A review of microbial and chemical assessment of indoor surfaces

Abstract

This review emphasizes the importance of classification of surface reservoirs represented by building and indoor materials. Models developed in the past 45 years to study mainly non-reactive thermodynamic partitioning of air pollutants to surfaces as well as their validation through analytical measurements are discussed. Analytical techniques aiming at surface characterization have been summarized. Applications relevant to surface chemistry performed indoors in the last decade are also given. Finally, techniques for microbiological sampling to characterize the hygienic state of surfaces have also been compiled. Readers will be introduced to the state-of-the-art of analytical techniques that are being used indoors for (physico)chemical and microbiological characterization of surfaces. The focus is on real-time in situ instrumentation used to understand chemical processes rather than air quality monitoring indoors. Our aim was to give an overview of useful instrumentation and a tutorial on analytical approaches and challenges represented by surfaces.

KEYWORDS:

• Bioaerosol
• cavity ring-down spectroscopy
• FTIR
• mass spectrometry
• real-time analysis
• VOC

1. Rationale of the present review

Humans breathe daily about 10m³ air and spend most of their time indoors (home, commuting, office) as the global population becomes more and more urban.^[1–3] Even though the air chemistry is not different just for changing the location, it is important to highlight some differences between the processes that happen indoors compared to those outdoors. Indoors, there is less light and due to air exchange (AE), only short times are available for reactions involving gaseous and particle constituents. Besides, gas and particle compounds can be significantly different in composition and/or concentration.^[4] Dwellings are not the same than in the previous decades. Due to global warming, there is the need to minimize AE with the outdoors in order to save energy. Moreover, plenty of gas–surface partitioning and multiphase reactions occur in the large surface reservoirs indoors, since some volatile organic compounds (VOCs) behave as semi-volatiles (SVOCs) indoors due to the partitioning processes. In the indoor atmosphere, most chemical constituents reside in surface reservoirs, rather than in the gas phase.^[5] Multiphase chemical reactions occur both indoors and outdoors. One important difference is that indoors, several reactions occur mainly on macroscopic surfaces, months or even years are needed for multiphase chemical reactions to occur in indoor surface reservoirs.^[5] The reaction times indoors are controlled by the exchange rate of indoor air with outdoor air. Only reaction times that compete with the exchange rate can have an impact on the indoor air chemistry.^[6] Indoors, the volatility of compounds is strongly dependent of the sorption capacity of static indoor surfaces. The much larger abundances of fixed sorptive surfaces in the indoor environment favor partitioning of gas pollutants to these surfaces.^[7] Species that have primarily indoor sources will decrease in concentration as the air exchange rate (AER) increases, whereas the converse is true for those species that have predominantly outdoor sources.^[8]

The objective of this review was two-fold: i) to collect all the information from the last years on sampling methods and instrumentation that allows to determine the transformation of pollutants on indoor surfaces and ii) to help planning a measurement campaign in real places (houses, offices, etc.) aiming at monitoring surface transformations of gaseous pollutants indoors. However, emissions or building material aging, surface transformation of particulate matter or humans as surfaces will be considered less, as these large topics go beyond the scope of the present review. Up to our knowledge, this is the first attempt to review chemical and microbiological surface characterization together.

2. Characterization of surface reservoirs

The accessible parts of the building materials and furnishings, the organic- and water-rich surface films formed on them, and the gas–surface interface are referred to as surface reservoirs, or simply surfaces.^[5] The most common surfaces indoors have been compiled in Table 1^[9–37]. Wang et al.^[32] defined the surface reservoir as a condensed-phase material containing chemical constituents, especially molecules, undergoing exchange with the gas phase through i) adsorption to clean glass^[38] and other building materials, ii) dissolution into thin surface organic films formed from uptake of gaseous species or particle deposition,^[24] or iii) diffusion into the underlying building materials and furnishings, such as wallboard and paint.^[39] Surfaces are referred to the exposed outside parts of building materials, coatings, and furnishings, as well as those within these materials that are accessible to indoor air via relatively rapid mass transport.^[11] Weschler and Nazaroff^[11] considered the air-surface interface of bulk water and organic films. The third dimension of surfaces is the thickness that is much larger than molecular scale. The extent of surface thicknesses that can interact considerably with indoor air components remains not well understood.^[11] The time needed for diffusive transport through a thin layer of a homogeneous substance is the inverse square of the thickness of the layer.^[11]

Table 1. Characteristics of construction and interior surface materials influencing partitioning of indoor contaminants between gas-phase and surface reservoirs.

Feature	Classification	Examples	Ref.
Chemical composition	inorganic	metallic (Al, Fe, Zn) and alloys (steel) mineral-based: cement, ceramics, glass, gypsum, lime, mud, stone	^[Citation9^] ^[Citation10^] ^[Citation11^] ^[Citation12^] ^[Citation13^]
	organic	natural organic: wood; polymeric: linoleum, nylon, PVC
	mixed (composites)	ceramic/metallic/polymer matrix; concrete	^[^13^,^14^]
Coatings	inorganic	metallic, enamel paint	^[Citation15^]
	organic	acrylic, epoxy, paints (enamel, emulsion, oil, etc.), polyurethane, wallpaper, wood stain	^[^9^,^12^]
Heterogeneity	large	both chemically and physically	^[Citation5^]
Hygroscopicity	water adsorption	by metal oxides (silica); several monolayers of water at 50% RH with growing average thickness as the RH increases	^[Citation16–18^]
		by mineral-based construction materials moisture content (MC) <1%	^[Citation11^]
		by wood span, MC = 8–11% at RH = 50% RH	^[Citation11^]
	water adsorption by materials	gypsum (i.e., interior of wallboard); cellulose (e.g., in cotton fabrics)	^[^19^,^20^]
	competition for adsorption sites between water and contaminant	n.a.	^[^21^,^22^]
	changing species diffusivity in matrices	organic matrices (e.g., paint affecting species viscosity)	^[Citation23^]
	hydrophilic / hydrophobic	cotton fabric, glass, gypsum in wallboard, nylon, stainless steel / candle wax, chromium, polypropylene, silver	^[Citation11^]
Orientation	horizontal	Aluminium and zinc surfaces; upward facing surfaces	^[^24^,^25–28^]
	vertical	Aluminium and zinc surfaces; window glass	^[Citation24^] ^[^24^,^25–29^]
Porosity	porous	upholstery fabric; wallboard	^[Citation5^]
	impervious	glass; stone
Permeability	permeable	gypsum (water vapor permeability, VOCs, etc.); paints (VOCs), painted plaster, painted wallboard, plush (fabrics, carpets and upholstery), wall-papered wallboard	^[^11^,^10^,^30^]
	impermeable	ceramics, glass, metal, porcelain, window glass; permeable hard surfaces (finished laminate and wood, unfinished wood, painted wood, plastic); almost impermeable: enamel paint, glazed ceramic tiles	^[^24^,^10^,^31^]
pH	direct effect	alkalinity of concrete	^[Citation5^]
	partitioning of acids and bases in the polar reservoir on the surface	carboxylic acids; NH3	^[Citation32^]
Polarizability	non-polar	carpet as a strong sorptive sink for non-polar organics	^[Citation33^]
	polar	virgin gypsum board as sink for highly polar organics
Refractoriness	refractory	granite; gypsum; silica; stainless steel,	^[Citation5^]
	flammable	insulation materials; upholstery components wood
Room element	basic room envelope	ceiling, floor, skylight, wall (divider)	^[Citation10^]
	envelope features	windows, doors, area rug
	large standing objects	furniture, appliances and fixtures
Room type	conceptual	house, office, residence	^[Citation9^]
	experimental	(Al, stainless steel) chamber, test room	^[Citation9^]
	miscellaneous	1. common areas (living-dining room, kitchen, hallway/foyer as common rooms); 2. bedrooms/offices; 3. bathroom	^[Citation10^]
		1. kitchen; 2. bedrooms; 3. offices	^[Citation9^]
Shape	irregular	block, stone-shaped	^[Citation34^]
	regular	cylinder, flat, open top container, rectangular prism, sphere	^[Citation9^]
Surface thickness	> 10 mm	gypsum board (13 mm)	^[Citation11^]
	<1 mm	paint (30–150 μm); 25–63 μm paint on tin-plated steel; 75–125 μm primer & two layers paint on drywall	^[^30^,^35–37^]
Surface (S) volume (V) ratio	large	typically, S/V = 3 m^−1 (S/V = 1.7–2 m^−1without contents and S/V = 3.0–3.6 m^−1 with contents)	^[^32^,^9^]
Roughness	rougher materials, fibrous materials	carpets
	smooth, flat materials	aluminium; glass; plastic; stainless steel	^[Citation9^]

Abbreviation: n.a. = not available; RH = relative humidity

As Nazaroff and Weschler pointed out,^[11] there is a lot of literature addressing chemistry and catalysis on ideal surfaces, but information on real-world surfaces is still scarce. In a typical room, indoor surfaces comprise floor, walls, ceiling, and furnishings. The surface in contact with room air usually differs from the underlying material that constitutes the bulk of the indoor material because floors may be varnished, covered by carpets, tiles or synthetic flooring (e.g., vinyl, linoleum, laminate) as well as polished, waxed, or coated with a polymer. External walls have windows having blinds, shades or drapery. Ceilings may be painted or covered with tiles. Occupants also contribute to indoor surfaces with their clothing, skin, and hair.^[11] The main characteristics of indoor materials influencing partitioning of an indoor contaminant between the gas-phase and surface reservoirs have also been compiled in Table 1.

Some comments in addition to the contents of Table 1 and the aforementioned definition are discussed below. Firstly, water is an important constituent in indoor environments (Table 1). Besides water vapors, it may be present indoors in condensed forms such as bulk, sorbed and particle-phase water as well as aqueous surface films. Bulk condensed water can be as large in abundance as its vapors.^[11] Acids and bases can partition into bulk condensed water from the gas-phase and undergo acid-base or redox reactions. Water sorption by nylon and for wool increased specific surface areas by 145 and 215 time, respectively, as compared to N₂.^[40]

Competition for adsorptive sites between water and contaminants (e.g., trimethylamine, TMA) on carpet, latex-painted wallboard and mineral surfaces has been observed not only for water but also for common indoor gases such as carbon dioxide (CO₂) and ammonia (NH₃).^[41,42] For example, the capacity of the carpet and mineral surfaces to sorb TMA was twice and 40–50%, respectively, when the CO₂ mixing ratio was increased to 1000 ppm at high relative humidity (RH) values, but NH₃ decreased it for all surfaces. The aqueous acid-base chemistry of TMA appeared to also contribute to the overall sorptive capacity of carpet at high RH. Site competition appeared to follow the Langmuir model.^[41]

Chemical analyses of nm-thick films developing on windows have identified several carboxylic acids, aromatic acids and alkanes.^[43–45] Weschler and Nazaroff^[24] estimated that the thickness growth rate of SVOCs with 10 ≤ logK_oa ≤ 13 on window films after 2 month-long exposure would be about 9 nm, or about 2× the associated contribution from sorbed water at 50% RH. On the other hand, diffusion of contaminant molecules into and out paint possesses relevant timescales.^[30,46] For example, latex paint is one of the most commonly used such material in residential environments.^[9] The resulting film after evaporation of its high water content^[47] usually consists of an organic binder, TiO₂ (to enhance light scattering and to hide the backing material), fillers (e.g., CaCO₃), preservatives, antimicrobial agents and surfactants.^[48] The organic binders used are often methyl methacrylate, butyl acrylate or styrene.^[49] For such films, the diffusion coefficients vary between 10⁻¹⁴–10⁻⁶ cm² s⁻¹ depending on the polymer and organic diffusant.^[50–52] The paint can be formulated to be sufficiently porous to permit AE with the walls and ceilings.^[46] Dried paint thickness is about 50 μm.^[53] In this case, the diffusion time through the paint layer is about 1 h. Several paints like casein or lime paint and fine finishing mortar from clay do not have influence on water vapor permeability of paint-plaster system.^[54] Weschler and Nazaroff^[7] introduced the term persistence defined as the ratio of the total mass (mass sorbed to fixed surfaces + mass sorbed to airborne particles + mass in gaseous state) to the removal rate (rate at which mass in the gas phase and sorbed to airborne particles is removed from the room by means of ventilation). The strongly sorbed compounds can persist for months if their removal is possible only via AE.^[55] The diffusion of molecules through thin surface films is fast, i.e., about 1 s for 10 nm-thick films if the diffusion constants, D_f are 10–12 cm² s⁻¹. The D_f values are much larger for many organic substrates (e.g., cooking oil).^[56] Only 10 nm-thick films of highly viscous materials will lead to diffusion times longer than 1 s.^[57] Also, the desorption for molecules adsorbed via van der Waals forces or H-bonding are also expected to be short for low molecular weight (LMW) SVOCs (e.g., calculated desorption timescale limonene sorbed on silica for to be on the order of tens of μs^[38]). The diffusion time of molecules within a building material may be longer. Mass transfer timescales may be even longer for vinyl flooring or concrete.^[33,58] There may be SVOCs in enclosed spaces with insulation in a wall cavity that exchange with indoor air on a slower timescale.^[59]

Surface roughness was indicated as a key factor in studies with painted tubes.^[60,61] The effect of surface undulations on the generation of turbulence in a laminar flow-tube experiment has already been studied.^[62] For the height of surface undulations (h), the following equation was proposed:(1) $$h =\frac{\mathrm{d}}{\sqrt[3/4]{\mathrm{Re}}}$$(1) where d is the diameter of the flow tube, and Re is the Reynolds number. If this critical height, h is of the same magnitude or greater than the applied film thickness, partitioning to the surface may be faster than predicted for fully laminar flow.^[30]

Surface to volume (S/V, m⁻¹) ratios are orders of magnitude larger indoors than outdoors (Table 1). The calculation of this ratio is not obvious as some studies considered only macroscopic surface area, no attempt was made to account for the additional area associated with rough (e.g., carpeted) or porous surfaces (e.g., wood, paint, polymers).^[9,10] Microscopic surface area is much larger than these reported S/V values.^[63] The surface area of carpet samples exceeded the floor area covered by factors of 30–66.^[63] The surface area of indoor airborne particles contributes negligibly to total indoor surface area.^[64] Finished/painted surfaces accounted for a substantial fraction of total surface area in different types of rooms (e.g., bedroom, bathroom, common areas, offices) equipped with metal, glass, ceramic/porcelain/tile, finished wood, unfinished wood, painted wood, PVC, other plastic, painted/papered plaster and wallboard, thin fabrics, upholstery/carpet, and paper in the following order: painted/papered plaster and wallboard (about 30% of the sum of the medians) vinyl (PVC) and other plastic surfaces (13–20%) > finished wood (5–19%) > textiles and fibrous materials (carpets, fabrics, and upholstery) (9–15%) ≅ painted wood (9–13%).^[10] Impermeable surfaces were extensive in bathrooms, with median S/V for metal, glass, ceramic, porcelain, and tile summing to 1.3 m⁻¹ (27%).^[10] Although Manuja et al.^[9] used somewhat different material categories (i.e., cardboard, concrete, fabric/fiber, glass, metal, paint, paper, plastic, wood, etc.), painted surfaces and stained/finished-wood surfaces accounted for almost two-thirds of the surfaces in the rooms investigated.^[9,10] Since the substrates beneath the paint are often gypsum wallboard or pressed wood composites, both the paint^[30] and the substrate are somewhat permeable, suggesting that painted surfaces and stained wood could serve as substantial sinks for gas-phase species [e.g., nitrous acid (HONO), formic acid (HCOOH) and acetic acid (AcOH)] taking also into account that the moisture content at 50% RH of painted gypsum board is 0.5–1.1%, while that of wood is 8–10%. Human occupants can contribute considerably to the total surface area of the rooms. An adult has a total body surface area of ≅ 2 m².^[65] If two adults occupy a 30-m³ room with an S/V of 3.5 m⁻¹, the human surfaces contribute about 4% to the total surface area omitting the fact that human skin has a 4.5 to 6 pH^[66,67] and hair are covered by surface lipids (about 25% of which are organic acids).^[68] Moreover, human bodies are covered by clothes that can have substantial moisture content at typical indoor humidity.^[11]

3. Sources and transformations of the main indoor contaminants

Indoor pollutants – chemical, physical (ionizing and non-ionizing radiation) and biological (microorganisms, allergens, viruses, etc.) type of agents – acting individually or combined with other factors, can cause a decrease in environmental comfort and can represent a risk for human health.^[69] Since the main gaseous contaminants indoors are well-known, we are going to focus only on their transformation processes involving surface reservoirs. Then, classification of biological agents relevant for indoor environment is presented.^[70–180]

3.1. Transformations of chemical contaminants indoors

Emission indoors or species infiltration from outdoors are considered primary sources for chemical contaminants indoors. Heterogeneous or photochemical reactions occurring indoors constitute the secondary sources. Sources can emit either continuously or episodically. For example, hydrogen peroxide and chlorine (Cl₂) cleaning agents are used episodically houses that were flooded after the Hurricane Katrina.^[70] Several contaminants have multiple sources. For example, indoor NH₃ is emitted from cleaning products, tobacco smoke, building materials, and humans. Also, PAHs can arise indoors from incomplete combustion (e.g., candle burning, cooking, and smoking).

Some transformations of indoor air pollutants give rise to appearance of new contaminants. Therefore, it is sometimes difficult to discriminate between sources and transformations of pollutants indoors. A typical example for that is the water-mediated disproportionation of nitrogen dioxide (NO₂) that can be considered as a way for its removal but also a secondary source for HONO and HNO₃. With the evaporation of water on indoor surfaces containing aldehydes in addition to amines and amino acids, low volatility imidazoles and other nitrogen-containing oligomers may form.^[71]

Another example is the sorption (e.g., by paper) and slow conversion of sulfur dioxide (SO₂) to sulfuric acid.^[72–74] The conversion of S(IV)→S(VI) on both Fe and Zn was found to have a half-life of ≅ 24 hours.^[75] The surface uptake of SO₂ is common to be characterized as first order kinetics (i.e., the uptake rate is proportional to the gas-phase SO₂ concentration).^[76] On the other hand, Edwards et al.^[72] reported that the rate was proportional to the square root of the gas-phase SO₂ concentration. This might be expected if the rate-limiting step is governed by the concentration of HSO₃⁻ that is proportional to the square root of the gas-phase SO₂ concentration.^[11]

Transformation of chemical constituents indoors may proceed not only via heterogeneous chemistry with gas-phase oxidants or photochemistry but via non-reactive partitioning. This means, for example, dry deposition. Deposition velocity (v_d) is defined from F_S,  flux to a surface that is directly proportional with v_d and the gas-phase concentration of the contaminant. The rate of the v_d is influenced by the ability of the target compound to undergo gas-phase mass transfer and by the uptake coefficient (γ_S).^[63] This latter corresponds to the collisions with the surface of a gas-phase molecule that leads to removal from the gas phase.

