Advanced instrumental approaches for chemical characterization of indoor particulate matter

Abstract

Particulate matter (PM) is an important player of indoor air quality and a topic of great interest in terms of public health. Deciphering the complex chemical composition of indoor PM is critical to understand the association between particles components and a wide range of adverse health effects. Over the last decades, advanced analytical instrumentation has been produced capable of providing various levels of information on the chemical features of indoor PM. This article reviews the most promising of these sophisticated analytical techniques that could be employed in the identification of organic and inorganic constituents of indoor PM, including (high-resolution) mass spectrometry, organic, carbonaceous and ions analytical techniques, elemental analysis techniques such as atomic spectrometry and X-ray based methods, and surface analysis techniques. A distinction is made between online and offline instrumentation, focusing on their capabilities and how they are currently being used in the targeted and untargeted analysis of PM components. This review aims to provide the indoor air chemistry community with insights into the power of the different techniques available today, so that they can be used advantageously in future studies.

Keywords:

• Indoor particulate matter
• chemical composition
• targeted and untargeted analysis
• online and offline analysis
• high-resolution analytical techniques

1. Introduction

Increasing research on the chemistry, composition, and microphysics of atmospheric particulate matter (PM) has been critical to uncover many aspects of its diverse functions, namely those related to air quality issues,^[1,2] gas–particle atmospheric reactions,^[3] biosphere–atmosphere interactions,^[4] climate change,^[2] and human health.^[5] Among these different aspects, air quality issues and, consequently, air pollution and its impact on human health are probably the most pressing examples of the environmental implications related with atmospheric PM.^[6] This is even more critical indoors, where people spend a significant portion of their day and receive most of the exposure to both outdoor and indoor-generated PM pollution.^[7] Furthermore, indoor aerosols (mostly in the size range below 100 nm) can also favor airborne transmission of pathogens, such as the SARS-CoV-2, which is responsible for the COVID-19 infections.^[8] The coalescence of indoor aerosol particles and virus-laden droplets or droplet nuclei are suggested as plausible players that could facilitate the transport and persistence of SARS-CoV-2 in indoor environments.^[8] The potential impact of indoor aerosol particles on COVID-19 transmission and infection in confined spaces highlights the need of a better understanding of the characteristics and sources of indoor aerosols.

Ambient PM comprises a complex mixture of solids and/or liquids with different size particles, whose chemical composition is commonly described as being multi-component within each size bin (referred as “internally mixed”) or distinguished between different size bins (referred as “externally mixed”).^[9] Regardless of the mixing state of a population of air particles, understanding the environmental implications of atmospheric PM depends on how well its chemical composition and sources are deciphered. As highlighted by Cassee et al.,^[10] additional air quality metrics related to the chemical composition of atmospheric PM [e.g., black carbon (BC), secondary organics and inorganics, metals, and polycyclic aromatic hydrocarbons (PAHs), among others] are more valuable for evaluating the health effects of mixtures of air pollutants from a variety of sources than if only the PM mass is taken into account. This is also valid for indoor air, where the chemical composition of indoor PM is also extremely dynamic and far less well characterized than its outdoor counterpart.^[7]

A comprehensive characterization of the complex mixture of organic and inorganic constituents within PM, to understand its origin and assess on abatement strategies, requires the application of multidimensional and complementary analytical methods, able to provide data ranging from bulk measurements such as mass loadings of specific PM components [e.g., BC/elemental carbon (EC), organic carbon (OC), major inorganic ions, organic speciation and major and trace elements] to in-depth properties of individual particles (e.g., elemental mapping of internally mixed air particles),^[1,11] and molecular-level characterization of either specific organic constituents [e.g., PAHs, organophosphate flame retardants (OPFRs), and polychlorinated biphenyls (PCBs)] or other complex organic molecules.^[12–18]

Over the last two decades, the rapidly-evolving field of analytical instrumentation has produced sophisticated tools capable of meeting this challenge, either by providing field and real-time information on the chemical composition of atmospheric PM, or by affording a high chemical resolution, but at the expense of a low temporal resolution. In either way, when delving into the chemical analysis of PM, it is important to match the level of knowledge on the chemical species to the problem being addressed. Figure 1 stem from the review works of Nozière et al.^[12] and Duarte and Duarte^[19] on the state-of-the-art of atmospheric organic compound analysis, and illustrates the levels of PM components identification related to the research needed on the atmospheric behavior, properties, and sources of PM components. This problem-oriented approach can help avoid redundancy of information and enables an easier coupling to appropriate chemometric tools for data treatment. Although developed for outdoor PM studies, such strategy can also be applied indoors, focusing attention onto the real needs of pollution studies in confined spaces.

Figure 1. Levels of airborne PM components identification related to the research need on atmospheric behavior, properties, and sources of PM components (n is the number of PM constituents identified and/or measured). n ≥ 100: identification of organic PM compounds according to their properties, such as functional group analysis; 10 ≥ n ≥ 100: real-time identification and monitoring of PM chemical composition (e.g., different organic aerosol categories and inorganic ions); n ≤ 10: identification of molecular markers of PM sources or formation processes (e.g., organic acids, saccharides, and aliphatic amines); and n ≤ 2 to 3: identification and quantification of specific molecular markers for assessing secondary organic aerosol (SOA) source processes (e.g., isoprene, α-pinene, and biomass burning SOA markers). Adapted from the review works of Nozière et al. ^[12] and Duarte and Duarte^[19].

Hence, this review aims to provide an overview of the most promising of these avant-garde analytical techniques that could be also employed into the chemical analysis of indoor PM. A distinction is made between online and offline instrumentation, focusing on their capabilities and how they are currently being used in the chemical characterization of PM. Where techniques have yet to be applied to indoor environment, their potential adaptability is presented. The main purpose of this review is to provide the indoor air chemistry community with insights into the capabilities of these high-resolution technologies, so that they can be used advantageously in the future.

2. Online instrumentation available for indoor PM chemical analysis

As recently reviewed by Duarte et al.,^[16] online or in situ methodologies are typically used in field observations to provide real-time (in the order of seconds/minutes) information on the chemical and physical properties of PM. These online measurements show several strengths, including: 1) in specific cases, avoiding potential sampling artifacts typically associated with offline analysis methods, such as evaporation and chemical reactions during long PM sample collection periods; 2) minimizing the handling of PM filter samples; 3) reducing time of analysis and cost consumption in respect to conventional offline chemical analysis; and 4) allowing the online monitoring hourly and daily changes of pollutants concentrations in order to deepen the study of dispersion/dilution/infiltration phenomena in indoor environments. Due to the tremendous evolution which took place on this field over the last decades, this section provides an overview of the online instrumentation that have been developed and used in real-time chemical analysis of ambient PM, distinguishing the instrumentation that allows the online measurement of the bulk PM chemical composition from those that are employed for measuring online specific components/properties of PM.

2.1. Instruments for measuring online the bulk PM (or size segregated) chemical composition

2.1.1. Online mass spectrometry in PM chemical analysis

Many of the significant advances in our understanding of the atmospheric particles composition can be attributed to the application of mass spectrometry (MS). As described by Pratt and Prather,^[15] on-line mass spectrometers can be distinguished in two broad categories: those that measure the bulk or size segregated chemistry of aerosol particles and those that measure the chemistry of individual particles (single-particle measurements). There are two representative instruments of each of these categories: the Aerodyne aerosol mass spectrometer (AMS) and the aerosol time-of-flight mass spectrometer (ATOFMS), respectively.^[15] Figure 2 shows an illustration of the chemical information provided by each method. The AMS combines thermal vaporization with subsequent electron impact ionization and a one-polarity TOFMS, providing real-time analysis (i.e., 1 minute time-resolution) of size-resolved mass concentration of non-refractory components, such as organic matter (OM) or organic aerosols (OAs), nitrate (NO₃⁻), and sulfate (SO₄²⁻), chloride (Cl⁻) and ammonium (NH₄⁺) in aerosol particles with aerodynamic diameters lower than 1 µm (PM₁), more precisely between approximately 60 and 600 nm.^[20] In the case of OAs, and with the aid of multivariate analysis tools (e.g., Positive Matrix Factorization, PMF), the OAs component is further partitioned into several categories based on the m/z: hydrocarbon-like OA (HOA), low-volatility oxidized OA (LV-OOA), semi-volatile OOA (SV-OOA), cooking OA (COA), and biomass burning OA (BBOA).^[15,21,22] The HOA is usually attributed to primary combustion products, whereas LV-OOA and SV-OOA to secondary organic aerosols (SOA), although the OOA is not always of secondary origin.^[21] The ATOFMS, on the other hand, detects and characterizes the mixing state, or distribution of chemical species, within individual particles with aerodynamic diameters between 30 and 300 nm in near real time. Unlike AMS, this method combines the laser ionization with a double polarity TOFMS.^[15] Thus, mass spectral data from the AMS are representative of the ensemble of particles sampled, whereas those from the ATOFMS are representative of individual particles. The later allows obtaining the distribution of secondary species {e.g., ammonium nitrate (NH₄NO₃), ammonium sulfate [(NH₄)₂SO₄] and oxidized organic carbon (OC)} on different primary particles [e.g., dust internally mixed with NH₄NO₃, sea salt internally mixed with NO₃⁻ and SO₄2, or soot internally mixed with NH₄NO₃, (NH₄)₂SO₄, and oxidized OC]. Therefore, the ATOFMS provides size-resolved chemical composition, including both refractory and non-refractory species, of atmospheric aerosols.^[15]

Figure 2. Comparison of data provided by AMS and ATOFMS measurements. AMS provides bulk size-resolved non-refractory species [e.g., organics, sulfate (SO₄²⁻), nitrate (NO₃⁻), and ammonium (NH₄⁺)] mass fractions and concentrations for PM₁. Chemometric tools (i.e., PMF) has been used to resolve the organic component into different organic aerosol (OA) categories: hydrocarbon-like OA (HOA), low-volatility oxidized OA (LV-OOA), semi-volatile OOA (SV-OOA), and biomass burning OA (BBOA). ATOFMS also reports size-resolved number fractions/concentrations and provides the mixing state of individual particles, allowing to discern the distribution of secondary species on primary particles: schematic examples of primary particles are shown for sub-µm (organic carbon (OC), elemental carbon (EC), biomass burning (BB)) and super-µm (sea salt and dust) particles, including examples of mixing with secondary species, including SO₄²⁻, NO₃⁻, NH₄⁺, and oxidized OC. Modified and reproduced with permission from Pratt and Prather.^[15] Copyright 2012 John Wiley & Sons, Inc.

