Reactions and Products of Squalene and Ozone: A Review

Abstract

This critical review describes the squalene-ozone (SqOz) reaction, or squalene ozonolysis. Ambient ozone penetrates indoors and drives indoor air chemistry. Squalene, a component of human skin oil, contains six carbon–carbon double bonds and is very reactive with ozone. Bioeffluents from people contribute to indoor air chemistry and affect the indoor air quality, resulting in exposures because people spend the majority of their time indoors. The SqOz reaction proceeds through various formation pathways and produces compounds that include aldehydes, ketones, carboxylic acids, and dicarbonyl species, which have a range of volatilities. In this critical review of SqOz chemistry, information on the mechanism of reaction, reaction probability, rate constants, and reaction kinetics are compiled. Characterizations of SqOz reaction products have been done in laboratory experiments and real-world settings. The effect of multiple environmental parameters (ozone concentration, air exchange rate (AER), temperature, and relative humidity (RH)) in indoor settings are summarized. This critical review concludes by identifying the paucity of available exposure, health, and toxicological data for known reaction products. Key knowledge gaps about SqOz reactions leading to indoor exposures and adverse health outcomes are provided as well as an outlook on where the field is headed.

1. Introduction

The indoor environment has a significant impact on human health since a typical person in an industrialized country spends ∼90% of their time indoors, with ∼70% inside their residence.¹⁻⁴ Indoor air chemistry, often driven by ambient ozone that has been transported indoors, is an important source of compounds people are exposed to indoors because the pollutant sources are in close proximity to the inhabitants.⁵⁻⁸ Low air exchange rates (AERs) result in pollutants emitted or formed indoors having increased air concentrations due to a lower transport of indoor pollutants outdoors.⁹⁻¹² The chemical composition that we are exposed to inside our homes is also affected by a variety of other parameters including relative humidity (RH), light, temperature, and air mixing. One source of precursors that reacts with ozone indoors is bioeffluents emitted from occupants, which includes skin oils.^8,13−16 Human skin oil includes compounds with unsaturated carbon–carbon bonds. Thus, the skin and surfaces that skin oil has been deposited on are among the most reactive surfaces within the home, thereby impacting the indoor air quality.¹⁷

Ozone generated outdoors typically enters the indoor environment through three pathways: natural ventilation, mechanical ventilation, and infiltration.^18,19 While indoor ozone concentrations are lower than outdoor concentrations, they depend on outdoor air concentrations since the primary source of indoor ozone is penetration of ozone from outdoor air.²⁰⁻²² Ozone then reacts with unsaturated hydrocarbons on multiple surfaces within a building, such as people’s skin, hair, and clothing, causing a reduction in its concentration.^16,23−26 Ozone concentrations measured in buildings in the United States (US), Poland, Japan, Taiwan, Cyprus, and the United Arab Emirates range from below detection to exceeding 100 parts per billion (ppb) and are typically highest during the summer months and in buildings with higher AERs.^{15,19,27−33}

A notable indoor ozone reaction is with squalene.³⁴⁻³⁶ Squalene is a 30-carbon chain molecule that contains six trans-oriented carbon–carbon double bonds, accounts for 5%–15% of the total human sebum composition, and accounts for approximately 48% of the unsaturated double bonds in skin oil, which makes it one of the most reactive components in human sebum.³⁷⁻⁴¹ The other main components of sebum are fatty acids, wax monoesters, triglycerides, cholesterol, and cholesterol esters.^{37,39,42−48} Squalene is a semivolatile unsaturated compound that is continuously produced by the body’s sebaceous glands and is on the surface of the skin, hair, and clothing, as well as in dust as a component of dead skin cells and in indoor particulate matter.^{42,43,49−51} In addition, squalene is found in fingerprints, which places squalene on most surfaces around the home.^52,53 Sebaceous glands produce squalene to beneficially protect the lower layers of the skin’s structure from oxidation by atmospheric oxidants such as free radicals and ozone.⁵⁴ The amount of squalene on one’s skin does not vary significantly throughout the day due to continual reactions and replenishment from the sebaceous glands.^40,41 Since the SqOz reaction occurs on people’s skin and clothing that contains squalene while being worn, there is high potential for exposure to the reaction products. Thus, understanding what products are formed, their potential health effects, and whether they remain on the surface of the skin or become airborne is important for understanding how people are exposed to SqOz reaction products and the potential health risks associated with them.

Due to the highly reactive nature of squalene and its presence throughout the home, which results in potential multiroute exposures indoors, there has been extensive literature published recently on squalene ozonolysis. Thus, a review of the subject, distinct from the many other ozonolysis reactions that occur indoors, is warranted. This critical review examines the products formed by SqOz reactions within the indoor environment. The broad sections covered are an overview of the reaction mechanism, volatile reaction products and their concentration in different enclosed environments, particle phase reaction products, reaction probability and kinetics, conditions that affect SqOz reactions, exposure to reaction products, and their toxicological and potential health effects.

2. Methodology

A literature search was performed through July 2021 using the Scopus database. The search contained combinations of the following words: ozone or ozonolysis AND squalene, ozone or ozonolysis AND skin lipid(s), ozone or ozonolysis AND skin oil(s), and ozone or ozonolysis AND sebum.

This yielded a total of 378 search results. After removing conference proceedings and duplicate articles, 260 unique articles remained. Titles and abstracts were read for all remaining search results and used to determine if the scope of the article fit within the research paper topic. Among them, 59 articles contained information relevant to the literature review on the reaction between squalene in human skin oil and ozone. This critical review incorporated articles that discussed squalene as part of total skin oil, its reaction with ozone, and human exposures to SqOz reaction products. Articles primarily focused on the other components of skin oil were not included in this critical review. Articles were also identified and included in this critical review through cross-references between articles and the use of a cocitation tool, ResearchRabbit.⁵⁵ ResearchRabbit is an artificial intelligence tool that uses a list of papers the user has marked as pertinent to suggest papers that were published prior to and after the previously selected articles that are in the same area of research.

3. The SqOz Reaction

3.1. Ozonolysis Mechanism

The mechanism whereby ozone reacts with alkenes, such as squalene, is known as ozonolysis and has been described in detail by Rudolf Criegee.⁵⁶ Ozone, a strong oxidizer, can react at multiple sites on the squalene molecule through the Criegee ozonolysis mechanism. Ozone does not appear to react preferentially with any specific double bond contained in the squalene molecule,^57,58 but this remains a subject of study. The addition of ozone across a double bond forms a primary ozonide which decomposes into a Criegee intermediate and a stable carbonyl compound (Scheme 1). The presence of Criegee intermediates was first described as short-lived structures that are formed and then rapidly rearranged during the course of the ozonolysis reaction.⁵⁶ The Criegee intermediates can either stabilize or further dissociate, isomerize, or undergo H migration to form aldehydes, ketones, and alcohols.⁵⁸⁻⁶² If the stable primary products formed contain unsaturated carbon–carbon bonds, they can continue to react with ozone to form secondary ozonolysis products.⁵⁸ The broad functional groups of primary and secondary products are ketones, aldehydes, hyperoxides, dicarbonyl compounds, carboxylic acids, and radicals.^58,62

Scheme 1. Structure of Squalene and General Ozonolysis Mechanism of Alkenes Reacting with Ozone.

