Sampling and analysis techniques for inorganic air pollutants in indoor air

Abstract

Effective evidence-based actions to reduce indoor air pollution must be derived from data obtained using accurate means for assessment of the critical air pollutants and for the identification of their sources. This is of paramount importance to provide robust evidence for establishing effective policies or preventive measures. Nevertheless, designing a reliable sampling plan for assessing concentration of inorganic pollutant in the indoor air requires expertise in conducting rigorous sampling campaigns and proper knowledge on the existing standards and methodologies for assessing concentration of the target substances. Therefore, this review focuses on the relevant information and recommendations that should be considered when designing a sampling plan to collect complementary data on the indoor environment under study, to properly define the criteria for establishing the details for the work and to ensure the quality of the assessments. In particular, comprehensive information on the most commonly used methodologies for the determination of a list of critical inorganic pollutants for indoor air quality monitoring has been compiled. Thus, inorganic gaseous pollutants such as CO₂, CO, O₃, NO₂, NO, NH₃, SO₂, H₂O₂, H₂S, HNO₃ and HNCO are included in the present review.

Keywords:

• indoor air
• inorganic pollutants
• quality assurance and control
• sampling plan
• analytical methods

1. Introduction

According to the World Health Organization (WHO), air pollution is one of the greatest environmental risks to health.^[1] Although most of the attention has traditionally been focused on outdoor air pollution, humans spend most of their time inside built environments, where they may be exposed to hazardous pollutants present in the indoor air. Air pollutants can be a mixture of both outdoor air infiltration and emissions from indoor sources (such as building materials, consumer products, indoor activities, etc.).^[2–5] Indoor air pollutants include particulate matter, biological pollutants and over 400 gaseous chemical pollutants including organic and inorganic compounds, whose concentrations are governed by multiple outdoor and indoor factors.^[6] The chemical mechanisms occurring indoors are very diverse and often poorly understood and characterized. In addition, most regulations on air pollutant levels reference ambient outdoor air, setting limit values for several compounds.

The most studied inorganic substances in the indoor air are carbon dioxide (CO₂), carbon monoxide (CO), nitrogen oxides (NO_x) and ozone (O₃).^[6] However, studies that assessed indoor air pollutant concentrations have provided evidence that other inorganic compounds such as ammonia (NH₃), sulfur dioxide (SO₂), hydrogen peroxide (H₂O₂), nitric acid (HNO₃), isocyanic acid (HNCO), and hydrogen sulfide (H₂S) are also important inorganic pollutants found in indoor air.^[7–12] The sources of inorganic compounds in indoor air are diverse. Carbon dioxide is a bioeffluent resulting mainly from the breath of occupants^[13] and along with nitrogen oxides (NO_x) and carbon monoxide (CO) can also be produced from combustion processes like candle burning or smoking, wood-burning stove or gas water heater or stove.^[14–16] Concentrations of CO₂ are also monitored as a proxy to assess the effectiveness of fresh air.^[17] Buildings with high CO₂ concentration levels due to insufficient ventilation rates present ideal conditions to promote accumulation of indoor pollutants. For instance, because CO₂ concentration levels are indicative of the amount of the air exhaled by occupants that is accumulating in the indoor air, in the context of COVID-19 pandemic, CO₂ has been also recognized as an useful indicator of risk of transmission of SARS-CoV-2 and other airborne infectious diseases indoors.^[18–20] Some office equipment such as laser printers, copiers and electrostatic air cleaners produce O₃.^[21] As well as being generated indoors, inorganic pollutants as O₃ and NO_x can also come from outdoors, infiltrating the building envelope, compromising air quality. Ammonia can be emitted by building occupants or from cleaning products.^[7] The sources of other inorganic acids such as HNO₃ and SO₂ include outdoor to indoor transport^[7] and other acids such as HNCO have been identified as products from cigarette smoke.^[9] Hydrogen sulfide has a nuisance odor and most H₂S detectors also measure mercaptans.

The methods employed to assess inorganic pollutants in the indoor air are diverse, and designing the best Indoor Air Quality (IAQ) assessment plan for inorganic pollutants requires specialist expertise. Unfortunately, there is little comprehensive robust, evidence-based guidance for designing an assessment plan for inorganics in indoor air in the literature.