Smaller and more portable than AMS instruments (but with lower resolution and without providing size-resolved chemistry), the Aerosol Chemical Speciation Monitor (ACSM) is another instrumental approach that has been developed for long-term unattended deployment and routine monitoring applications, with minimal user interference. Thus, under ambient conditions, ACSM provides quantitative particle mass loading and chemical composition of non-refractory submicron aerosol particles, including OA, SO₄²⁻, NO₃⁻, NH₄⁺, and Cl⁻, with a detection limit less than 0.2 µg m⁻³ and 30 minute time-resolution.^[23] PMF analysis of the OA spectra obtained from the ACSM can be also pursued to further deconvolve the OA into the above categories.^[12,23]

A new instrument based on AMS and ACSM technology, the time-of-flight ACSM (TOF-ACSM), has been further developed by Fröhlich et al. ^[24] for continuous online measurements of chemical composition and mass of non-refractory organic fraction of submicron aerosol particles. This instrument retains the advantages of the ACSM such as compact design, semi-autonomous operation, and relatively low cost, while greatly improving mass resolution and detection limits (< 30 ng m⁻³ for a time resolution of 30 minutes). However, in contrast to AMS, the TOF-ACSM does not feature particle sizing, which is similar to the widely used quadrupole-ACSM. According to Fröhlich et al.,^[24] the associated software packages (single packages for integrated operation and calibration, and analysis) provide a high degree of automation and remote access, minimizing the need for trained personnel on site.

The Filter Inlet for Gases and AEROsols (FIGAERO) coupled with a time-of-flight Chemical Ionization Mass Spectrometer (TOF-CIMS) allowed online measurements of both gas-phase and particle-phase chemical constituents of ambient OA.^[25] It must be mentioned, however, that the TOF-CIMS allows for the determination of molecular ion composition but not molecular structure, which means that it cannot distinguish between, e.g., peroxy nitrates and multifunctional alkyl nitrates.^[26] Nevertheless, due to its low backgrounds, this instrumental approach provides detection limits of ppt or lower for compounds in the gas-phase and in the pg m⁻³ range for particle phase compounds,^[25] being well suited for both chamber studies as well as measurements in locations where OA and gas loadings are low, namely in indoor environments. Moreover, different types of ionization sources can be used as a function of the target compounds. For example, the FIGAERO-TOF-CIMS using iodide-adduct ionization allowed the identification of 88 particle-phase organic nitrates, arising from first- or second-generation gas-phase products of biogenic hydrocarbons oxidation.^[26] Using a similar iodide-adduct ionization scheme, FIGAERO-TOF-CIMS was also successfully used for the online identification and quantification of 17 sulfur-containing organics (i.e., organosulfates, organosulfonates, and nitrooxy organosulfates) in the gas- and aerosol-phase in Beijing, with limits of detection in the ng m⁻³ range.^[27] The acquired data also allowed to quantitatively assess the contribution of both biogenic (monoterpene and isoprene) and anthropogenic (PAH, and aromatic precursors, such as benzene and methyl naphthalene) precursors to the total sulfur-containing organics determined in this study.^[27] Recently, the FIGAERO-TOF-CIMS using an acetate ionization scheme allowed the identification of 4 to 46% of fresh and aged bus emissions of total particulate through identification of 61 species.^[28] According to Roberts et al.,^[29] the acetate ionization scheme is a sensitive approach for measuring semi-volatile organic compounds (SVOCs), particularly carboxylic acids, owing to its relatively weak acidity, and therefore ability to abstract a proton from many ambiently more acidic SVOCs. Le Breton et al.^[28] also reported that CIMS was able to probe chemical regimes and markers within fuel emissions, but it is also blind to many of the organics due to the high selectivity of the acetate ionization and inability to detect functional groups such as aldehydes and ketones.^[28] The ability of FIGAERO-TOF-CIMS to simultaneously measure the particle- and gas-phase also enables the oxidation process forming SOA to be probed into further detail. In this regard, Buchholz et al.^[30] employed the FIGAERO-TOF-CIMS to investigate chemical composition changes during isothermal particle evaporation and particulate-water driven chemical reactions in α-pinene SOA of three different oxidative states. The thermal desorption data was then analyzed with PMF to identify the drivers of the chemical composition changes observed during isothermal evaporation.^[30] As highlighted by the authors, this data treatment approach can be also applied to ambient FIGAERO-TOF-CIMS thermal desorption data, allowing the identification of OA sources (e.g., biomass burning or oxidation of different precursors) and types (e.g., HOA or OOA).^[30] Siegel et al.^[31] used an offline strategy to assess the chemical composition of secondary aerosol components and semi-volatile organic fraction of primary aerosols collected close to the North Pole. In this study, the aerosol samples were collected on Teflon® filters, which were subsequently analyzed in the laboratory using an FIGAERO-TOF-CIMS with an iodide-adduct ionization scheme. The authors were able to identify 1586 different ions, whereof 519 were clustered with I⁻ and only these were considered in the analysis.^[31] The collected submicron aerosol samples were abundant in oxygenated organic and sulfur-containing compounds, indicative of marine aerosol sources, with a wide range of carbon and oxygen numbers.^[31] Future work with FIGAERO-TOF-CIMS in this pristine atmosphere is expected to include gas-phase measurements for direct comparison between gas- and particle-phase chemical composition.^[31] This strategy could be also employed in indoor environments, where the levels of OA and gas-phase species are also low.

New particle formation (NPF) represents the first step in the complex processes leading to the formation of new secondary ultrafine particles (UFP).^[32] The newly formed nanoparticles affect human health and air quality, besides weather and climate, being therefore of importance for indoor air quality studies. The atmospheric pressure ionization (APi) and the chemical ionization‐atmospheric pressure ionization (CI‐APi), both based on the high-resolution (HR) TOF‐MS, have been key tools for the molecular-level identification of ion clusters involved in atmospheric NPF, as well as the discovery of new types of reaction mechanisms in the atmosphere, leading to the formation of extremely low-volatile organic compounds (e.g., see the works of Lee et al.^[32] and Passananti et al.,^[33] and references therein). In the APi‐TOF-MS, no chemical ionization is applied to the sample flow, and it detects ions (either positive or negative) that naturally form in the atmosphere.^[32] As reviewed by Lee et al.,^[32] the use of the HR‐TOF‐MS allows the acquisition of complete mass spectra at rates greater than 100 Hz, with mass resolving power between 3000 and 7000 m/z, and mass accuracy better than 10 ppm. The nitrate CI‐APi‐TOF-MS allows to detect sulfuric acid (H₂SO₄), NH₄⁺, iodic acid (HIO₃), and highly oxygenated molecules (HOMs) in the gas-phase.^[32,34] The HOMs have generated considerable interest, both from the perspective of providing a new mechanistic pathway to rapid oxidation of volatile organic compounds (VOCs), and also a ready source of SOA precursor molecules.^[34–36] The quantitative determination of HOMs has been achieved using the same calibration factor for the H₂SO₄detection, under the assumption that HOMs are detected near the collision limit and have the same transmission efficiencies as H₂SO₄.^[32] The lack of direct calibration and peak identifications of HOMs species is perhaps the major technical issue for CI-APi‐TOF-MS.^[34] Nevertheless, the sophisticated CI-APi‐TOF-MS tool has been used to study summertime NPF in urban areas of Barcelona by Brean et al. ^[37], allowing the determination of oxygenated volatile organic compounds (OVOCs) and HOMs in high concentration. Of these, the SVOCs arose mostly from isoprene and alkylbenzene oxidation, whereas low-volatility organic compounds (LVOCs) and extremely low-volatility organic compounds (ELVOCs) arose from alkylbenzene, monoterpene, and PAH oxidation. The detection of HOMs in the particle-phase is another analytical challenge, involving additional procedural steps compared to gas-phase, e.g., filter extraction, sample derivatization or thermal evaporation, all processes that may lead to HOMs decomposition.^[34]

2.1.2. Semi-real time analyzers of gases and soluble ions

In the last years, a different set of online analytical techniques, not intended for long term monitoring, has been also developed. One of these are the semi-real time analyzers, which enables the measurement of soluble ions and specific aerosol gaseous precursors.^[38,39] These analyzers consist of an aerosol extraction unit where particles, passing through a growth chamber saturated with water vapor, grow and form liquid droplets that are subsequently collected and transferred into an ion chromatography (IC) system, for anion and cation detection and quantification.^[38,39] As reviewed by Rodríguez et al.,^[38] if denuders are available, time-resolved direct measurements of gaseous precursors such as NH₃, HCl, NO₂, HNO₃, HONO and SO₂ may be carried out and, thus, provide relevant information for the understanding of secondary particle formation processes in the atmosphere. Commercial versions of this analytical technique are available, namely the Monitor for Aerosols & Gases in Air (MARGA) and Ambient Ion Monitor (AIM), online analyzers for semi-continuous detection of gases and soluble ions in ambient air particles. This automatic sampling system allows a time-resolution of 1 h and it has been used to determine the inorganic ions Cl^–, NO₃^–, SO₄^2–, Na⁺, NH₄⁺, K⁺, Mg²⁺, and Ca²⁺ in the PM_2.5 aerosol phase and the corresponding inorganic gases HCl, HNO₂, SO₂, HNO₃, and NH₃ present in the gas-phase at an industrialized area.^[40] According to Thomas et al.,^[39] this automated instrumental approach online coupled to IC has very good linearity and accuracy for liquid and selected gas phase calibrations over typical ambient concentration ranges, and it provides sufficient precision (3 − 9%), even at low ambient concentrations. These features make the automatic MARGA method of great potential for short term online automated measurements of gases and soluble ions in aerosols in indoor environments.

2.1.3. Online measurements of water-soluble organic carbon and ions

A particle-into-liquid sampler (PILS) coupled with a total organic carbon (TOC) analyzer and two IC detection systems (PILS-TOC-IC) is another example of an analyzing system that was developed to acquire high time-resolution measurements of water-soluble ions and water-soluble organic carbon (WSOC, taken as a measure of SOA) by a single sampling and analytical setup.^[41] The time resolution of the PILS-TOC-IC measurements depends on the sample volume needed for the subsequent analysis and sensitivity of the detection system. Therefore, according to Timonen et al.,^[41] a 6-minute time resolution is recommended for WSOC measurements, whereas a 15-minute time resolution is used for inorganic ions measurements. The limits of quantification for the WSOC and major ions was 4 µg L⁻¹ and 2.5 ng mL⁻¹, respectively, which equals to air concentrations of 0.15 µg m⁻³ and 0.1 µg m⁻³, respectively.^[41] Timonen et al.^[41] further compared the PILS-TOC-IC data to those acquired from AMS measurements, and for NO₃^–, SO₄^2–, and NH₄⁺, the correlations between the PILS-TOC-IC and AMS were 0.93, 0.96, and 0.96 (Pearson correlation), respectively. Furthermore, the particulate OM (deduced from AMS data) and the water-soluble particulate OM (deduced from PILS-TOC-IC data) results exhibited a relatively good correlation (r = 0.66, Pearson correlation), highlighting that both methods can provide high time-resolved, quantitative data.^[41] The quantum leap provided by the PILS-TOC-IC together with HR-TOF-AMS for the investigation of the characteristics, sources and water-solubility of ambient submicron OA has been further demonstrated by Timonen et al. at an urban background area, in Helsinki.^[42] The simultaneous measurements of WSOC (in PILS-TOC-IC) and OM from the HR-TOF-AMS provided a possibility to evaluate the solubility of organic fractions recognized by PMF, with the LV-OOA and BBOA (typical of highly oxidized and aged fractions of OA) being found as the most water-soluble OA fractions.^[42] The authors concluded that the PMF combined with the PILS-TOC-IC and HR-TOF-AMS data provided an important tool for OA source apportionment,^[42] which could be of interest to be employed in indoor air chemistry studies. In order to further understand the water-solubility and sources of OA factors, Xu et al.^[43] developed a novel system by coupling a PILS to a HR-TOF-AMS (PILS-HR‐TOF‐AMS) for the direct and online characterization of water-solubility of OA in contrasting urban and rural environments in the Southeastern United States. A 2 D map of the mass defect spectrum (calculated by subtracting the ion’s nominal mass from its exact mass) vs. m/z of HR-TOF-AMS organic ions [C_x⁺ (carbon clusters), C_xH_yN⁺, C_xH_y⁺, C_xH_yO⁺, and C_xH_yO > 1⁺] allowed to conclude that the water-solubility of OA factors (less-oxidized and more-oxidized OOA, as resolved by PMF analysis) varies with time and location,^[43] which suggest that the same OA fractions could have different formation pathways or precursors among different locations. Yet, this online PILS-HR-TOF-AMS methodology has also large uncertainties associated with the estimative of the ambient concentrations of HOA or OA from sources with low water-solubility, mostly because HOA is largely water-insoluble and it only accounts for a small fraction of water-soluble OA.^[43]