Secondary Criegee intermediates form at least 10 times faster than primary Criegee intermediates.⁶³ On the basis of these reaction rates, even though secondary products with low carbon numbers (C₃, C₈, C₁₃) were measured, mechanistically the larger corresponding compounds of C₁₇, C₂₂, and C₂₇ aldehydes that form from secondary Criegee intermediates should be present. This is because the primary ozonide decomposes asymmetrically to produce a small Criegee intermediate (C_3,8,13) and a large carbonyl (C_17,22,27), which could be either a ketone or an aldehyde product.^58,64 This product formation is highly dependent on RH, as is further discussed in Section 5.2, due to interactions of the intermediate with water vapor and resulting in a shift in the formation pathway toward volatile product formation.⁶⁵ Branching ratios can be used to define the preferential way a reaction proceeds in relation to a different pathway, with the primary pathways producing either gas-phase primary ozonide reaction products or particle-phase secondary ozonide reaction products. While other branching ratios are possible, it was reported that the 10:90 product branching ratio and the 50:50 product branching ratio are likely branching ratios, but experimentally, products associated with the 10:90 branching ratio are observed to be favored over those with a 50:50 branching ratio.⁶³ Under controlled laboratory conditions, it is also possible for Criegee intermediates to react with each other to form additional carbonyl products.⁶⁶ While reactions of Criegee intermediates are not expected to occur in the gas phase under outdoor atmospheric conditions due to high convection conditions, reactions in liquid phases and on particles are possible.⁶⁷ Indoors the air velocity is lower than outdoors, particularly near surfaces, such as the body, where squalene loadings are higher than outdoors, potentially leading to higher Criegee intermediate concentrations and resulting in interactions between them to produce new products. Many research groups have proposed similar reaction schemes outlining the general process of ozonolysis as it pertains to the SqOz reaction and the formation of the various volatile reaction products and are the basis of our Scheme 1.^{46,57,58,62,65,68} By simulating the oxidation of squalene using a series of predefined rules based off of their proposed reaction scheme, a Computational Brewing Application (COBRA) model predicted and tentatively identified 83% of reaction products measured by mass spectrometry of a SqOz reaction.⁶⁹

3.2. Volatile Reaction Products (VRPs)

The majority of the reported research studying the reaction of ozone with skin oils has examined VRPs. The specific products formed depend upon which double bond of the squalene molecule ozone reacts with. Unsaturated primary or secondary SqOz reaction products will continue to react with ozone in the air or on surfaces to produce secondary or tertiary reactions products. Fruekilde et al.⁷⁰ studied the reaction of skin lipids and ozone and identified several VRPs emitted that included 4-oxopentanal (4-OPA), 6-methyl-5-hepten-2-one (6-MHO), geranylacetone, and acetone. This was the first study to directly link the presence of various specific VRPs to having precursors from humans. When investigating the total volatile compounds present in a simulated office space, Bako-Biro et al.⁷¹ also linked precursors of ozone VRPs to humans when they noticed that there were higher air concentrations of oxygenated compounds in a room containing ozone when people were present. Ozone-generated air chemistry in aircraft cabins was studied in the early 2000s because of the potential adverse effects surrounding exposure to high ozone levels measured along some aircraft routes and elevated levels of ozone-initiated reaction products.⁷² Wisthaler et al.³⁵ used a chamber to simulate an aircraft cabin and found that out of the total 110 ppb of volatile organic compounds (VOCs) detected, approximately 20 ppb were from ozone reactions with skin oils found on the soiled clothing placed in the simulated aircraft cabin. Of the total 20 ppb VOCs measured, acetone, 4-OPA, 6-MHO, and geranylacetone were the main contributors and were specifically linked to squalene ozonolysis. Similar ozone reaction products were identified in a simulated aircraft cabin study that had human occupants.³⁴ Wisthaler and Weschler³⁶ later identified in real time five prominent primary squalene ozonolysis VRPs and seven prominent secondary squalene ozonolysis VRPs (Table 1). All the compounds reported are formed through Criegee’s ozonolysis mechanism and are consistent with being produced from ozone reactions with carbon–carbon double bonds in skin oil components, including squalene.

Table 1. Potential Volatile SqOz Reaction Products and Concentrations Grouped by Source of Squalene.

Source of squalene	Products	Concentrations	ref
Human skin in small enclosure	Acetone, 6-MHO, geranylacetone, hydroxyacetone, OH-6-MHO, 4-OPA, 4-MON, 4-MOD, 1,4-butanedial, 5-hydroxy-4-oxopentanal, levulinic acid, 4-oxobutanoic acid	2.5 ppb 6-MHO	Wisthaler and Weschler36a
		0.15 ppb geranylacetone
		0.2 ppb 4-OPA
Human forearm, small enclosure	Geranylacetone, 6-MHO, 4-OPA	4.5 ppb geranylacetone	Morrison et al.83a
		6.2 ppb 6-MHO
		0.2 ppb 4-OPA
Human forehead, small enclosure	Acetone, geranylacetone, 6-MHO, 4-OPA, 1,4-butanedial	15 ppb acetone	Lakey et al.63
		11 ppb 6-MHO
		3 ppb geranylacetone
		1 ppb 4-OPA
Human hands enclosed in bags	Geranylacetone, 6-MHO, 4-OPA, 4-MOD, 4-MON, OH-6MHO, OH-geranylacetone	8.03 ppb geranylacetone	Zeng et al.87
		4.09 ppb 6-MHO
		0.84 ppb 4-OPA
		0.77 ppb 4-MOD
		1.64 ppb 4-MON
		0.32 ppb OH-6MHO
		0.06 ppb OH-geranylacetone
Human hair samples	Geranylacetone, 6-MHO	—	Pandrangi and Morrison41a
Human oil from fingers rubbed together	6-MHO, 4-OPA, geranylacetone, acetone	—	Fruekilde et al.70a
Pure squalene	6-MHO, 4-OPA, geranylacetone, acetone, C17-trienal, C-22 tetraenal
Pure squalene	Glyoxal, 6-MHO, 4-OPA, geranylacetone	—	Wells et al.46
Pure squalene	Acetone, formaldehyde, 4-OPA, glyoxal, pyruvic acid	—	Petrick and Dubowski57
Pure squalene	Hydrogen peroxide	(35 ± 0.5)% molar yield	Zhou and Abbatt84
Human-worn clothing	6-MHO, 4-OPA, acetone	20 ppb attributed out of 110 total measured	Wisthaler et al.35
Human-worn clothing	6-MHO, acetone	9.49 ppb acetone	Yang et al.81a
		1.46 ppb 6-MHO
Soiled t-shirts	Geranylacetone, 6-MHO, 4-OPA, 4-MON, 4-MOD, 1,4-butanedial	0.027–1.1 ppb geranylacetone	Salvador et al.88
		0.04–0.65 ppb 6-MHO
		0.06–2.2 ppb 4-OPA
Soiled clothing and aircraft materials	6-MHO, acetone, C1–C10 aldehydes	—	Coleman et al.26
Commercial flights	Acetone, 6-MHO	10.9 ppb acetone	Gao et al.74a
		1.23 ppb 6-MHO
Commercial flights	6-MHO, C6–C10 aldehydes	(1.8 ± 2.7) ppb 6-MHO	Weisel et al.79
		(0.9 ± 3.1) ppb C6–C10
Commercial flights	6-MHO	—	Wang et al.89a
24 middle school students	6-MHO, 4-OPA	0.12 ppb 6-MHO	Fischer et al.73
		0.7 ppb 4-OPA
5 male volunteers	6-MHO, geranylacetone	—	Tsushima et al.82a
Students in a university classroom	Acetone, 6-MHO, 4-OPA	Humans contributed 40% VOCs during daytime	Liu et al.75
Students in a university classroom	6-MHO, 4-OPA, geranylacetone, hydroxyacetone, 1,4-butanedial	0.2–0.6 ppb 6-MHO	Tang et al.76
		0.2–0.4 ppb 4-OPA
Two groups of 16 female volunteers in a simulated aircraft cabin	C4–C8 aldehydes, 6-MHO, 4-OPA, acetone, geranylacetone	1.5 ppb C4–C8 aldehydes	Weschler et al.34a
		4–6 ppb 6-MHO
		5–7 ppb 4-OPA
		5–21 ppb acetone
Humans in simulated office setting	Acetone, 6-MHO, geranylacetone, 4-OPA, 1,4-butanedial. Hydroxyacetone, OH-6-MHO	2.3 ppb 6-MHO	Wisthaler and Weschler36a
		2.0 ppb 4-OPA
Two occupants living in their residence	6-MHO, 4-OPA	0.036–0.36 ppb 6-MHO	Liu et al.85a
		0.056–0.46 ppb 4-OPA
Outdoor football stadium occupancy	6-MHO	0.78 ppb 6-MHO	Veres et al.77a