It is common to model the irreversible loss of a reactive gas on indoor surfaces through the use of a deposition velocity, v_d.^[11] In the absence of indoor emission sources, and assuming that i) penetration from outdoors with ventilation occurs without loss, and ii) the primary removal processes from indoors are by means of ventilation and irreversible surface uptake, the average indoor/outdoor concentration ratio, f, is reasonably well-modeled by Eq. (2)(2) $$\mathrm{f}=\frac{\mathrm{AER}\times \mathrm{V}}{\mathrm{AER}\times \mathrm{V}+{\mathrm{v}}_{\mathrm{D}}\mathrm{S}}$$(2) :(2) $$\mathrm{f}=\frac{\mathrm{AER}\times \mathrm{V}}{\mathrm{AER}\times \mathrm{V}+{\mathrm{v}}_{\mathrm{D}}\mathrm{S}}$$(2) where AER is the air-exchange rate, V is the interior volume of the building, and S is the area of surfaces on which irreversible uptake occurs.^[11] However, the calculated v_D of 0.18 m h⁻¹ based on experimental S/V^[77] for residential rooms and communities^[78] and known AER values was somewhat smaller than what might be inferred from values reported by Grøntoft and Raychaudhuri.^[79]

Ozone uptake by surfaces is the most studied. The O₃ uptake has been studied on carpets,^[63,80] clothing and fabric,^[80–82] green building materials,^[83] human hair, cooking and skin oil,^[84–87] insulation materials,^[59] ventilation filters and ducts.^[88–90] Its deposition velocities are frequently high.^[79] Generally, newly purchased items (e.g., carpets) or freshly painted surfaces have a high reactivity toward O₃.^[91,92] By contrast, kitchen tables remained reactive over 2 yrs due to the constant use of cooking oils.^[91] The O₃ deposition velocity is large for two reasons. First, O₃ decomposes to form O₂ on mineral dust,^[93] although some degree of the reactivity regenerates with time. On the other hand, O₃ is a potent indoor oxidant. The RH-dependent deposition velocities (cm s⁻¹) for O₃, NO_x and SO₂, usually determined with the decay method, have been reported for several (in)organic indoor materials.^[79] The OH radical is formed in the presence of O₃ but it cannot be transported from outdoors due to its short lifetime. The NO₃ radical is formed in the reaction between NO₂ and O₃.^[94] Indoors, the gas-phase photolytic lifetime of NO₃ can be longer than 104 s^[95]).

The v_D values for LMW organic acids are comparable to those for SVOCs.^[11] Since the indoor gas-phase concentrations of HCOOH and AcOH are orders of magnitude larger than those of indoor SVOCs, the flux of LMW organic acids to an aqueous surface is orders of magnitude larger than that of SVOCs.^[11] However, SVOCs accumulate much faster than particle-associated water-soluble salts on impermeable indoor surfaces.^[11] It was observed that the initial SVOCs in the film oxidize increasing the oxygen-to-carbon (O/C) ratio and favoring water uptake during time periods with higher indoor humidity, eventually rendering the surface film more hygroscopic and less viscous.^[96,97] This phenomenon also results in the acquisition of inorganic salts. Oxidation of surface films may also lead to phase separations^[98] that might further influence acid/base reactions. According to Nazaroff and Wechsler,^[11] there may exist pockets of aqueous solutions and hydrophobic organics within porous surfaces. In this case, organic acids and bases could partition to both aqueous and organic substrates, and the relative amounts in each phase could influence the surface chemistry.^[11]

3.2. Microorganisms and bioaerosols

Biological agents most commonly present in indoor environments include pollen, bacteria, filamentous fungi and viruses as well as microbial metabolites and fragments. Biological particles released from terrestrial and marine ecosystems are carried in the air as bioaerosols that can be suspended in the air for a long time.^[99–102]

Aerosolized particles are mostly aggregated to each other or to inert, suspended particles and incorporated in the droplet-nuclei whose size determines their fate. Thus, air represents the vehicle through which biological agents are transported, reach the surfaces and settle there. Inhaled by humans, bioaerosols can be deposited in the respiratory tract, and potentially cause irritation, allergies, contagious infectious disease, acute toxic effects, and even death if the concentrations are high.^[103–106] For example, legionellosis, influenza, measles, and tuberculosis, are often spread by aerosols in poorly ventilated environments.^[107–109]

Microbes are carried to indoors both through airborne transmission and along with soil, clothes and pets. In addition to the transfer from outdoors, there are significant indoor sources such as humans themselves, organic materials and microbial proliferation on wet building materials or water reservoirs.^[110,111] Environmental microbes are ubiquitous and mostly harmless or even beneficial to health as seen in the diminished cases of asthma and allergy in farming environments.^[112]

Molds are microscopic, primarily multi-celled and filamentous fungi composed of hyphae and mycelium.^[113] Cladosporium, Penicillium, Alternaria and Aspergillus are the most prevalent fungal genera.^[114–117] The various species of Cladosporium are of particular interest as they also colonize indoor intermittently-wetted surfaces, for example, due to winter condensation on exterior walls.^[118] Molds are normally present indoors at levels which do not have any impact on healthy individuals.^[119] Between 20% and 40% of the indoor climate in European and North American buildings is affected by mold.^[120] Mold does not damage the building material directly, but its presence refers to high RH in the structure and possible risks such as decomposition.^[121] There is a strong relationship between excessive moisture, mold growth and respiratory health effects.^[122,123]

Humans themselves are significant contributors to the indoor microbiome. People both shed and resuspend particularly bacteria such as Streptococcaceae, Lactobacillaceae, and Pseudomonadaceae, usually considered as weak pathogens.^[124] However, some pathogenic bacteria such as Legionella and Mycobacterium proliferating in water systems or fecal bacteria Escherichia coli and Salmonella spp. can cause severe infections particularly in immunosuppressed individuals.^[125] The bacterial and fungal communities have a complex relationship with surfaces. For example, the pH of aqueous surface films influences which bacterial or fungal species thrive. In return, bacterial and fungal species can alter the pH of surface moisture to promote their own growth.^[124]

Viruses are a special group of biological agents. The first reviews on viruses in indoor environments were published in the 1980s.^[126,127] Since then, viruses belonging to different families have been identified in the air and on the surfaces of different types of indoor environments. First viruses that were identified in the air of confined environments were adenoviruses causing respiratory symptoms in military recruits.^[128] Other examples of viruses relevant for public health are influenza- and rhinoviruses, which are the most common cause for upper respiratory tract infections, as well as coronaviruses including SARS-CoV-2.^[125]

Microorganisms present indoors are able to produce constitutive and inducible hydrolytic enzymes such as amylase, keratinase, collagenase, chondroitin sulphatase, elastase, phosphatase, phospholipase, gelatinase, hyaluronidase, lecithinase, lipase, pectinase, proteinase, urease.^[125] They play a physiological role in morphogenesis, but also, during microbial metabolism, SVOCs and VOCs can be released from microbial cells.^[129] Up to now, >200 microbial volatile organic compounds (MVOCs) have been identified. Ninety-six of them are regarded as typical MVOCs, but none are considered to be exclusively of microbial origin.^[130,131] Indoor concentrations of single MVOCs encompass about six orders of magnitude, ranging from several ng m⁻³ to 1 mg m⁻³, the accumulation enhanced by favorable growth conditions and low AERs.^[131] Exposure to MVOCs can irritate eyes and upper airways and enhance airway inflammation, possibly contributing to the development of asthma and allergies.^[132,133]

Structural components of microbes include potent bioactive compounds such as endotoxins (pyrogenic cell wall component of gram-negative bacteria) and glucans (proinflammatory fungal cell wall component).^[112] The components of biological agents can act also as allergens, eliciting allergic response on repeated contact. Indoor allergens come from various sources: house dust mites, cockroaches, pets (dog, cat), fungi and pollen.^[125]

4. Indoor environments investigated

Past studies on partitioning of gas-phase molecules to surfaces have primarily been conducted in chambers equipped with (acrylic painted) indoor wall, carpet, furnishing, or window.^{[30,33,134–136]} Emissions of (S)VOCs from specific materials present indoors have most commonly been studied in such controlled environments.^[134–137] These experiments have demonstrated the ability of the gas-phase molecules to partition to indoor surface reservoirs.^{[30,135,136,140]} The decay to steady state during sorption is monitored after injection of the target analyte into the chamber^[30] (Figure 1). Moreover, desorption can be studied after cease of the exposure^[30] (Figure 1). Field studies under human occupancy constitute a new trend. Thus, in a recent study, partitioning of contaminants to textile materials (e.g., clothing) in flow-through chambers was studied.^[141]

Figure 1. Schematic of the portable surface reactor (a) and the painted tube apparatus suitable to determine absorptive partitioning of VOCs into paint. Adapted from Ref.^[30]

There are only a few non-reactive partitioning studies in genuine indoor spaces (e.g., model room, residential rooms, daycare centers) carried out with HONO emitted from gas stoves^[142] and VOCs (e.g., terpenes, large aromatics),^[77,134,143] these latter by either injection^[77,134] or surface wipes.^[144] The dynamic, time-resolved nature of the sorption of select organic compounds in residential rooms has been demonstrated by Singer et. al,^[77] although relatively slow time-response, off-line chemical analysis was used.^[77]

Recently, the dynamic response of SVOCs was reported through measurements performed every hour.^[137] One of the most important collaborative field investigations of recent years involving monitoring of surface transformation of indoor air pollutants was the House Observations of Microbial and Environmental Chemistry (HOMEChem) study conducted in a 110-m² single-story manufactured home equipped with two independent heating ventilation and AC (HVAC) systems. The HOMEChem study was aimed to study how everyday activities influence the emissions, chemical transformations and removal of gaseous contaminants and particles indoors.^[29] An overview of the HOMEChem study including the experimental design and instrumentation used has been described elsewhere.^[29]

5. Models and proxies for thermodynamic partitioning of air pollutants to particulates and indoor surfaces

Due to the challenges arisen at the experimental determination of partitioning of gaseous contaminants to particulate matter (PM) and indoor surfaces, these processes have been modeled in depth for gas phase, particles and, recently, for organic films or their surrogate, 1-octanol (Table 2). The processes described below refer mainly to non-reactive conditions. Thus, equilibrium partitioning of an SVOC between air and surfaces has been described using i) adsorption models (i.e., partitioning to a surface/interface) and ii) absorption models (i.e., partitioning into a bulk phase). Evolution of the main theoretical approaches for study partitioning of SVOCs between air and condensed phase is compiled in Table 2.

Table 2. Summary of model approaches published in the past 45 years on partitioning of (semi)volatile pollutants between gas and solid phases.