2.1.4. Online measurements of trace elements in PM

The elemental composition of ambient aerosols can also be measured with high temporal resolution using online X-ray fluorescence (XRF) spectrometry. The Ambient Metals Monitor (AMM) using XRF spectrometry was designed to measure trace elements concentrations with a time-resolution of at least 15 minutes (and up to 1 hour). This system employs a reel-to-reel design to move a filter tape firstly over the PM collection area and secondly over the XRF analysis area. About 24 elements with a 15-minute time resolution can be simultaneously determined using the AMM,^[38,44] with detection limits as low as 0.014 ng m⁻³.^[44] Nonetheless, as reviewed by Rodriquez et al.,^[38] this online technique has limitations for determining some key dust light elements (Z ≤ 15) such as Si, Al or P due to self-attenuation-induced artifacts, which could be a limitation of this method when dust is mixed with trace anthropogenic metals. Although the online XRF spectrometer can be used for continuous (months, years) monitoring of trace elements at a given site, their cost may restrict the simultaneous deployment of a larger number of devices for different aerosols size fractions or at different sites. The benefit of long-term, quasi-real-time data access, advantageous, for example, for air quality monitoring,^[44] contrasts with the possibilities of using the conventional low-cost, multi-site, and multi-size aerosol samplers used so far in episodic field studies.

2.2. Instruments for measuring online specific PM physicochemical properties

2.2.1. Measurement of atmospheric black carbon (BC)

Carbonaceous compounds constitute a significant fraction of atmospheric PM. One important form of carbonaceous aerosol is soot, which is light-absorbing and whose effects are especially significant in high traffic areas and in those that rely on biomass burning and coal for heating and cooking. Soot contains both OC and EC/BC, with the proportions varying as a function of the origin (higher and lower OC/EC in, respectively, biomass burning and diesel soot for example). Although the term soot is sometimes used as equivalent to BC or EC, this is not technically correct, even if BC/EC is a major component. This soot component is made by graphitic carbon arising from incomplete combustion processes. Depending on the analytical procedure employed, this graphitic-BC can be operationally defined as: i) BC, when its light-absorbing properties are measured, and ii) EC, when its mass concentration is measured by means of thermal-optical methods. Despite intensive efforts over the last decades, no widely accepted standard measurement method exists for the determination of BC. Several analyzers based on the principle of light attenuation through a filter and of photoacoustic oscillation have been developed to determine BC mass. The Multi Angle absorption photometer (MAAP) and aethalometers are two instrumental approaches typically used for measuring online aerosol BC.^[45,46] Since BC by definition cannot be unambiguously measured with these instruments, it is customary to convert the absorbance units into mass ambient concentrations using factors obtained by comparison with EC measurements. Therefore, when using these instruments, the measured carbonaceous light absorbing aerosol constituent is often referred as equivalent BC (BCe) or light-absorbing carbon (LAC).^[45]

The aethalometer measures the change in optical transmission of a filter as aerosol is deposited on the filter over time. This measurement is made at either two (880 nm and 370 nm) or seven (370 nm, 470 nm, 520 nm, 590 nm, 660 nm, 880 nm, 950 nm) wavelengths.^[46] The ability of these aethalometers to measure light absorption of air particles from near ultraviolet (UV) to near infrared wavelengths offer the opportunity to distinguish between different origins of BC. Two wavelengths instruments allow the distinction between brown carbon produced by biomass burning (at 370 nm), and traffic related BC (at 880 nm). Seven wavelengths instruments further allow a deeper source investigation, including for example the presence of Sahara dust.^[46–48] In this equipment, the filter change is performed automatically, which is an advantage for long term monitoring programs. On the other hand, since the aethalometers continuously collects air particles on a quartz-fiber filter tape, this might also cause various systematic errors, including: (1) multiple scattering by the filter fibers, (2) scattering of the aerosols embedded in the filter, and (3) increase of the light attenuation due to light absorbing particles accumulating in the filter, which reduce the optical path for a loaded filter. Therefore, this online instrument requires the application of specific correction algorithms, mostly accounting for the optical properties of the aerosol particles embedded in the filter.^[49] In this regard, a new aethalometer has been equipped with a real-time loading effect compensation algorithm, based on a two parallel spot measurement of optical absorption.^[50] Intercomparison studies have shown excellent reproducibility of the acquired data, and very good agreement with post-processed data obtained using earlier aethalometer models and other filter-based absorption photometers.^[50]

In contrast to the aethalometer, the MAAP detects not only the transmitted, but also the backscattered light at two angles to minimize the influence of light-scattering aerosol components on the angular distribution of the backscattered radiation.^[38,49] The aerosol BC mass concentration is thereafter obtained at a single nominal wavelength (670 nm).^[45] One of the main advantages attributed to the MAAP instrument is that it does not require the use of any aerosol-related correction factors, since the instrumental artifacts are reduced in comparison with aethalometer.^[38,49] However, a measurement artifact in the MAAP has been reported for locations with high BC concentrations.^[45] According to Hyvärinen et al.,^[45] this artifact seems to be related to erroneous dark counts in the transmitted light photodetector, in combination with an instrument internal averaging procedure of the photodetector raw signals. This can be overcome by applying a specific algorithm to correct the BC estimation.^[45]

A portable, reliable, and more affordable equipment for online BC monitoring is also available, namely the microethalometer. For example, Rivas et al. employed a microethalometer for real-time measurements of BC concentrations in schools (in classrooms and outdoors), yielding a good correlation with EC concentrations from filter samples simultaneously collected in situ.^[51] The microethalometers are particularly well suited for on-person exposure monitoring programs and, therefore, can be used to take measurements of cookstoves and tobacco smoke exposure in indoor environments. Nevertheless, no automatic filter change is available on microethalometer, so the measurement period is limited to one day or less in room with high concentrations.

2.2.2. Measurement of atmospheric carbonaceous aerosols

The continuous monitoring of carbonaceous aerosols, namely its OC and EC constituents, is also essential for long-term air quality assessment. The Semi-Continuous OC-EC Field Analyzer, developed by Sunset Laboratories, provides measurements of OC and EC on a customizable sampling time (usually varying from 1 to 3 h intervals, although it could be as low as 30 minutes). This allows for the semi-continuous sampling with online analysis immediately after sample collection without requirement of offline sample treatment or laboratory analysis.^[52] When interferences from refractory material are considered, the Semi-Continuous OCEC analyzer is suitable for routine monitoring networks.^[53] According to the manufacturer, this OC/EC analyzer is also suitable for indoor air exposure assessment monitoring. However, one must consider that the OC/EC concentrations at indoor environments can be on the levels of instrumental noise, which means that the instrument will be working at the threshold of their analytical possibilities. In this regard, Vodička et al. employed field OC/EC Sunset Laboratory analyzers for online measurements of very low EC and OC concentrations in aerosols at a subarctic remote station.^[54] The authors reported that even though the OC concentrations were always above the detection limit of the analyzer, up to 70% of the OC concentration was due to the dynamic blank (gaseous phase).^[54] In these cases of consistently low OC concentrations, it is suggested to apply a dynamic blank correction in order to avoid an overestimation of the carbonaceous aerosol concentrations, particularly of the most volatile fraction of OC (< 200 °C).^[54] On the other hand, the main bias at low EC concentrations originated from the automatically determined split point between OC and EC, which usually led to an overestimation of the real EC concentrations. This bias was solved by employing the new RTCalc703 software package provided by Sunset Lab.^[54]

The photoelectric aerosol sensor (PAS) has been also the technique of choice for real-time measurements (10 second time resolution) of the total particle-bound PAHs concentrations close to traffic emissions,^[55,56] as well as for monitoring outdoor tobacco smoke particles.^[57] This sensor works on the principle of photoionization of particle-bound PAH. Using an Excimer lamp, the aerosol flow is exposed to UV radiation. The wavelength of the light is chosen such that only the PAH coated aerosols are ionized, while gas molecules and non-carbon aerosols remain neutral. The aerosol particles which have PAH molecules adsorbed on the surface emit electrons, which are removed when an electric field is applied. The remaining positively charged particles are collected on a filter inside an electrometer, where the charge is measured. This electric current provides a signal which is proportional to the concentration of total particle-bound PAH. The background signal linked to the proximity of a combustion source can be excluded by operating the Excimer lamp in a chopped mode.

2.2.3. Measurement of atmospheric PM light absorption

One of the most common approaches for measuring online PM light absorption includes the particle soot absorption photometer (PSAP), which is a filter-based, real-time analyzer.^[38,58] The PSAP has a principle of measurement very similar to that of the aethalometer and it measures, in continuum, the light absorption (at 550 nm) by monitoring the transmittance across two areas on the filter: a particle deposition area and a reference area. The absorption reported by the PSAP is calculated with a nonlinear equation corrected for the magnification of absorption by the filter medium and for nonlinearity response as the filter is loaded. As highlighted by Rodríguez et al.,^[38] the PSAP equipment requires a manual filter change, which is a drawback in long-term monitoring programs.