^aReported additional products that were derived from ozonolysis of other components of skin oil.

Various combinations of VRPs have been identified in the air of occupied and unoccupied classrooms, actual and simulated aircraft cabins, simulated office settings, a football stadium, and laboratory settings, with acetone, 4-OPA, 6-MHO, and geranylacetone being the most frequently identified compounds.^{23,26,34,36,41,46,48,57,63,73−83}Table 1 contains a summary of VRPs consistent with SqOz reactions and their respective concentrations, when measured. It is possible that a portion of the observed concentration could have been formed by the reaction of ozone with other unsaturated precursors present at the site studied. Acetone was typically present at the highest concentrations as it can be formed by ozone’s attack on the terminal double bond as well as from disassociation of intermediate compounds but was not reported by some groups due to too high of concentrations that would interfere with the measurement of other compounds and instrument function. Several studies also identified C6 to C10 aldehydes, but these have other sources beside SqOz reactions. Interestingly, a study done by Morrison et al.⁸³ found that concentrations of geranylacetone were lower than concentrations of 6-MHO for all sampling periods, which is unexpected given the structure of squalene and the expectation that both compounds would have similar mixing ratios. The discrepancy could be due to the lower volatility of geranylacetone coupled with its sorption to the sampling line. Those authors also found high variability in geranylacetone and 6-MHO concentrations between participants, suggesting that the composition and thickness of skin oils are variable.⁸³

When ozone initially reacts with one of squalene’s double bonds, a primary ozonide is formed and rapidly decomposes into a Criegee intermediate. Since the Criegee intermediate reacts with water vapor to produce an α-hydroxyhydroperoxide, which then decomposes to form an aldehyde and hydrogen peroxide, the stability of the Criegee intermediate is dependent upon the RH.⁸⁴ This mechanism and the production of hydrogen peroxide have been confirmed for SqOz reactions. Water vapor in the air facilitates the decomposition of α-hydroxyhydroperoxide, as it is a stable under low RH conditions.

Recently, concentrations of 6-MHO were observed to persist within a single-family home for 7 days after the occupants left the home.^85,86 This is likely due to ozone reacting with squalene contained within the house dust or deposited on surfaces, such as furniture. It is estimated that there is more squalene on surfaces and in the house dust than on the occupants themselves, except for when a large number of occupants is present, such as during large gatherings or parties.⁸⁵

3.3. Factors Affecting Volatile Product Concentration

The air concentrations of squalene ozonolysis VRPs have been reported to increase with time until reaching steady state and to be positively correlated with the number of occupants present.^{34,36,75,79,81} Two studies have found that 6-MHO and 4-OPA, along with other VRPs, were still elevated 3.5 h after students left their school classroom for the day.^73,76 This could be from the VRPs formed while the students were present and remained if the AER was sufficiently low, from secondary product formation, or from continued ozone reactions with soiled surfaces after the students left. While it is likely that each mentioned process contributed to the overall VRP concentrations measured, it has been estimated that the amount of off-body squalene present on the room’s surfaces can be a larger contributor to indoor air chemistry than the people themselves.⁸⁵

Increasing the ozone concentration or the amount of squalene resulted in an increase in VRP concentrations. At low to moderate ozone concentrations from 25 to 50 ppb, steady-state mixing ratios have been observed to scale linearly with increasing ozone concentration.³⁶ However, when the ozone was increased from 40 ppb to approximately 150 ppb, an almost 4-fold increase, the emissions of 4-OPA, acetone, C6–C8 aldehydes, and formaldehyde from a shirt soiled with human skin oils disproportionally increased by a factor of 2, while at lower ozone concentrations a proportional increase in products with ozone had been observed.⁹⁰ It was observed that as the ozone concentration increased, its penetration through the clothing was greater. Thus, a possible reason for the lower production of SqOz VRPs at higher ozone concentrations is that a greater percentage of the ozone is consumed by other skin oil components or the cotton fibers in the t-shirt, while at lower ozone concentrations the more reactive squalene molecules dominate. A second possibility is a shift within the Criegee pathways at the higher ozone concentrations to reaction products that were not being measured. Materials with a higher degree of soiling (i.e., higher squalene loadings), such as unwashed hair, exhibit higher emission rates of products, such as acetone and 6-MHO, with yields increasing per unit of ozone consumed by human surface reactions.^{26,41,90−93} Ventilation type, such as mixing or displacement ventilation, can affect the concentration gradient and distribution of ozone in the room by changing the air flow pattern and speed.^93,94 Wang et al.⁹³ found that the maximum variation range for the breathing zone to bulk air ratios for the overall SqOz reaction product concentrations were 1.54–2.00 and 1.12–1.69 under mixing and displacement ventilation systems, respectively. The authors did not report the ratios for individual reaction products. For mixing ventilation, concentrations of volatile compounds are relatively uniformly distributed throughout a room, while for displacement ventilation volatile compounds were found to be vertically stratified.⁹⁵ Overall, SqOz VRP formation is positively correlated with ozone concentration and the amount of squalene available to react, as estimated by the number of occupants or degree of soiling on surfaces, with the distribution of VRPs in an indoor environment altered by ventilation type.