This review presents guidance for best practice when planning a monitoring campaign aiming at assessment of IAQ with general considerations depending on the sampling objective and considering the general procedures for a quality assurance and quality control program. Another objective of the present study is to revisit the techniques and methodologies used to sample and analyze the important toxic inorganic compounds in the indoor air, considering technical characteristics of instrumentation, cost, limit of detection (LOD) and quantification (LOQ), and the advantages and disadvantages of each method are included with the aim of helping to choose the most appropriate technique matching the selected sampling strategy. This review specifically excludes two low-cost technologies: metal oxides and amperometric electrochemical gas sensors. A method based on electrochemical cells for monitoring CO is included because they are the ubiquitous technology for this gaseous contaminant, but otherwise low-cost sensors are not included because their validity in terms of uncertainty is still being discussed in the literature and by standards committees.

2. Method

This work is part of the activities conducted in the scope of INDAIRPOLLNET (INDoor AIR POLLution NETwork; Cost action: CA17136) that aims to advance significantly the field of indoor air pollution science. In particular, the project intends to highlight future research areas and to bridge the gap between research and business in order to identify appropriate mitigation strategies that optimize IAQ.^[22] Recently, working group (WG) 3 (Which should be the key species we aim to measure indoors in the future and why?) formed within INDAIRPOLLNET has identified a list of pollutants, including inorganic chemicals, namely CO₂, CO, O₃, NO, NO₂, NH₃, SO₂, H₂O₂, H₂S, HNO₃, HNCO, H₂SO₄, HF, HCN and CS₂, as being relevant to be assessed in the indoor environment. Nevertheless, indoor studies of some substances, such as H₂SO₄, HF, HCN and CS₂, are limited and their occurrences were associated to very specific events (e.g., malodor for CS₂), justifying why these compounds will not be considered in this review. In addition, other inorganic compounds classified as free radicals or their precursors (including OH·, HO₂, HONO, RO₂, NO₃·, N₂O₅, HCl, Cl, Cl₂, ClNO₂, HOCl, ClO and OClO) will not be also explored here as they are a topic of study from a different working subgroup. Table 1 shows the list of inorganic substances that have been selected for this study, together with the respective CAS number, sources, reported indoor concentrations and guideline values. These pollutants have been selected due to their occurrence in indoor spaces and being of concern in terms of human health. The techniques described in this paper include the most common methods used in the literature for determining the concentrations of the selected inorganic pollutants in the air. In fact, the objective is not to provide an exhaustive list with all existing methods, but rather to identify which well-established methods have been commonly used or have potential to be employed in indoor environments. Many of these procedures are described: i) in international standards such as the standard methods of analysis in indoor air according to International Organization for Standardization (ISO) or the European Committee for Standardization (CEN), and/or ii) in the existing literature including some advanced techniques to monitor the pollutants reported in publications of peer reviewed journals. Some methods from National Institute for Occupational Safety and Health (NIOHS) and from American Society for Testing and Materials (ASTM) have also been considered.

Table 1. List of inorganic compounds identified in indoor air, their sources, reported indoor concentrations and indoor guideline values.