The photoacoustic spectroscopy also allowed in situ measurements of particle light absorption, and it is based on the measurement of the effect of absorbed electromagnetic energy (particularly of light) on matter by means of acoustic detection. Once particles absorb laser radiation, they heat up and rapidly transfer this heat to the surrounding air causing an expansion or an increase in pressure. If the aerosol sample is illuminated with a modulated laser beam, its heating will produce a periodic variation in pressure (i.e., a sound wave) with the same frequency as the laser modulation and will be detected as an acoustic signal by a sensitive microphone.^[58] According to Davies et al., biases associated with photoacoustic spectroscopy include a lack of proportionality between the photoacoustic signal and the aerosol absorption cross section for particles with radii higher than 0.7 µm.^[59] This would not be an issue if using an impactor to remove particles with radii higher than this size threshold. Photoacoustic instruments also require calibration, which can be achieved by measuring the photoacoustic signal generated by known quantities of gaseous ozone.^[59]

3. Offline instrumentation available for indoor PM chemical analysis

Analytical procedures for assessing offline the chemical composition of indoor PM would be similar to those applied to outdoor PM chemical analysis. Typically, measurements of PM composition using offline methods involve the collection of air particles on substrates (i.e., filters and impaction surfaces) for a certain period of time (hours to days), depending on the ambient loading, and using different sampling techniques, such as impactors. In this regard, the European Standard gravimetric method (EN 12341:2014) for the determination of PM₁₀ or PM_2.5 mass concentrations specifies a 24 h sampling period. High-volume samplers are usually employed outdoor for guaranteeing the collection of a significant amount of PM mass, crucial for determining most PM components in a unique filter sample. In indoor spaces, it is recommended to locate the pump out of the room to reduce noise pollution. In terms of collection substrate, quartz fiber filters are typically used for OC, EC and organic compounds analysis, but not for elemental analysis using, e.g., XRF spectrometry. In the latter case, the reason is twofold: (1) higher concentration of the analyzed species in the blank quartz filters; and (2) X-ray based methods are prone to attenuation effects of the emitted X-rays inside the quartz filter itself for low atomic number elements (Z ≤ 15), beside self-attenuation inside single aerosol particles.^[60] In this case, Teflon filters and a low-volume sampler are usually employed in PM collection for subsequent analysis of major and trace elements, whereas nylon filters are recommended for analysis of ions (nylon is the only filter type able to sample NO₃⁻ without negative sampling artifacts). Nevertheless, the nylon filters are usually avoided in many PM field studies, and the ions concentrations are determined in leachates of the Teflon filter after the nondestructive XRF analysis. When using quartz fiber filters for PM collection, these filters are typically divided according to the purpose of the analysis (OC/EC or organic compounds analysis). The following subsections provide an overview of the analytical methodologies commonly employed to unravel offline the chemical composition of PM.

3.1. Analysis of OC, EC, and WSOC concentrations in PM samples

Thermal-optical transmittance methods are typically used for measuring OC and EC concentrations in PM samples collected in quartz fiber filters. In these methods, the carbonaceous species are firstly thermally desorbed in an inert atmosphere (He) and then in an oxidizing atmosphere (mixture of He and O₂). The OC should desorb in the inert atmosphere, while EC combusts in the oxidizing atmosphere at high temperature. However, thermally unstable organic compounds might be pyrolytically converted into EC (char) when heated up in inert atmosphere. The formation of pyrolytic carbon (PC), which darkens the filter, is used to correct the measurement for charring, by continuously monitoring the transmittance or reflectance of the filter during the analysis. The point where the laser transmittance or reflectance reaches its initial value indicates that PC has completely evolved and defines the split point between OC and EC. The most employed thermal protocols for the analysis of OC and EC in PM are the IMPROVE, NIOSH-like protocols, and the EUSAAR2 (European) protocol.^[61–63] The EUSAAR2 protocol is nowadays the standardized thermal-optical method (EN 16909:2017) for measuring OC and EC in PM that minimizes the potential positive bias in EC determination resulting from the incomplete evolution of OC.^[64]

The measurement of the WSOC component in indoor air particles is usually carried out following procedures similar to those applied to outdoor PM samples.^[61,65–67] The PM samples are collected (e.g., using cascade impactor samplers, personal exposure monitors, or a cyclone with an aluminum filter cassette) on filters (e.g., quartz, Zefluor, or Teflon filters) during either 24 h or 48 h, followed by extraction with ultrapure water, filtration of the aerosol aqueous extracts to remove any residue, and measurement of the dissolved organic carbon content in the aqueous extracts using a total organic carbon analyzer.^[68–70] Due to the low amount of air particles typically collected in indoor sampling campaigns, the filters are usually assembled first (in weekly composite samples) before WSOC analysis. Using this methodological approach, the obtained WSOC data result in a weekly overview of the indoor aerosol WSOC concentrations. A daily assessment of WSOC concentrations in indoor air particles will require the use of high-volume samplers, thus allowing the collection of a high amount of air particles for WSOC analysis. Furthermore, the studies published thus far on indoor aerosol WSOC do not express the volume of ultrapure water that should be employed in filter extraction. In this regard, Psichoudaki and Pandis suggested the use of a ratio of water (in cm³) per volume of air sampled (in m³) on the analyzed filter of 0.1 cm³ m⁻³ for the extraction of WSOC for OA concentrations between 1 and 10 μg m⁻³.^[71] This ratio was designed considering the chemical features of outdoor aerosol WSOC; however, this ratio can be used as a guideline for offline indoor aerosol WSOC measurement.

3.2. Gas chromatography-mass spectrometry for the analysis specific organic PM constituents

Assessing the risks of indoor exposure to organic pollutants requires a thorough understanding of the complex organic chemistry within the airborne PM. The organic component covers a very wide range of compounds of both primary and secondary origin. The characterization of those individual organic PM constituents at the molecular level in highly complex mixtures generally requires chromatographic separation. Gas chromatography (GC), often with prior derivatization, has been the technique of choice for the separation of analytes of interest (i.e., semivolatile organic compounds, SVOCs) from other constituents of the complex organic PM matrix. Commonly used detectors in combination with GC include the flame ionization detector (FID), often for quantification, the electron capture detector (ECD), sensitive to halogenated species, the nitrogen phosphorus detector for nitrated compounds, and a range of MS techniques enabling more detailed structural characterization (see the review work of Nozière et al.^[12] and references therein). Moreover, GC coupled to MS (GC-MS) has been applied to the identification and quantification of PAHs, phthalates, organophosphate flame retardants, brominated flame retardants, polychlorinated biphenyls (PCBs), chlorinated paraffins, pesticides, alkylphenols, parabens, key organic species as source tracers (e.g., levoglucosan, galactosan, mannosan, squalene, oleic acid, cholesterol, nicotine), and synthetic musk. Given the atmospheric indoor relevance of some of these classes of compounds that can be selectively detected and quantified by GC-MS, the following subsections will address some specificities of the analytical procedures for their targeted analysis. It is also worthy to mention that the World Health Organization (WHO) has recently published a document about methods for sampling and analysis of chemical pollutants in indoor air, in which most of the organic PM constituents listed below are included in the aforementioned publication.^[72] Furthermore, indoors, the partition of SVOCs between gas-phase, airborne particles and available surfaces depends on the ambient temperature, humidity, types and concentration of SVOCs and PM, and residence time in the air. Therefore, the sampling of the gas-phase has also been considered in the following sections in case the atmospheric phase distribution or the total concentration of SVOCs in indoor air has to be assessed.

3.2.1. Polycyclic aromatic hydrocarbons (PAHs)

A group of ubiquitous environmental pollutants generated from the incomplete combustion of organic substances such as fossil fuels and biomass are PAHs.^[73,74] The sampling strategy for PAHs in indoor air is described in ISO 16000-12, whereas ISO 12884 reports the protocol for the analysis of total (gas- and particle-phase) PAHs using GC-MS, although it could be used exclusively for the analysis of PAHs in the particle-phase. It is important to highlight that depending on the vapor pressure, some PAHs (especially those with vapor pressure above 10⁻⁸kPa) will tend to vaporize from particle filters during sampling.^[75] It should be also mentioned that ISO 16362 specifies the protocol for the analysis of particle-bound PAHs, but using high performance liquid chromatography (HPLC) coupled to either fluorescence or Diode Array detector (DAD), although the combination of both detectors is also possible. According to ISO 16000-12, the same measurement protocol described in ISO 16000-13 for the determination of total PCBs and polychlorinated dibenzodioxins (PCDDs)/dibenzofurans (PCDFs) must be used for PAHs.^[72] This method incorporates a sampling procedure using a low-volume sampler for collecting the PM on a fine-particle filter backed up by a sorbent trap consisting of polyurethane foam (PUF) to collect the vapor-phase fraction. The sampling volume per hour must not be higher than 5–10% of the volume of the room to maintain to the minimum the changes produced in the room atmosphere due to sampling. Alternative sampling devices, easier to handle, are described in the literature to collect an array of SVOCs, including PAHs, PCBs, pesticides, phosphoric esters, musk, phthalates, and polybromodiphenylethers (PBDEs).^[17,18] For the extraction of the PAHs, the filter and the sorbent trap (PUF) can be extracted together using Soxhlet, accelerated solvent extractor or other extraction techniques (e.g., ultrasonic bath and microwave-assisted extraction) after method validation. The extraction solvent is usually dichloromethane (DCM) and before extraction, recovery standards must be added to the sample. A clean-up method can be applied through column chromatography on silica gel. The PAHs are then analyzed by means of GC-MS, operating in SIM mode or, if a triple quadrupole is available, operating in multiple reaction monitoring (MRM) mode, as described by Mercier et al.^[76] and Villanueva et al.^[77] A detailed description of the analytical procedure can be found in Raffy et al.^[18] and in ISO 12884. As for the quality assurance/quality control (QA/QC) procedure, three major practices should be taken into account regarding blanks and that could be applied to the determination of other SVOCs: 1) in order to minimize blank contamination, plastic materials must be avoided, and glass materials must be solvent-rinsed prior to use; 2) if the concentration of a field blank sample or a laboratory blank sample shows a concentration greater than 30% of the concentration found in the associated samples, the samples should be discarded; and 3) if the concentration in the blanks is less than 30% of the concentration found in associated samples, the concentration of the samples can be reported without blank correction. A detailed description of the procedure for determining the sampling efficiency is available in Raffy et al.^[18] and in ISO 16000-13.

3.2.2. Polychlorinated biphenyls (PCBs)

Polychlorinated biphenyls is one category of flame retardants (FRs) that have a wide range of application in industry, construction and electrical devices.^[78] Because of their persistence, bioaccumulation and toxic characteristics, PCBs have been listed as Persistent Organic Pollutants in the Stockholm Convention, adopted in 2001. There are several advanced techniques available for PCBs analysis in air. ISO 16000-12 provides a sampling strategy for PCBs in indoor air, ISO 16000-13 specifies sampling and preparation of sampling media for dioxin-like PCBs, and the corresponding analytical method is included in ISO 16000-14.