Mathematical expression(s)	Target analyte	Surface/bulk phase	Notes	Ref.
$\mathrm{\phi }=\frac{{\mathrm{c}}_{\mathrm{J}}{\mathrm{\theta }}_{\mathrm{J}}}{{\mathrm{p}}_0+{\mathrm{c}}_{\mathrm{J}}{\mathrm{\theta }}_{\mathrm{J}}} (3)$	Trace constituents	n.d.	BET isotherm as a starting point in the development of the equation, the simplifying assumptions equivalent to a linear Langmuir isotherm	^[Citation145^]
$\lg \frac{[{\mathrm{C}}_{i,\text{ vapor phase}}\times {\mathrm{C}}_{\mathrm{TSP}}] }{{\mathrm{C}}_{i,\text{ particulate phase}}}=-\frac{\mathrm{const}}{\mathrm{T}}+\mathrm{B}\mathrm{ }(4)$	1) Three- to five-ring PAHs; 2) Organochlorine compounds.	particulate matter (PM)	Langmuir adsorption concept	^[Citation146,Citation147^]
$\mathrm{\phi }=\frac{{\mathrm{b}}_{\mathrm{L}}\mathrm{R}\mathrm{T}{\mathrm{\theta }}_{\mathrm{J}}{\mathrm{N}}_{\mathrm{s}}}{1+{\mathrm{b}}_{\mathrm{L}}\mathrm{R}\mathrm{T}{\mathrm{\theta }}_{\mathrm{J}}{\mathrm{N}}_{\mathrm{s}}} (5)$	n.d.	PM	Langmuirian gas-particle distribution (isotherm linearization when 1 ≫ bLP)	^[Citation148^]
$\frac{\mathrm{dM}}{\mathrm{dt}}={\mathrm{k}}_{\mathrm{a}}\mathrm{C}-{\mathrm{k}}_{\mathrm{d}}\mathrm{M}\mathrm{ }(6)$	1) Tetrachloroethylene (perchloroethylene); 2) Ethylbenzene.	1) carpet; 2) gypsum drywall with an interior latex paint;3) composite ceiling tile; 4) window glass; 5) upholstery	1) Langmuir adsorption; 2) Chamber testing.	^[Citation135^]
${\mathrm{K}}_{\mathrm{p}}=\frac{\mathrm{F}}{{\mathrm{C}}_{\mathrm{TSP}}\times \mathrm{A}} (7)$ $\text{log }{\mathrm{K}}_{\mathrm{p}}={\mathrm{m}}_{\mathrm{r}}\text{log p}°+{\mathrm{b}}_{\mathrm{r}}$ (8)	1) C18-C24 alkanes; 2) PAHs; 3) PCBs.	urban & rural PM	1) br values should be similar, mr values should be near − 1; 2) Thermodynamic sources of variability: compound-to-compound differences in adsorption, event-to-event variability in the specific surface area of the aerosol and ambient T; 3) Non-thermodynamic sources of variability: sorption of gaseous analytes to the filters used in differentiating between F and A, non-exchangeable components for F, within-event adsorption/desorption kinetics, contaminant levels and T	^[Citation149^]
$\lg {\mathrm{K}}_{\mathrm{p}}=-\lg {\mathrm{p}}_{\mathrm{L},\mathrm{i}}^0+\lg \frac{{\mathrm{f}}_{\mathrm{OM}}760\mathrm{R}\mathrm{T}}{{\mathrm{MW}}_{\mathrm{om}}{\mathrm{\zeta }}_{\mathrm{i}}{10}^6}\mathrm{ }(9)$	1. PAHs: chrysene, benz[a]antracene + triphylene; 2. organochlorines	urban PM, TSP	1) Regardless of the adsorptive vs absorptive partitioning, log Kp vs $\lg {\mathrm{p}}_{\mathrm{L},\mathrm{i}}^0$ expected to have a slope of near −1.0. 2) Increasing atsp and $e^{(Q_i-Q_V)/RT}$ favor adsorption; 3) Increasing fom, decreasing MWom, and decreasing ζ favor absorption.	^[Citation150^]
cf. Eq. (9)	1) PAHs: naphthalene, acenaphthane, fluorene, phenanthrene, anthracene, fluoranthene, pyrene, chrysene; 2) C16-C25 n-alkanes.	1) ammonium sulfate, dioctyl phthalate (DOP) and secondary organic aerosol (SOA) from the photooxidation of gasoline vapor in a Teflon chamber; 2) urban PM during summer smog episodes	1) DOP aerosol good surrogate for ambient aerosol consisting of organic compounds from primary emissions; 2) Environmental tobacco smoke particles good surrogate for SOA; 3) The sorption properties of ambient smog aerosol and the chamber-generated SOA are very similar; 4) Literature SOA yield data from smog chamber studies suitable to predict the extent of SOA formation during summer midday smog episodes.	^[Citation151^]
$\ln {\mathrm{K}}_{\mathrm{i}}^{\mathrm{abs}}=-\left(1+\mathrm{s}\right)\mathrm{l}\mathrm{n}{\mathrm{p}}_{\mathrm{iL}}^0+\mathrm{const}\mathrm{ }(\mathrm{ }10)$ $\mathrm{s}=\frac{\Delta \ln {\mathrm{\gamma }}_{\mathrm{i}}^{\mathrm{abs}}}{\Delta \ln {\mathrm{p}}_{\mathrm{iL}}^0} (\mathrm{ }11)$	1) n-alkanes; 2) chlorobenzenes; 3) PCBs.	1-octanol as surrogate for condensed phase	1) Absorption from the gas phase into a bulk phase; 2) Within a given compound class, both ln p°iL and the log of the solubility (ln γ^iabs) proportional to the molecule size.	^[Citation152^]
$\ln {\mathrm{K}}_{\mathrm{i}}^{\mathrm{ads}}=-0.133\sqrt{{\mathrm{\gamma }}^{\mathrm{vdw}}}\mathrm{l}\mathrm{n}{\mathrm{p}}_{\mathrm{iL}}^0+2.09\sqrt{{\mathrm{\gamma }}^{\mathrm{vdw}}}+2.08{\mathrm{\beta }}_{\mathrm{i}}\sqrt{{\mathrm{\gamma }}^{+}}+1.37{\mathrm{\alpha }}_{\mathrm{i}}\sqrt{{\mathrm{\gamma }}^{-}}-19.5$  (12) $\log {\mathrm{K}}_{\mathrm{i}}^{\mathrm{ads}}=\mathrm{mlog}{\mathrm{p}}_{\mathrm{iL}}^0-6.83\mathrm{m}-14.6\mathrm{ }(13)$ cf.^[^153^]	a) alkanes, chlorobenzenes; b) groups containing O or N or aromatic rings), which differ by molecular entities that interact by van der Waals forces only (typically alkyl groups); c) aromatic rings of the PAHs	a) Teflon, paraffin, polyethylene; b) quartz; c) electron accepting surfaces.	1) Eq. (12) describes adsorption from the gas phase to a surface when the adsorbates cover all possible interactions: the first two terms of the equation consider van der Waals interactions, while the next two, Lewis acid-base interactions; 2) Cf. Eq. (12): the acid-base interactions between the compounds i and surface are (a) 0; (b) constant for all compounds; (c) change proportionally with their lnp°L.	^[Citation152,Citation154^]
$\mathrm{V}\frac{\mathrm{d}{\mathrm{C}}_{\mathrm{g}}}{\mathrm{dt}}={\mathrm{Q}}_{\mathrm{g}}{\mathrm{C}}_{\mathrm{g},\mathrm{in}}-{\mathrm{Q}}_{\mathrm{g}}{\mathrm{C}}_{\mathrm{g}}-{\mathrm{k}}_{\mathrm{a}}{\mathrm{C}}_{\mathrm{g}}{\mathrm{A}}_{\mathrm{s}}+{\mathrm{k}}_{\mathrm{d}}{\mathrm{M}}^{\mathrm{n}}{\mathrm{A}}_{\mathrm{s}}$ (14) $\frac{dM}{dt}={\mathrm{k}}_{\mathrm{a}}{\mathrm{C}}_{\mathrm{g}}-{\mathrm{k}}_{\mathrm{d}}{\mathrm{M}}^{\mathrm{n}}$ (15)	VOCs (isopropanol, MTBE, cyclohexane, toluene, ethylbenzene, tetrachloroethene, 1,2-dichlorobenzene, 1,2,4-trichlorobenzene)	carpet, gypsum board, upholstery, vinyl and wood flooring, acoustic tiles, and fruit in stainless-steel chambers	1) Linear adsorption/desorption model used; 2) Carpet and virgin gypsum board as significant sorptive sink for non-polar and highly polar VOCs, respectively; 3) Increases in RH increased the degree of sorption of isopropanol to carpet.	^[Citation33^]
$\frac{{\mathrm{K}}_{\mathrm{PA}}}{{\mathrm{f}}_{\mathrm{OM}}}={\mathrm{K}}_{\mathrm{OM}/\mathrm{A}}=\frac{{\mathrm{V}}_{\mathrm{oct}}{\mathrm{\gamma }}_{\mathrm{oct}}{\mathrm{K}}_{\mathrm{OA}}}{{\mathrm{V}}_{\mathrm{OM}}{\mathrm{\gamma }}_{\mathrm{OM}}} (\mathrm{ }16)$ $\log {\mathrm{K}}_{\mathrm{OM}/\mathrm{A}}=\log {\mathrm{K}}_{\mathrm{OA}}+\log \frac{{\mathrm{\gamma }}_{\mathrm{oct}}}{{\mathrm{\gamma }}_{\mathrm{OM}}}+\mathrm{c}\mathrm{ }(\mathrm{ }17)$	non-polar SVOCs (chlorobenzenes, PCBs, naphthalenes, dibenzo-p-dioxins, dibenzofurans, chlorinated, brominated diphenyl ethers)	1-octanol	1) γOct = 1 − 10 suggesting that PL and KOA are very highly correlated; 2) The estimated standard deviation of γOct tends to be so large that PL and KOA are indistinguishable within the measurement uncertainty.	^[Citation155^]
${\mathrm{K}}_{\mathrm{part}}=\frac{{\mathrm{C}}_{\mathrm{part}}}{{\mathrm{C}}_{\mathrm{a}}}=\frac{{\mathrm{f}}_{\mathrm{om}-\mathrm{part}}{\mathrm{C}}_{\mathrm{om}}}{{\mathrm{C}}_{\mathrm{a}}}={\mathrm{f}}_{\mathrm{om}-\mathrm{part}}{\mathrm{K}}_{\mathrm{oa}} (\mathrm{ }18)$ ${\mathrm{K}}_{\mathrm{oa}}=\frac{{\mathrm{C}}_{\mathrm{o}\_\mathrm{sat}}}{{\mathrm{C}}_{\mathrm{a}\_\mathrm{sat}}}=\frac{{\mathrm{C}}_{\mathrm{o}\_\mathrm{sat}}}{\frac{{\mathrm{P}}_{\mathrm{s}}}{\mathrm{RT}}}$ (19) ${\mathrm{K}}_{\mathrm{dust}}={\mathrm{f}}_{\mathrm{om}-\mathrm{dust}}{\mathrm{K}}_{\mathrm{oa}}$ (20) ${\mathrm{K}}_{\mathrm{surf}}={\mathrm{K}}_{\mathrm{oa}}$ (21) ${\mathrm{K}}_{\mathrm{hum}}={\mathrm{K}}_{\mathrm{oa}}$ (22)	SVOCs(biocides/preservatives, combustion byproducts, flame retardants, heat-transfer fluids, microbial emissions, personal care products, pesticides, plasticizers, sealants, stain/water, oil repellents, surfactants, emulsifiers)	OM sorbed to all indoor surfaces, including dust and airborne particles	1) Ideal solution behavior; 2) Even the most impenetrable indoor surfaces have films at their interfaces with room air; 3) For Koa >108, partitioning into the OM in surface films is a key fate; 4) Affinity of an SVOC for sorbed OM similar to its affinity for 1-octanol;5) The equilibrium dermal concentration of an SVOC estimated as the product of its Koa, the thickness of the organic film on the skin surface, and its gas-phase concentration.	^[Citation156^]
${\mathrm{\tau }}_{\mathrm{s}}\sim \frac{{\mathrm{M}}_{\mathrm{s}}}{{\mathrm{F}}_{\mathrm{s}}}=\frac{{\mathrm{K}}_{\mathrm{oa}}\mathrm{X}}{{\mathrm{v}}_{\mathrm{d}}} (\mathrm{ }23)$		sorption reservoirs modeled as thin films or slabs of large extent; slab thicknesses in indoor environments vary from ∼10 nm (e.g., organic film on a window) to ∼1 mm (e.g., grease on a kitchen surface).	1) Affinity of the SVOC for the sorption reservoir is the same as its affinity for 1-octanol; 2) Partition into the sorptive reservoir is uniform; gas-phase concentration is maintained constant indefinitely; 3) The mass-transport-limited vd cf.^[^156^]; vd, ∼3 mh^−1. For Koa = 10^7–10^12 and varying X: a) fast sorption: τs < 1 h, competition with pollutant removal from air by ventilation; b) moderate-rate sorption, reversible sorption: 1 h < τs < 1 y; c) slow sorption: τs > 1 y, irreversible sorptive uptake; 4) The rate-limiting process for partitioning is external mass transfer, internal transport time scales are low.	^[Citation7^]
${\mathrm{k}}_1=\frac{{\mathrm{k}}_{\max }\left[{\mathrm{O}}_3\right]{\mathrm{K}}_{\mathrm{ozone}}}{1+\left[{\mathrm{O}}_3\right]{\mathrm{K}}_{\mathrm{ozone}}}(24)$	benzo[e]pyrene (PAH)	glass	1) pseudo first-order rate coefficients, k1 extracted from BeP decay curves due to oxidation with O3; 2) Langmuir − Hinshelwood adsorption.	^[Citation157^]
${\mathrm{K}}_{\mathrm{s}}=\left(\frac{{\mathrm{C}}_{\mathrm{inc}}}{{\mathrm{C}}_{\mathrm{ss}}}-1\right){\mathrm{K}}_{\mathrm{f}} (\mathrm{ }25)$ ${\mathrm{C}}^{\mathrm{*}}=\frac{\mathrm{\gamma }{\mathrm{M}}_{\mathrm{W}}\mathrm{P}}{\mathrm{RT}}(\mathrm{ }26)$ $\mathrm{logP}\mathrm{ }\left(\mathrm{atm}\right)=\log {\mathrm{C}}^{\mathrm{*}}\left(\frac{\mathrm{\mu }\mathrm{g}}{{\mathrm{m}}^3}\right)-10.0\mathrm{ }(\mathrm{ }27)$ $\log {\mathrm{K}}_{oa}=-0.988\mathrm{l}\mathrm{og}{\mathrm{C}}^{\mathrm{*}}\left(\frac{\mathrm{\mu }\mathrm{g}}{{\mathrm{m}}^3}\right)+11.58$ (28)	series of 2-ketones or carboxylic acids, or single carboxylic acids	Teflon chamber studies with painted wallboard surface	1) Partitioning of VOCs to glass and painted gypsum wallboard using a portable surface reactor; 2) P calculated cf. ^[^158^]; 3) Plotting Ks vs. C^* 4) Using the Koa, cf. Xiao and Wania^[^155^] (Eqs. (11) and (12)).	^[Citation30^]
${\mathrm{D}}_{\mathrm{f}}=\frac{4{\mathrm{L}}^2}{{\mathrm{\pi }}^2\mathrm{r}} (\mathrm{ }29)$ ${\mathrm{d}}_{\mathrm{p}}=\frac{\mathrm{\lambda }}{2{\mathrm{n}}_2\mathrm{\pi }\sqrt{{\text{sin}}^2\mathrm{\theta }-(\frac{{\mathrm{n}}_1}{{\mathrm{n}}_2}{)}^{2}2}}\mathrm{ }(\mathrm{ }30)$	1) C1-C6 carboxylic acids; 2) C2, C4, and C6 1-alcohols; 3) C6, C9, and C12 n-alkanes; 4) C6, C8, C10, and C12 2-ketones.	paint on poly(methyl methacrylate) polymers	1) Df for VOCs in paint extracted from time profiles of the C–H stretch at ∼2900 cm^−1 plotted against time: exponential curve fitting to extract τ; 2) Df extracted from τ depends on the penetration depth of the IR waves; when the polymer film is much thicker than dp, Eq. (29) is considered ^[^159^].	^[^30^]
${\mathrm{k}}_{\mathrm{f}}=\mathrm{ }\frac{\mathrm{Q}}{{\mathrm{V}}_{\mathrm{Bin}}}$(31) ${\mathrm{k}}_{\mathrm{a}}=\frac{8{\mathrm{D}}_{\mathrm{g}}}{{\mathrm{d}}_{\mathrm{t}}^2}$(32) ${\mathrm{k}}_{\mathrm{d}}=\mathrm{ }\frac{{\mathrm{k}}_{\mathrm{a}}}{{\mathrm{K}}_{\mathrm{gw}}}$(33) ${\mathrm{K}}_{\mathrm{gw}}=\frac{{\mathrm{C}}_{\mathrm{w}}}{{\mathrm{C}}^{\mathrm{*}}}$(34) ${\mathrm{C}}_{\mathrm{wS}}=\mathrm{ }{\mathrm{L}}_{\mathrm{S}}\left(\frac{{\mathrm{C}}_{\mathrm{w}}}{\mathrm{L}}\right)$ (35) $\frac{\mathrm{d}\left[{\mathrm{G}}_{\mathrm{i}}\right]}{\mathrm{dt}}=\mathrm{ }{\mathrm{k}}_{\mathrm{f}}{[\mathrm{G}}_{\mathrm{i}-1}]-{\mathrm{k}}_{\mathrm{f}}[{\mathrm{G}}_{\mathrm{i}}]+{\mathrm{k}}_{\mathrm{d}}[{\mathrm{SL}}_{\mathrm{i}}]-{\mathrm{k}}_{\mathrm{a}}{[\mathrm{G}}_{\mathrm{i}}]$ (36) $\frac{\partial {\mathrm{C}}_{(\mathrm{x},\mathrm{t})}}{\partial \mathrm{t}}=\mathrm{ }{\mathrm{D}}_{\mathrm{f}}\frac{{\partial }^2\mathrm{C}}{{\partial \mathrm{x}}^2}$(37) $\frac{{\mathrm{C}}_{(\mathrm{i},\mathrm{ }\mathrm{t}+1)}-{\mathrm{C}}_{(\mathrm{i},\mathrm{ }\mathrm{t})}}{\text{Δt}}={\mathrm{D}}_{\mathrm{f}}\left(\frac{{\mathrm{C}}_{(\mathrm{i}+1,\mathrm{ }\mathrm{t})}-2{\mathrm{C}}_{(\mathrm{i},\mathrm{ }\mathrm{t})}-{\mathrm{C}}_{(\mathrm{i}-1,\mathrm{ }\mathrm{t})}}{{\mathrm{\Delta }\mathrm{x}}^2}\right)$(38)	series of 2-ketones or carboxylic acids, or single carboxylic acids	painted tube containing several bins	1) Diffusion of VOCs into the paint film using a numerical approach cf.^[^160^]; 2) To determine the concentration gradient of VOC, the film was split into layers, and mass transfer between layers cf. Fick’s 2^nd law of diffusion; 3) The coefficient 8 in Eq. (32) changed to 24 to account for surface roughness of the paint layer.	^[Citation30^]
${\mathrm{k}}_{\mathrm{f}}=\mathrm{AER}$(39) ${\mathrm{k}}_{\mathrm{a}}=0.59\mathrm{ }\frac{{\mathrm{A}}_{\mathrm{P}}}{\mathrm{V}}\frac{2}{\mathrm{\pi }}\sqrt{{\mathrm{k}}_{\mathrm{e}}\mathrm{D}}$(40)		house model (single longitudinal-bin painted tube box with AE, gas-wall partitioning, diffusion in the paint, and irreversible uptake by the wallboard)	1) The governing equations for the house model are Eqs. (31)–(38). 2) Cw determined by multiplying A/V by a 40–240 µm range of paint film thicknesses and D calculated for VOCs using the relationship with C^*; 3) Compounds allowed to diffuse across the paint-wallboard barrier, but not to return (infinite sink).	^[Citation30^]
${\mathrm{F}}_{\mathrm{gas}}=\frac{1}{\left(1+K_{oa}\frac{V_{\mathit{org}}}{V_{\mathit{gas}}}+K_{wa}\frac{V_p}{V_{\mathit{gas}}}\right)}(\mathrm{ }41)$ ${\mathrm{F}}_{\mathrm{p}}=\frac{1}{\left(1+\frac{1}{{\mathrm{K}}_{\mathrm{wa}}\frac{{\mathrm{V}}_{\mathrm{p}}}{{\mathrm{V}}_{\mathrm{gas}}}}+\frac{{\mathrm{K}}_{\mathrm{oa}}}{{\mathrm{K}}_{\mathrm{wa}}}\frac{{\mathrm{V}}_{\mathrm{org}}}{{\mathrm{V}}_{\mathrm{p}}}\right)}(\mathrm{ }42)$ ${\mathrm{F}}_{\mathrm{org}}=\frac{1}{\left(1+\frac{1}{{\mathrm{K}}_{\mathrm{oa}}\frac{{\mathrm{V}}_{\mathrm{org}}}{{\mathrm{V}}_{\mathrm{gas}}}}+\frac{{\mathrm{K}}_{\mathrm{wa}}}{{\mathrm{K}}_{\mathrm{oa}}}\frac{{\mathrm{V}}_{\mathrm{p}}}{{\mathrm{V}}_{\mathrm{org}}}\right)}(\mathrm{ }43)$ $\frac{{\mathrm{V}}_{\mathrm{org}}}{{\mathrm{V}}_{\mathrm{gas}}}={\mathrm{X}}_{\mathrm{org}}\frac{\mathrm{S}}{\mathrm{V}} ; \frac{{\mathrm{V}}_{\mathrm{p}}}{{\mathrm{V}}_{\mathrm{gas}}}={\mathrm{X}}_{\mathrm{p}}\frac{\mathrm{S}}{\mathrm{V}} ;\frac{{\mathrm{V}}_{\mathrm{org}}}{{\mathrm{V}}_{\mathrm{p}}}=\frac{{\mathrm{X}}_{\mathrm{org}}}{{\mathrm{X}}_{\mathrm{p}}}(\mathrm{ }44–46)$ for acids: ${\mathrm{K}}_{\mathrm{eff},\mathrm{wa}}={\mathrm{K}}_{\mathrm{wa}}(1+{10}^{\mathrm{pH}-{\mathrm{pK}}_{\mathrm{a})}}$ (47) for bases: ${\mathrm{K}}_{\mathrm{eff},\mathrm{wa}}={\mathrm{K}}_{\mathrm{wa}}(1+{10}^{{\mathrm{pK}}_{\mathrm{a}}-\mathrm{pH})}$ (48)	1) low molecular weight acids (nitrous, C1-C9 carboxylic and isocyanic acids); 2) phenol; 3) NH3; 4) terpenes, isoprene, furfural ; 5) D5 siloxane.	1) 1-octanol as surrogate for paint (a weakly polar sorbing material); 2) Water as surrogate for hydrated gypsum in wallboard (a more polar medium).	1) Polar and weakly polar partitioning media of variable volumes present indoors; 2) Molecular geometry irrelevant; 3) Partitioning among indoor air, the polar and the weakly polar reservoirs assumed to be the only processes affecting gas-phase concentrations; 4) In indoor environments Vgas ≫ Vp and Vgas ≫ Vorg, Vgas = V (volume of the house); 5) Typically, S/V = 3.	^[Citation32^]
${\mathrm{k}}_{\mathrm{a}}^{\mathrm{\text{'}}}=\frac{{\mathrm{k}}_{\mathrm{a}}{\mathrm{h}}_{\mathrm{m}}}{{\mathrm{h}}_{\mathrm{m}}+{\mathrm{k}}_{\mathrm{a}}}$ (49) ${\mathrm{k}}_{\mathrm{d}}^{\mathrm{\text{'}}}=\frac{{\mathrm{k}}_{\mathrm{d}}{\mathrm{h}}_{\mathrm{m}}}{{\mathrm{h}}_{\mathrm{m}}+{\mathrm{k}}_{\mathrm{a}}}$ (50)	1) formaldehyde; 2) toluene.	PVC, rubbers, and linoleum flooring	1) Models demonstrated that sorption on floorings can modify the indoor concentrations under certain conditions of low AER; 2) Rubber as a sink of formaldehyde characterized by a negligible desorption constant.	^[Citation161^]

Observations: 1. For nomenclature and Greek symbols in equations, see below; 2. For abbreviations other than mathematical ones, see text.

Nomenclature (dimension when provided) :A = gaseous concentrations of the compound of interest, ng m⁻³; AER = air exchange (rate); A_P = surface area of the walls (here the surface area of paint); A_s = sink area (m²); a_tsp = specific surface area of suspended particles, m²/g; B = constant; b_L = constant in Langmuir isotherm, atm⁻¹; b_r = intercept in a plot of log K_p vs log p°_L; C = concentration, mg m⁻³ ; C_a = steady gas-phase concentration of OM; C_{a_sat} = saturated concentration in gas-phase; C_i = individual concentrations expressed as ng m⁻³; C_inc = incoming concentration after the cell has been passivated but before the plunger is raised$; {\mathrm{C}}_{(\mathrm{i}-1,\mathrm{ }\mathrm{t})}$ is the concentration in layer i −1 at time t$; {\mathrm{C}}_{(\mathrm{i},\mathrm{ }\mathrm{t})}$ = concentration of VOC in layer i at time t; ${\mathrm{C}}_{(\mathrm{i},\mathrm{ }\mathrm{t}+1)}$ = concentration in layer i at time t + 1; C_g = VOC concentration in chamber air, mg m⁻³; C_g,in = VOC concentration in inlet air, mg m⁻³; C_part = equilibrium concentration of an SVOC in airborne particles (i.e., the particle-associated SVOC mass per volume of airborne particles); C_om = equilibrium concentration of an SVOC in the OM sorbed to indoor surfaces; C_{o_sat}= saturated concentration in octanol; C_ss = steady-state concentration of VOC; C_TSP = concentration of total suspended particulate material, μg m⁻³; C_w = equivalent absorbing mass of the painted walls; C^* = saturation vapor concentrations; c_J = constant in Junge’s (1977) equation; D = the diffusion coefficient for VOCs in the laminar boundary layer at the walls ; D_f = diffusion coefficient; D_g = gas phase diffusivity ; d_p = penetration depth; d_t = inner diameter of the tube applied; F = the particulate-associated concentration of the compound of interest, ng m⁻³; F_gas = fraction of a chemical in indoor air at equilibrium; F_org = fraction of a chemical in weakly-polar organic reservoir at equilibrium; F_p = fraction of a chemical in polar reservoir at equilibrium; F_S= flux to a surface; f_om = weight fraction of OM in the sorbing phase; f_{om_part} = OM fraction in the airborne particle-phase, F_s = initial SVOC mass flow rate, mass per time; [G_i] = gas phase concentration of VOC in bin i;[G_i⁻¹] = concentration in the bin immediately upstream of bin i; h_m = convective mass transfer coefficient; K_aw = air-water partition coefficient; K_dust = partition coefficient for settled dust; K_f = rate constants for flow; K_gw = gas-wall partitioning equilibrium constant; K_hum = partition coefficient associated with human occupants; K_i^abs = absorption constant of compound i into a bulk phase; K_i^ads = adsorption constant of compound i into a bulk phase; K_{eff, wa} = effective Henry’s law constant ; K_oa = octanol-air partition coefficient ; K_OM = OM–air partition coefficient ;K_p = temperature-dependent partitioning constant, m³μg⁻¹ ; K_PA = bulk phase–gas partition coefficient ; K_part = partition coefficient for airborne particles ; K_s = rate constant for loss to the surface; K_surf; = partition coefficient for fixed surfaces; K_wa = water-air partitioning coefficients, m³ air m⁻³ water; k_a (’)= (apparent) adsorption rate, m h⁻¹ ; k_d (’)= (apparent) desorption rate, h⁻¹$; {\mathrm{k}}_{\mathrm{e}}\mathrm{ }$= coefficient of eddy diffusion; k_f = rate constant for transfer between tube bins; k₁ = pseudo first-order rate coefficient; k_max = maximum pseudo first-order rate coefficient; L = film thickness; L_S = thickness of the surface layer bin; M = mass per unit area in sink, mg m⁻²; MW = molecular weight of the polymer (assumed to be 1 and 250 g/ mol); MW_OM = molecular weight of the OM; m = slope; N_s = moles of sorption sites per cm² of aerosol, mol cm⁻² ; n = constant that accounts for non-linearities in the desorption process; n₁, n₂ = refractive indexes of the crystal and paint taken to be 45°; OM = organic matter; P = vapor pressure; P_s = saturation vapor pressure; p⁰ = saturation vapor pressure at T of interest, torr; p⁰_L = sub-cooled liquid vapor pressure, torr; Q = volume flow rate, Q_i = enthalpy of desorption from adsorbing surface, kJ/mol; Q_V = enthalpy of vaporization of the liquid, subcooled if necessary, kJ/mol; R = universal gas constant, 8.314 Pa m³ mol⁻¹ K⁻¹, S = surface area; [SLi] = concentration of VOC in the surface layer of the paint film in bin i.; s = constant; T = temperature; t = time, h; V = volume of chamber/room; V_bin = volume of bin; V_gas = volumes of indoor air (i.e., house volume), m³; VOM = average molar volume; V_org = volume weakly-polar organic (represented by octanol) reservoir, m³; V_p = volume of polar (represented by water) reservoir, m³; v_d = deposition velocity, mh⁻¹; X = thickness of the large surface area.