No.	Pollutants	CAS number	Sources	Indoor concentrations	Guideline Values
Carbon oxides
1	Carbon dioxide (CO2)	124-38-9	Human breath, combustion sources^[^13^,^23^]	269 − 4960 ppmv (schools)^[^24–26^] 294 − 10000 ppmv (residences)^[^23^] 607 − 2249 ppmv (infant bedrooms)^[^27^]	It depends on building type and occupancy^[^28–31^]
2	Carbon monoxide (CO)	630-08-0	Tobacco smoke, wood and gas stoves, kerosene heaters, malfunctioning gas appliances, car or truck exhaust from attached garages^[^32^]	0.4 − 33.0 mg m^-3 [14]^a LOD − 1.6 mg m^-3 (infant bedrooms)^[^27^] LOD − 4.5 mg m^-3 (schools)^[^26^]	100 mg m^-3 (15 min)^b / 35 mg m^-3 (1 h)^b10 mg m^-3 (8 h) / 7 mg m^-3 (24 h)^[^14^] 11.5 mg m^-3 (24 h) 28.6 mg m^-3 (1 h)^[^33^]
Reactive oxygen species
3	Ozone (O3)	10028-15-6	Household air cleaners generating O3, soldering or welding-related hobbies, photocopy machines, outdoor air^[^32^]	0.8 − 114 μg m^-3 (as average, schools)^[^34^] 0.0 − 96.8 μg m^-3 (offices)^[^34^] 0.0 − 32.0 μg m^-3 (infant bedrooms)^[^27^]	100 μg m^-3 (8 h)^[^35^] 0.07 ppmv (8 h)^[^36^] 40 μg m^-3 (8 h)^[^33^]
4	Hydrogen peroxide (H2O2)	7722-84-1	O3-initiated indoor chemistry, air cleaners, cleaning products^[^37^]	–	–
N-containing compounds
5	Nitric oxide (NO)	10102-43-9	Fuel combustion, kerosene and gas stoves, malfunctioning gas appliances^[^32^]	4.9 ± 3.1 μg m^-3 (home with gas appliances)^[^38^] 39.4 μg m^-3 (museum)^[^39^] 25.0 μg m^-3 (office)^[^40^]	–
6	Nitrogen dioxide (NO2)	10102-44-0	Fuel combustion, kerosene and gas stoves, malfunctioning gas appliances, tobacco smoke^[^32^]	13 − 62 μg m^-3 [14]^c 8.5 − 65.8 μg m^-3 (homes)^[^7^] LOD − 88 μg m^-3 (schools)^[^24^] 26.1 μg m^-3 (6 − 68.5μg m^-3) (schools)^d 23 μg m^-3 (3.4 − 56.5 μg m^-3) (offices)^d	200 μg m^-3 (1 h)^[^14^] 40 μg m^-3 (1 year)^[^14^] 20 μg m^-3 (24 h)^[^33^] 170 μg m^-3 (1 h)^[^33^]
7	Ammonia (NH3)	7664-41-7	Building materials, cleaning products, human emissions, tobacco smoke, outdoor air^[^7^]	5.7 − 43 μg m^-3 (homes)^[^7^]^e	200 µg m^-3 [41]
8	Nitric acid (HNO3)	7697-37-2	Complex reactions from nitrogen oxides^[^7^]	< 1.5 μg m^-3^e	–
9	Isocyanic acid (HNCO)	75-13-8	biomass burning, diesel and gasoline vehicles, industrial processes, tobacco smoke^[^9^]	0.15 ppbv (<LOD − 0.55 ppbv; residence)^[^9^]	–
S-containing compounds
10	Sulfur dioxide (SO2)	7446-09-5	Coal combustion, kerosene^[^32^]	1.2 − 21.9 μg m^-3 (homes)^[^7^]^e	500 μg m^-3 (10 min) 125 μg m^-3 (24 h) 50 μg m^-3 (annual) ^[^42^]
11	Hydrogen sulfide (H2S)	7783-06-4	Human bioeffluents, crude Petroleum, natural gas, volcanic gases, hot springs, groundwater, industry^[^43^,^44^]	86 − 600 pptv (museums)^[^45^] < 60 ppbv (households)^[^10^]f 2.5 − 6.3 ppbv (schools)^[^8^]	150 μg m^-3 (24 h) 7 μg m^-3 (30 h odor annoyance)^[^42^]

^a Data varies greatly between studies and in the absence of indoor sources, concentrations of CO in indoor air in European and North American cities are well below the levels of existing air quality guidelines and standards.^[14]

^b Exceedance to this level should not occur more than once per day.

^c Maximum levels associated with the use of gas appliances in European homes are in the range 180–2500 μg m⁻³.^[14]

^d Median concentrations and range of average concentrations in brackets.^[46]

^e Median concentration in different indoor environments.^[7]

^f Median concentration but corresponds also to the limit of detection.

General considerations defined in ISO 16000-1^[47] can be applied to most circumstances. Nevertheless, due to the large variety of situations and pollutants in the different indoor environments, the specificities of the building under investigation and the aim of the investigation should be considered in the study design.

This paper presents general considerations on the complete process of designing an assessment plan for inorganic gases in indoor air, ensuring that the following topics are considered:

• Defining the objective(s) of the measurements and the target pollutants;

• Collecting information related to building characteristics, location and potential emission sources;

• Selecting the most suitable procedures for assessing the selected pollutants;

• Selecting the sampling sites within the building;

• Selecting the correct sampling device location within a selected room;

• Determining the sampling timing, duration and frequency;

• Assessing the outdoor surrounding the buildings.

Noteworthy, the criteria for establishing the details of a sampling plan presented in this review are in general not restricted to inorganic pollutants. For instance, although the paper is focused on the methods for assessing a list of critical inorganic pollutants, most of recommendations that are provided can also be applicable for the assessment of other groups of pollutants (e.g., volatile organic compounds and particulate matter) in the indoor air.