3.4. Secondary SqOz Product Formation

Secondary reaction products are formed when primary products that contain one or more double bond, such as 6-MHO or geranylacetone, continue to react with ozone and produce new reaction products. These continued reactions also consume ozone.^9,96 4-OPA is a secondary reaction product that results from the reaction of ozone with 6-MHO, a primary SqOz product, as well as geranylacetone, C27-pentaenal, and C22-tetraenal. 4-OPA does not react further with ozone since it does not contain any double bonds. There is a lag time for the formation of secondary products, since they are a product of reactions between ozone and the primary products and subsequently take longer to reach steady state than the primary products.³⁶ As the ratio of ozone to squalene progresses from less than one to greater than one, which indicates the consumption of squalene, the ratio of secondary products to primary products increases because the production of primary products decreases. Increases in ozone concentration result in a shift from predominantly primary aldehyde reaction products toward secondary oxidation products consisting of aldehydes, ketones, and alcohols.⁵⁷

3.5. Particulate-Phase and Condensed-Phase Product Formation

Squalene ozonolysis reaction products can be found in both the particulate phase, particles that are emitted into the air, and the condensed phase, particles that remain on the reaction surface. Particles were observed to be formed by Rai et al.⁹⁷ approximately one hour after ozone was introduced into a chamber containing shirts that had been previously soiled with human skin oil from wear by a volunteer. The average particle diameter measured was around 18 nm, with the mean diameter increasing with time. This is consistent with initial particle formation via nucleation, which produces particles in the low nanometer diameter range, followed by subsequent condensation or coagulation of ultrafine particles onto themselves or other particles present within the air resulting in the observed particle growth. A secondary burst of the number of particles with a small mass medium diameter (∼20 nm) was seen after 4.7 h, which the authors suggested to also be from nucleation due to insufficient numbers of seed particles being present in the air and available for condensation.⁹⁷ Both an increase in particle number and mass have been observed as the SqOz reaction progresses.⁷⁸ Particle formation or growth was observed in real-world settings, as an ∼15% increase in total particulate mass occurred when people entered a previously unoccupied indoor space.⁹⁸ This percent particle mass increase corresponded to 0.29 μg/m³ and a 25% increase in organic aerosol mass, which the authors attributed to secondary organic aerosols (SOAs) containing compounds consistent with SqOz reactions based on aerosol mass spectrometer measurements. Overall, an increased amount of squalene produces a greater amount of particulate reaction products and affects the particle size generated by shifting product formation toward condensation mechanisms.^97,99 Condensation particle formation mechanisms, also known as partitioning mechanisms, would result in larger particle sizes due to the additional material in the air that can act as seeding materials and result in larger particle diameters compared to the very small particle diameters observed for homogeneous nucleation mechanisms.

Mechanistic nucleation models based on SVOC vapor-phase processes with subsequent condensation, but not coagulation processes, along with removal by deposition and ventilation were developed to predict particle formation from SqOz reactions. Physical modeling constants and inputs were derived from literature values and fit with controlled experimental particle generation data. Model parameters, which were applied for different particle diameter ranges and ozone concentrations, included changes in the air SVOC concentration, production of SVOCs from ozone reactions, ventilation and deposition removal mechanisms, and particle formation rates from condensation and nucleation. Several theoretical nucleation models were compared against controlled laboratory data and while they all overpredicted the particle number concentration, their temporal profile matched the initial rise in particle number concentration but did not predict the decline in particle number concentration.^35,94,96,100 Differences in particle generation were also apparent between what the authors referred to as “excess skin-oils” and “depleted skin-oils” models. It is likely that a condensation mechanism began to have a greater role than the nucleation mechanism as the reaction proceeded, resulting in less agreement between the model and controlled measurement data. The models also predicted the particle size distribution for earlier time points better than at later time points, with a skewness to a smaller particle range, and did not capture a secondary nucleation burst observed in the measurement data at a later time point. The authors used the thermodynamic nucleation model, which had the closest fit to their laboratory data, to predict ultrafine particle generation from ozone reacting with skin oils for conditions measured by others in several buildings and aircraft cabins. For classroom, office, and home settings, the predicted particle generation initially increased with AER until two to four air changes per hour, since higher AERs initially resulted in increased ozone levels indoors and more reactions, followed by a decline at higher AERs since the precursors were being removed more quickly than they could react with the ozone.⁹⁴ However, no particle data were available in these settings for comparison to the model outputs. A comparison of the model to measured ultrafine particles (UFP) was done for the aircraft cabin setting.³⁴ The model predictions were approximately 40% of the measured particle number concentration and had an ∼15 min longer lag and slower rise in particle number concentration than was measured. This may be due to a partitioning mechanism forming the particles in the aircraft rather than the nucleation mechanism that the model based on, which will produce particles more rapidly.

3.6. Characterization of the Particulate and Condensed Phases

The identity of condensed-phase reaction products was first determined through analysis of a squalene film after it had been exposed to ozone in a controlled reaction chamber, extracted from the reaction surface, and analyzed by gas chromatography–mass spectrometry (GC-MS).⁴⁶

Fewer studies have been done characterizing the chemical composition of the condensed-phase reaction products compared to the VRPs (Table 2). Zhou et al.⁶² has identified numerous compounds resulting from this reaction including: levulinic acid, succinic acid, 4-oxopentanal, secondary ozonides, and C17, C22, and C27 aldehydes. The C22 and C27 aldehydes were formed during the initial reaction phase, with their formation rates decreasing upon further oxidation, due to secondary reactions and conversion to other compounds. This would also be expected for many other unsaturated reaction products. In a squalene-limiting reaction, the concentrations of saturated lower-mass compounds increase before their concentrations level off, since they do not undergo further oxidation. A nontargeted, qualitative GC-MS analysis of SVOCs that included reaction products of ozone was done on a series of forehead skin wipe samples taken from human subjects.¹⁰² The authors observed the aforementioned SqOz reaction products previously identified and expanded the list to include C17 trienoic acid, C22 tetraenoic acid, hydroxyacetone, and geranylacetone. They further found that the top surface of the skin had the highest level of the nonvolatile ozone reaction products compared to the subsequent two wipes (an ∼0.6 μm thick layer of skin cells is removed with each wipe). Particles and skin surfaces contain volatile SVOC and nonvolatile compounds. The more volatile species gave less consistent levels, likely due to losses from the skin or from the wipes during analysis. Less volatile reaction products produced from ozone reacting with skin oils are likely to remain on the skin or can be transferred to clothing. Duncan et al.¹⁰³ points out that some of the SqOz reaction products found indoors are likely to be soluble in water, to interact with indoor dampness, and to be involved in aqueous chemistry. This is particularly true for the various indoor acids produced. The impact that particulate-phase and semivolatile squalene ozonolysis products have on indoor environments, the resulting dermal and inhalation exposures, identification of the reaction products produced, and the factors that control the formation and of particulate phase products are areas of active research.

Table 2. Particulate and Condensed-Phase SqOz Reaction Products.

Source of squalene	Compounds	Reference
Pure squalene	Geranylacetone, 6-MHO, 4-OPA, glyoxal, C17-trienal	Wells et al.46
Pure squalene	6-MHO, geranylacetone, C17-trienal, C27-pentaenal, secondary ozonides	Petrick and Dubowski57
Pure squalene	Levulinic acid, succinic acid, 4-OPA, C17-trienal, C22-tetraenal, C27-pentaenal, secondary ozonides	Zhou et al.,62 Zhou et al.101
Pure squalene	C20, C25, C30, C35, C40 secondary ozonides	Heine et al.,58 Heine et al.64
Human skin wipe samples	6-MHO, C17-trienal, C22-tetraenal, C17 trienoic acid, C22 tetraenoic acid, hydroxyacetone, geranylacetone	Garrido et al.102

