 ppbv h-1¬. These indoor NO3 production rates and reactivities are much higher than is typical outdoors. At low O3 levels indoor combustion suppresses nitrate radical chemistry, but when sufficient O3 enters residences from outdoors or is emitted directly from indoor sources, gas stove combustion emissions promote indoor NO3 chemistry. Therefore in polluted regions with high levels of ozone, indoor NO3 chemistry will take on a greater importance.

Chapter 3 presents a laboratory study of the ozonolysis of squalene, a major component of human skin oil. Rather than directly study skin oil, oxidation experiments are carried out on pure squalene particles passing through a flow tube reactor, allowing for the quantification of gas phase products from a single starting material with varying conditions of water vapor and O3. Previous work examining the condensed-phase products of pure squalene particle ozonolysis in a flow tube reactor found that an increase in water vapor concentration led to lower concentrations of secondary ozonides, increased concentrations of carbonyls, and smaller particle diameter, suggesting that water changes the fate of the Criegee intermediate. To determine if this loss of volume corresponds to an increase in gas-phase products, we measured gas-phase VOC concentrations via proton transfer reaction time of flight mass spectrometry (PTR-TOF-MS). Studies were conducted at atmospherically relevant O3 exposure levels (5-30 ppb h).

An increase in water vapor concentration led to strong enhancement of gas-phase oxidation products at all tested O3 exposures. An increase in water vapor from ~0% to 70% relative humidity (RH) at high O3 exposure increased the total mass concentration of gas-phase VOCs by a factor of three. The observed fraction of carbon in the gas phase correlated well with the fraction of particle volume lost. Experiments involving O3 oxidation of shirts soiled with skin oil confirms that the RH dependence of gas-phase reaction product generation occurs similarly on surfaces containing skin oil under realistic indoor conditions. Relative humidity changes the fate of the Criegee intermediates and the volatility of the oxidation products, resulting in a RH dependence of the product distribution and amount of gas-phase VOCs emitted from ozonolysis of unsaturated carbon bonds in skin oil. Similar behavior is expected for O3 reactions with surface bound organics containing unsaturated carbon bonds.

Chapter 4 presents VOC emission profiles from scripted cooking, cleaning and human occupancy experiments performed during the HOMEChem study at the University of Texas, Austin test house. Quantifying indoor VOC speciation and emissions is critical to understanding and modeling the processes controlling indoor concentration dynamics, human exposure, and the chemistry of indoor air. Much of previous research on indoor VOCs has utilized broad surveys of different structures with low time resolution measurements of concentrations for specific target compounds. Such studies necessarily average over the dynamics driven by events and processes controlling variability in concentrations, and focus on the resulting concentrations rather than the actual emission rates. This study was designed to quantify VOC emission profiles from the building and its contents, and from the scripted experiments with multiple replicates. Measurements of VOCs were performed with a PTR-TOF-MS which continuously measured the time-resolved mass spectrum of indoor and outdoor air. Continuous tracer releases enabled determination of air change rates (ACR) and thus calculation of speciated, time resolved net VOC emissions. The building and its contents were the dominant emission source into the house, with large emissions of acetic acid, methanol, and formic acid. Cooking emissions are greater than cleaning emissions, and comprised mainly of ethanol. Bleach cleaning leads to high emissions of reactive chlorinated compounds, while cleaning with a natural product emits predominantly monoterpenes and terpenoids. Emissions from occupancy experiments show large enhancements of siloxanes from personal care products in the morning which are nearly depleted by the afternoon when the products had already mostly evaporated. These results are used to construct VOC emissions for a hypothetical 24 hours, and show emissions from the house and its contents make up nearly half of the indoor VOC emissions, while the rest come from occupants and their typical activities.

In Chapter 5 we provide preliminary evidence that indoor emissions escaping to outdoors are an increasingly important fraction of the fuel for outdoor air pollution in urban areas of developed countries, and suggest this as a promising new research direction for the field of atmospheric chemistry and air pollution control.