3. Establishing the objective(s) of the assessments

It is of utmost importance to clearly define the purpose(s) of the IAQ assessment plan before starting any sampling work in order to select the procedures that are more suitable to accomplish the defined objectives. Possible objectives for conducting sampling works include:

• To assess the levels of exposure and estimate health risks;

• To evaluate the compliance of airborne concentrations with existing guidelines and standards;

• To identify and characterize emission sources, defining their spatial and temporal trends;

• To study the cause of complaints from the occupants- sick building syndrome (including the identification of emissions sources);

• To identify risk factors to guide the establishment of recommendation on risk reduction measures;

• To assess the effectiveness of an intervention adopted in reducing pollutant concentrations.

• To establish an IAQ baseline before refurbishment or rebuild.

For instance, the specifications of the study design for a sampling plan should be adjusted in accordance with the defined objective(s) and the nature of the substances that are targeted for analysis. As an example, if acute effects on health (eye irritation, headache, bad odor perception, etc.) is the aim, it is best to determine the maximum concentration of the pollutants (short-term monitoring, preferentially in expected peak concentration periods) in a room, while in the case of chronic effects, the average concentration of the pollutants by sampling over long periods would be more appropriate^[48]. For the determination of the average concentration of a pollutant over long periods (from 8 hours to many days) passive or diffusive samplers can be used; however, for short-term monitoring, typically up to 1 hour, an active sampling approach is more appropriate. When the investigation aims to characterize an emission source, understand chemical processes, or collect information on concentration evolution throughout a time period, then real time monitoring based-strategies are the best choice.

Some of the most commonly reported sampling plan objectives and information on recommended study design specifications are summarized in Table 2. The sampling approach in each case was included.

Table 2. Common objectives for an indoor air quality assessment plan and different aspects to be considered in defining the most appropriate strategy for sampling (adapted from Seifert et al.^[49]).

Sampling objective	Sampling condition	Sampling timing and frequency	Sampling duration	Minimum number of samplers	Minimum number of sampling points	Sampling method
Assessing exposure and health risks
Average exposure concentration	Real-life conditions	Two seasons (winter and summer)	Long-term	2–4	1–2^b	Passive or active sampling
Maximum exposure concentration (peak of exposure)	Worst case	Any time	Short-term	1–2	1–2	Active sampling (continuous monitoring)
Studying the sources of symptoms and complaints
Chronic effects	Worst case	Any time^a	Long-term	1–2	1–2	Active or passive sampling
Acute effects	Worst case	Any time^a	Short-term	1–2	1–3	Active sampling (continuous monitoring)
Investigating the effectiveness of a remedy adopted	Real-life conditions	Before and after	Short/long-term	1–2	1	Active or passive sampling
Testing the compliance with indoor air quality guidelines or standards	Real-life conditions/Worse case	Any time	Period of reference value	2–3	1	Active (continuous monitoring) or passive sampling
Characterizing emission source(s) (e.g., cleaning activity, cooking etc.)	Real-life conditions	Before, during and after activity	Short-term	1	1	Continuous monitoring

^a According to activities.

^b Personal sampling.

4. Building survey: Information on building characteristics, location and potential sources of emission

An important step of an IAQ assessment plan is the development of a questionnaire to gather relevant complementary qualitative and quantitative data in accordance with the objectives of the evaluation. Recommendations about the information to be collected during indoor air sampling are available in ISO 16000-1^[47], ASTM D7297^[50] and the EPA Protocol for Office Buildings.^[51] The questionnaire can be prepared in a checklist format and usually includes questions about building location and its characteristics (year of construction and renovation, typology, use, etc.), the existence of potential sources of air pollution including outdoor (as traffic, industrial and agricultural-related sources) and indoor sources (e.g., materials, consumer products as cleaning products, equipment), human activities, HVAC system, ventilation conditions and possibility to open windows.^[52,53] The questionnaire should also include specific information for the characteristics of the spaces where the sampling will be carried out. The questionnaire can be filled in onsite during the audit based on walkthrough inspection and may require consultation with building managers/users.

Different types of questionnaires have been developed and published in the scope of international projects such as SINPHONIE project (School Indoor Pollution and Health Observatory Network in Europe)^[24] or OFFICAIR project (On the reduction of health effect from combined exposure to indoor air pollutants in modern offices)^[54].