4. SqOz Reaction Kinetics

Chemical kinetics of a reaction describe the rate at which and the likelihood that a reaction will occur. This information can help elucidate the reaction mechanism and factors that can influence the mechanism. The kinetics of the SqOz reaction have multiple parameters including ozone deposition velocity (v_d), reaction probability (γ), mass accommodation coefficient (α), and rate constants which may vary as a function of ozone concentrations and squalene loading. For squalene, it can be assumed that all of the ozone molecules absorbed to the surface will react, and the mass accommodation coefficient is represented by the reaction probability as under these conditions: 1/ γ ≈ 1/ α. Further, these parameters are not independent. The deposition velocity relates the dynamics of the ozone transfer and uptake to the surfaces containing squalene. The transport to the surface depends on the thickness of the boundary layer, which is a function of the near-surface airflow, while the uptake is reflected by the reaction probability and is expressed as 1/v_d = 4/(⟨v_b⟩ γ) + 1/v_t, where ⟨v_b⟩ is the Maxwell–Boltzmann velocity (for ozone: 3.62 × 10⁴ cm/s at 296 °K), and v_t is the mass transport-limited deposition velocity.¹⁰⁴ The boundary layer thickness is typically limiting when airflow conditions are stagnant and the surface contains ozone reactive species, such as squalene. These parameters impact the calculated rate constants. Studies of ozone reactions with a variety of indoor surfaces, some of which will contain squalene in addition to other unsaturated compounds that react with ozone, have been evaluated for the overall reaction kinetics of ozone indoors. It was found that as the indoor surfaces become soiled, the reaction rate decreases presumably because the substrate that ozone is reacting with is either consumed or no longer present at the surface of the material. This would occur for squalene, as well in real-world conditions, although this has not been studied for squalene^105−110 as was summarized by Weschler.⁴⁸ While these studies have reported the ozone deposition velocity, reaction probability, and reaction rates, only one study of the SqO reaction has reported multiple parameters, so each is discussed separately for squalene in the following sections. In addition, various environmental conditions can also impact these parameters as described below.

4.1. Ozone Deposition Velocity

Ozone deposition velocity is defined as the flux of ozone to the reaction surface, normalized by the ozone concentration. Mass transfer of molecules from the gas phase and the surface uptake efficiency controls the magnitude of deposition velocity for a given set of conditions.¹⁶ Ozone deposition velocity is a function of the ozone diffusion rate, as governed by Fick’s Law (ozone concentrations in the bulk air and at the surface where it reacts), whose constant typically has temperature dependence and can be affected by air flow velocity near the reaction surface. Reported ozone deposition velocities to various squalene sources range between 5.4 and 30 m/h (Table 3). The deposition velocity can be affected by ventilation type and AER, but both are less influential than the reactivity of the surface.⁹⁴ More specifically, ozone deposition velocity is larger for mixing ventilation than displacement ventilation, and ozone deposition velocity increases as AER increases. SqOz reaction products can differ by as much as 30%–40% in the breathing zone between mixing and displacement ventilation strategies.⁹⁵ Ozone concentration did not affect ozone deposition velocity.¹⁰⁵

Table 3. Ozone Deposition Velocity to Various Squalene Sources.

Source of squalene	Initial ozone concentration (ppb)	Ozone deposition velocity (m/h)	ref
16 human subjects in a simulated aircraft cabin	60–80	7.2–8.3	Tamas et al.111
2 human subjects in a simulated office setting	32–33	14.4–18	Wisthaler and Weschler36
24 children and 1 teacher in a real-world classroom	7–35	16.2	Fischer et al.73
15–64 students in a university classroom	0.9–30	9.0	Xiong et al.112
18–20 human subjects in a simulated office setting	35–60	14.4–22.3	Fadeyi et al.113
People visiting an art museum	4.5–6.6	13.68	Pagonis et al.114
Human skin, hair, and clothing for a variety of reaction probabilities	120	6.6–30	Pandrangi and Morrison41
Soiled cotton vest worn by a subject	15–165	28.44	Di et al.105
17 soiled t-shirts in a simulated aircraft cabin	60–80	6.8–9.7	Tamas et al.111
1 soiled t-shirt placed over a human simulator	22–148	5.4–10.4	Rai et al.90
4 soiled t-shirts stretched across chairs	30–60	10.8	Salvador et al.88
Simulated human figure	100	8–9	Rim et al.94
Simulated passengers in an aircraft cabin	60–80	10.4–11.9	Rai and Chen96

Direct reactions on human skin or their clothing were responsible for increased deposition and reaction rates, along with secondary reactions (on surfaces and in the gas phase) between unsaturated primary products and ozone. The presence of air circulation, such as the use of fans, appeared to increase the deposition velocity due to enhanced mixing.⁸⁸ Ozone surface deposition and gas-phase reaction rates are higher when people are present in an indoor space than for soiled surfaces alone due to greater surface areas containing the skin oils, i.e., the skin and the hair in addition to the clothing, and the continual production of skin oils by the body.^88,98 A computational fluid dynamic (CFD) model was used to estimate the ozone deposition velocity on the passengers and various surfaces in a simulated aircraft.⁹⁶ The investigators were able to show that simulated passengers, modeled by both a simple shape and a complex shape, were the most important sink for ozone removal based on higher ozone deposition velocities of 10.4 and 11.9 m/h, respectively, compared to 2.2 m/h for the carpet and 3.6 m/h for the airline seats. The passengers have more exposed surface area containing squalene to react with ozone compared to the carpet and the seats, which resulted in the increased ozone deposition velocity modeled. For some surfaces, ozone deposition velocity decreased with an increase in exposure time if the squalene was consumed completely or consumed faster than it was replaced.⁴¹

4.2. Reaction Probability

A reaction probability reflects how likely it is for a reaction to occur when reactants are mixed together. A successful reaction depends on the molecules or atoms colliding in the proper orientation, with sufficient activation energy to break existing bonds, and then form new bonds. The reaction probability is a function of the number and orientation of the reaction compounds, squalene and ozone, and is positively correlated to the ozone deposition velocity for the range measured. Additionally, SqOz reactions are on the higher end of deposition velocities measured for other indoor surfaces, which range from 0.9 to 2.7 m/h,¹¹⁵ with sufficient ozone depositing onto the surface at indoor concentrations from 10 to 50 ppb, which supports the rapid reaction rates reported. A summary of SqOz reaction probabilities can be found in Table 4. The deposition velocity was also reported by Pandrangi and Morrison.⁴¹

Table 4. Ozone Reaction Probability for Various Sources of Squalene.

Source of squalene	Ozone concentration (ppb)	Reaction probability	Reference
1.2 × 1013 molecules/cm2 pure squalene	50	(4.5 ± 1.4) × 10–4	Wells et al.46
	100	3.0 × 10–4
3.7 × 1015 molecules/cm2 pure squalene	50	(4.0 ± 2.2) × 10–4	Zhou et al.62
Pure squalene thin film	40	1 × 10–5	Petrick and Dubowski.57
Pure squalene thin film	250	1.7 × 10–4	Fu et al.116
Pure squalene aerosols	0–6000	(1.0 ± 0.2) × 10–3	Jacobs et al.117
Human hair samples	120	(0.5–4) × 10–4	Pandrangi and Morrison41
Soiled t-shirts (cotton, wool, polyester)	160	(2.1–2.7) × 10–4	Coleman et al.26
Indoor accreted mass from a residence and office building	100	(2–9) × 10–7	Gall and Rim118

SqOz reaction probability has been studied using pure squalene, and most studies have reported reaction probabilities ranging from (1.7–4.5) × 10^–4.^46,62,116 Two research groups, Jacobs et al.¹¹⁷ and Petrick and Dubowski,⁵⁷ reported larger ((1.0 ± 0.2) × 10^–3) and smaller (1 × 10^–5) values, respectively, due to the high sensitivity focused on spectral double bond decay in ATR-IR spectroscopy, as opposed to monitoring all types of bond decay, and the inclusion of thicker squalene films, respectively. The thicker squalene films used by Petrick and Dubowski⁵⁷ did not allow for all of squalene molecules to be available for reaction with ozone as they would if a monolayer was used, which resulted in the smaller reaction probability. The reaction probability of ozone and skin oils contained on human scalp hair and a soiled t-shirt was found to be (0.5–4) × 10^–4, which is similar to the value obtained from pure squalene experiments.^26,41 Squalene accounts for approximately 48% of the unsaturated double bonds in skin oil, making it one of the most reactive components found in human sebum and likely contributing approximately half of the overall reaction probability of ozone with the various compounds containing unsaturated double bonds with total human skin oils.^40,41 Gall and Rim¹¹⁸ found a much lower reaction probability of (2–9) × 10^–7 from indoor accreted mass in a residence and office building in Singapore. Possible reasons for their lower value are due to the presence of nonreactive compounds in the accreted mass and the small amount of accreted mass present that maintained potential for oxidation throughout the duration of their field deployment.