Greek symbols α_i = parameter that describes the electron acceptor properties, dimensionless; β_i = parameter that describes the electron donor properties, dimensionless; γ = the activity coefficient of the VOC in the polymer; γⁱ_abs = activity coefficient of compound i in the absorbing phase; γ_Oct = activity coefficient of the chemical in octanol at infinite dilution; γ_OM = activity coefficient of the chemical in the OM phase; γ_S = uptake coefficient; γ^vdW = van der Waals parameter of the surface, mJ m⁻²; γ⁺ = electron acceptor parameter of the surface, mJ m⁻²; γ^- = electron donor parameter of the surface, mJ m⁻²$; \mathrm{\Delta }\mathrm{x}$ = width of each layer$;\mathrm{\Delta }\mathrm{t}$ = time step; ζ_i = activity coefficient of compound i in the OM phase on the mole fraction scale; θ = angle of incidence of the light beam; θ_J = concentration of aerosol surface area, cm² cm⁻³ $; {\mathrm{\tau }}_{\mathrm{s}}$= time scale to achieve equilibrium; φ = fraction sorbed on aerosol particles, dimensionless.

Thus, the first partitioning models developed by Junge^[145] in 1977 was applied by Yamasaki for explaining adsorption of PAHs onto PM.^[146] Indoor chemical constituents can also partition into aerosol particles, although these latter possess much lower total volumes than the surface reservoirs. Moreover, timescales for partitioning with aerosol particles are much shorter than with surface films. The initial idea was that simple physical adsorption would dominate the sorption process for solids such as atmospheric PM consisting of mainly mineral materials.^[162] However, suspended PM may contain organic matter (OM) (e.g., particles of plant wax). Also, urban PM always contains amorphous organic carbon from primary emissions,^[163] and from the formation of secondary organic aerosol.^[164] While partitioning from air to bulk water and to a water surface is well understood,^{[154,165,166]} there is still a considerable lack of knowledge on the partitioning between the atmosphere and indoor surfaces. In most cases it is not even clear whether adsorption or absorption is the prevailing process. Knowledge of the type of sorption is important since, in the case of absorption, the capacity of the sorbent depends on the bulk mass or volume of the sorbent, while in the case of adsorption, the capacity depends on the surface area. Another milestone in the development of partitioning models constitutes the publications of K-U. Goss who proposed 1-octanol as surrogate for condensed phase^[152] when describing absorption of alkanes and polychlorinated biphenyls (PCBs) from the gas phase into bulk ones. For the low concentrations typically found under ambient conditions, the partitioning of organic compounds between air and condensed phases often follows a linear isotherm and can be characterized by a partition constant.^[152] Experimentally determined partition constants (K) between air and condensed phases of a series of compounds were evaluated as a function of their (subcooled) liquid saturation vapor pressure p^°_L.^[152] Generally, linear free energy relationship cf. Eq. (8) (Table 2) has been found for structurally related compounds.^[152] The slope m of such linear relationships (Table 2, Eq. (8)) for true equilibrium partitioning is independent of the units of the sorption coefficients and vapor pressures used in the ln K_i^abs vs ln p°_iL plots. The m values should be near −1 regardless of whether absorption or adsorption takes place. The reason behind this is that there is an inverse linear relationship between the size of molecules showing van der Waals interactions (e.g., homologous series of organic molecules differing in CH₂ units) and their ln p°_iL. Slopes deviating from −1 have mostly been attributed either to deviations from true equilibrium or to experimental problems.^{[149,167–171]} Goss and Schwarzenbach^[152] demonstrated by theoretical considerations as well as experimental data taken from the literature that such deviations do not necessarily indicate nonequilibrium effects. The affinity of molecular entities having van der Waals interactions to any other liquid phase should be equal to or smaller than their own liquid phase.^[172] Therefore, the change in the activity coefficient Δlnγ_i^abs that is caused by adding such an entity to the molecule is ≥0. Thus, in general, s is smaller or equal to zero, and consequently the slope m = -(1 + s) in Eq. (8) (Table 2) will be equal or less negative than −1. The more the slope deviates from −1, the smaller is the affinity of the molecules to the absorbing phase as compared to their pure liquid phase.^[152] In contrast to CH₂ units, Cl substituents engage in H-bond interactions (e.g., chlorobenzenes^[173]). This effect is not proportional to the number of Cl substituents and experimental partitioning data scatter within a compound class.^[152] Moreover, this slope, m may help to evaluate and characterize unknown absorbents or sorption processes that are not yet understood very well.^[152] The performance of different sorbents can only be compared if the conditions of the sorption process is the same (e.g., set of compounds, temperature, RH) and if the sorption constants are normalized to that property which appropriately describes the amount of sorbent, i.e., interfacial area or a bulk property.^[152] Different slopes for two absorbents – one or both of which are unknown – unambiguously indicate a difference if all other conditions are the same. Nevertheless, from slopes alone, it is not possible to conclude whether adsorption or absorption dominates the gas/particle partitioning in a given case.^[152] However, for compound classes for which sorption is determined by van der Waals interactions only (e.g., alkanes, PCBs), information provided by the slope m and the absolute sorption constants, K should be combined. The calculated adsorption constants, K^ads from Eq. (13)^[151], can be compared to the experimental ones. However, in many cases, the experimental K values proved to be one to several orders of magnitude higher than predicted. According to Goss and Schwarzenbach,^[152] this suggests that sorption from the gas phase to particles is governed primarily by absorption and not adsorption.

The attempt to gain information about unknown sorbents from m requires a linear relationship between ln K and ln p°_L. However, this linearity is most likely obtained for compounds showing van der Waals interactions. However, when a direct comparison of two different sorbents is desired, the sorption coefficients of compounds i for both sorbents, ln K_i1 and ln K_i2, should be plotted against each other.^[152] If the sorbents are the same, K_i1 and K_i2 must be identical or, if they have different or not appropriate dimensions, they differ only by a constant factor independently of the target compounds. Such a plot does not require the use of compounds of one class. However, a slope of 1 in a plot of ln K_i1 vs ln K_i2 provides no final evidence for the identity of two sorbents, but it is an indicative for a high similarity of both sorbents if very different sorbates have been considered for the comparison.^[152]

A new direction in the gas-surface partitioning models has been established by Weschler and Nazaroff evolving from their particle deposition model^[156] and SVOC diffusivity in air.^[174] It was observed that even the most impenetrable indoor surfaces have films at their interfaces with room air (e.g., windows).^[43,175] Such films are mixtures of sorbed water, organic and inorganic molecules and ions, and deposited PM. Hence, interactions between gas-phase SVOCs and indoor surfaces can be reduced to interactions with surface films and not just with either mineral or bare surfaces. Thus, the extent of the partitioning depends on the equilibrium partition constant and on the kinetics of the partitioning process. The partitioning of organic compounds from both equilibrium and kinetic aspects in indoor environments has been conceptually modeled by Weschler and Nazaroff^[7] using 1-octanol as a thermodynamic surrogate for organic films. Depending on the thickness of the film present, Weschler and Nazaroff^[7] predicted that species with log K_oa values (where K_oa is the octanol-air partition coefficient) larger than about 5 or 6 will reside mostly on surfaces. For SVOCs with larger values of K_oa (e.g., for typical indoor conditions, >10⁸), partitioning into the OM in surface films appears to play a key role, influencing indoor SVOC dynamics.^[24,176] In addition, a wide range of building materials sorbs VOCs, often taking many hours to attain equilibrium partitioning.^[58,177] These equilibria can be described by partitioning coefficients specific to the different sorptive compartments (Table 2, Eqs. (18)–(23)). The appropriateness of using one or the other reduces to whether the interaction of a given SVOC with sorbed OM is more like its interaction with itself, or more like its interaction with 1-octanol.^[7] Xiao and Wania^[155] have addressed this point in an extensive review. Finizio et al.^[171] and Pankow^[178] authored pioneering papers that led to current understanding of the importance of K_oa for this partitioning. Recently, the concept of Weschler and Nazaroff^[7] has been extended to acids and bases leading to the introduction of partition coefficients taking into account the acid/base dissociation constants of the target analytes.^[32] The result of calculations cf. Eqs. (41)–(48) (Table 2) is a partitioning space plot to display where a chemical resides under thermodynamic equilibrium conditions indicating the shift of partitioning of acids and bases according to the pH of the polar reservoir.^[32]

The ability to characterize partitioning of gas contaminants between the gas and condensed phases is still a hot topic. A proof for that is the recent work of Qiao et al. in which theoretical predictions developed by Yamasaki,^[146] Pankow and Bidleman^[149] were confronted with monitoring data for SVOCs taken for about 40 different studies.^[179] Distribution of (S)VOCs between air and solid surfaces can currently be estimated by applying increasingly sophisticated instrumental approaches.^[30,32] Complex mathematical calculations now make possible to estimate the rate of diffusion of selected VOCs into painted surfaces, and to estimate their rate of adsorption and desorption.^[30,32] In the future, it is expected development of complex approaches modeling not only surface processes but indoor air chemistry as a whole. For example, the INdoor air Detailed Chemical Model (INDCM) developed by Kruza et al.^[180] to investigate the impact of O₃ reactions with indoor surfaces (including occupants), considers a single well-mixed environment and includes irreversible deposition to surfaces, photolysis from both attenuated sunlight through windows and artificial lighting, exchange with outdoors and internal emissions. The INteraction with Chemistry and Aerosols (INCA) model^[181] made the well-mixed assumption and included photochemistry, deposition and emissions processes, as well as surface processes indoors. The surface interactions were determined by the O₃ that ingressed from outdoors, emissions from building materials, sorption processes and heterogeneous chemistry reactions at surfaces.

6. Analytical techniques

Outdoor air pollution research is relatively more advanced than indoor the air research, with measurements made over a longer period of time and better established and validated models. Indoor surfaces act as labile reservoirs of all the gas-phase species measured. Taking advantage of the recent advances in instrumental analysis, high time-resolution measurements provide the characterization of the kinetics of gas-surface dynamic partitioning in a real indoor environment.

6.1. Surface characterization techniques

The properties of the materials that sorb chemicals need to be understood. The physico-chemical properties of several indoor materials (e.g., silica, gypsum, stainless steel, granite) have been studied, but much less is known about less refractory substances such as wood, upholstery components, and insulation materials.^[5]

Adsorption isotherms used to describe the quantity of adsorbate on the surface as a function of its pressure at constant temperature are still widely used. Raw data for recording adsorption isotherms are provided by very diverse instrumental techniques depending on the chemical nature of the adsorbate. Specific surface area of material can be estimated via gas adsorption by using the BET test.^[182] Nitrogen physisorption measurements can be performed at −196 °C using automated volumetric adsorption analyzer to determine porosity of solids.^[183]

Detailed molecular-level descriptions of deposition and surface-mediated reactions of adsorbed species on indoor surfaces are difficult to obtain also because of the ever-increasing types of surfaces and variety of different indoor emission sources.^[184] However, modern analytical techniques permit detailed characterization of surfaces such as hygroscopicity, morphology and homogeneity.

Metal oxides are either basic or acidic (e.g., silicon oxide), or amphoteric (e.g., rutile).^[185] Thus, most metal surfaces acquire a charge and, consequently, an electrical double layer at an aqueous interface, or react with acidic and basic gases. Protons and hydroxide ions in the aqueous layer influence this surface charge.^[11] The most common methods to characterize surface acidity have been reviewed by Sun and Berg.^[186] Briefly, the isoelectric point as indicator to characterize the acidity of solid surfaces can be determined by electrokinetic titration, a type of colloid titration.^[11] Auroux and Gervasini^[187] applied microcalorimetry to determine the number and character of basic and acidic sites on 20 metal oxides using NH₃ and CO₂ as probe molecules. Recently, Rindelaub et al.^[188] determined directly the pH in individual particles using a Raman microspectrometer coupled with a confocal optical microscope. Wei et al.^[189] have applied a related method using 2 D and 3 D confocal Raman microscopy to determine the pH of suspended aerosol droplets smaller than 50 μm diameter. According to Nazaroff and Weschler,^[11] these methods might be adapted to probe surface acidity.

In a recent study, window glass was placed vertically in different locations including a copier room, office, kitchen, and garage to assess differences in particle and coating depositions by atomic force microscopy – photothermal infrared (AFM-PTIR) spectroscopy.^[184] The detailed microspectroscopic imaging method showed deposition of particles and the formation of organic thin films that increased the surface area and roughness of glass. For AFM and AFM-PTIR studies, different settings have been applied. The PTIR spectra taken from the corresponding images of organic depositions found on the kitchen surfaces contained a carbonyl stretching frequency associated with an aldehyde compound, identified by ν(C = O) (1734 cm⁻¹). The authors proposed that these films were formed from deposition of aldehydes and long chain fatty acids. Besides deposition of organic particles on glass exposed in the copier room, the PTIR spectra also indicated the presence of a mixture of calcium carbonate from ν(CO₃^2-) (1466 cm⁻¹) and carbon black from ν (C═C) (1540, 1580 cm⁻¹) as major component of paper and the primary coloring component of black toners, respectively. To complement the microspectroscopic imaging and physicochemical characterization of AFM-PTIR spectroscopy, analysis of metals present in the different samples was also performed by inductively coupled plasma mass spectrometry.^[184]

Studying a thin film of benzo[e]pyrene (BeP) with bis(2-ethylhexyl) sebacate (BES) added as reference formed on the outside of glass capillary tubes after evaporation of the solvent by scanning electron microscopy (SEM), the registered images demonstrated that the films contained crystalline BeP cubes covered in a liquid BES film. According to the estimation of the authors considering the geometric area and the amounts of deposited BeP, the films would have been close to monolayer thickness if they were uniformly thick.^[157] Besides imaging, X-ray spectra acquisition and thus, information on the elemental composition of the surface scanned with SEM can be achieved by coupling it to energy dispersive X-ray spectroscopy. Other techniques suitable for surface characterization of different materials are Auger electron spectroscopy , X-ray photoelectron spectroscopy (XPS), AFM, laser desorption ionization – Fourier transform ion cyclotron resonance mass spectrometry ^[190] and time-of-flight secondary ion mass spectrometry .^[191,192] However, these techniques are still applied at a higher extent for aerosol research.

Depoorter et al.^[193] applied ¹H nuclear magnetic resonance for characterization of the furfural films at a frequency of 400.13 MHz using a 30° flip angle with a recycle delay and an acquisition time of 1 s and 5 s, respectively. Scans of single pulse ¹H excitation were collected over a spectral width of 20 ppm. Additionally, UV spectrophotometric analyses were performed on films further extracted in deionized water to understand the influence of the film preparation process on the structure of furfural. The recorded UV spectra showed the conservation of the furan structure while the aldehydic band was shifted, indicating a transformation, later confirmed by ¹H NMR and proton transfer reaction time time-of-flight mass spectrometric (PTR-TOF-MS) analyses.^[193] A new band registered at 340 nm could correspond to maleic anhydride and/or 2-acetylfuran. Prolonged irradiation at 340 nm with LED lamp showed degradation of the new product. The ¹H NMR analyses showed broad peaks, thus confirming the complex nature of furfural, since this molecule undergoes self-condensation under anhydrous conditions or thermally at 100–250 °C. Based on the results, the authors concluded that the structure of the film could corresponded to polymeric structures.

6.2. Conventional analytical techniques for indoor surface chemistry

Sampling, sample preparation and quantitative determination of indoor air pollutants aiming at studying surface transformations is always subject of the advances in instrumental analysis. Common sampling procedures from surfaces such as wiping with nylon membrane filters and subsequent extraction are also widely used.^[142,194] In early works, the majority of indoor VOC measurements have been conducted through time integrated (i.e., off-line) sampling, which typically uses adsorbents to retain VOCs followed by analysis by gas chromatography (GC)^[129] or liquid chromatography.^[195] However, determination of indoor SVOCs is relatively challenging. For example, in the work proposing the theoretical approach for the partitioning of three- and four-ring PAHs on PM in 1982, sampling of particulate-associated material collected on a filter, gaseous material collected on an adsorbent after the filter as well as total suspended particulate (TSP) by (high-volume) air samplers was performed. A gas chromatograph-mass spectrometer (GC-MS) was used to identify the PAHs after performing Soxhlet extraction with cyclohexane. A GC with a flame ionization detector (GC-FID) was used to determine quantitatively the PAHs identified.^[146] Tichenor et al. used GC with electron capture detector for determination of tetrachloroethylene and GC-FID for ethylbenzene.^[135] Liang et al.^[151] collected gas and particulate-phase organic compounds in the frame of a chamber study. The sampler utilized glass fiber filter (GFF) to collect particulate-phase organic compounds. An identical backup filter was used to correct for the adsorption of gas-phase SVOCs to the front filter.^[196] Following the filters, gas-phase compounds were collected on Tenax cartridges and determined by GC with adsorption/thermal desorption. In a second train, two sequential polyurethane foam plugs were used to collect the less volatile compounds that were Soxhlet-extracted with methylene chloride. The TSP was determined by collecting particles on Teflon-coated GFF. Complete aerosol number and size distribution measurements were recorded using a cylindrical scanning electrical mobility spectrometer equipped with a condensation nuclei counter. As it was mentioned, the pioneer study by Singer et al.^[77] used relatively slow time-response, off-line chemical analysis. In similar works conducted by depositing PAHs on different supporting materials, such as filters, silica gel, and glass beads, followed by extraction, analysis of the PAHs and their oxidation products was achieved by IR spectroscopy, spectrophotometry, or MS.^[157] Zhou et al.^[157] pointed out at the difficulties in the monitoring of specific reaction products. Later, a further step forward was achieved by application of GC and high performance liquid chromatography (HPLC) hyphenated to UV/fluorescence detectors or MS.^[157] For example, GC and HPLC have been used for studies performed on several surfaces.^[197–204] However, these techniques may result in artifact formation due to extraction and preconcentration.^[205] Moreover, analysis of some products, such as carbonyls or carboxylic acids, also requires derivatization.^[206,207]

In some cases, conventional techniques could be successfully adapted to surface chemistry. Thus, Wang et al.^[32] collected surface samples from glass surfaces daily by wiping it with nylon membrane filters. The filters were subsequently extracted and analyzed for nitrite concentration using a UV-vis spectrophotometer following the Griess assay.^[142,194] A HONO calibration was also conducted daily on-site during the field campaign by collecting gas-phase HONO from a known flow of indoor air bubbled through deionized water and analyzed similarly with the Griess assay. The overall sensitivity of the instrument was calculated by averaging the sensitivity from the on-site calibration, and the pre- and post-campaign calibrations.^[142] The measurement of HONO indoors provides clear evidence that there is sufficient liquid water to drive aqueous chemistry in residences, and that aqueous chemistry can alter the composition of indoor air.^[71]

In conclusion, these off-line analysis approaches have been widely used for monitoring of indoor air quality,^[208] but they are not useful for field studies to understand indoor chemical processes since the reaction rates are fast enough to compete with typical AERs.^[6] However, a possible solution to this has been offered by Liu et al.^[209] investigating VOC emission in a wood-frame structure single-family house by PTR-TOF-MS placed in a detached garage next to the house. In this study, air was continuously drawn through separate 30‐meter‐long 6.4‐mm Teflon tubes at a flow rate of 2 L min⁻¹ from six locations (outdoors, kitchen, crawlspace, basement and attic). The sampling lines were bundled together and heated in order to minimize wall-effects due to temperature changes in the line. However, PTR-MS is not a conventional technique and the analysis are not cost-effective. Another option was presented by Adams et al.^[210] by analyzing removable kitchen coupons and bathroom tiles installed in an occupied home for 4 weeks for their microbial and chemical content after their removal and transport to the laboratory, where the VOC characterization was done by using a PTR-TOF-MS instrument. In this case, PTR-TOF-MS was applied to analyze VOC emissions from coupons and tiles.