As reviewed by Reddya et al.,^[78] particulate sampling can be accomplished using high-volume air samplers or dust samplers. Absorption of gaseous pollutants into a liquid medium is another method for the collection of air samples. Impinger and midget type devices can be used to emphasize a high degree (pollutant) of gas-liquid contact. Air sampling for PCBs is typically carried out by using a XAD resin or PUF, although PUF combined with granular adsorbents (Porapak®R, Chromosorb®102, Amberlite XAD-2, Florisil®PR and Tenax®GC) were also successfully employed in the air sampling of PCBs.^[79] Nevertheless, granular adsorbents are expensive and possess higher airflow resistance than PUF.^[78] As an alternative, Vorkamp et al. suggested the use of silicone for PCBs sampling in indoor air based on its higher sampling rates, greater sensitivity, and good precision.^[80]

Soxhlet extraction is also the technique of choice for PCBs extraction from air samples. A large variety of organic solvents, including DCM, mixture of DCM-acetone, DCM-hexane and acetone-hexane, have been employed for PCBs extraction. The major disadvantages of the method are their prolonged extraction times varying from 24 to 48 h, consumption of high organic solvent (up to 500 mL per sample), and the need for solvent evaporation after sample extraction. To overcome these drawbacks, some improvements have been made to the conventional Soxhlet extraction, including automated, high pressure, and microwave assisted variants. Sample clean up procedures are also performed using both nondestructive and destructive methods.

As for the identification and quantification of PCBs, the most common and generally considered techniques are GC-MS and GC-ECD. In recent years, industries have introduced more sensitive instruments for analyzing PCB congeners, among which high resolution GC coupled with HR-MS (HR-GC/HR-MS) is the most important hyphenated technique. A collection of MS, such as magnetic sector, orbitrap, quadrupole (QMS), ion traps (IT-MS), time-of-flight (TOF-MS) and triple quadrupole (QqQ-MS) detectors, are currently available and can be connected to either GC or LC.

3.2.3. Phthalates

Phthalates are typically used as plasticizers and are added to polyvinyl chloride (PVC) to soften it. Exposure to phthalates has also become a major public health concern, being classified as endocrine disrupting chemicals which impair the human endocrine system inducing fertility problems, respiratory diseases, childhood obesity and neuropsychological disorders.^[81] The ISO 16000-33 is specific for phthalates in indoor air and describes the sampling and analysis of phthalates in house dust and in solvent wipe samples of surfaces in indoor air by means of GC-MS. Two alternative sampling and processing methods are specified for indoor air. Phthalates are collected on an adsorbent tube with Tenax® TA and subsequent thermal desorption as described in ISO 16000-6 for VOCs, or in an adsorbent tube filled with Florisil® and extracted with a solvent as described in ISO 16000-33. In both procedures, the phthalates are analyzed by means of GC-MS.

Since phthalates can be present in the environment in both the particulate fraction and the gas phase, a combined sampler with filters and adsorption tube (PUF or XAD-2) is usually used for air sampling. Phthalates can be collected on particulate matter by using a glass fiber (GFF) or quartz fiber (QFF) filter in a low-volume sampler as described for PAHs or using a personal sampling pump at 2 L/min for 24 h or even days.^{[17,18,82,83]} The flow rate must be checked before and after sampling with an external calibrator. The filter and sorbent trap are usually conditioned before being used. Filters are pre-baked at 400 °C for 5 h and PUFs are cleaned with a solvent (e.g., DCM, DCM-acetone, and toluene) to remove organic compounds. After sampling, filters and sorbent traps are wrapped in aluminum foils (previously washed with hexane or heated at 450 °C for 1 h), sealed in a clean and hermetically container, and transported to the laboratory and refrigerated. The filter and sorbent trap can be extracted using pressurized liquid extraction, ultrasonic extractors or microwaved-assisted extraction.^[17,18,82] Before extraction, recovery standards have to be added to the sample. The phthalates are then determined using GC-MS operating in SIM mode or, if a triple quadrupole is available, operating in MRM mode.

3.2.4. Organophosphate flame retardants (OPFRs)

A detailed description of the sampling and analytical method for FRs based on organophosphorus compounds-phosphoric acid esters is given in ISO 16000-part 31. The sampling system for OPFRs recommended in ISO 16000-31 is described in detail in ISO 16000-13, and it is based in the same sampling device as for collecting PAHs. OPFRs are collected from air on a fine-particle filter (GFF) and a PUF by means of a low-volume sampler. The sampling time should be 1 h, and the sampling volume should not exceed 10% of the air exchange rate. The filter and PUF are extracted together in a Soxhlet extractor with DCM, and the analysis is carried out by means of high- or low-resolution GC-MS (GC/HR-MS or GC/LR-MS, respectively). The compounds are quantified by means of the internal standard method.^[72]

The OPFRs have also been collected in indoor air by using passive sampling. Passive samplers are cheap, easy to handle, and do not require external power. They provide time-averaged concentrations of pollutants over the sampler’s deployment period. PUF disks can be exposed to indoor air using a single-bowl (sampling rate of 1.6 m³/day) or double-bowl (sampling rate of 0.82 m³/day) passive samplers housing for 28 day.^[84] Before extraction, all samples must be spiked with known amounts of recovery standards. PUF are extracted with DCM using automated warm Soxhlet extraction, the extract is cleaned-up, spiked with injection standard, and analyzed by means of GC-MS/MS. The mass spectrometer should be operated in positive atmospheric pressure ionization mode (APGC⁺) using MRM mode.^[84]

3.2.5 (Short and medium) chlorinated paraffins

Chlorinated Paraffins (CPs) are complex mixtures of polychlorinated alkanes produced by chlorination of single chain alkanes. The commercially available CPs are generally divided into three groups: short-chain CPs (SCCPs) comprised by 10 to 13 carbon atoms, medium-chain CPs (MCCPs) comprised by 14 to 17 carbon atoms, and long-chain CPs (LCCPs) with 18 or more carbon atoms. The degree of chlorination of the alkane backbone varies but it is usually between 40% Cl and 70% Cl by weight. There is a wide range of industrial uses of CPs, including metallurgy, as high-pressure lubricants in metal processing fluids, and in PVC processing. Additionally, CPs are used as plasticizers and FRs in rubber, paints, glues, sealants, skin fluids, textiles and polymeric materials in order to replace PCBs. Therefore, the CPs are closely associated with the indoor use of these products in vinyl flooring, carpet backing, textiles and fabrics, floor polishes, furniture, wallpaper, and kitchen equipment and appliances. There is currently no evidence of natural sources of CPs in the environment, and the anthropogenic releases of CPs into the environment are mainly through volatilization, wash-off, and abrasion.^[85] The CPs are toxic compounds and have long-term carcinogenic activity, meeting therefore the criteria for being classified as persistent organic pollutants.

The same sampling system recommended in ISO 16000-31 for OPFRs could be also employed to collect CPs. It consists of a low-volume sampler equipped with fine-particle filter backed by a sorbent trap, although other sampling devices have been described in the literature.^[86,87] Analysis of CPs is, however, highly demanding, and the current methods are characterized by lack of precision and accuracy. Furthermore, congener specific analysis is not currently possible. Several different approaches for CPs analysis exist, yielding to very dissimilar results. Different analytical methods have been developed, relying on the use of two-dimensional comprehensive gas-chromatography (GC × GC) coupled to MS, or the use of faster methods with lower complexity, such as GC coupled to electron ionization tandem mass spectrometry (GC/EI-MS/MS), or GC coupled to negative chemical ionization with LR-MS (ECNI-LR-MS). However, GC-orbitrap/MS and GC-QTOF/MS offer higher selectivity and more sensitive results. Moreover, the use of high-performance liquid chromatography (HPLC) coupled to atmospheric pressure chemical ionization (APCI) or metastable atom bombardment (MAB) have been also reported for the analysis of CPs.^[85,88] Nonetheless, the development of an analytical method for the detection of CPs in the atmosphere is a real challenge due to the simultaneous presence of thousands of different isomers, enantiomers and diastereomers.

3.2.6. Brominated flame retardants (BFRs)

Brominated flame retardants (BFRs) such as polybrominated diphenyl ethers (PBDEs) have been collected in indoor air by passive and active sampling.^[72] The active sampling consists of a GFF or QFF filter, followed by a PUF. In general, sample volumes in the range of a few hundred liters to less than 30 m³ are enough to reach indoor limits of detection (LODs) in the low ng m⁻³ level for most compounds. However, lower detection limits (in pg m⁻³ level) can be achieved if sample volumes are higher (100–385 m³).^[89] Therefore, the low-volume sampler described in ISO 16000-13, and specified for PAHs, can also be used to collect BFRs.^[72]

Personal sampling pumps have been also used to collect PBDEs (BDE 85, BDE 99, BDE 100 and BDE 119) in schools.^[18] Sakhi et al. also used a similar device to collect BDE 47, BDE 85, BDE 99, BDE 100, BDE 153 and BDE 154 in indoor air with a flow of 12 L/min for 24 hours.^[86] BFRs are usually extracted using a Soxhlet due to its high extraction efficiency and general robustness. However, other extraction systems have also been used, such as pressure solvent extraction or ultrasound-assisted extraction. Typical solvents are DCM, n-hexane, toluene, acetone or a mixture of these. Treatment with concentrated H₂SO₄, and a sort of clean-up procedures using different sorbents (alumina, silica gel, Florisil® or a combination of these) are also usually employed.^[86,89]

Separation and identification of BFRs is generally performed by means of GC–MS. Details of the analytical method can be found in Melymuk et al.,^[90] Lim et al.,^[91] and Braouezec et al.,^[92] whereas the GC–MS/MS parameters are available in the study of Braouezec et al.^[92] Other techniques for quantifying PBDEs include GC–HR-MS,^[91,92] and GC-MS in electron capture negative ionization mode (GC–ECNI–MS).^[93] Commonly used internal standards of PBDEs are ¹³C-labelled BDE 28, BDE 47, BDE 99, BDE 153, BDE 183 and BDE 209, whereas the recovery standards are ¹³C-labelled BDE 77 and BDE 138.^[90,94] Moreover, ISO 22032 and United States Environmental Protection Agency (EPA) Method 1614 A for determining PBDEs in different environmental samples could help to establish and validate an analytical procedure for indoor air.^[72]

3.2.7. Synthetic musk

Synthetic musk has been usually collected using either low-volume or high-volume samplers with GFF or QFF to collect the airborne PM, and a PUF to collect the gas-phase. Typical flow ranges are between 2 L/min and 0.3–0.4 m³/min, and with sampling volumes of 2–100 m³. For example, Raffy et al. used an active sampler consisting of a QFF fitted in front of a PUF placed in a glass tube.^[18] The air was pumped through the device using a personal sampling pump at 2 L/min during five consecutive days.^[18] An alternative approach for the collection of synthetic musk is to use the same protocol as applied for sampling PAHs in indoor air.