4.3. Rate Constants

The SqOz reaction has been reported to be a pseudo-first-order reaction with a rate constant of roughly (0.012–1.6) × 10^–3 s^–1 in the presence of excess ozone at 25–250 ppb ozone.^46,57,62,116Table 5 presents the reaction rate findings for SqOz reactions.

Table 5. Squalene Ozonolysis Reaction Rate Constants.

Pseudo-first-order rate constant	Excess ozone concentration (ppb)	Second-order rate constant	ref
1.6 × 10–3 s–1	50	2.6 × 10–15 cm3 molecules–1 s–1	Wells et al.46
1.2 × 10–5 s–1	40	1.2 × 10–17 cm3 molecules–1 s–1a	Petrick and Dubowski57
(2.5 ± 0.3) × 10–4 s–1	250	(4.1 ± 0.5) × 10–17 cm3 molecules–1 s–1a	Fu et al.116
(6.3 ± 0.4) × 10–4 s–1	250	(1.0 ± 0.1) × 10–16 cm3 molecules–1 s–1a	C=C bonds
			C=O bonds
(6.0 ± 0.4) × 10–4 s–1	25	(9.7 ± 0.7) × 10–16 cm3 molecules–1 s–1a	Zhou et al.62
(1.3 ± 0.1) × 10–3 s–1	50	(1.1 ± 0.1) × 10–15 cm3 molecules–1 s–1a

^aValue calculated from ozone concentration and pseudo-first-order rate constant.

Second-order rate constants were calculated from pseudo-first-order rate constants by converting ppbv ozone to molecules per cm³ air, assuming ideal gas parameters, and then dividing the pseudo-first-order rate constant by that ozone concentration. The calculated second-order rate constants vary across 2 orders of magnitude, which likely reflects the different conditions that existed among the studies. This can indicate that the reaction kinetics are not fully understood and require further investigation. Petrick and Dubowski’s⁵⁷ estimated pseudo-first-order rate constant was converted to a second-order rate constant that is roughly 45 times smaller than Wells et al.,⁴⁶ which could be due to thicker squalene films used for analysis and the inclusion of unsaturated ozonolysis products in the kinetic analysis. The rapid reaction rates measured reflect the high reactivity of ozone with carbon–carbon double bonds. Compared to the six double bonds contained in squalene, cholesterol contains one double bond and has a slower rate constant of approximately (2.4 ± 0.4) × 10^–6.^68,119 This is also reflected in the reaction probability values reported, although multiple compounds are formed through the Criegee ozonolysis mechanisms due to the multiple double bonds present in squalene. The reaction rates are sufficiently high to exceed removal through indoor air exchange processes and exceed the rates for many gas-phase homogeneous reactions of ozone with volatile terpenes.¹²⁰

5. Conditions Affecting the SqOz Reaction

Indoor air chemistry can be influenced by multiple environmental conditions, such as ozone concentration, temperature, RH, and AER. These factors alter the amount and nature of the resulting products. Changes in RH, ozone concentration, and outdoor air composition affect reaction product formation and vary across regions and seasons. Typically, RH, ozone levels, and organic loading in particles are lower in the winter than summer in many parts of the US and globally.^14,121,122

5.1. Ozone Concentration

Because ozonolysis is a nonreversible chemical reaction, it consumes both the squalene and ozone during the course of the reaction. Reductions of ozone air concentrations in a defined space due to the presence of people have been demonstrated in environmental chambers, aircraft cabins, a football stadium, office settings, and classrooms.^{34−36,63,73,77,111,113} Reactions with skin oils can remove 10%–95% of the ozone present, dependent upon the number of people or amount of skin oil present and the ozone levels.^{9,34−36,70,81,88,111,112,123} Some of the variation in observed ozone reduction may be due to variations in the amount of squalene on surfaces such as shirts, seat cushions, hair, and carpet, in each setting studied; the AER; the dimensions of the room; and the number of and characteristics of the individual people present.

An ozone gradient has been observed near the human body, indicative of ozone consumption by skin oils.^9,124 The region above the head, where the loading of sebum and squalene are highest on the body, exhibits the greatest reduction in ozone concentration.^94,96,111 Smaller rooms will have larger surface-to-volume ratios than larger rooms and remove ozone via surface reactions at a faster rate.^15,63

Increasing the ozone concentration increases product formation of squalene ozonolysis in both the volatile and particulate phases until the squalene is depleted.^{17,63,78,88,90,92,97} The increased product concentrations occur both near the body in the breathing zone as well as in bulk room air.^92,93 In general, as the ozone concentration increased, there was a shift in the products formed from predominantly aldehydes to a mixture of various ketones, aldehydes, and alcohols.⁵⁷ This shift is possibly due to increased secondary oxidation products or secondary ozonides with various C–O functionalities due to additional ozone reactions with unsaturated aldehydes.

5.2. Temperature and RH

Temperature is not likely to be an important variable in SqOz reactions because there is little variation in skin temperature where much of the reactions occur, and the temperature range of indoor environments is typically controlled.¹²⁵ Squalene deposited onto surfaces, though, is likely to be affected by surface temperature. Temperature has been shown to have no effect on primary product formation or surface kinetics, but increased temperatures can enhance desorption of products from surface sites into the air and result in faster gas-phase secondary reaction kinetics.^57,116

RH can affect squalene ozonolysis in a variety of ways. The formation of VRPs is favored under higher humidity conditions through Criegee intermediates interacting with water vapor in the air, which shifts product formation toward volatile compounds and results in increased VOC concentrations.^{26,58,64,65,68,90} An RH of 70% increased the yields of two volatile SqOz reaction products, 6-MHO and 4-OPA, by 4 and 6 fold, respectively when compared to dry conditions (0% RH).⁶⁵ Through a carbon-balance approach, particle mass lost from the condensed phase was seen in the volatile phase when the RH was increased from 0% to 30%. At high RH conditions (RH ≥ 50%), squalene ozonolysis can produce hydrogen peroxide at levels that approach a few ppb and are potentially of concern due to its oxidative capacity.⁸⁴

As RH increases, reaction kinetics remain unaffected, but the functional groups of the compounds formed were altered.^57,58,116 The formation of secondary ozonides is favored under dry conditions.^58,69,87,117 Secondary ozonides are formed through reactions of Criegee intermediates with carbonyls and greatly influenced by RH through competing reactions of the Criegee intermediates with the water vapor in the air.¹¹⁷ Additionally, increases in RH resulted in a 30% decrease in secondary ozonide particle diameter as RH was increased from 5% to 60%.⁵⁸ Increased RH shifted the products from low-molecular weight ketones to higher-molecular weight aldehydes and ketones due to water vapor quickly reacting with the Criegee intermediates, thereby changing the formation pathway. Arata et al.⁶⁵ measured the gas-phase concentrations of geranylacetone, 6-MHO, 4-OPA, acetone, and 1,4-butanedial and showed that all product concentrations had a strong dependence on RH. When the RH increased, the concentrations of each of the four products also increased. 1,4-Butanedial had an initial increase in concentration with increasing RH, but its production leveled off at higher RH values with the magnitude of the change varying with ozone. Wang and Waring⁷⁸ showed that the deposition velocity of ozone to squalene decreases as RH increases because of competition with water vapor for the reactive surface sites. Particles increased in size when formed under higher RH conditions, while the particle number remained unaffected.⁹⁷