6.3. Recent advances in spectroscopic technologies for surface chemistry indoors

As pointed out by Abbatt and Wang,^[5] development of fast-time response instrumentation with excellent detection limits, such as on-line MS makes it now possible to study how the indoor environment dynamically responds to transient behavior during window or door opening, cleaning, cooking, or changing human occupancy. Up to our knowledge, no real-time measurements on indoor surface composition have been reported in genuine indoor species. This is needed to follow the dynamics of surface composition change and to assess the SVOCs that will not be present if indoor surface samples are taken back to the lab for analysis.^[5] In the sections below, we are going to provide an overview on the principle of operation of these novel techniques and some relevant applications related to indoor surface chemistry.

6.3.1. Cavity ring-down spectroscopy (CRDS)

Cavity ring-down spectroscopy (CRDS) is an optical spectroscopic technique that enables measurement of absolute optical extinction by samples that scatter and absorb light. A typical CRDS setup consists of a laser that is used to illuminate a high-finesse optical cavity that is usually equipped with two highly reflective mirrors. When the laser is in resonance with a cavity mode, intensity builds up in it due to constructive interference. When the laser is turned off, the intensity inside the optical cavities decreases exponentially as measured by photomultiplier tubes that detect the light intensity transmitted through the rear mirrors.^[211] During this decay, light is reflected back-and-forth many times between the mirrors giving an effective path length for the extinction. If a light-absorbing material is placed in the cavity, the mean lifetime decreases. A CRDS setup measures how long it takes for the light to decay to 1/e of its initial intensity. This so-called ringdown time can be used to calculate the concentration of the absorbing substance in the gas mixture in the cavity.

Indoor gas-phase NO, NO₂, and O₃ were monitored using a custom-built CRDS.^[211,212] In brief, NO₂ is directly measured by CRDS at 405 nm supplied by a laser diode. Then O₃ is quantitatively converted to NO₂ by reaction with NO in excess, NO conversely with O₃, and the NO₂ formed is determined by CRDS. The calibration of the instrument was conducted by converting O₃ with known concentration to NO₂, which was then measured by CRDS. The 1σ precision was < 30 pptv at 1 s and <4 pptv at 1 min time resolution.^[211] The surface removal rate and deposition velocity of O₃ in a residence^[212] was determined by dual beam O₃ monitor. This latter uses two detection cells to improve precision, baseline stability, and response time. In the UV-based dual beam instrument, the UV light intensity measurements I⁰ (O₃-scrubbed air) and I (unscrubbed air) are made simultaneously. This instrument makes it possible to reduce the time between O₃ measurements to 2 s, while still retaining the small size, weight, and power requirements. Gas-phase CO₂, CO, CH₄ and water can also be determined by CRDS.^[206]

Ammonia is difficult to be quantitatively determined due to its high reactivity, solubility in water, and tendency to sorb to a variety of surfaces.^[214] During a comprehensive indoor chemistry study conducted in a test house, 30-s mean concentration of NH₃ was determined indoors using CRDS.^[76] Measurement accuracy was 0.5 ppb ± 5% of reading with a precision of 0.10 ppb + 0.1% of reading for 30-s of data. The molar absorptivity, or extinction coefficient, of the instrument was calibrated. Based on these results, the authors reported that the NH₃ concentrations determined indoors during the cooking, cleaning with NH₃-based cleaning agents, and occupancy experiments never dropped to typical outdoor values. Thus, large source of NH3 on indoor surfaces could rapidly be liberated into the gas-phase to maintain equilibrium. This surface-gas equilibrium mechanism is similar to the proposed HONO equilibrium mechanism described in detail by Collins et al.^[142] Plots of ln NH₃ concentration vs the inverse of the temperature using the peaks of NH₃ concentrations recorded during the response time activity, plus the average of peak concentrations from the background measurements resulted to be linear (R² = 0.97) in the case of cooking, cleaning, and occupancy experiments suggesting one of two potential mechanisms occurring simultaneously or alone: i) NH₃ emissions from building materials responded to temperature in an exponential way; ii) standard sorption kinetics were governing the NH₃–surface behavior. In the absence of activity-based sources, the HVAC operation was the main modulator of NH₃ concentration indoors. A possible explanation of this trend is that the loss of NH₃ due to the solubility in the water film of the coils and/or other surfaces in the house reached saturation and were not as effective at taking up additional NH₃. Gas-phase mixing ratios of VOCs, HONO, NH₃ dropped significantly upon enhanced-ventilation and then rebounded to their initial steady-state values upon closing the doors and windows. This effect also indicates that labile surface reservoirs are a source of these gaseous contaminants.^[32]

6.3.2. Attenuated total reflectance-Fourier transform infrared (ATR-FTIR) spectroscopy

Attenuated total reflectance (ATR) used in conjunction with IR spectroscopy enables samples to be examined directly in the solid or liquid state without further preparation.^[215] Recently, ATR-FTIR revolutionized solid and liquid analysis not only due to its easy sample preparation but also because of i) faster sampling; ii) improved sample-to-sample reproducibility; iii) minimizing user-to-user spectral variation.^[215]

6.3.2.1 Principle of operation

The principle of this technique has been comprehensively described elsewhere.^[215] Briefly, changes occurring in an IR beam (e.g., Xe UV-IR lamp^[216,217] directed onto an optically dense crystal (e.g., diamond, ZnSe or Ge) with a high refractive index at a certain angle (e.g., 2.38–4.01 at 2000 cm⁻¹) with a sample placed horizontally on top of it is measured. Under these conditions, the beam is totally reflected. The ZnSe crystal is a cost-effective used for analyzing liquids and non-abrasive pastes and gels, but it is not robust with a working pH range of 5–9.^[215] Among the abovementioned materials, Ge crystal has by far the highest refractive index. Moreover, it can be used to analyze weak acids and bases. Diamond crystal is the most expensive, but it is robust and durable, while ZnSe and Ge crystals can easily scratch.^[215] The internal reflectance creates an evanescent wave that penetrates only a few micrometers (0.5–5 µm) into the sample held in contact with the crystal. The evanescent wave will be then oriented toward the detector (e.g., mercury-cadmium-telluride detector cooled by liquid N2).^[216,217] The number of reflections is 5–10, depending on the length and thickness of the crystal and the angle of incidence.^[215] This multiple reflection modality is suitable for elastomers and fine powders, but many solids give very weak spectra because the contact area is small. This latter problem has been overcome by using small diamond crystals. In this case, only a single reflection is generated but this is sufficient due to the very low noise levels provided by FTIR spectrometers.^[215]

6.3.2.2 Applications of ATR-FTIR in surface characterization

In early works, ATR-FTIR was applied to studying heterogeneous oxidation of n-hexane soot and anthracene adsorbed on TiO₂ and mineral dust particles with O₃.^[216,217] In terms of indoor bases, Destaillats et al.^[218] used ATR-FTIR spectroscopy to investigate molecular associations of pyridine with cellulose and gypsum surfaces as surrogates for paper (wallboard panel and wallpaper, cotton used in furnishings) and wallboard panels, respectively. Their results indicated that there were at least two sorptive states for pyridine, a chemisorbed state identified by the presence of IR signals in the N–H and O–H stretching region of the spectrum (2900–3600 cm⁻¹) besides the aromatic C–H stretching modes (2950–3100 cm⁻¹) and a more labile physisorbed state. During desorption under a stream of N₂, surface enriched with the chemisorbed species showed a slower reduction of the absorbance of the broad band at 2900–3600 cm⁻¹ in relation to the total pyridine absorbance change^[287] This spectroscopic evidence was consistent with the desorption behavior observed in their previous work for nicotine from model surfaces.^[219]

Recently, ATR-FTIR spectroscopy^[159,220] has been used to determine diffusion coefficient (D_f) of gases and solvents in polymer matrices^[21] as well as that of VOCs in paint.^[30] The schematic representation of the experiment conducted by Algrim et al.^[30] can be seen in Figure 2. Experiments were conducted with C₁–C₆ carboxylic acids; C₂, C₄, and C₆ 1-alcohols; C₆, C₉, and C₁₂ n-alkanes; and C₆, C₈, C₁₀, and C₁₂ 2-ketones. After acquisition of the paint film background spectrum, and subsequent opening of the VOC vial and placement of the glass enclosure on the film, spectra were collected every 8 min up to 24 h. Spectra had a range of 400–4000 cm⁻¹ and 4 cm⁻¹ resolution. The IR beam only penetrates about 1 µm into the paint film. The mathematical background of the calculation of D_f is presented in Table 2. Although the approach of Algrim et al.^[30] is novel, long path FTIR has already been applied to determination of formic and acetic acids.^[222]

Figure 2. Schematic of the ATR-FTIR apparatus used to determine diffusion coefficients of VOCs in paint. The IR beam penetrates about 1 µm into the paint film. Adapted from Ref.^[30]

6.3.3. Mass spectrometric surface characterization

About two decades ago, mass spectrometric secondary ionization mass spectrometry (SIMS),^[223,224] desorption electrospray ionization^[225–227] and extractive electrospray ionization^[228–230] have predominantly been used for surface characterization and especially to investigate the chemical composition of the particle interface.^[231] Despite recent advancement such as TOF-SIMS as a high spatial resolution mass spectrometry (HR-MS) for surface analysis of solid samples,^[232] according to Nah et al.,^[231] these techniques tend to either require analysis under high vacuum conditions,^[223,224] off-line filter sampling prior to analysis^[225–227] or the use of charged solvent sprays.^[228–230] Moreover, their major field of application seems to be limited mainly to aerosol chemistry. Therefore, we focus in the following subsections on real-time in situ measurements by novel mass spectrometric methods using PTR, negative ion chemical ionization and direct analysis real-time (DART) modalities to understand chemical processes occurring on indoor surfaces. The main ionization mechanisms as well as their advantages and drawbacks have been summarized in Table 3. It is still unpredictable what does future hold for these research-grade instrumentation in terms of their use for field studies.

Table 3. Overview on the ionization mechanisms and advantages/drawbacks of mass spectrometric techniques used for surface characterization.

Acronym	Ionization mechanism	Advantages	Drawbacks	Ref.
CIMS	Acetate as reagent ion	i) no sample preparation; ii) soft chemical ionization at reduced pressure; iii) high mass spectra acquisition rates (<1 Hz); iv) study of the temporal behavior of previously unknown compounds; v) different reagents can be used for different compounds	i) complicated ionization mechanisms; ii) ^210Po is a radioactive source for generation of the ionization reagents, although the best one for ionization reagents; iii) I^− adducts only with neutral analytes; iv) use only in negative ionization mode; v) difficult calibration;	^[Citation208^,^233^] ^[^234^] ^[Citation235^] ^[Citation236^]
	Proton abstraction: CH3COO^− + R − H → CH3COOH + R^− Cluster reaction: CH3COO^− + R − H → [CH3COO + R−H]^− Ligand exchange: [CH3COO + R−H]^− + R´−H →[CH3COO + R´−H]^− + CH3COOH Declustering via collisional dissociation: [CH3COOH + R]^− + ECDC → CH3COOH + R^− Fragmentation: R + ECDC → r1^− + r2
	Iodide as reagent ion			^[Citation237^]
	Ligand exchange reaction with analyte molecule (M): [I + H2O]^− + M → [I + M]^− + H2O
	Nitrate as ionization source (HNO3)n+1 → (HNO3)n·NO^−3 (HNO3)n·NO3^− + HX → (HNO3)n + HNO3 + X^−→ (HNO3)n + (NO^−3)·HX , where HX is H2SO4, oxidized organic acids, etc.	i) optimized for measuring low volatility compounds; ii) capable of measuring sulfuric acid and its clusters with dimethylamine; iii) detection of VOCs with six or more oxygen atoms playing role in the SOA formation; iv) use of a soft X-ray source for the ionization instead of the ^210Po source.	i) operation only at ambient pressure; ii) highest quantifiable concentration of about 1 ppb; iii) complicated calibration.	^[^238–240^]
DART	Positive ionization mode	i) no sample preparation ii) soft chemical ionization at AP; iii) real-time information; iv) study of heterogeneous kinetics; v) identification of hardly detectable reaction products; vi) no use of radioactive sources; vii) suppressed fragmentation with NH4^+, but corrosive NH3 is formed; vii) wide range of applications.	i) for calibration with PEG 1000, m/ z = 16–2000; ii) qualitative analysis; ii) only laboratory use, costly equipment.
	Formation of protonated water clusters via Penning ionization: ${\mathrm{He}}^{\mathrm{*}}\left(\mathrm{g}\right)+n{\mathrm{H}}_2\mathrm{O}\left(\mathrm{g}\right)\to \mathrm{He}\left(\mathrm{g}\right)+[({\mathrm{H}}_2\mathrm{O}{)}_{n-1}\mathrm{H}{]}^{+}\left(\mathrm{g}\right)+{\mathrm{OH}}^{-}(\mathrm{g})$ Formation of molecular ions with the analyte molecule (M): $[({\mathrm{H}}_2\mathrm{O}{)}_{n-1}\mathrm{H}{]}^{+}\left(\mathrm{g}\right)+\mathrm{M}(\mathrm{g})\to [\mathrm{M}+\mathrm{H}{]}^{+}(\mathrm{g})+n{\mathrm{H}}_2$ Hydride abstraction: ${\mathrm{He}}^{\mathrm{*}}\left(\mathrm{g}\right)+{\mathrm{O}}_2\left(\mathrm{g}\right)\to \mathrm{He}\left(\mathrm{g}\right)+{\mathrm{O}}_2^{+}\left(\mathrm{g}\right)+{\mathrm{e}}^{-}$ ${\mathrm{O}}_2^{+}\left(\mathrm{g}\right)+\mathrm{M}\left(\mathrm{g}\right)\to {\mathrm{M}}^{+}\left(\mathrm{g}\right)+{\mathrm{O}}_2\left(\mathrm{g}\right)$ ${\mathrm{O}}_2^{+}\left(\mathrm{g}\right)+\mathrm{M}\left(\mathrm{g}\right)\to [\mathrm{M}-\mathrm{H}{]}^{+}(\mathrm{g})+\mathrm{H}{\mathrm{O}}_2^{·}(\mathrm{g})$			^[Citation231^] ^[Citation241^] ^[Citation242^]
	NH4^+ + M → [M + H]^+ + NH3, and/or [M + NH4]^+ (cluster)			^[Citation157^]
	Negative ionization mode
	Electron capture by oxygen atoms: ${\mathrm{e}}^{-}+{\mathrm{O}}_2(\mathrm{g})\to {\mathrm{O}}_2^{-}(\mathrm{g})$ Proton abstraction and charge transfer reactions: ${\mathrm{O}}_2^{-}\left(\mathrm{g}\right)+\mathrm{M}\left(\mathrm{g}\right)\to [\mathrm{M}-\mathrm{H}{]}^{-}(\mathrm{g})+\mathrm{H}{\mathrm{O}}_2^{·}(\mathrm{g})$ ${\mathrm{O}}_2^{-}\left(\mathrm{g}\right)+\mathrm{M}\left(\mathrm{g}\right)\to {\mathrm{O}}_2\left(\mathrm{g}\right)+{\mathrm{M}}^{-}(\mathrm{g})$			^[Citation231^]
PTR	Proton-transfer reaction: H3O^+ + R → RH^+ + H2O Cluster ion formation: ${\mathrm{H}}_3{\mathrm{O}}^{+}+n{\mathrm{H}}_2\mathrm{O}\to {\mathrm{H}}_3{\mathrm{O}}^{+}({\mathrm{H}}_2\mathrm{O}{)}_n$ $\mathrm{R}{\mathrm{H}}^{+}+n{\mathrm{H}}_2\mathrm{O}\to {\mathrm{RH}}^{+}({\mathrm{H}}_2\mathrm{O}{)}_n$ but: ${\mathrm{H}}_3{\mathrm{O}}^{+}\left(H_{2}O\right)+\mathrm{R}\to \mathrm{R}{\mathrm{H}}^{+}+2{\mathrm{H}}_2\mathrm{O}$ SRI + transfer reaction with other ions: NH4^+; Kr^+; NO^+; O2^+ CHARON (aerosols)	i) no sample preparation; i) soft chemical ionization at AP; iii) high mass spectra acquisition rates (1 Hz), m/z = 1.000 − 500.000; iv) time-resolved measurements; v) study of non-acidic VOCs.	i) use only in positive ionization mode; ii) not suitable for organic nitrates, organosulfates, phosphates; iii) large, costly equipment.	^[Citation129^] ^[Citation243^]
NB	Assuming that time-of-flight mass spectrometer is hyphenated, further universal advantages: i) possibility of non-targeted analysis; ii) identification of chemical compounds due to the accurate mass measurements.

Abbreviations: AP = atmospheric pressure; CIMS = chemical ionization mass spectrometer; DART = direct analysis real-time; PEG = polyethylene glycol; PTR = proton transfer reaction; SOA = secondary organic aerosol; SRI = selective reagent ionizatio; VOC = volatile organic compound.

6.3.3.1 Proton transfer reaction time mass spectrometer (PTR-MS)

The design of the PTR-MS instrument was discussed in detail by Lindinger et al.^[242] and de Gouw and Warneke.^[243] Thus, only a shorter description is given here. This instrument is based on the proton-transfer reactions from H₃O ⁺ to gaseous organic compounds with a higher proton affinity (PA) than H₂O. Usually, the set-up consists of i) a discharge ion source to produce the H₃O ⁺ ions, ii) a drift-tube reactor, in which the proton-transfer reactions take place, and iii) an MS.^[243] At the early stage of instrument development, quadrupole mass spectrometers (QMS) were used. Nowadays, TOF-MS is the preferred mass analyzer and detector.^[129] The ion source consists of a hollow cathode discharge with a constant flow of water vapor. Ions from the hollow-cathode are further converted into H₃O ⁺ that are transported toward the drift tube by a fast flow of gas.^[243] By varying the potential values of the discharge chamber, the H₃O ⁺ signal can be increased but higher fraction of O₂⁺ and NO ⁺ impurity ions will also be formed. The latter are formed from air back streaming from the drift tube into the ion source. The O₂⁺ and NO ⁺ ions undergo unwanted charge-transfer reactions for example with VOCs such as alkanes that do not react with H₃O⁺.^[244] For some compounds with PA close to the value of water, the measurements are tricky. An example is the case of formaldehyde that presents interferences from methanol and ethanol reacting with O₂⁺.^[245] Besides, this compound shows a significant humidity dependence due to the backward reaction of protonated formaldehyde with water, therefore corrections have to be made in order to obtain correct results.^[246–248] Another group of compounds that present difficulties to be determined are hydroperoxides that can decompose to carbonyls in the PTR-MS sampling lines, leading to a measurement bias.^[249] Modifications to the original PTR-MS instrument are currently being developed using other chemical compounds for the ionization that allow the measurement of compounds with PA lower than that of water or to distinguish isomeric compounds under the name of selective reagent ionization.