Analysis of synthetic musk in the laboratory implies their extraction from both the filter and PUF, which can be carried out using the conventional extraction techniques, such as Soxhlet or pressurized liquid extraction, by means of different solvent mixtures (although DCM has been the most common solvent). Details of the extraction and clean-up procedures can be found in previous studies.^{[18,89,95,96]} GC-MS has been the technique of choice for the analysis of synthetic musk, and the details on the method parameters can be found in the work of Melymuk et al.^[90], while those of GC–MS/MS are described in the works of Raffy et al.^[18] and Mercier et al.^[76] Mass-labelled PAHs have been also successfully used as internal standards in polycyclic musk analysis; moreover, d10-p-terphenyl could be used as a recovery standard, and d10-fluoranthene as an internal standard.^[90,97,98]

3.2.8. Other key organic species as airborne PM source tracers

GC-MS has been also widely used for the determination of different airborne PM source traces, namely levoglucosan, galactosan, and mannosan (biomass burning), squalene and cholesterol (wood burning, vegetation and other natural inputs, meat smoke, skin surface lipids), oleic acid (microbial sources, processing, degradation and combustion of plant and animal constituents), and nicotine (tobacco smoke) (e.g., see the works of Simoneit et al.,^[99], Weschler et al.,^[100] and Alves^[101] for additional details). These organic species should be major constituents of settled dust in any indoor environment that is occupied by humans and, therefore, their analysis should be included in any indoor air quality assessment.

3.3. Multi-elemental analysis of airborne PM

The presence of elements in airborne PM can often occur at low or extremely low concentration levels (typically in the order of ng/m³, but also pg/m³ for some trace metals).^[99] Therefore, analytical techniques capable of a low-level of detection and a robust quantification are needed, to ensure sensitive and accurate measurements of these pollutants in the atmosphere and assess their compliance with legislative limits, if required. Teflon-membrane filters and polycarbonate membrane filters are typically used in PM sampling for subsequent elemental determinations mainly because of their low elemental blank levels.^[102] Nevertheless, since these filters are extremely resistant to dissolution if a sample preparation step is required, e.g., as in the case of a conventional inductively coupled plasma mass spectrometry (ICP-MS) analysis procedure, then quartz fiber filters or cellulose filters (cellulose nitrate filters or mixed cellulose ester membranes) are preferably employed.^[103,104]

The elemental quantification is mostly performed by techniques that can be classified into three main groups: atomic spectrometric techniques, X-ray methods, and activation analysis. Atomic spectrometry, and especially wet chemical analysis via ICP-MS (trace elements) and ICP-optical emission spectrometry (ICP-OES, for major elements and specific trace ones) combined with different types of digestion, has become one of the most applied techniques for the elemental characterization of airborne PM in the last two decades. Moreover, the use of X-ray based techniques increased since the early 2000s, mainly with energy dispersive X-ray fluorescence (EDXRF) and particle-induced X-ray emission (PIXE),^[105] whereas the instrumental neutron activation analysis (INAA) showed a marked decline and nowadays its use is greatly reduced.^[106]

3.3.1. Atomic spectrometric techniques

Among the atomic spectrometry-based techniques, ICP-MS and ICP-OES are currently the most employed techniques for elemental analysis of PM. In ICP-MS, the spectral interferences are predictable and number less than 300, exhibiting also a very low background, typically below 10 counts/second, which does not pose a problem for elements detection. In ICP-OES, on the other hand, the spectral interferences are more numerous (> 50000 spectral lines documented) and, therefore, difficult to eliminate. The background in ICP-OES is also high, thus requiring an offline correction. Taken together, these interference problems contribute for the superior detection limits of ICP-MS compared to those of ICP-OES. The former instrumental approach provides LODs and limits of quantification (LOQs) in the order of ppb and ppt, and it is characterized by a wide linear dynamic range, multi-element capability, adequate precision and accuracy, rapid scanning and a high sample throughput, which are all excellent features if a trace or ultra-trace analysis is needed. Nonetheless, these traditional ICP methods require the dissolution of the PM sample prior to analysis, which typically means laborious, time-consuming, and quite expensive sample preparation procedures (e.g., high reagent consumption is sometimes needed). The conventional ICP analysis is indeed a destructive technique, which generally involves different pretreatment steps to guarantee a complete sample digestion. The most popular preparation method is the dissolution of filter-based samples in an acid extraction medium, with or without microwave assistance. Compared to conventional heating, the microwave-assisted digestion has some advantages, including the use of less chemicals, reduced time for sample dissolution, and lower losses of some volatile elements, such as Sb.^[107] Indeed, NIOSH and CEN standards recommend the microwave acid digestion as the reference method for a rapid and effective multi-element analysis in airborne dust, through ICP-MS or ICP-OES (EN 14902:2005; NIOSH:2014). Nevertheless, since a single reference-validated method for the digestion of various types of filters and for the simultaneous analysis of various elements is not available, several methodologies have been proposed in the literature, using different heating programs, digestion time, and working acid mixtures.^[108–112] Yet, this digestion step is a source of uncertainty as elements recovery is never equal to 100% (in this regard, EN 14902:2005 specify a recovery between 90 to 110% for Pb, Cd and 85 to 115% for As, Ni).

The choice of the acid mixture is a key issue since reagents must digest the PM sample as completely as possible, keeping elements stable in the final solution. Nitric acid (HNO₃) or conventional aqua regia (HNO₃ + HCl) can be useful for many elements and are often used in routine analysis, in combination with H₂O₂, as their use is able to destroy and oxide most of the organic compounds, ensuring recoveries generally > 90% for metals such as Pb and Cd. Nevertheless, this acid mixture cannot completely digest silicon-containing compounds and elements generally bonded to siliceous material [e.g., Al, Zr, Hf, Rb, and rare earth elements (REEs)], for which a combination of HF-HNO₃ digestion is mandatory for improving elements recovery, with or without the addition of various oxidative agents (e.g., HClO₄, H₂O₂ or H₃BO₃). HF can be employed in different percentages and some authors, in the past, used relatively large amounts of HF,^[107,113] regardless of HF being extremely hazardous, corrosive, difficult to work with, and it has to be evaporated before the introduction into the ICP equipment (the presence of HF, even in trace amounts in the analyzed sample, will create permanent deterioration of the quartz glassware of the ICP equipment). Therefore, if the silica content in airborne PM has to be measured, the use of HF should be minimized, as performed by Pekney and Davidson,^[114] who examined the use of trace amounts of HF for a complete digestion of PM samples. When using HF for digestion, B and Si are lost during evaporation of HF prior to ICP-OES analysis. Thus, Si might be determined in parallel using Teflon filters and XRF analysis or by implementing HF digestion but buffering the dissolutions with H₃BO₃ instead. Moreover, in addition to the effects of HF, the ICP-MS sensitivity suffers from some spectral interferences due to an overlap of molecular or polyatomic ions (mainly formed at plasma conditions or derived from sample matrix) at the same nominal mass as the analyte of interest.^[115,116] For example, the use of HClO₄ or HCl in the digestion process promotes the release of molecular ions, such as Cl⁺, ClO⁺, ArCl⁺, which may strongly interfere with the detection of V, Cr, Se, and As, yielding values higher than those actually present in the PM sample.^[117] To overcome this problem, some modifications to the conventional ICP-MS introduction system, such as the use of collision/reaction cell (CRC) technologies or HR-MS, need to be applied to significantly improve the removal of interfering ions and the detection of trace and ultra-trace interfered elements.^[118]

Direct solid-samples analysis techniques, such as the laser-ablation combined with ICP-MS (LA-ICP-MS), are getting increasing interest as an alternative method to determine elements directly in solid samples with minimal sample preparation. LA-ICP-MS analysis does not need laborious sample pretreatment steps, thus i) overcoming the time-consuming approaches required for wet chemical analysis, and ii) reducing the possibility of sample losses or contamination. It is well suited when the amount of PM material is limited (few µg) and, under the optimized operating parameters, it can provide very low detection limits, sometimes one order of magnitude lower than those obtained from conventional ICP-MS analysis.^[119,120] Nonetheless, three main problems may occur with LA-ICP-MS analysis: i) the difficulty in obtaining matrix-matched standards for quantitative analyses, which implies the design of in lab-made standards, as performed by Rovelli et al.;^[119] ii) the need of ensuring a homogeneous sample distribution throughout the filter surface; and iii) the transport efficiency of the ablated material from the ablation chamber to the MS detector, that has to be as complete as possible to ensure accurate and reliable results for quantitative analyses.

Compared to ICP (-MS or -OES), atomic absorption spectrometry (AAS) is a more traditional technique based on the atomization or ionization of a given element, and it can be differentiated into three variants, viz. the flame AAS (FAAS, flame up to 2600 °C), graphite furnace AAS (GFAAS, atomization temperature between 2000 and 3000 °C) , and cold vapor AAS (CDAAS, dedicated to Hg analysis without vaporization system).^[120] AAS devices are cheaper and easier to use than the ICP technologies, and they can be applied in combination with X-ray analysis for the quantitative detection of low-Z elements (Z < 12), such as Be, Na and Mg. FAAS is characterized by little interferences, although some refractory elements (e.g., B, W, Ta, Zr, As and Sn) cannot be quantitatively determined because of the too low flame temperature that does not induce a complete atomization. The FAAS technique only works with relatively large volume of liquid samples – this implies the need of a sample digestion step, as in ICP analysis – and it is generally used when single or few elements within a sample need to be quantified. Although far less applied than ICP into elemental analysis in indoor PM, the FAAS has already been applied to assess the contribution of severe outdoor pollution sources (i.e., wind-blown crustal dust and traffic emissions) in the concentration of some elements indoors (namely, Mg, Ca, Zn, Cd, Sb, Pb, Cr, Mn, Fe, Co and Ni).^[121,122] This sort of analysis was possible due to the high load of PM collected in the studied indoor sites. To the best of author’s knowledge, GFAAS has never been applied into trace element analysis in indoor PM. GFAAS exhibits lower LODs (10–100 times better) compared to FAAS (ppb vs sub-ppb for GFAAS and FAAS, respectively) and it requires relatively small amount of a liquid or solid sample, allowing also a single-element detection.^[123] Nevertheless, despite the higher atomization temperature of GFAAS, refractory elements are still difficult to be determined with an acceptable level of precision, and many spectral interferences cannot be overcome with a matrix or background correction.^[123]

3.3.2 X-ray methods

X-ray based methods, such as particle-induced X-ray emission (PIXE) and XRF, are nondestructive techniques that involve minimum sample manipulation, being able to directly analyze the airborne PM sample. Although PIXE exhibits LODs lower than those of XRF, it requires a more complex instrumentation, being therefore more expensive and difficult to use. XRF is a widespread and versatile technique for the quantification of elements with atomic number Z > 10.^[60] The quantification of elements is allowed considering the PM filter “thin” and thus, excluding matrix and secondary/tertiary fluorescence effects.^[60] In fact, in this way, a linear relation between the concentration of an element in the sample and the number of fluorescence photons per second is found. The instrument calibration, energy adjustment, and instrumental drift check have to be performed prior to analysis using pure Cu and thin film standards in order to establish the elemental sensitivity factors, as well as the magnitude of the interference or overlap coefficients.^[60] Thin film standards closely resemble the layer of particles on a filter, and they are of two kinds: thin films deposited on Nuclepore substrates and polymer films. The first ones are available for almost all the elements analyzed, ranging in atomic number from 11 (Na) to 82 (Pb) with deposit masses gravimetrically determined to ± 5%.^[60] The second ones consist of films containing known amounts of two elements in the form of organo-metallic compounds dissolved in a polymer. These standards are prepared for elements with atomic numbers above 21 (titanium and heavier).^[124]