5.3. AER

AER, or the number of times the air volume in an enclosed space is replaced per hour, influences indoor air chemistry. Gas-phase reactions are sensitive to AER, consuming more ozone at lower AERs compared to higher AERs and resulting in higher SqOz volatile and particle product concentrations.^{57,63,78,88,90,92,93,95,97,126} Many authors note that while the buildup of indoor reaction products in the indoor air, and therefore the resulting exposure, would decrease with increased AER since there is more dilution of indoor emissions by outdoor air at higher AER, but ozone exposure would increase due to increased replenishment from outdoor air.^91−93,126 As AER decreases, indoor product concentrations derived from gas-phase ozone chemistry generally increase even though indoor ozone concentrations decrease.¹²⁷

6. Exposure, Toxicity, and Health Effects

6.1. Exposure

Reactions of ozone with human skin oils have been identified as a source of VOCs and SOAs in indoor environments where people spend most of their time and are exposed to SqOz reaction products.^8,16,113 Once emitted into the indoor air, SqOz reaction products can remain airborne, become incorporated into the “personal cloud”, or sorb onto exposed surfaces, which includes skin, hair, and clothing, resulting in inhalation and dermal exposures.^63,78,102 The “personal cloud” is composed of VOCs, skin flakes, particles intrinsic to the fabric, or exogenous particles from the environment and can be released into the breathing zone by mechanical abrasion when a person moves.¹²⁸ Further, since the reactions are on the skin, clothing, and other indoor surfaces, reaction products are in very close proximity to the breathing zone and directly contact the skin.^5,129,130 Thus, the exposure concentrations for both inhalation and dermal exposures will be higher than concentrations typically estimated from background indoor air levels.^91,131 When inhaled, the aerosol diameter determines where it is likely to deposit within the respiratory tract, so particle size plays an important role in related health effects.^11,132 Product concentrations are subject to the changing parameters of the house, such as AER, air movement, and human activities. In addition to inhalation and dermal exposures, SOAs that partition into dust can be a source of ingestion exposure, which is particularly relevant for babies and toddlers because they ingest significantly more dust than adults, are much smaller, and are more vulnerable to toxins.^49,133

6.2. Toxicity and Health Effects

Approaches to evaluate potential adverse health effects to or toxicity of SqOz reaction products include the following: (1) epidemiological studies comparing levels of the SqOz reaction products in homes to symptom reports/health outcomes, (2) human-controlled exposure studies to SqOz reaction products and evaluating symptoms such as irritation, discomfort, or lung function (following ethical protocols such as using levels and durations similar to that occurring indoors that result in only transitory responses), (3) in vitro or in vivo (if data present a sufficient rationale for conducting animal studies) experiments on the mixture of VOCs and particles produced by SqOz reactions, and (4) a literature review of health effects or toxicity associated with individual compounds or mixtures of SqOz reaction products.

No epidemiological study specific to SqOz reaction products was identified, although general relationships between indoor air quality (IAQ), which would include SqOz reaction products, and health have been conducted and multiple reviews published.^134−138 While initial indoor air studies have focused on the role of common outdoor air pollution, soil gas penetration (radon), and emissions from major indoor sources such as smoking, consumer products, and building materials,^135,139 within the last couple of decades there has been recognition of the role of indoor air chemistry on IAQ and health.^8,16,138,140 However, the ability to separate the effects of SqOz products on health in a large-scale epidemiological study of indoor air is challenging, since people are exposed to many different chemicals and biological agents indoors besides SqOz reaction products. Additionally, several of the SqOz reaction products have multiple production sources that could contribute to their indoor air concentrations. Thus, epidemiological studies focused on residential indoor settings currently have not provided definitive data on the health effects associated with SqOz reactions.

One setting whose air quality may be impacted by SqOz reaction products more than in other indoor locations is a commercial aircraft cabin. This is because ozone levels can exceed those typically in residential indoor settings, and the high density of people can result in higher production and concentrations of SqOz reaction products.^72,79 Beko et al.¹⁴¹ conducted a survey on the prevalence of adverse health and comfort symptoms among commercial aircraft passengers and found a series of symptoms increased with increasing ozone levels: eye, mouth, and upper respiratory irritation, and poor rating of air quality, which may be associated with exposures to SqOz reaction products. It is possible, though, that the symptoms reported may not solely be related to SqOz reaction products since ozone would also react with other skin oils, cleaning agents used, or materials present on aircraft. Additionally, the symptoms could be caused by the elevated ozone level itself or the low RH (<15% RH) typically present in aircraft cabins. In a second study that reviewed symptoms of aircraft crew members and office workers, tired and irritated eyes were reported, which the authors suggested were linked to ozone or its reaction products from squalene or other skin oils.¹⁴²

As with the epidemiological studies, the first controlled exposures examining effects of IAQ on people, animals, or cells evaluated effects of VOCs commonly present indoors because of their use in commercial and consumer products^143,144 or outdoor pollutants that penetrated indoors¹⁴⁵ and did not include the majority of oxygenated compounds that are produced by SqOz reactions. Additional controlled studies on the effects of IAQ containing high concentrations of oxygenated compounds due to occupant activities are needed.^146−150 Subjects in a controlled human study conducted in a simulated aircraft cabin where ozone was introduced in their presence reported increased numbers of eye and nasal irritation, lip and skin dryness, headache, dizziness, mental tension, and claustrophobia.¹⁵¹ The symptoms are most likely related to the ozone reaction products from skin oils, particularly squalene, since the study was done at two different AERs, and the same ozone concentration and a greater number of symptoms were reported when there were higher oxidation and ozone–squalene product concentrations.

A literature review of reported toxicological effects of the individual or mixtures of compounds produced by SqOz reactions via in vitro or in vivo experiments also does not provide a complete picture potential health concerns from SqOz reaction products. Some of the SqOz products are not unique to this reaction and have been studied in regard to other potential exposures. However, other SqOz products have few, if any, other sources in the environment, and available toxicology data on those products are limited.^152,153Table 6 lists the SqOz reaction products reported in the literature and potential adverse health outcomes.

Table 6. Potential Adverse Health Effects of SqOz Reaction Products.