The drift tube consists of stainless steel rings connected to a resistor network, which divides the overall drift voltage into a homogeneously increasing voltage and establishes a homogeneous electric field inside the reactor.^[243] Increasing the ion kinetic energy using the drift field limits the degree of cluster-ion formation and simplifies the interpretation of the mass spectra.^[243] Contrary to the flow tube, the electric field is also responsible for the transport of ions over the length of the drift tube.^[243] The drift tube is separated from the second intermediate chamber by a small orifice, since only a small part of the air goes into the instrument, the rest of the air is pumped away by a turbomolecular pump. A fraction of the ions is extracted through the nose cone into the MS. To improve the response time of the instrument, the diameter and volume of the drift tube were reduced to decrease the residence time of air.^[243] For indoor VOC analysis, the sensitivities of the masses were calibrated using calibration standards (from formaldehyde, m/ z = 31 to α-pinene, m/z 137).^[250] For MVOCs, calibration on two specifically designed gas mixtures contained typical MVOCs such as 1-octen-3-ol, furfural, methional, pentylthiophene, indole, acetophenone or crotonolactone has been reported.^[129] The self-design calibration method was described by De Gouw et al.^[251] and De Gouw and Warneke.^[243] The calibration used must ensure the entry of known concentrations of the gases into the instrument. Holzinger et al.^[252] developed a new calibration method based on fast injection of a gas standard. This methodology is very useful for instruments with pulse durations lower than 1 s. During calibration, the constant flow of the calibration gas mixture was diluted by VOC-free air before entering the PTR-MS and measured. The VOC-free air measurement is achieved by passing ambient air through a catalytic converter that effectively removes the VOCs.^[212] The sensitivity was determined as the slope of the linear regression of the instrument response (counts/s) vs input gas concentration (ppb).^[212] For m/z values that were not represented in the calibration gas, the sensitivity was assumed to be the average sensitivity of a variety of previously analyzed compounds, with an expected uncertainty of ∼50%.^[253] Nevertheless, transmission can be modeled based on a previous injection of a gas mixture, and concentrations of unknown compounds or compounds without a direct calibration, can be obtained. Transmissions vary between instruments but also over time for individual instruments.^[243,252] Sensitivities can be determined experimentally using calibration gas mixtures or calculated using the rate constant for the PTR, which depends on the kinetic conditions in the drift tube. However, just some compounds have experimentally determined rate constants,^[254] therefore different theoretical approaches can be used.

In PTR-MS, only the mass of product ions is determined,^[243] thus significant mass overlap can arise. This hampers the interpretation of the product ion signals, especially when no prior knowledge of the VOC composition exists. To overcome these disadvantages, Warneke et al. developed a proton-transfer ion-trap mass spectrometry instrument (PIT-MS).^[255,256] In the PIT-MS, ions of one particular mass can be collected in the ion trap consisting of a ring and two end-cap electrodes supplied with a radio-frequency (RF) voltage, and then fragmented in collisions with the buffer gas.

The use of selective (also called switchable) reagent ion mass spectrometer (SRI) integrated in the PTR allows the use of other ions for determining VOCs in the air. Initially the switchable instrument used a modified hollow cathode design that allows the fast switch among the different ions: H₃O⁺, NO ⁺ and O₂⁺.^[257] Therefore, VOCs with PAs smaller than that of water and isomeric compounds can be determined. Aromatic compounds show highest sensitivities for NO ⁺ ions;^[258] on the other hand, the use of O₂⁺ allows to distinguish between some isobaric compounds such as aldehydes/ketones,^[259] or the detection of n-alkanes^[260] or alkenes.^[261,262] The latest advances in SRI use Kr⁺ as reagent ion.^[263] Compounds that cannot be distinguish or measured with H₃O⁺, NO ⁺ or O₂⁺ are still important for ambient air chemistry (CO₂, methane). Introducing Kr ⁺ as reagent ion helps to overcome this drawback, since it has a higher ionization energy than that of O₂⁺. More recently, a new model of hollow cathode has been developed using NH₄⁺ as reagent ion for the detection of amines and labile oxygenated compounds.^[264] With the development of TOF mass detectors, one of the main drawbacks of the PTR technology – the low mass resolving power – was overcome. Therefore, with PTR-TOF-MS is possible to discriminate isobaric compounds since it takes into account the decimals of the molecular masses to identify the compounds. However, it is still not possible to distinguish between isomeric compounds, that is compounds with the same formula but different chemical structure.^[259,265] In a PTR-MS instrument a constant current of primary and product ions is generated, and injecting them into a TOF-MS as ion pulses can only be done at the cost of some ion losses.^[243]

For a better understanding of the factors affecting the abundance and time-resolved variability of VOCs indoors, the use of time-resolved measurements is needed^[266] and PTR-TOF-MS is an option.^[129] The PTR-TOF-MS is particularly sensitive to benzenoid compounds, the majority of unsaturated linear, branched and cyclic hydrocarbons, oxygentated hydrocarbons, alcohols, aldehydes, ketones, acids, lactones, esters, cyclic and heterocyclic compounds, amides, sulfides, thiols, ethers, etc.^[129] Detection of some compounds such as organic nitrates, organosulfates, phosphates and some amines are more challenging since they are too sticky to get through the inlet.^[129] Short n-alkanes have PA lower than water and are therefore not ionized efficiently by H₃O⁺, although their detection is still possible via ionization by the abovementioned SRI + with O₂⁺/NO⁺/Kr⁺/NH₄⁺ ions. However, the sensitivity to n-alkanes is about two orders of magnitude lower than that for other VOCs, since the abovementioned ions are only present at trace level (ca. 1–3% of H₃O⁺).^[129] Short-chain alkanes therefore do not typically affect the total VOC data unless their abundance is extremely high. The detector relative transmission is verified using multicomponent VOC gas standard mixtures.

By coupling GC to the PTR-QMS, the chromatographic column separates the VOCs in a sample prior to the injection into the PTR-MS, thus VOCs that are detected at the same mass to be separated. However, this is achieved at the cost of the rapid response time.^[243] This hyphenation was successfully applied to urban air samples.^[253,267]

Proton transfer reaction instruments can also chemically characterize in a quantitative manner aerosol (submicrometer atmospheric particles) and gas phases using the Charon-PTR-TOF-MS instrument^[268,269] as well as it can be done with any PTR-MS. However, this newly developed instrument modification can work together with SRI+, therefore the range of gas compounds measured is each day wider.^[253,267]

6.3.3.2 High-resolution time-of-flight chemical ionization mass spectrometer (HR-TOF-CIMS)

The HR-TOF-CIMS instrument has been described in earlier publications.^{[237,271–273]} Briefly, the CIMS consists of a series of chambers that ionize, transfer, and analyze the target molecules.^[274] The main components in a TOF-CIMS with reagent ionization are a HR-TOF-MS, an ion molecule reaction cell (IMR), an atmospheric pressure interface (API), an atmospheric pressure sampling inlet, a data acquisition system and a vacuum system. With the selection of the convenient ionization scheme,^[275] different groups of compounds can be detected. Ethanol as reagent ion can be used for the characterization of amines.^[276,277] For the detection of peroxy acids, iodine ionization can be used.^[278] in general iodine-adduct chemical ionization is suitable for the determination of oxygenated volatile organic compounds.^[279] Nitrated phenols are detected using acetate chemical ionization^[280] as well as organic acids.^[233] In each case, the air flow first enters the ion-molecule reaction chamber where the target analyte (R–H) is ionized through reactions with reagent ions. Acetate (CH₃COO^-) and iodide (I⁻) ions are the most commonly used reagent ions for SVOCs. Acetate ions, generated by flowing acetic anhydride through a ²¹⁰Po radioactive source^[281] introduced into the ion-molecule reaction chamber of the HR-TOF-CIMS. Acetate ions ionize R–H through either proton transfer or clustering reactions, followed by declustering before detection by TOF-MS (Table 3). Proton abstraction and cluster reaction have been reported in the literature for acetate CIMS assuming that carboxylic acids are detected.^[272] The MS can be operated at a maximum mass resolution and tuned in “declustering mode” to ensure that CH₃COO⁻ are in the form of bare ions.^[272,282] The analyte ions are transferred and accelerated through three chambers containing various ion optics, followed by MS analysis with a TOF-MS.

The I-CIMS is another rapidly developing method being applied to TOF-CIMS.^{[237,283–286]} Iodide ions are generated by passing a ultrahigh purity N₂ flow over a permeation tube filled with methyl iodide and then through the ²¹⁰Po ion source.^[237] In this case, mainly I ⁻ adducts with neutral species are formed; I ⁻ is not expected to substantially abstract protons or transfer electrons.^[287] The ion optics of the MS must efficiently transmit these clusters to the mass analyzer. Thus, I-CIMS method operates in a cluster mode because the I ⁻ holds most of the negative charge. The I-CIMS arrangement can be located in the same trailer and share the indoor and outdoor air inlet line with the acetate CIMS.^[32]

The advantages and drawbacks of both acetate and I-CIMS have been compiled in Table 3. Briefly, compared to quadrupole-MS affected by duty cycle effects, the TOF-CIMS collects a continuous mass spectrum at high (<1 Hz) acquisition rates. However, calibrations remain as important for TOF as quadrupole systems. Liu et al. pointed at the fact that carboxylic acids other than HCOOH and AcOH are challenging to be measured indoors.^[274] For example, highly polar carboxylic acids elute poorly on typical GC columns. As CH₃COO ⁻ has the lowest gas-phase acidity of the common atmospheric carboxylic acids,^[281] it deprotonates gas-phase carboxylic acids leading to little or no fragmentation during CIMS measurement. Therefore, the molecular identity of the parent ions can be preserved.^[281] Moreover, aldehydes and alcohols do not react with CH₃COO⁻, making the ionization process selective for carboxylic acids.^[274] Liu et al. demonstrated that as low as 1.0 ppb could be determined for carboxylic acids by acetate CIMS HR-TOF-MS.^[274] They also applied an empirical parametrization of the sensitivity (ion counts per ppt of the analytes) of the HRTOF-CIMS to the acids, since it is impossible to calibrate the sensitivity for all studied carboxylic acids. For carboxylic acids, sensitivity is either calculated as the average^[233,234] or assigned to be the maximum^[235] of the sensitivities measured for a subset of target analytes.

Recent studies show that gas-phase nitrophenols^[288] as well as peroxy acids^[235] and benzoyl peroxide^[235] as secondary aerosol components are detectable with acetate CIMS due to their gas-phase acidity relative to CH₃COO⁻.^[288] However, CH₃COO ⁻ can also form adducts with levoglucosan, detected as [levoglucosan + CH₃COO]⁻ clusters and not deprotonated due to their low gas-phase acidity relative to the CH₃COO⁻.^[352] Similarly, isoprene epoxy diols and isoprene hydroxy hydroperoxides have also been reported to cluster with CH₃COO^−.[289] Bertram et al.^[272] demonstrated that a distribution of acetate clusters exists but can be dissociated through collisions during their transfer through the API by applying stronger electric fields across the ion optics.

Another ionization source that can be used for the determination of highly oxygenated compounds is the nitrate ion source (NO₃⁻). The ionization mechanism as well as the advantages and drawbacks of this technique have been compiled in Table 3. One of the main advantages of the instrument is the possibility to combine gas phase measurements with particle ones. In one of the first studies published,^[290] gas-phase low volatility compounds generated from oxidation of an isoprene hydroperoxide product were determined simultaneously to the particle phase composition. That study managed to investigate the link between organic compounds with low volatility and SOA formation and composition using a CIMS instrument with nitrate source in combination with an aerosol mass spectrometer. There are other studies in which the I—CIMS is used combined with FIGAERO module to study gas and particle phase composition.^{[285,291,292]} More recently, an extractive electrospray ionization instruments (EESI) has been developed for identification and quantification of aerosols^[293] in combination with CIMS-TOF instruments, for example using nitrate as reagent ion for the gas phase characterization. Some recent works point out the importance of this technology, EESI-TOF, for the molecular characterization of ultrafine particles^[294] or characterization of highly oxygenated molecules.^[295] Thanks to the advances in mass spectrometry, CIMS devices are continuously increasing the number of compounds both in gas and in particle phase, that can be quantified. On the other hand, the reagent ion selection is the other important point since that selection depends on the unambiguous determination of compounds. That factor is also a limitation that can be overcome using multiple reagent systems.^[275,296] Recently, a new inlet, named multi-scheme chemical ionization inlet (MION) that allows a quick switch between different reagent ion schemes has been developed^[341] working at ambient pressure. Previous multiple reagent ion instruments ionized the sample at reduced pressure introducing issues related to dilution of the sample, recombination processes, etc.^{[257,259,296]} The MION has been tested studying the oxidation of α-pinene by O₃ by applying bromide and nitrate chemical ionization consecutively in a second time scale^[275] and combined with FIGAERO inlet for aerosol characterization.^[292]

6.3.3.3 Direct analysis real time mass spectrometer (DART-MS)

Direct analysis in real time mass spectrometry (DART-MS) is a fast, non-contact atmospheric pressure ionization technique first developed by Cody et al.^[241] Originally, DART-MS proved to be suitable for the identification of a myriad of molecules found on concrete, asphalt, human skin, currency, airline boarding passes, business cards, fruits, vegetables, spices, beverages, body fluids, plastics, horticultural leaves, cocktail glasses, and clothing but related mainly to forensic science.^[297]

Samples, placed between the ion source and the atmospheric interface inlet of an MS, are exposed to a thermal stream N₂ or He gas containing metastable atoms or molecules (e.g., He* or N₂*) emitted by the ion source to induce desorption and ionization.^[241] The principle of DART-MS has been extensively described by Cody et al.^[241] Briefly, the ion source consists of a tube divided into several chambers through which N₂ or He flows. First, the gas is introduced into a discharge chamber containing a cathode and an anode. Gas flow rates are typically 1 L min⁻¹. An electrical potential of several kilovolts initiates an electrical discharge producing ions, electrons, and excited-state species in a plasma. Gas exiting through a third perforated electrode or grid acting as an ion repellent removing ions of the opposite polarity, thereby preventing signal loss by ion-ion recombination, is directed toward the MS sampling orifice. The reactions of the ionization mechanism can be seen in Table 3. Although optimum geometries exist for specific applications, the exact positioning, distance, and angle of DART with respect to the sample and the MS are not critical.^[241] The DART position is adjustable on a wide range of angles and distances. A typical DART/sample/orifice distance was 5 to 25 mm. However, ions could be detected even when DART was positioned 1 m away from the mass spectrometer.^[241] For mass scale calibration, Cody et al.^[241] used poly(ethylene glycol) (PEG). For PEG1000, calibrated range was between m/z 16 and 2000.^[241] The DART is usually installed on a HR-TOF-MS that provides exact mass measurements. The DART-MS provides qualitative information, although some compounds have been detected at relatively low levels. For example, 2 pg of ethyl palmitate deposited on a glass rod could be detected with a signal-to-background ratio of 70.^[241] This compound is easily analyzed in a vacuum by conventional electron ionization, but it is difficult to detect by atmospheric pressure ionization sources. Signal strength is related to sample quantity. A semiquantitative example was demonstrated for capsaicin in pepper pod tissues.^[241]

6.3.3.4 Recent mass spectrometric applications in surface chemistry indoors

Relevant MS applications in surface chemistry indoors have been compiled in Table 4. The actual number of works are higher, but studies conducted on target analytes as components of SOA in test chambers have not been considered. However, they can be considered as pioneer studies for data compiled in Table 4. For example, DART-MS was used to study submicrometer aerosol particles by introducing a stream of model single component organic particles consisting of alkanes, alkenes, acids, esters, alcohols, aldehydes and amino acids in the region between the DART ionization source and the MS inlet (Figure 3).^[231,298] For equivalent aerosol mass concentrations, the ion signal scaled with particle surface area, with smaller diameter oleic acid aerosols yielding higher ion signals relative to larger diameter aerosols. For the aerosols of the same size, but different vapor pressures, the ion signal was larger for more volatile succinic acid aerosols than less volatile adipic and suberic acid particles.^[231] The reaction of submicrometer oleic acid particles with O₃ was also used to demonstrate the ability of DART-MS to identify products and quantify reaction rates in heterogeneous reactions.^[231]

Figure 3. Schematic representation for sample introduction in the direct analysis real-time mass spectrometry with some typical operating conditions cf. Cody et al.^[241] for a) organic aerosol^[231] and b) polyaromatic hydrocarbon (PAH) film deposited outside a glass capillary.^[157]

Table 4. Overview on relevant mass spectrometric surface characterization studies conducted between 2017 and 2020.

Technique	Target analytes	Surface	Notes	Ref.
CIMS-TOF-MS	mono-, di-, hydroxy-, carbonyl, aromatic RCOOH	university classroom & human skin	Mean indoor concentration of RCOOH 6.8 ppbv, of which 87% formic and lactic acids	^[Citation274^]
	HONO	test chamber	Discussion on interferences to nitrite ion signal, sources of measurement uncertainty, and the calibration method	^[Citation142^]
	highly oxygenated VOCs	art museum & box model	The lifetimes of oxidation of VOCs with O3 as well OH and NO3 radicals >5 – >15 h	^[Citation95^]
	HONO, HNCO, and C1 to C9 monocarboxylic acids	test house	Surface reservoirs dominate dynamic gas-surface partitioning by enhanced ventilation or surface cleaning with vinegar- and NH3-based cleaners.	^[Citation32^]
	RCOOH	glass and latex painted gypsum wallboard	Partitioning to paint films through reversible absorption that depends on the absorption capacity of film and VOC Df.	^[Citation30^]
DART-MS	BeP	BeP film on capillary glass	Kinetics of the O3 − BeP oxidation in an off-line reaction cell	^[Citation157^]
PTR-Q-MS	oxygenated and aromatic VOCs	university classroom & human skin	1) Human increased VOC emission in 40% in a similar way to CO2; 2) Identification of skin lipid ozonolysis products.	^[Citation212^]
	VOCs	art museum & box model	1) Deposition to surfaces and ventilation the dominant VOC sinks in the museum; 2) Emission rates below the guidelines for museums.	^[Citation95^]
	MVOCs	kitchen coupons and bathroom tiles from an occupied home	Biomass estimates based on qPCR correlated with determined MVOCs (e.g., products associated with fatty acid production).	^[Citation210^]
PTR-TOF-MS	MVOCs	painted drywall	Emissions of MVOCs (>400) varied between microbial taxa, life stage and substrate type.	^[Citation129^]
	MVOCs	Petri dishes with agar	1) C4-C9 MVOCs tracers of the microbial activity highest in the early stage of growth; 2) Emissions decreased after 10–14 d and highly affected by the RH; 3) Some MVOCs sensitive at 340 nm.	^[Citation299^]
	VOCs	wood-frame single‐family house	Large contribution to VOC (>200) emission from wood decomposition.	^[Citation209^]
	ethanol, furfural, isoprene, mono- & sesquiterpenes, D5-siloxane	test house	Surface reservoirs dominate dynamic gas-surface partitioning by enhanced ventilation or surface cleaning with vinegar- and NH3-based cleaners.	^[Citation32^]
	2-ketones	glass and painted gypsum wallboard in a test chamber	Partitioning of VOCs to paint films through reversible absorption that depends on the absorption capacity of film and VOC Df.	^[Citation30^]
	furfural (alone and with NO3^−)	glass	High concentrations of 2(5H)-furanone/2-butenedial, maleic anhydride formation after UV irradiation.	^[Citation193^]

Abbreviations: BeP = benzo[e]pyrene CIMS = chemical ionization mass spectrometry; DART = direct real time analysis; D_f = diffusion coefficient; HONO = nitrous acid; HNCO = isocyanic acid; Q-MS = quadrupole mass spectrometer; qPCR = quantitative real-time polymerase chain reaction; PTR = proton transfer reaction; RCOOH = carboxylic acids; RH = relative humidity; MVOC = (microbial) volatile organic compound; TOF-MS = time-of-flight mass spectrometer.