5. Methods for the assessment of inorganic pollutants in indoor air

The available options for determining concentrations of inorganic pollutants in the air include both offline and on-line monitoring methods. Regarding offline sampling, different approaches can be employed to sample the compounds of interest: active, i.e., using a pump to collect the samples and passive, i.e., by diffusion.

Passive samplers are based on molecular diffusion, where the materials used as the collecting medium can be a variety of adsorbents:

• material is impregnated with reagents that interact with the target contaminant(s), leading to the formation of a quantifiable product. In this case, pollutants are measured indirectly since they form a product that is later analyzed;

• material that does not react with the contaminant, e.g., activated carbon. In this case the contaminant is adsorbed onto the surface of a high surface area material.

The difference between available passive samplers is the material used as collecting support and the reagents, most of which are the same as those used in active sampling. Nevertheless, the geometry of the available passive samplers can differ and influence the diffusive uptake rate of the pollutant, which must be calculated by the manufacturer in a standard atmosphere. This type of sampler is cheap and noiseless, does not require power supply and is easy to install. This sampling approach can be used to simultaneously determine the mean concentration of the target compound(s) in several places over long periods. Passive samplers can be specific, i.e., designed to measure a compound or a restricted number of compounds, or nonspecific, i.e., covering a wider set of compounds. Depending on the geometry, the main types used to collect inorganic compounds are:

• Filter type (e.g., Ogawa, Figure S1): based on impregnation of the material with a suitable solution like nitrite ion or triethanolamine to react with the contaminant to quantify, e.g., for sampling O₃ and NO₂, where the product of reaction is determined by ion chromatography (IC) or by spectrophotometry.

• Tube type (e.g., Palmes, Figure S2): based on diffusion of the molecules along an inert tube toward an absorbent. They are used to collect NO₂, CO₂, sulfur dioxide (SO₂), and O₃ mainly. Spectrophotometers or IC are used to quantify their concentrations.

• Radial type (e.g., Radiello®, Figure S3): its radial geometry allows higher diffusion rates, providing a great sensitivity to subsequently determine the concentration of inorganic compounds such us O₃, NO₂, H₂S, SO₂, HF, HCl, NH₃, etc. In this passive sampler the adsorbing material is contained in a coaxial cylindrical cartridge. The analysis is carried out by means of spectrophotometry or IC.

Passive sampling is suitable for daily and weekly sampling, providing average values for long-term monitoring but not for high frequency sampling.

In turn, active sampling approach requires a system including a pump to suck the air and force it to pass through the sampler. Active sampling requires access to a power supply (pumps) and some degree of expertise, but the cost is relatively low. As stated above, most of the collecting supports and the reagents are the same as those used in passive systems. Sampling time is short-term, typically sampling for a few hours.

Both active and passive samples require analysis in the laboratory for their quantification. Actually, offline sampling can be quite labor-intensive and time-consuming for sample collection and laboratory analysis procedures. In addition, conventional offline methods lack high temporal resolution and cannot capture rapid changes of various chemical species in the indoor air.

Regarding the online instruments used to measure and analyze the compounds of interest, there is a variety of analysis techniques ranging from optical to chromatographic and also dedicated monitors and sensors. These techniques have greatly improved and can measure a higher number of compounds with better reliability. They need a power supply, and, in general, they require expertise to operate and to analyze the output data. This kind of equipment can sample at high frequency, typically with one minute logging intervals and, depending on the technique, they can cover specific compounds, as in the case of the monitors or systems using lasers as the light source, and also a wide range of compounds, as in the case of techniques based on mass spectrometry. Spectroscopy-based methods, which monitor the levels of the compound directly, i.e., without sample handling, are less prone to be affected by interferences because they are based on first principles (e.g., Lambert-Beer Law and known absorption cross sections), while mass spectrometry techniques can measure a wide variety of compounds with good sensitivity, but they are bulky and difficult to deploy on-site. In general, real-time on-line monitors can obtain high time-resolved data, offering better understanding of pollutants’ formation, sources, and transport, especially for those episodes that exhibit relevant temporal variations.