Compound	Health effect(s) reported	Dose range	Study type (animal, cell, human)	ref
6-MHO	Sensory irritation, airflow limitation, and pulmonary irritation	34–1186 ppm	Animal-Mouse	Wolkoff et al.154
4-OPA	Sensory irritation, decrease in tidal volume, airflow limitation, and pulmonary irritation	3.4–444 ppm	Animal-Mouse	Wolkoff et al.154
	Increases in mRNA expression of the inflammatory mediators IL-6 and IL-8, GM-CSF, and TNF-α	15–65 ppm	Human epithelial lung cells	Anderson et al.155
	Dermal exposure: irritant and sensitizer; Pulmonary exposure: nonspecific airway hyperreactivity, irritant, and allergic responses	50 μL of 0.5%–10% 4-OPA in acetone	Animal-Mouse	Anderson et al.156
Acetone	Irritation of eyes, nose, throat, and trachea, increased response times, increased white blood cell counts, headache	0–12,000 ppm	Human	Agency for Toxic Substances and Disease Registry (ATSDR)157
Glyoxal	Irritation of mucous membranes, skin, eyes, respiratory organs, gastrointestinal tract	30%–40% glyoxal in water solution	Humans (skin), Animal-Rabbit (eye), Animal-Rat (remaining)	Kielhorn et al.158
Formaldehyde	Carcinogenic		Human	International Agency for Research on Cancer (IARC)159
	Mild irritation to rabbit skin	10% formaldehyde solution in water	Animal-Rabbit	Phillips et al.160
	Severe eye irritation (epithelium, deep stroma, endothelium)	37% formaldehyde solution in water	Animal-Rabbit	Maurer et al.161
	Moderate to no skin irritation, severe to mild intradermal reaction, moderate to no eye irritation	0.75%–37% formaldehyde solution in water	Animal- Rat, Mouse, Guinea Pig	Sekizawa et al.162
	Eye and nose irritation, pulmonary fibrosis, lung injury	Vapor from 40% formaldehyde solution in water	Animal-Rabbit	Chinedum et al.163
Levulinic acid	Capable of inducing serious eye damage	750 μL levulinic acid solution	Bovine cornea	European Chemical Agency (ECHA)164
Succinic acid	Severe ocular irritation and corneal opacity	0.1 mL into the conjunctival sac of 1 eye	Animal-Rabbit	Hartwig and MAK Commission165
Hydrogen peroxide	Corrosive damage (eye, skin, mucous membranes), oxygen gas formation (ingestion), lipid peroxidation (cytotoxic)	3%–35% solutions in water	Literature review-human	Watt et al.166
Pyruvic acid	a	—	—	No data available
Geranylacetone	a	—	—	No data available
Hydroxyacetone	a	—	—	No data available
1,4-Butanedial	—	—	—	No data available
5-Hydroxy-4-oxopentanal	—	—	—	No data available
4-Oxobutanoic acid	—	—	—	No data available
OH-Geranylacetone	—	—	—	No data available
OH-6MHO	—	—	—	No data available
4-MON	—	—	—	No data available
4-MOD	—	—	—	No data available
C17 aldehyde	—	—	—	No data available
C22 aldehyde	—	—	—	No data available
C27 aldehyde	—	—	—	No data available
C17-trienal	—	—	—	No data available
C22-tetraenal	—	—	—	No data available
C27-pentaenal	—	—	—	No data available
C17 trienoic acid	—	—	—	No data available
C22 tetraenoic acid	—	—	—	No data available

^aListed as a skin, respiratory and/or ocular irritant on multiple material safety data sheets but primary reference source for effects study is not given.

In summary, the most frequent human health effects reported to be associated with SqOz reactions include respiratory inflammation, dermal and ocular irritation, and cardiovascular limitations. However, long-term studies of health effects from chronic exposure to many SqOz reaction products are not available, so more persistent or long-term adverse health outcomes are not known. In addition, no toxicological or adverse health data were identified for the majority of the known reaction products, so the overall potential for adverse health effects associated with SqOz reaction products still needs to be determined. Once the effects of individual compounds are identified, it is also necessary to consider if the compounds produced have similar modes of action, and if so, exposures to multiple compounds concurrently should be considered when evaluating the potential risk of SqOz reaction products.

7. Knowledge Gaps and Future Directions

The SqOz reaction proceeds through Criegee ozonolysis mechanisms forming multiple oxygenated compounds, although our knowledge of which formation pathway is preferred under various environmental conditions is incomplete. Formation of more volatile species occurs as the RH increases and increases in secondary Criegee intermediates with higher ozone levels. However, future studies are needed to elucidate the affect environmental conditions (squalene loading, RH, AER, ozone concentration) have on the amount of (on a molar yield basis) and chemical identity of products formed. Multiple studies have measured volatile reaction products (e.g., 6-MHO, 4-OPA, geranylacetone) in controlled settings utilizing pure squalene and skin oil deposited on clothing as the substrate for the reaction with ozone and in a variety of environmental settings with a range of occupancy. However, fewer studies, predominantly in controlled laboratory settings, have studied SqOz products in a particulate phase or on surface or skin wipe samples. Additional elucidation of the formation of secondary ozonides and organic hydroperoxides are needed. Future studies should be done to evaluate the compounds produced by SqOz reactions in complex real-world environments, where exposures occur and the specific compound profiles produced differ from controlled laboratory settings. These studies should consider that squalene is on multiple indoor surfaces (e.g., skin, clothing, hair, furniture, household surfaces, bedding, and in house dust) which can potentially affect the mechanisms and kinetics of product formation and release into the indoor air. Such studies should examine the lifetime of squalene and measurements of SqOz products on actual indoor surfaces in homes, schools, and offices, including the orientation and distribution of squalene molecules on surfaces which could affect its ability to react with ozone. The orientation of the squalene molecules at the surface of the dust will alter the efficiency of the SqO reaction. Additionally, if the layer of skin oil on the surface is thick or is below other components of the dust, ozone may not be able to penetrate the skin flakes so would not react with squalene molecules in the layers underneath the surface.

A major knowledge gap is the lack of information on the potential adverse health effects of the majority of individual SqOz reaction products and at realistic environmental mixtures. Toxicity studies should be conducted evaluating realistic environmental mixtures using in vitro approaches and controlled human studies for both inhalation and dermal acute exposures to assess irritation responses and other acute end points. Further, no toxicity studies of chronic exposure to most individual SqOz reaction products or mixtures have been conducted. Thus, as exposure profiles are determined in indoor settings, in vitro or in vivo chronic toxicity studies of them should be done. Consideration should also be given to evaluate health outcomes in populations that experience potentially higher exposures to SqOz products due to continually being in locations containing high density of people and elevated ozone levels. These may include service attendants in aircraft or public transportation modes in locations with elevated ozone levels and teachers in crowded classroom. As with most epidemiological studies, it will be important to access the actual exposure to SqOz products, since indoor sources can be highly variable and dependent on the exact conditions of the indoor settings being evaluated and the presence of confounding exposures.

On the basis of our assessment of the current state of knowledge regarding SqOz chemistry, and the difficulty of reproducing the complex conditions indoors that likely affect SqOz formation, it is important that future research emphasizes real-world studies. Real-world systems will require measurements of the multiple variables that potentially alter the SqOz reaction, such as the number and time profiles of people present, estimates of surface area potentially soiled with skin oils, AERs, ozone levels indoors, temperature, and RH profiles. Ideally, the many dynamic existing sources and sinks of the SqOz reactants and products will be identified to provide key information for indoor air chemistry models. Collaborative efforts, such as those of HOMEChem¹⁶⁷ and ICHEAR,¹⁶⁸ have begun to look at more complex environments, and some individual groups began to look at an occupied residence,⁸⁵ classrooms,^73,75,76 an art museum,¹¹⁴ and commercial flights,^74,79,89 which should provide more information about the air chemistry within the dynamic indoor environments with human subjects.

Author Present Address

^§ B. Coffaro: Eastman Kodak Company, Specialty Chemicals, Inks, and Dispersions Division, 343 State Street, Rochester, New York, 14650, United States

The authors declare no competing financial interest.