Some comments to Table 4 are hereby presented. Zhou et al.^[157] applied DART-MS to study the oxidation kinetics and product formation associated with gas − surface heterogeneous reactions occurring on organic films (i.e., O₃ reacting with a condensed-phase PAH, benzo[e]pyrene, BeP) (Table 4). Using He as the reagent gas and with the DART heater temperature of 500 °C, nanogram quantities of BeP deposited on the outside of glass melting point capillary tubes were analyzed in positive ion mode with a limit of detection of 40 pg (Figure 3). Using bis(2-ethylhexyl) sebacate as an unreactive internal standard, the surface-bound BeP decays of the O₃–BeP reaction were determined after oxidation in an off-line reaction cell. The overall validation of the proposed analytical method was demonstrated because the plots of pseudo-first-order rate coefficients, k₁ derived from the analytical data as a function of gas-phase O₃ concentrations were consistent with the Langmuir − Hinshelwood adsorption that typically characterize this type of heterogeneous reactions. Moreover, a wide array of oxygenated, condensed-phase products such as lactone, quinones, diphenols, dialdehydes, and dicarboxylic acids has been observed in agreement with literature data that can otherwise only be detected after solvent extraction and perhaps derivatization or chromatographic separation.

Algrim et al.^[30] used PTR-MS and I-CIMS for validation of partitioning model of VOCs (series of 2-ketones or carboxylic acids in a concentration of about 25 ppb) to glass and painted gypsum wallboard using a portable surface reactor (PSR), and to paint films coated onto the inner walls of a glass flow tube.^[30] The PSR was connected to an 8 m³ fluorinated ethylene propylene (FEP) Teflon environmental chamber (RH < 1%, 295 K). Air from PSR was sampled through a FEP Teflon tube at 0.5 and 2.1 L/min. The precision of PTR-QMS and I-CIMS measurements was ±2.5% and ±0.5%, respectively, based on the standard deviation for 200 data points measured when sampling from the chamber at a constant VOC concentration. The calibration accuracy was about ±10% based on the uncertainties in the chamber volume and amount of VOC added. The same instrumental arrangement was applied to a painted tube apparatus used to determine absorptive partitioning of VOCs (i.e., carboxylic acids) into paint (Figure 1). One end of the painted tube was then connected to the chamber containing air with VOCs at 30% RH using an FEP Teflon tube, and the other end was connected to the I-CIMS to monitor VOC concentrations. The paint film was passivated and depassivated for 3 h each.

Wang et al.^[32] used CIMS-TOF-MS to determine the mixing ratios of gas-phase acids, and PTR-TOF-MS for non-acidic VOCs to study their dynamic gas-surface partitioning (Table 2). Data derived from CIMS, PTR-TOF-MS proved that the ability of these technique to quantify dynamic response times (typically between 1 and 5 s) needed to reestablish steady-state conditions after ventilation perturbations.^[32] In a recent review, Farmer pointed out the problem represented by instrument inlets transferring air from the study space into detectors valid not only for CIMS and PTR-MS but also for DART-MS.^[208] Teflon tubing can act as a chromatography column, causing delays in detection time for SVOCs.^[300] Fortunately, these gas-phase delays can be described by relatively simple models.^[300] Flexible silicone tubing contains siloxane-containing organic species that can interfere with target analytes.^[301]

6.4. Supporting measurements

Like in every indoor air quality field study, data on temperature, RH, CO₂ and AERs are needed also for experiments involving surfaces indoors. In this section, some highlights are provided. Water vapor buffering, adsorption, desorption, hysteresis is well-known and widely used for porous materials. These parameters are strictly RH-dependent and can be estimated most easily within climatic chambers.

For continuous monitoring of CO₂ and water, a convenient gas non-dispersive IR (NDIR) gas analyzer has been designed.^[302] The equipment, operating between −20 and 45 °C, is a based upon a field proven single path, dual wavelength, and thermally controlled IR detection system. The CO₂ and water concentration ranges are up to 20,000 ppm and up to 60 ppt, respectively. The gas analyzer was calibrated using CO₂ standard during the campaign, but an additional correction might be necessary. Liu et al.^[212] applied this NDIR gas analyzer to study the contribution of human-related sources to indoor VOCs in a university classroom. Indoor temperature and RH were recorded using a Vaisala HMP60 temperature and humidity probe suitable for integration into other manufacturers’ equipment.^[95] Indoor RH was used as a tracer for ventilation/infiltration.^[212] The RH can easily be determined by using different sensors operating on capacitive, resistivity or thermal principles. Schwartz-Narbonne and Donaldson^[303] studied water uptake by indoor surface films using quartz crystal microbalance. Schwartz-Narbonne and Donaldson^[303] exposed gold-coated quartz crystals horizontally for periods of approximately two months in homes, so they would accumulate both settling dust and organic vapors. After the exposure period, the crystals were exposed to conditions in which the RH could be controlled. The humidity was systematically varied from 5% to 85% at a rate of 1% min⁻¹. By measuring the change in oscillation frequency, the mass change associated with water uptake could be evaluated as a function of RH.

For the determination of hygroscopic sorption properties of materials, ISO 12571^[304] can be used. Rode et al.^[305] proposed the concept of moisture buffer value (MBV) to indicate the amount of moisture uptake/release by a material when it is exposed to diurnal RH variations between two given values. The MBV can be used as one tool for evaluating the potential of recycled material like paper waste as paper plaster for indoor use.^[306] The amount of water vapor ab/desorbed by a hygroscopic material can be directly measured.^[307] Materials under the same name could have different properties. Altmäe et al^[308] found that product “clay plaster” can offer water vapor sorption properties differentiating up to 20 times.

A simple hygrometric method has been developed by Zandens and Goossens^[23] with the use of which the sorption isotherm of water could be recorded in three alkyd as well as three latex paint films. The principle of the technique is that a known amount of water was added to a system of paint and air placed in a chamber made of stainless steel, then the RH was determined using a sensor. The amount of water in the paint follows from a mass balance.

Several methods have been proposed for the determination of the AER such as the CO₂ concentration decay method,^[309] CO₂ or sulfur hexafluoride (SF₆) constant concentration method,^[310] SF₆ tracer gas method^[311] and perfluorocarbon tracer (PFT) method,^[312–314] continuous release of a tracer gas (butane-d3, Cambridge Isotope Laboratories)^[209] or the decay of methane injected into it over time.^[315] The advantages and drawbacks of these methods are well documented.^[316]

To estimate airflow on site EN 16211 and EN 12599^[317,318] can be applied by using equipment meeting the requirements^[319] indicated in these standards. Also a method using tracer gases is also in use (ISO 12569).^[320]

6.5. Effect-based analysis

The motivation behind studying the properties of indoor air or materials is ultimately protecting the occupants of the building from the possible adverse effects of the exposure to the emitted or produced chemicals. In that sense it is more important to know the overall effect or health risk of the exposure instead of concentrations of specific chemicals. Considering the high number of different chemicals typically present in indoor environments, the possible combined effects of the mixture have to be taken into account when assessing the risk of the exposure. For environmental sample matrices such as wastewater, using panels of bioassays to determine the potential health risks has already been established,^[321] whereas for indoor air only sporadic efforts of using individual bioassays have been reported. The main principle behind the bioassays is to use cultured reporter organisms for flagging the activation of particular toxicological mechanisms such as mutagenicity.^[322] For example, the endocrine disrupting activity of indoor dust of kindergartens has been studied with human ovarian carcinoma cell line transfected with estrogen responsive luciferase reporter gene, reporting associations both with presence of phthalates and relative amount of plastic in the rooms.^[323] Although they are not directly related to toxic effects in exposed humans, the strength of bioassays is in their ability to measure toxicological potential of the combined effects of the chemicals in the sample, including unidentified compounds.

In the indoor air context one of the main issues is the choice of sampling material and strategy: the decision to focus on airborne particles, volatile compounds or material samples each bring their own complications and affect also the generalizability of the results. So far, the reported studies of indoor air toxicity have used actively or passively collected particulate matter or extracts of it, condensed water vapor or direct contact with indoor air.^[323–325] There are some positive results on the association between the toxicity of the sampled material and building status and/or health effects in the occupants, but also contradictory reports on inability of effect-based methods to identify e.g., buildings with moisture problems. Studies of other environmental exposures show great potential of the effect-based methods in health risk assessment of combined exposures, but currently the applications for indoor environments lack successful validation studies and, in some cases adequate proof of concept -studies.

7. Microbiological analyses indoors

7.1. General considerations

There is a need to validate and harmonize methods for the characterization and evaluation of biological pollutants (bioaerosols) in closed environments to be able to identify the measures necessary to prevent and/or reduce the indoor concentration levels of these pollutants. Available water in building materials and air is the main limiting factor for microbial growth.^[326] Hence, it is also important to assess microclimatic factors and the activities that take place in the specific environment.^[327] Type of rooms, building materials, activities carried out and ventilation practices can lead to high levels of humidity indoors providing a favorable habitat for the survival and reproduction of fungi and bacteria on the different types of substrates available. In the indoor environments, surfaces can provide an ideal substrate due to the potential presence of nutritional elements capable of supporting development of the microbial flora. Biological agents can also be produced or transported indoors by the occupants themselves or their activities. The resulting bioaerosols are deposited on indoor surfaces as settled dust, which can be considered as an integrated sample of airborne particulate matter. Thus, monitoring of the surfaces is essential to know the composition of the microbial reservoirs, especially when there is potential for health risk due to pathogenic organisms.

7.2. Microbiological surface sampling

Different methods have been developed for the assessment of microbial exposure, since a unique reliable and standard quantitative method has not been established yet.^[328–330] Material surfaces are sampled directly to identify organisms that are producing visible growth, to investigate whether staining is due to organism growth, or to determine whether organisms are still present post-remediation. Thus, the main methods for microbiological surface characterization are removing the material itself (bulk sampling), the contact plate, contact or dip slides techniques, the sponge method, the tape-lift and the swab techniques. It is also possible to sample dust settled on surfaces and/or to collect spores that can be present in unfavorable conditions for microbial growth. Several authors have suggested that air sampling is a better measure of fungal exposure in terms of studying respiratory adverse effects rather than settled dust. However, air sampling is restricted by short sampling periods, representing only a snapshot of fungal exposure and observed, whereas MVOC concentration are generally low, much lower to other VOCs.^[331,332] Continuous monitoring of fungi and bacteria can be expensive and time-consuming.^[333] The main capabilities of these sampling procedures have been compiled in Table 5.

Table 5. Overview of advantages and drawbacks of microbiological surface sampling methods.

Name	Advantages	Drawbacks	Ref.
Bulk samples	Both bacteria and fungi may be recovered from bulk samples; Samples preserve the fungal structures allowing identification of similar organisms to genus. Samples are compatible with direct microscopy and culturable methods.	Directly cutting the material to be tested is undesirable or impossible in many cases.	^[Citation111,Citation134^]
Contact plate	Ready-to-use contact plates are available containing culture media suitable for the isolation of various microorganisms; Application to psychrophilic bacteria, mesophilic bacteria, fungi and dermatophytes. Applicable for both active (puming, e.g., BioStage impactor) and passive (sedimentation) sampling.	Different sampling protocols cf. target germ: Carrying out duplicate withdrawals on the same surface and identification of three significant points to be sampled; Need of individual verification of sampling time; Not suitable for most pathogenic genera and species that require (pre)enrichment; Application only on smooth surfaces (if used without Petri dish).	^[Citation111,Citation125,Citation332^]
Contact/dip slides	Two different types of media (e.g., agar) on the two sides of the slides; Quick and ready for use (also for liquids), shelf-life up to 9 months; Possibility for two checks with one slide for detecting yeast and molds, enterobacteria, E. coli and coliform;	If two different kinds of microorganism are tested at the same time, incubation time and temperature to quantify them may differ; Samples must be kept refrigerated and checked constantly for exposure to contaminants.	^[Citation335,Citation336^]
Sponge method	Benefits in recovery organisms; Suitable for sampling larger surfaces. Applicable also as a swirling liquid collection method, which can be defined.	Use of expensive and specialized microbiological equipment; Lower sensitivity compared to the swab method.	^[Citation332,Citation337^]
Swabs	Suitable for uneven surfaces, moist surfaces or hard to reach areas; Application to heterotrophic bacteria (psychrophiles and mesophiles), fungi, dermatophytes; Swabs are compatible with direct microscopy, culturable methods and qPCR.	Sampling is not simultaneous with the inoculation of the growth medium, the latter to be done in the laboratory after transport in refrigerated conditions in special containers; The culture medium must allow the growth of the target organisms; Soils containing selective substances should not be used; Different incubation protocols cf. target germ; Growth of organisms may artificially amplify quantitative results; With direct microscopy, swabs may not preserve the fungal structures required for differentiation of similar genera.	^[Citation111,Citation125,Citation334^]
Tape-lift (adhesive)	Samples preserve the fungal structures during sampling, allowing identification of similar organisms to genus (e.g., Penicillium and Aspergillus spp.); No growth of organisms during shipping, so, less time sensitive analysis; Non-fungal particulates (e.g., fibers, pollen, soot, spores) may also be identified. Some defined sampling devices (e.g., spore trap cassettes) are available, which makes this technique suitable for both (active and also passive sampling).	Direct microscopy for fungi and non-fungal particulates is the only analysis compatible with tape-lift samples. Culturable analysis is not possible; Bacteria cannot be identified; If the surface is wet, the adhesive may not hold the sample.	^[Citation125^] ^[Citation334^]
Settled dust	Samples are compatible with direct microscopy for fungi and for non-fungal (e.g., fibers, pollens) particle identification, qPCR and culturable methods; Allergen, mycotoxin and endotoxin analysis can be performed; It allows a larger area to be sampled; Dust samples are relatively stable and do not require rush shipping.	Direct microscopy may not preserve the fungal structures required for differentiation of similar genera (e.g., Penicillium and Aspergillus); Dust sampling is not advisable for collecting bacteria; Filter type used is important for qPCR analysis.	^[Citation334,Citation335,Citation338^]

7.2.1. Bulk samples

Bulk sampling is a destructive method in which samples are directly removed from the surface to be analyzed, by scratching, scraping or coring of small pieces of the material (cca. 0.3–5 g). Thus, the bulk sample refers to a portion of material, for example pieces of wallboard, duct lining, carpet, etc. that is removed from a site (e.g., cutting with a knife or a razor blade). It is the most widely used sampling technique in microbial assessment of building materials.

Microorganisms can be isolated by bulk sampling in two ways: 1) direct plating of the bulk sample onto a culture medium, and 2) microbial solution plating onto a culture medium.^[111] Bulk samples can be than analyzed for both bacteria and fungi, by direct microscopy (for fungi only), quantitative real-time polymerase chain reaction (qPCR), phospholipid-derived fatty acids (PLFA) and culturable methods (for both bacteria and/or fungi). Bulk samples may not have necessarily visible microbial growth on it upon collection, they may look completely clean or show some discoloration.^[334,338]

7.2.2. Sampling by contact plates

Contact plates are typically round culture plates in which the agar is poured so that the top of the agar forms a meniscus slightly above the top rim of the plate. A surface sample is collected by inverting the plate and pressing the agar directly onto a flat surface of interest. The plate is then removed, protected from air contamination by a lid, incubated and inspected for microbial growth.^[339] Nitrocellulose membranes were reported to be slightly more effective than contact plates at surface sampling, and easier to use on curved surfaces.^[340] Some studies have shown that the extractability of microorganisms depends on various parameters, notably time and pressure on the plate. For this reason, the commercial applicators are usually designed for a defined time and pressure,^[111] e.g., the BioStage cascade impactor that meets NIOSH and ACGIH recommendations for sampling indoor and outdoor mold and bacteria. It comprises an inlet cone, precision-drilled 400-hole impactor stage, and a base that holds a standard-size agar plate, into which a high flow sample pump pulls microorganisms in air through the holes.^[341]

7.2.3. Sampling by the sponge method

The sponge method consists of rubbing of sponges soaked in sterile solution (mostly saturated physiological solution, NaCl or Ringer’s solution) on the delimited surface to be tested. This step is followed by homogenization and elution. Finally, samples are sown in a plate with agar medium.^[334] Similarly, a swirling liquid collection method could be included into the sponge method group. It can be sampled used by a specific equipment (e.g., BioSampler, that reduces particle bounce, minimizes re-aerosolization, and preserves bioaerosol viability^[341] and/or by using a standard glass washer. The most important is to use sterile equipment including liquid solutions as well as to ensure sufficient concentration of bioaerosols in a solution (i.e., flow-time-liquid ratio).

7.2.4. Swab and wipe samples

Swabs are widely used to collect airborne material that has settled onto surfaces and to identify microbial contaminants that may be colonizing building materials. The ASTM International has a standard for collecting fungal material by swab.^[342] The USP and ISO have standards that include swab sampling for microbiological contamination in clean rooms.^[343–345] Sterile swab in a tube containing a sponge soaked with transporter buffer are used to sample both bacteria and fungi on wet surfaces and/or difficult to reach as well as for non-porous materials that cannot be cut for sampling.^[334] If the area to be sampled is dry, the swab should be moisturized by dipping it into the tube, so the tip contacts the sponge within. The elution media must both remove the biological material from the surface and subsequently elute it from the swab to be effective. A plastic template or a ruler and masking tape may be used to delineate the area to be sampled.^[346–348] The swab should be thoroughly rolled over the sample area. Swabs must be kept cool post-sampling and delivered to the laboratory quickly to keep their vitality and to prevent organism growth during transport.^[334,349] For this purpose, buffered physiological solution, or even special media are used that exclude the presence of sources of carbon, nitrogen and organic growth factors to avoid microbial multiplication (e.g., Stuart Transport Medium and Ames medium).^[125]

All of the considerations and limitations of swabs also apply to wipe sampling. Swabs are typically more useful for small surfaces and hard-to-reach locations, while wipes are more effective at collecting dust from large non-porous surfaces.^[348] Electrostatic wipes have been used to collect settled dust for studies of mold and endotoxin.^[350,351]

7.2.5. Tape-lift samples

The main purpose of the tape-lift sampling method is to confirm the existence of mold growth and to identify the type of fungi at the sampling site. The secondary purpose is to obtain fungal spore distribution data to determine if the sampling site is normal or contaminated.^[334] Tape-lift sampling is best used in conjunction with air sampling for verification of indoor mold growth by direct fungal examination. Tape-lifts are performed by gently pressing a piece of clear adhesive tape (e.g., transparent Scotch tape) onto a surface of interest, then removing with a slow steady force pressing.^[352–354] Then, it is attached to a glass slide or placed in a vial. Upon transportation to the laboratory, the tapes are stained and directly examined for identification of the fungal spores. The results depend upon the ability of the examiner to identify microorganisms and their fragments, and do not provide a quantitative assessment of exposure.^[334,339] It is also possible to cultivate microorganisms by applying the tape-lift sample to solid or liquid culture medium.^[111]

7.2.6. Sampling of settled dust

Settled dust sampling allows collection of large quantities of material, provides a long-term sample, and does not require a dedicated sampling device for each location. Dust assays allow quantitative data to be generated per weight and surface area of dust. In addition to settling from the air, dust can be produced by other mechanisms, making it difficult to distinguish the source (e.g., floor and carpet dust containing outside material brought in by shoes, skin flakes, clothing fibers and animal dander). Therefore, sampling locations well above floor level (e.g., 1 m) are often chosen.^{[110,333,355]} Air sampling should be completed before collecting surface samples by vacuuming.^{[341,356,357]}

In the case of allergens, it is possible to evaluate accumulation, compared to the point value that can come from sampling of the air which gives an indication of the quantity suspended allergen in the volume of air sampled. The choice of surface sampling depends mainly on the characteristic of the indoor allergens. For example, mites, proliferate inside heavy fabrics (sofas, mattresses, carpets and curtains). The cleavage residues of cockroaches are found mainly in the settled dust. Allergens deriving from pets can easily adhere to clothing, resulting in transport from one place to another (carry over). This same peculiarity allows these particles to travel associated with the settled dust. The importance of surface sampling is also linked to the size of the particles on which the allergens are transported, sizes that generally vary from 5 to 40 μm, with the possibility of depositing quickly.