It should be also mentioned the work of Lucarelli et al., who showed that PM_2.5 and PM_2.5–10 samples collected by means of a low-volume two-stage streaker sampler followed by PIXE or XRF analysis, can provide elemental analysis data with time resolution of 1 h.^[125,126] The advantage of using these types of streaker samplers is that measurements with PIXE can be performed by scanning the beam automatically over the multi-hours’ samples and in a relatively short time. Szigeti et al. also employed PIXE for the quantitative determination of Al, Si, S, Cl, K, Ca, Ti, Cr, Mn, Fe and Zn in indoor PM_2.5 samples (at ng m⁻³).^[127] The authors also reported that quartz fiber filters were less adequate than Teflon filters for the PIXE measurements due to the greater thickness and filamentary structure of the former.^[127]

Overall, and on the quest to select between XRF and ICP-MS or ICP-OES for multi-elemental analysis of ambient PM, one must consider that:

XRF is fast and cheaper, but XRF analysis is only possible to apply when PM samples are collected in Teflon filters, which requires the use of low-volume samplers. This analytical tool allows Si determination, but usually does not exhibit adequate LODs to determine As, Cd, Sb, Se, Sn and other tracers that might have relevance in source apportion analysis and health studies;ICP-MS or ICP-OES require sample digestion, is time consuming and it is more expensive, but the LODs are very good for more than 50 elements. Nonetheless, the major drawback is that digestion of quartz filters for ICP analysis does not allow determining Si, and indirect data must be obtained from theoretical Si/Al ratios.

3.4. Ion chromatography (IC) for water-soluble inorganic ions analysis in airborne PM

The inorganic ionic fraction [mainly secondary inorganic aerosol (SIA)] represents one of the major airborne PM components, after the carbonaceous fraction. It is mostly made up of SO₄²⁻, NO₃⁻ and NH₄⁺, usually present as (NH₄)₂SO₄ and NH₄NO₃ deriving from the interaction of H₂SO₄ and HNO₃with NH₃, and commonly found in air masses influenced by anthropogenic emissions.^[40] These compounds are mostly of outdoor origin and their presence in the indoor environment mainly occurs when outdoor PM infiltrates indoors.^[128,129]

Within the context of the EU Air Quality Directive 2008/50/EC, the European Committee for Standardization (CEN) released a procedure for the determination of inorganic ions in PM_2.5 samples (CEN/TR 16269:2011). This is not a standard method, but a CEN Technical Report that specifies a methodology for the measurement of water-soluble NO₃⁻, SO₄²⁻, Cl⁻, NH₄⁺, Na⁺, K⁺, Mg²⁺, and Ca²⁺ in PM_2.5, following a water extraction process and subsequent analysis of anions and cations by IC coupled to a conductivity detector. The CEN Technical Report also indicates that cations, excluding ammonium, can be alternatively analyzed by ICP-OES. Although the CEN Technical Report has been firstly proposed to provide guidance for the monitoring of inorganic ions at rural background locations, it may be considered equally applicable to all site types, including the indoor environment. Nevertheless, one should be aware that this procedure has not been validated in the field for these specific monitoring sites.

As described in the Technical Report, the PM sample preparation for IC analysis generally involves a 30-min water (or, occasionally, the IC eluent) extraction procedure in an ultrasonic bath, followed by a filtration through 0.45-µm PTFE (or acetate cellulose) membranes to remove the insoluble fraction. Nevertheless, Emma et al. assessed the possible influence of the extraction time on the recovery of water-soluble inorganic ions from ambient PM_2.5 samples, by performing an ultrasonication-assisted water extraction for 30 min, 1 h, 3 h and 6 h on different pieces of the same filter.^[130] The authors concluded that there was no significant trend attributable to the extraction time, and 30 min was enough to obtain a complete dissolution of all water-soluble cations and anions in PM_2.5 under the selected experimental conditions.^[130] Anions and cations can subsequently be separated by using suitable mobile and stationary phases (e.g., Na₂CO₃/NaHCO₃ or KOH and PDA/HNO₃ or tartaric acid/dipicolinic acid for anions and cations, respectively), in either isocratic or gradient elution conditions, with the latter providing better baseline separation between the analytes of interest in a lower time of analysis.^{[111,112,130]} Nevertheless, since no standard methodologies are actually available, simple, selective and sensitive methods have been developed – and are being developed – for the simultaneous determination of a range of water-soluble inorganic ions as well as selected organic acids in airborne PM. This implies the optimization of chromatographic conditions for a rapid and effective separation and a method validation (quality assurance procedures, i.e., linearity, sensitivity, recovery, accuracy and precision tests) using standard reference materials.^{[111,112,131,132]} The IC technique is able to guarantee: i) satisfactory LODs and LOQs (e.g., typically from few to tens of ng/m³ (4–40 ng/m³), even though LODs and LOQs are strictly dependent on the noise level of detector, the variability of blank laboratory values, and the concentration levels in the ambient air); ii) acceptable dynamic ranges (0.01–15 mg/L, depending on the investigated ionic species) and good linearity; iii) measurements uncertainties in the 5–10% range for most of the ions (NO₃⁻ and NH₄⁺ may have a larger uncertainty); and iv) a minimum sample manipulation prior to analysis, thus reducing the risk of contamination.^{[112,133–135]}

3.5. Spectroscopic methods for the analysis of fibrous material in indoor PM

3.5.1. Man-made vitreous fibers in indoor environments

Man-made vitreous fibers (MMVF) are widely used in the construction sector for thermal and acoustical insulation, and thus they can be found in office environments as major constituents of ceiling tiles. All the commercially available MMVF are mainly constituted by silica and alumina, with different proportions of alkaline and earth-alkaline oxides or Ti and Zr oxides. In Europe, MMVs are classified in two different categories based on the alkaline and alkali earth oxides content (Na₂O + K₂O + CaO + MgO + BaO) (1272/2008 Regulation, Annex VI). If higher than 18% by weight, one refers to “mineral wools”, otherwise to “refractory ceramic fibres” (RFCs). Rock wool, stone wool, slag wool, glass fibers are typical examples of mineral wools, while RFCs are typically referred to as aluminum silicate or calcio-aluminum silicate glasses. Respirable RCFs are classified in category 1B by EU (1272/2008 – CLP regulation), as substances presumed to have carcinogenic potential for humans, and in Group 2a by the International Agency for Research on Cancer (IARC). The respirable fibers composed by mineral wools, except for certain special-purpose glass wools, are classified in category 2 EU (suspected human carcinogens) and "not classifiable as to carcinogenicity in humans" (Group 3) by IARC. Non-respirable fibers of any kind (1272/2008 Regulation, Annex VI, Note R conditions) must be regarded as non-carcinogenic, in the case of the fulfillment of the length-weighted mean fiber diameter threshold determined by the official A22 method (Annex II, Directive 761/2009/EC). The human health hazard posed by MMVFs depends not only on their dimensions, and therefore respirability, but also on their durability (1272/2008 Regulation, Annex VI, Note Q conditions), assessable by toxicologists via short-term biopersistence or inhalation tests or suitable long-term inhalation tests.

The determination of MMVs concentrations in indoor air is substantially based on the classical methods developed and applied for asbestos measurements. Basically, a sampling of at least 4 filters is needed (ISO 16000-7:2007) to actively collect a known volume of air on membrane filters according to the chosen analytical technique, namely cellulose ester membranes for phase contrast optical microscopy (PCOM), and polycarbonate filters for scanning electron microscopy (SEM). Total suspended PM should be collected using asbestos cassettes to achieve homogeneous deposition, and the collected respirable fibers (i.e., those having length ≥ 5 µm; diameter ≤ 3 µm and aspect ratio ≥ 3) should be counted by PCOM or SEM, and eventually analyzed by energy-dispersive X-ray (SEM-EDS). To obtain reliable and specific results, with the possibility to discriminate between mineral wools and RCFs, ISO 14966:2019 indications are recommended (sampling on polycarbonate filters, and SEM-EDS analysis at low magnification − 1000 X). Using this method, fibers with diameters > 0.2 µm can be easily detected, while the ability to detect fibers with widths lower than 0.2 µm is limited. In such a case, which can happen for asbestos fibers but is not expected for the most common MMVFs, a transmission electron microscopy (TEM) method, such as ISO 10312:2019, can be used to analyze these smaller fibers. However, it is worth noting that SEM analysis is characterized by longer analysis times and higher costs (up to one order of magnitude) than the nonspecific PCOM analyses, which could then be regarded only as screening information on the total airborne fiber contamination of indoor air.

Usually, low airborne levels of respirable MMVFs are expected indoor, but in some specific cases a particular attention should be paid to the potentially high emissions from friable and accessible building materials, such as sprayed insulations, as well as on installation, maintenance or removal interventions on MMVF-containing materials. By way of example, in the case of modernized office buildings with false ceilings made by panels containing MMVFs, indoor concentrations of total airborne MMVFs reached 3500 ff/m³,^[136] which was linked with the high prevalence of self-reported symptoms mainly referable to skin and eye irritation. Moreover, up to 6230 respirable ff/m³ were measured in premises with MMVFs sprayed insulations^[137], and up to 15000 ff/m³ just after the installation of MMVFs insulation products.^[138]

There are some literature indications suggesting that health complaints due to MMVFs correlate more with surface contamination than airborne concentrations,^[139] nevertheless air sampling remains the gold standard for risk assessment of airborne fibrous PM. The long-term contamination of indoor air by MMVFs can be indirectly assessed by surface sampling according to ISO 16000-27:2014. This allows the quantification of fibrous structures with 0.2 μm diameter or greater in settled dusts, and their classification according to specific groups of substances (e.g., asbestos or other inorganic fibers) through sampling with aluminum, copper or carbon conductive tapes, and subsequent SEM-EDS analysis. Results are given as number of MMVFs per area unit (e.g., cm²). Surface samples can be collected from dry areas considering the cleaning frequency; one possible approach is to perform the sampling on frequently cleaned surfaces (desks and low shelves) vs. seldom cleaned surfaces (e.g., the tops of bookcases, high shelves and top doorposts) to obtain information about cleaning effectiveness. The use of standardized adhesive tapes has been applied as a standard collection technique for surface particles since the 1920s and made into some national and international standards in the last decades (ASTM E1216-11: 2016). In the literature, stereomicroscopic analysis at low magnification (100X) was used as a low-cost alternative to SEM analysis for the counting of coarse MMVFs (aspect ratio ≥3) on the tape lifts used as surface sampling media for desks, shelves and supply air ducts. The detection limit depends on the selected sampling and analytical methods; in the case of low magnification analyses it can be about 0.1 MMVFs cm⁻².^[140]