The following sections summarize the most common methods employed for sampling and analysis of each inorganic chemical pollutant listed in the Table 1. Some techniques have only been used in outdoor air, although they can be used also for sampling indoor air, with the caveat that equipment employed is suitable for the measurement purposes and does not adversely affect the normal use of the rooms where they are used due to their size, sampling rate and noise. This is especially relevant in works conducted in residences, schools or other socially sensitive environments. The instrument should be low noise, and the sampling rate should not interfere with the normal ventilation/air circulation patterns of the room.

5.1. Carbon oxides

5.1.1. Carbon dioxide (CO₂)

Carbon dioxide has been widely assessed as an important indicator of IAQ, in particular as a proxy for ventilation conditions with respect to fresh air supply, number of occupants and occupants’ metabolic activity. In the absence of combustion sources, the main source of CO₂ indoors is, indeed, the occupants. In this regard, the ASTM provides a standard guide for using indoor CO₂ concentrations to evaluate IAQ and ventilation: D 6245^[28], currently under revision. To ensure adequate ventilation rates, some international standards and guidelines have established threshold levels for CO₂, as a tracer gas for ensuring fresh air supply of at least 7.5 L/s/person. The correct CO₂ level depends on the number, activity and profiles of the occupants, as well as the volume in the space and type of ventilation system.^[28–31] Depending on the total environment, these threshold limits generally range from 700 to 1500 ppmv, with the common assumption that concentrations above 1000 ppmv are likely to indicate insufficient fresh air.^{[23,28,55,56]} Generally high concentrations of CO₂ (above 1000 ppmv) can result in health complaints (as headaches and breath issues), with impacts in productivity, learning performance and absenteeism in case of working and educational environments.^[57–60] Controlling the indoor levels of CO₂ is an indirect way of ensuring adequate fresh air which improves air quality. A description of the recommended sampling strategy for indoor CO₂ can be found in ISO 16000-26:2012^[61] and ASTM D6245^[28]. Annex A of the referred ISO standard describes the calculation of the ventilation requirements of a room.

The most commonly employed techniques to measure CO₂ include the following:

• Non-dispersive infrared spectroscopy (NDIR), according to ISO 16000-26:2012^[61] standard. Several equipment is available, including passive low-cost sensors (e.g., SenseAir S8 or LP8, Winson Electronics MH-Z19B, Sensirion SCD30, Alphasense IRC-AT^[62,63]). Typical uncertainty is either ±30 ppmv or ±50 ppmv plus ±3% or ±5% of reading at 400 ppmv CO₂ and the support electronics corrects for temperature, with some models autocalibrating and also correcting for ambient pressure. Best LOD for low-cost sensors is 10 ppmv CO₂. Active sampling instruments pump the sampled air, using either proprietary optical cells or commercially available low-cost sensors.^[63,64] The LICOR-850 (LiCOR Inc., USA) is a high performance CO₂ monitor with measurement range of 0–20000 ppmv, accuracy within 1% of reading and with a LOD of 1.5 ppmv. It includes automatic pressure, temperature and water correction and is thermostatically stabilized to achieve more accurate measurements. The instrument response is confirmed through exposure to calibration gases.^[63] As a complete instrument and not just a sensor system, it is more expensive than low-cost sensors. Carbon dioxide monitoring systems mounted with low-cost sensors are cheap, easy to operate and commercially available with high reliability, such as Aranet4 sensor (Figure S4, https://aranet4.com/). This sensor costs about 200 € and allows additional concurrent measurements of temperature, relative humidity, and atmospheric pressure. The device has a data logger and a working range of 0-9999 ppmv with an accuracy of ± 50 ppmv or 3% of reading for 0-2000 ppmv and 10% of reading for 2001-9999 ppmv. It records data at a frequency of 1-5 min. This kind of sensors requires an automatic and manual verification every week consisting of locating the device in a well ventilated area outside to adjust the sensor to ambient air CO₂ reference value. Other examples of instruments more expensive and well-established due to their broader use^[25,27] are Delta Ohm HD21ABE17 (see Figure S5) and IAQ-CALC device (model 7545, TSI) with a working range from 0 to 5000 ppmv and accuracy of ±3% of reading or ±50 ppmv able to measure CO, temperature, relative humidity among other parameters. In general, NDIR gas sensors suffer from spectral interference^[65] and higher detection limits than other techniques, although are appropriate for indoor measurements.