Vacuum cleaners and electrostatic dust fall collectors (EDC) are the two most common methods of settled dust sampling.^[333] The collection and analysis of dust that has settled onto floors, carpets, and other surfaces is widely used as a means of identifying not only allergens but also endotoxin and molds in buildings.^{[110,353,354,356,358]} The US Department of Housing and Urban Development has developed a protocol for the vacuum collection of home dust samples to test for allergens.^] Vacuum collection of settled dust from floors and carpets has been used to determine the Environmental Relative Moldiness Index (ERMI), which is a measure of mold contamination in homes.^{[350,360–364]}

A common vacuum cleaner is usually used to collect the settled dust from the surfaces. There are also commercially available plastic nozzles that can be adapted to the end of the suction tube of almost all vacuum cleaner models (e.g., sampling).^[341] After weighing, specific filters are adapted to trap the dust collected. It is recommended to use the vacuum cleaner at 1600 W, since this power value is generally suitable to collect dust samples homogeneously.^[125] To test the effect of applying a certain anti-allergen treatment (e.g., acaricides), it is more appropriate to limit the sampling to the surface on which the treatment was carried out and to express the result µg m⁻² min⁻¹. On the other hand, to know the allergens present in a specific environment emerges, it is advisable to select representative points where it is more appropriate to sample based on the environment being examined and express the result in µg of allergen per gram of dust (µg g⁻¹).^[125] The samples should be transported at refrigerated temperature (5 ± 3 °C). The main allergens can be determined in dust samples, after appropriate extraction. Dust samples taken from the filters can be stored at −20 ° C until analysis after having weighed the loaded filters.^[125]

The EDCs are simple and cost-effective tools to quantify fungal exposure (e.g., CFU) in settled dust and may better represent inhalable air with respect to vacuum floor dust.^[365–368] After sampling, the collectors are stored at −20 °C until further processing, since this temperature is suitable for preservation of fungal spore suspensions in culture collections for up to 5 yrs with very good survival ability.^[369] In a recent work, fungi were isolated from the EDC by pressing the surface onto a Petri dish with potato dextrose agar as a nutrient medium. The colonies grown on plates incubated at 25 °C for 7–10 days were counted and studied.^[333]

7.3. Sampling procedures for virological monitoring of surfaces

Airborne viruses in bioaerosols are more difficult to study (e.g., culture) than bacteria and fungi,^[128,370] because they are intracellular parasites that require a host cell for reproduction.^[371] Pathogenic viruses have been found to be present indoors in low concentrations that can render their detection difficult.^[372,373]Viruses also are generally more susceptible to damage during aerosol collection compared to bacteria or fungi, although this varies widely with the collection method and species.^[374–376]

For virological monitoring of surfaces, the same sampling procedures are applied as those used for classical microbiology, i.e., swabs, sponges and slides. Contact plates are rarely used. The most widely used protocols for the identification of surface viruses have been described in a review by Julian et al.^[377] who compared the results of 59 studies. The most used materials in swabs are cotton (60% of studies) and polyester (16%), followed by rayon and other antistatic swabs. As for elution, the most used eluents are broth (beef extract, minimum essential medium, tryptose / phosphate broth) and Ringer’s solution. Generally, the most effective method for recovering viruses is with polyester swabs pre-moistened and eluted with Ringer’s solution (¼ concentration).^[158] Macrofoam swabs performed best when recovering wet or dried norovirus from stainless steel surfaces, followed by cotton, rayon and polyester swabs.^[378]

7.4. Methods of analysis

In order to protect humans and the environment from biological threats, bioaerosol concentration should be determined in a reliable way. As it was already mentioned, harmonized analytical standards have not been established yet. The microbiological analyses can be performed either on samples taken directly from the air or from surfaces. The main methods of investigating air samples include direct culture methods and biological, biochemical and immunological analyses. To evaluate the hygienic condition of surfaces mainly i) microbiological; ii) chemical and iii) biochemical methods are preferred. Results may be quantitative or semiquantitative depending on the analysis performed. All methods assume that a surface, especially if properly cleaned and disinfected, must not have high quantities of those molecules considered indicators of a poor degree of hygiene. First of all, microorganisms can be identified with a microscope on the basis of their morphological characteristics. Microscopic observations utilize the bright-field, light, phase contrast, fluorescence or even electron-based approaches.^[379,380] These methods enable the enumeration of both viable and non-viable microorganisms^[380] as well as other non-culturable bioaerosols including cell wall fragments, plant pollen, etc.^[381,382]

For example, swabs can be analyzed by direct microscopy (for fungi only), by qPCR (fungi and bacteria), PLFA (fungi and bacteria) and culturable methods (fungi and bacteria). The number of microorganisms grown in the culture media is calculated as CFU per cm² cf. Eq. (51)(51) $$\mathrm{CFU}\mathrm{ }\left({\mathrm{cm}}^{-2}\right)=\frac{\mathrm{D}\times \mathrm{V}\times \mathrm{N}}{100}$$(51) ^[125].(51) $$\mathrm{CFU}\mathrm{ }\left({\mathrm{cm}}^{-2}\right)=\frac{\mathrm{D}\times \mathrm{V}\times \mathrm{N}}{100}$$(51) where: N = number of colonies counted; V = volume of eluent solution used; D = possible dilution factor 100 = corresponding buffered surface.

To obtain consistent and reliable results from swab sampling, material, elution media and method of swabbing should be chosen carefully. If swab samples are to be cultured, aseptic technique is needed to avoid contamination. Moore and Griffith^[349] tested eleven different swab (e.g., cotton, nylon-flocked and rayon swabs) wetting solutions containing various combinations of salts, surfactants and nutrients. They found that the recovery efficiency varied widely depending upon the species of bacteria, type of swab, and whether the surface was wet or dry.

The number of microorganisms grown on the surface of the culture media is calculated as colony forming units (CFU) per cm² cf. Eq. (52).$$\mathrm{CFU}\mathrm{ }\left({\mathrm{cm}}^{-2}\right)=\frac{\mathrm{N}}{\mathrm{S}} \left(52\right)$$where N = number of colonies counted; S = surface in cm² of the contact plate used.^[125]

The microbiological methods allow identification of microorganisms, requiring at least 24–48 hours^[383,384] to provide results due to cultivation of viable microorganisms on agar media^[384,385] in samples transported to dedicated laboratories. Besides its time-consuming nature, another drawback of the microbiological methods is that not every microorganism is culturable. Moreover, viable bacteria sampling is limited to short sampling times for reducing the loss of viability and may introduce significant measurement error. To the contrary, the chemical and biochemical methods (which requires an instrument for analysis) provide results in a few minutes, with the possibility of data acquisition. The microbiological methods have the possibility of being applied with numerous techniques such as the dip slides, contact plates and swabs.^[125]

For a quick evaluation of the hygienic condition of a surface, (bio)chemical methods such as the adenosine triphosphate (ATP) bioluminescence assay can be used. This method was originally designed to determine the bacterial concentration on the surface without incubation, and it is based on the good correlation between cellular ATP measurement and the number of viable bacteria present.^[386] Apart from being a universal energy carrier in all living organisms, ATP is involved in several enzymatic reactions. This method uses luciferin as substrate, ATP and luciferase enzyme (Eq. (53)(53) $$\text{Luciferin}+\text{ATP}\underset{+{\text{Mg}}^{2+}}{\overset{\text{Luciferase}}{\to }}\text{Luciferyl}\mathrm{ }\text{adenylate}\mathrm{ }\text{complex}+\text{Pyrophosphate}$$(53) ). The light emitted cf. Eq. (54)(54) $${\text{Luciferyl}\mathrm{ }\text{adenylate}\mathrm{ }\text{complex}+\mathrm{O}}_2\stackrel{}{\to }\text{Oxyluciferin}+\mathrm{h}v$$(54) is detected by a luminometer.^[387,388](53) $$\text{Luciferin}+\text{ATP}\underset{+{\text{Mg}}^{2+}}{\overset{\text{Luciferase}}{\to }}\text{Luciferyl}\mathrm{ }\text{adenylate}\mathrm{ }\text{complex}+\text{Pyrophosphate}$$(53) (54) $${\text{Luciferyl}\mathrm{ }\text{adenylate}\mathrm{ }\text{complex}+\mathrm{O}}_2\stackrel{}{\to }\text{Oxyluciferin}+\mathrm{h}v$$(54)

Another advantage of the method is that the reaction is activated with extremely low ATP levels (after appropriate preparation protocols). A drawback is that a damaged or dead cell is no longer able to produce it. Therefore, the bioluminescence assay does not allow to discriminate either the type or the species of contaminant, but it can quickly give an evaluation comparable to a “total count” test.^[125] The ATP bioluminescence assay has found different application fields.^[389–391] The efficiency of the ATP bioluminescence assay has been evaluated also by comparison with the colony counting method.^{[383,389,392]}. Recently, a novel biosensor consisting of a condensation system, microfluidic channel, and bioluminescence transducer to detect ATP from aerosols in real-time has been developed^[386] (Figure 4).

Figure 4. Schematic diagram of a system consisting of a pneumatic aerosol generation system, high-efficient condenser and a microfluidic chip for ATP extraction and detection of bioluminescence adapted from Lee et al.^[386]

Mold levels were quantified in a housing estate after sampling of surfaces with sterile cotton swabs and an adhesive template using a chemical method based on the quantification of the β-N-acetylhexosaminidase activity. Briefly, an enzyme substrate containing 4-methyl umbelliferyl as fluorophore was added to the swab samples. After a reaction time, the resulting fluorescence formed was measured using handheld fluorometer.^[393]

Endotoxins are heat-stable lipopolysaccharide molecules associated with the outer membranes of certain Gram-negative bacteria. When these bacterial cells die, endotoxins are released. For example, in a recent study comparing schools with and without reported indoor air problems, dust was sampled using a vacuum cleaner with disposable wipers as filters and the endotoxin concentration was determined after extraction of endotoxins into 0.9% NaCl supplemented with 0.025% Tween 20 via a chromogenic signal on a microplate absorbance reader at 405 nm. For quantitation, a standard curve was created using the E. coli endotoxin standard.^[394] Although widely used, endotoxin assays lack comparability between results obtained in different laboratories because of differing sampling, extraction, and analytical methods^[395–398] and water-insoluble endotoxins are not detected.^[379]

For living fungal biomass, typically the membrane lipid ergosterol is frequently used as a biomarker. It is based on the assumption that ergosterol is labile, and therefore rapidly degraded after the death of fungal hyphae.^[399]

The chemical methods rely on kits that are based on a chemical reaction that involves a color change in the presence of proteins.^[125] Indirect or proxy chemical-based approaches are available for the detection and quantification of bacteria and fungi in environmental samples.^[356,400] Common approaches consist of off-line sampling and thermal desorption – GC-MS,^{[130–133,400]} GC-MS/MS, and matrix-assisted laser desorption/ionization MALDI-TOF.^[338,400] Mycotoxins can be determined by liquid chromatography-based methods.^[339] Although MS detectors possess excellent sensitivity, crude extracts cannot be directly investigated. The most used cleanup procedures are filtration and centrifugation. Both methods will lessen the likelihood of clogging or damaging the system, but neither of them offers target analyte enrichment. For this purpose, either liquid-liquid extraction or solid-phase extraction should be considered.^[339] Recently, Misztal et al. demonstrated that PTR-MS^[129] can be used for detection of fungal VOCs.

Immunoassays are also commonly used for bioaerosol analysis,^[339] e.g., enzyme-linked immunoassay (ELISA).^[339] From the four ELISA variants developed, the “sandwich” one is used for detection of allergens. Briefly, there is need of two antibodies, the first is used to capture the antigen present in the material obtained from the sample extraction process. The second one has the function of recognizing the antigen captured by the first antibody and of then act as a “bridge” with an appropriately labeled secondary antibody or with a system capable of further amplifying the signal (of the biotin-avidin type) or directly with an enzyme such as peroxidase.^[125]

To overcome challenges associated with traditional methods, real-time sensor technologies are being developed for the detection of bioaerosols.^[401,402] The sensor technologies are based on a variety of signal detection strategies that include optical, mechanical, electrical, or magnetic sensing approaches. Savory et al.^[118,403] proposed the use of optoelectronic sensor for in situ monitoring of mold growth in concealed spaces in real-time by measuring changes in light reflectance from the active element of the sensor. This latter was a permeable, hydrophilic membrane inoculated with mold spores affixed to a test surface over which a miniaturized optoelectronic illumination (white light emitting diode) and sensing (photodiode) device was positioned within a plastic housing.^[118] Other devices use indirect (or proxy) methods, such as VOC detection using an electronic nose.^[404]

Nucleic acid-based molecular diagnostics are becoming popular for detecting and quantifying bioaerosols.^[405] There are three steps involved in gene-based assays: i) extraction and purification of nucleic acids; ii) amplification of the gene target; and iii) detection of the amplicon. Several kits are commercially available for viral, bacterial, plant and fungal nucleic acid extraction and purification. These kits are generally based upon either silica adsorption (spin-column) or affinity purification (magnetic separation) methodologies.^[339] The PCR and real-time qPCR detect specific genetic sequences in the sample DNA or RNA. Therefore, it has been possible to provide standardized assays, reduce analysis time, and enhance sensitivity and detection specificity.^[406–408] For example, swabs can be analyzed by qPCR for fungi and bacteria.^[125] The PCR method is quantitative if a known area is swabbed. Otherwise, results may be given a semi quantitative rating. Recently, differences in survival for a synthetic community comprised of 5 non‐model bacteria (i.e., Bacillus timonensis, Enterococcus hirae, Kocuria rosea, Microbacterium oleivorans, and Pantoea allii, cultivated from dust samples collected from athletic facilities) on three chemically and physically distinct materials (e.g., conventional and microbicidal acrylic paint as well clay paint) has been demonstrated by qPCR analysis.^[409]

Similarly, the PLFA measurements are widely used in microbial ecology as chemotaxonomic markers of bacteria and other organisms. Phospholipids are the primary lipids composing cellular membranes, which can be saponified. Once the phospholipids of an unknown sample are saponified, the composition of the resulting PLFA can be compared to the PLFA of known organisms to determine the identity of group of organisms in the sample (such as Actinomycetes, gram-positive and gram-negative bacteria, Archaea and fungi).^[410]

8. Conclusions

It is challenging to characterize the surface transformations of indoor air pollutants and to provide information that is valid for all types of buildings, since most of the literature found in this topic deal with dwellings in the US equipped with HVAC systems (e.g., HOMEChem studies). Moreover, these buildings are often made of materials different from other regions of the World. Mainly due to global warming, the effect of air conditioning on surface transformations will definitely have a higher impact. By analyzing the corresponding literature data, a major recommendation to the building scientific community is that surface reservoirs indoors should be clearly defined. Their classification could be achieved considering the physico-chemical characteristics of building and indoor materials affecting transformation of indoor air pollutants rather than according to room type or room elements. From the previous reviews in this field,^[5,11,29] it also stands out that there is a need of a better physico-chemical characterization of surface reservoirs. While chemical properties of the interfaces of many of refractory materials (e.g., silica, gypsum, stainless steel, granite) have been studied, little is known about other material such as wood, upholstery components or insulation materials. As surfaces of these less-refractory materials are not pristine having been subject of chemical aging and/or particle deposition after production and use, achievement of this goal is challenging. Not only the physico-chemical characteristics of the sorbed chemicals need to be understood but also that of the underlying materials. Fortunately, physisorption measurements, microscopic imaging as well as modern optical/molecular spectroscopic techniques are at easy reach nowadays. Costly and large benchtop instruments such as direct analysis real-time, proton transfer reaction-mass and negative ion chemical ionization mass spectrometers have been successfully used for monitoring of air contaminants, notably (microbial) volatile organic compounds and low molecular weight (in)organic acids (e.g., HONO, carboxylic acids) in laboratory and chamber studies conducted especially in the North America to understand chemical processes (e.g., oxidation kinetics of contaminants with O₃, response of contaminant concentration to perturbation experiments such as enhanced ventilation or floor washing with vinegar- and NH₃-based cleaners). In order to avoid that issue, the best option is to locate such instruments in a room nearby, nevertheless the effect of the sampling lines (length, material) is recommended to be studied for a correct chemical characterization of compounds. Another option is the analysis of removable building materials in laboratory. Since most of these real-time mass spectrometric results have been published in the past five years, it is still unpredictable what does future hold for this research-grade instrumentation in terms of their use for field studies. Development of new inlets/instruments that combine chemical characterization of gas and particle phases is promising. On the other hand, it is also recommendable to clearly define the analytical task to be solved. For example, surveys aiming at monitoring of indoor air quality in dwellings typically require unobtrusive, energy-saving, and cost-efficient instrumentation such diffusive samplers followed by off-line analysis. Hopefully, outcomes of studies conducted with these fast response, real-time mass spectrometric methods could be used to refine sampling and analysis protocols for optimizing indoor air quality monitoring. Surfaces act as sources and sinks of trace gases and particulate matter and are an important media for the water-mediated conversion of inorganic contaminants such as NO₂ (into HONO) or acid-base dissociation (e.g., NH₃, HCOOH, AcOH, etc.), since water is ubiquitously present. Deposition to surfaces, ventilation and exchange with outside air will affect the lifetime of indoor pollutants.

Models to study partitioning of air pollutants to surfaces as well as their validation through analytical measurements have been experiencing a huge progress since 1977, when attention was drawn for the first time to adsorption of semivolatile polyaromatic hydrocarbons onto particulate matter studied by time integrated (i.e., off-line) sampling using adsorbents to retain analytes followed by gas chromatographic analysis with flame ionization detection. Using benefit of recent advances in analytical instrumentation, predictive type of models will provide increasingly reliable results and better interpretation of chemical processes involving surfaces indoors. In the future, it is expected the development of complex approaches modeling surface processes.

Occupants are often interested in more tangible results such as spread of legionellosis or viral diseases. Moreover, biological processes occurring on surfaces can impact the chemical composition of the air surrounding these surfaces. Therefore, a holistic and transversal approach including physico-chemical and microbiological characterization of surfaces is also recommended with the harmonized cooperation of analytical chemists, building engineers and microbiologists. There is a strong need to provide harmonized methods for the characterization and evaluation of biological pollutants in closed environments. In the lack of them, the measures necessary to prevent and/or reduce the indoor concentration levels of the pollutants are limited to controlling the known sources and microclimatic factors supporting microbial growth. Continued characterization of pollutant deposition and reactive products generated on various indoor surfaces (e.g., carpets and furniture) will be important toward identifying or developing the most promising building and furnishing materials for improving indoor air quality. These could be zero/low emitting materials or materials that actively remove pollutants.

Acknowledgements

The authors would like to express their gratitude to Delphine Farmer (Colorado State University) for the inspiring plenary lecture on the HOMEChem study delivered at the annual meeting of CA17136 on December 13th, 2018 at the University of York (UK). The initial idea for developing this review given by Gabriel Bekö (Technical University of Denmark) is also warmly thanked.

Additional information

Funding

This publication is based upon work from COST Action CA17136 supported by COST (European Cooperation in Science and Technology) (www.cost.eu).