3.5.2. Microplastics in indoor environments

Microplastics refer to plastics with a size of < 5 mm that originate from primary sources (e.g., plastic beads and synthetic fibers deliberately manufactured in microscopic size) and secondary sources (e.g., environmental degradation of large-sized plastic pieces that yield microplastics).^[141] The most commonly manufactured plastics are polypropylene (PP), polyethylene (PE), polyethylene terephthalate (PET), polystyrene (PS), polyurethane (PUR), polyvinyl chloride (PVC), and polycarbonate (PC). The major sources of airborne microplastics are thought to be synthetic clothing, furniture, building materials, cleaning habits, abrasion of rubber tires, and city dust.^[142] Nonetheless, the number of studies evaluating microplastics concentration indoors is very limited.^[142–144]

Based on the review of Mbachu et al.,^[145] indoor deposited particles have been sampled by the pan and brush method from bedrooms and living room floors, or by vacuum cleaning. In the case of indoor suspended dust sampling, a filtration device is typically employed. In this regard, Dris et al.^[142] used a stand-alone sampling pump that absorbs dust through a filter paper, while Vianello et al.^[144] used a breathing thermal manikin for human breathing rate simulation. The suspended indoor air sampler was placed at different height to enable accurate evaluation of human exposure to microplastics, even at heights close to the average breathing height of an individual (1.2 m).^[142]

After sampling, it is necessary to remove extraneous and biogenic materials that could interfere with the identification of microplastics in the samples, but also to eliminate any inorganic contaminants with similar composition to the analyte of interest.^[145] Therefore, the preparation of samples for microplastic analysis consists of several stages: sieving, organic matter digestion, density separation, filtration, and air-drying.^[145–150] Regarding air samples, the wet peroxidation method was proved to be efficient in digesting organic matter. However, not all samples undergo an oxidation process, thus organic matter decomposition is performed for samples with organic matter content large enough to cause interference during microplastics identification. In the works of Abbasi et al.^[151,152] and Dehghani et al.,^[153] subsamples of between 10 and 30 g of composite bulk dust samples were treated with 35–100 ml of 30% H₂O₂ to remove the organic matter. Likewise, a solution of H₂O₂ with a FeSO₄ catalyst was recommended as an alternative method to breakdown organic matter within a shorter period.^[154] As described in the work of Mbachu et al.,^[145] common inorganic salt solutions used in the density separation of atmospheric microplastics include sodium iodide (NaI) and zinc chloride (ZnCl₂) due to their ability to produce solutions with densities within the range of 1.6–1.8 g/cm³, which is an optimum density required for the separation of microplastics from mineral particles.

The analysis of microplastics in the air is challenging as no standard technique exist for their identification. Of the current analytical techniques employed for microplastic identification, techniques based on physical and chemical properties are usually applied. Regarding the former approach, the analysis is based on physical characteristics such as elasticity, hardness, color, shininess and structure, which are identified by visual (i.e., naked eye, stereomicroscopy) or optical microscopy (i.e., fluorescence microscopy, binocular microscopy and polarized light microscopy) techniques.^{[146,151,152]} After their detection, the identified microplastics are quantified to determine their abundance in the environmental matrix. This is performed by counting the identified microplastics using binocular microscopy techniques with up to × 200 magnification.^[151–153] Usually, physical characterization methods are used as a prerequisite for subsequent chemical analysis. This is because microplastics extracted from various environmental matrixes may suffer interferences from different materials such as paper, wood, and vegetation, which exhibit characteristics similar to those of microplastics. ^[155] In such cases, chemical analysis based on particle composition is required. The main analytical tools for the chemical characterization of microplastics include Fourier transform infrared (FTIR) spectroscopy, micro-FTIR spectroscopy attenuated total reflectance (ATR)-FTIR spectroscopy, focal plane array (FPA)-FTIR spectroscopy, and scanning electron microscopy (SEM) techniques (for additional details see the review work of Mbachu et al.,^[145] and references therein).

3.6. Offline analysis of airborne PM samples using AMS

Field deployments of the AMS have significantly advanced real-time measurements and source apportionment of non-refractory ambient PM. However, the cost and complex maintenance requirements of the AMS make its deployment at many sites impractical. Therefore, many studies have employed the HR-TOF-AMS technique for the analysis of the aqueous and organic solvent extracts of PM samples collected on filters (e.g., see the works of Mihara and Mochida^[156] and Chen et al.^[157] for additional details). Although this offline analytical approach may greatly extend the ability of the AMS to collect spatially resolved long-term data sets, the results obtained are subject to inherent artifacts of filter-based measurements. Notwithstanding this experimental circumstance, the HR-TOF-AMS has been mostly used for the chemical characterization of the aqueous extracts of filter samples, particularly of water-soluble OA. PMF or non-negative matrix factorization (NMF) of HR-TOF-AMS organic mass fragments has been used to identify the different factors describing atmospheric OA, including factors for fresh biomass burning OA, and different types of oxygenated OA, as well as for OA source apportionment (e.g., Chen et al.^[157] and Brege et al.^[158]). Despite the different time resolution, it has been also demonstrated that online and offline AMS source apportionment yields comparable results for the different OA sources,^[159] which increases the spatial coverage accessible to AMS measurements, allowing for PM samples being routinely collected at many different locations.

3.7. Other advanced multidimensional analytical tools for aerosol WSOC analysis

Knowledge on indoor aerosol WSOC component is still limited to their quantification and contribution to air particles mass load.^{[68–70,160]} Based on the current knowledge on outdoor aerosol WSOC component, it is well recognized that the task of determining the detailed composition of this WSOC fraction is daunting. However, it offers unparalleled rewards since determining the molecular composition and structure of aerosol WSOC is warranted to increase the current understanding of its role in various atmospheric processes and on human health.^[161–163] Hence, the most promising analytical techniques in terms of molecular and structural elucidation include one- and two-dimensional nuclear magnetic resonance (1 D and 2 D NMR) spectroscopy^{[16,19,61,65–67]} and HR-MS [e.g., Fourier transform ion cyclotron resonance mass spectrometry (FTICR-MS)] (e.g., Willoughby et al.,^[164] Bao et al.,^[165] and Tang et al.^[166]). While 1 D and 2 D NMR spectroscopy is particularly suitable for the untargeted analysis of the complex aerosol WSOC fraction, the HR-MS allows for a more reliable target analysis with reference standards as well as a screening for unknowns within the WSOC matrix. Solution-state ¹H NMR spectroscopy has been the most usable and useful 1 D NMR technique for the past 18 years, by providing generic structural assignments and average parameters (e.g., aliphatic, oxygenated aliphatic, aromatic, and/or carboxylic acid content) of WSOC samples.^[19] Although the use of solution-state ¹H NMR has become a routine step for the structural profiling of aerosol WSOC, the future directions for solution-state NMR applications are clearly toward 2 D NMR, using ¹H-¹H and ¹H-¹³C correlation experiments.^[12,19] Nevertheless, the lack of analytical expertise in the use of 2 D NMR techniques, as well as handling and interpretation of 2 D NMR data are major challenges that once solved, will have a far-reaching impact in the study of aerosol WSOC. On the other hand, a reliable identification of aerosol WSOC constituents using HR-MS requires both high resolving power and high mass spectral accuracy to increase selectivity against the WSOC matrix background, as well as for a correct molecular formula assignment to unknown compounds. According to Krauss et al.,^[167] the identification and structure elucidation of unknown compounds in environmental matrices, within a reasonable time frame and with a reasonable soundness, requires the use of both advanced automated software solutions and improved prediction systems for theoretical fragmentation patterns, retention times, and ionization behavior.

As for the case of indoor aerosol WSOC, exploring the chemical composition of this complex mixture will require the use of complementary state-of-the-art analytical tools, such as HR-MS, 2 D NMR, and online or offline coupling between multidimensional LC and different detection systems (e.g., diode array detector, fluorescence detector, NMR, or HR-MS), assisted by chemometric tools [e.g., principal component analysis (PCA) or multivariate curve resolution-alternating least squares (MCR-ALS) methods].^[16] The combined use of multiple structural characterization techniques may prove indispensable to molecular-level identification of indoor aerosol WSOC. The structural information gleaned from the untargeted profiling of this indoor aerosol carbonaceous fraction will allow suggesting fruitful directions for resolving issues of aerosol WSOC effects on indoor air quality and human health.

4. Conclusions

Deciphering the chemical composition of indoor PM and the chemical processes involving these particles includes as many unknowns as nearly as many challenges. The complexity of compositional patterns and size distribution of indoor PM is very high due to multiple emission sources of both PM and gaseous precursors, secondary reactions (such as the formation of new ultrafine particles and oxidized volatile organic compounds), and the gas/particle partitioning equilibriums. Accordingly, state of the art studies on indoor PM, emission factors, infiltration, and secondary reactions require the use of a diversity of sophisticated analytical tools capable of providing high-time resolution and information on a wide range of organic and inorganic species both in the gaseous and solid/liquid states, from the nucleation to the coarse modes, and not focusing only on specific PM or gaseous components. The different online and offline analytical techniques presented in this review illustrate the tremendous importance for an in-depth knowledge on the identity of indoor PM constituents. The considerable progress accomplished in the analytical tools available have been fueled by a diversity of field and laboratory studies carried out by the atmospheric outdoor community. Nonetheless, the challenges of assessing indoor PM concentration and composition, requires not only the use of the existing techniques, but also their adaptation to indoor applications or the development of new tools. For example, a 12 h sampling of PM mass in not always suitable for indoor concentration estimation due to low mass loading. Moreover, not all the instrumentation is suitable for in situ indoor PM studies due to their noise and the space need for their installation. Additional concerns that may hamper the straightforward application of the avant-garde analytical tools discussed in this review includes the lack of analytical expertise in the use of these techniques, as well as the handling and interpretation of the voluminous and complex data sets. Avoiding these shortcomings require interdisciplinary collaborations between outdoor and indoor researchers, with different backgrounds to unravel the complex composition and size distribution of indoor PM. Moreover, the adaptation of existing real-time or offline analytical tools to microenvironments promises to have major impacts on the current understanding of indoor PM composition. Furthermore, epidemiological and toxicological studies combined with full characterization of PM and gas pollutants in indoor air would decisively contribute to constrain the species of environmental concern.

Acknowledgments

Thanks are due to the European Cooperation in Science and Technology (COST) through the financial support to COST Action INDoor AIR POLLution NETwork (INDAIRPOLLNET). One of the authors (Regina M. B. O. Duarte) also wishes to acknowledge FCT/MCTES for the financial support to CESAM (UIDP/50017/2020 + UIDB/50017/2020) through national funds, as well as to the Exploratory Research Project (IF/00798/2015/CP1302/CT0015, Investigator FCT Contract IF/00798/2015).