• A similar technology, the Infrared gas analyzer (IRGA) with tunable Fabry-Perot Interferometer (FPI) filter is also an on-line active monitor based on the detection in the infrared region (CARBOCAP CO₂ monitors: GMW90, Vaisala, Finland). This monitor shows a dynamic range from 0 to 5000 ppmv and an accuracy of ± 50 ppmv at 1000 ppmv.^[66,67] It records data at a frequency of 1-5 min and shows a good long-term stability, is reliable and accurate, with continuous internal reference measurement. Further, this system requires minimal maintenance, with calibration being recommended each 5 years.

• Cavity ring-down spectroscopy (CRDS). Active method based on absorption of the molecule in the infrared region.^[64] With a LOD <50 ppbv and a dynamic range of 0-1000 ppmv, it has a high sensitivity with a precision of 10 ppbv and negligible drift for months. Commercial CRDS analyzers include the Picarro-G2401 (https://www.picarro.com/g2401_gas_concentration_analyzer ), providing simultaneous measurement of CO₂, CO, methane (CH₄) and water vapor down to parts-per-billion (ppbv), although analytes are limited both by the availability of tunable lasers at the appropriate wavelength and the availability of high reflectivity mirrors at those wavelengths. In addition, the materials involved make the instrument orders of magnitude more expensive than other spectroscopic techniques.

Requirements and test methods for electronic portable and transportable apparatus for detecting and measuring of CO₂ and CO are listed in standard EN 50543.^[68]

5.1.2. Carbon monoxide (CO)

Carbon monoxide is typically emitted to air from incomplete combustion processes/events involving carbonaceous fuels as gas, gasoline, kerosene coal, tobacco or wood. This colorless, nonirritant, odorless substance is recognized to be highly toxic and can even cause death when a high concentration of CO is inhaled. WHO has included CO in the list of selected indoor pollutants as a substance which is hazardousness to health and often found indoors in concentrations of health concern.^[14] For this reason, WHO has established indoor limit values for exposure to CO: 100 mg m⁻³ for 15 min, 35 mg m⁻³ for 1 hour, 10 mg m⁻³ for 8 hours and 7 mg m⁻³ for 24-hour exposure. Recently, the new WHO Global Air Quality Guidelines 2021, aiming to save millions of lives from air pollution was published, announcing more strict limit values for exposure to CO (4 mg m⁻³ for 24 hr-exposure).^[69] Carbon monoxide is a problem in Nordic countries where wood burning or gas stoves are in use and detectors are mandatory in certain buildings ^[70]. Requirements for equipment are presented in standard EN 50291.^[71] Some examples are from Honeywell or Testo.^[72,73]

The methods that have been commonly used to measure this hazardous inorganic pollutant in indoor air are:

• Non-dispersive infrared spectroscopy (e.g., Teledyne Ins., Figure S6), according to the EN 14626:2012 standard.^[74] It is the reference method in accordance with the Directives 2008/50/EC^[75] and EU 2015/1480^[76]. This automatic active analyzer presents a LOD of <0.04 ppmv and a dynamic range 0-1000 ppmv, with a precision < 0.5% and accuracy < 1%. It is a highly sensitive method used in continuous monitoring, although it is expensive and needs periodic external calibration (every 3 months).^[77] EN 14626^[74] provides information about quality assurance (QA) and quality control (QC) for this analyzer.

• Cavity ring-down spectroscopy (CRDS). This method has already been explained in the previous section, since it can also be employed for determination of CO₂ levels. Typical performance of this technique for CO shows uncertainties of <2 ppbv (1 hour average) with precision of 1.5 ppbv for 5 min sampling time, and a dynamic range covering up to 5 ppmv.

• Electrochemical sensors usually rely on diffusion for monitoring but can also include a pump to sample and are commercialized as portable monitors. The typical accuracy is ± 2-3 ppmv^[78] for residential CO alarms, but commercial grade cells^[79,80] can resolve to 10 ppbv. Although the residential CO alarm typical LOD is higher than with other techniques, this equipment is easy to handle and not expensive. It can include a logger to register data being very useful for works requiring continuous monitoring. Normally, some of these equipment also measure temperature and relative humidity (an in some cases CO₂) in addition to CO, which can be acquired by a price of around 1500-3000 €(e.g., Delta Ohm, Gray Wolf, IAQ-CALC device, etc.). However, the readings may be affected by cross sensitivities with other compounds.

• Diffusive sampler using palladium chloride as the absorbing medium: the amount of the formed product, metallic palladium, can be photometrically determined (see Figure S7).^[81] Typical LOD is 0.6 mg m⁻³ for a 2-week sampling period and a dynamic range of 0.5-20 mg m⁻³. This relatively new technique is cheap and easy to handle. However, it has not been widely used since it is only suitable for long-term monitoring, a typical sampling time is from 1 to 2 weeks. This method provides an average concentration over the sampling period. Because high-level exposures (over several hundred mg m⁻³) can cause unconsciousness and death^[14], this method is not recommendable for cases where high CO concentrations are expected or if assessing for compliance with guidelines (e.g., WHO).

The methods for sampling and analysis of carbon oxides along with some characteristics are compiled in Table 3.

Table 3. Methods for sampling and analysis of carbon oxides.

Pollutant	Kind of Instrument	Mode of operation^a	Standards	Analysis Technique	Brand (e.g.,)	Accuracy/precision	LOD	Operation range	Ref.
Carbon oxides
Carbon dioxide (CO2)	High precision Automatic analyzer	OnM	ISO 16000-26	Non-dispersive infrared spectrometry (NDIR)	LICOR-Li 840	Accuracy better than 1.5% of reading	10.2 ppmv	0–20000 ppmv	^[^64^]
	Automatic portable indoor air quality monitor	OnM	ISO 16000-26	NDIR	Delta Ohm HD21ABE17 IAQ-CALC 7545	Precision ±3% of reading or ±50 ppmv	n.a	0–5000 ppmv	^[^25^,^27^]
	Low cost sensor	OnM	n.a	NDIR	Aranet4	Precision 0-2000 ppmv ± 50 ppmv or 3 % of reading	n.a	0–9999 ppmv	–
	Automatic analyzer	OnM	n.a	Cavity ring-down spectroscopy (CDRS)	Picarro G2401	Precision < 50 ppbv	<0.04 ppmv	0–1000 ppmv	^[^64^,^82^]
Carbon monoxide (CO)	Automatic analizer	OnM	EN 14626	NDIR	Teledyne Ins (T300)	Precision < 0.5 %; accuracy < 1%	<0.04 ppmv	0–1000 ppmv	^[^77^]
	Automatic Monitor	OnM	n.a	CDRS	Picarro-G2401	Precision 1.5 ppbv at 5 min	<2 ppbv (1 hour average)	0–5 ppmv	^[^64^]
	Automatic portable indoor air quality monitor	OnM	n.a	Electrochemical sensor (EC)	Delta Ohm HD21ABE17, Gray Wolf	Precision ±3 ppmv + 3% of reading	n.a	0–500 ppmv	^[^24^]
	Diffusive sampler (DS)	OffM	n.a	Focaltin Reaction, Photometric determination	Passam	Expanded uncertainty 24.3 % at levels of 10 mg m^-3	0.6 mg m^-3 for a 2-week sampling period	0.5–20 mg m^-3	-

n.a: not available.

^a OnM: on-line monitoring; OffM: offline monitoring

5.2. Reactive oxygen species

5.2.1. Ozone (O₃)

Outdoor air is an important source of O₃ to the indoor environments. Air purifiers, ionizers and electrostatic precipitators to remove particles from the air are among the indoor sources of O₃.^[21] Although O₃ is a priority compound outdoors, the typical low concentrations that are found indoors makes this pollutant less important from the toxicological point of view. Nevertheless, O₃ has been considered of interest for the indoor chemistry, as a compound that initiates the oxidation of other compounds resulting in a chain of secondary products to air.^[83] According to the WHO, the limit value of O₃ for human health protection, which can be applied to indoor air, is 100 µg m⁻³ for an average 8-hours exposure.^[69]

The main techniques that have been employed to measure O₃ are:

• Monitor based on ultraviolet (UV)-Photometry (e.g., Teledyne Ins. T400, Figure S8), is the reference method in accordance with the Directives 2008/50/CE^[75] and EU 2015/1480^[76], and it is described in the standards EN 14625:2012^[84] and ISO 13964:1998^[85]. This method for continuous monitoring of O₃ is based on the internal direct quantification of the UV absorption of O₃ according to Lambert-Beer’s Law. The dynamic range of the method is selectable between 0 to 100 ppbv and 0 to 10 ppmv and the precision is 0.5% of reading above 100 ppbv. It allows fast measurements (readings typically reported at 1 min intervals), and is easy to operate and maintain. This instrument requires an external power supply and periodic external calibrations (every 3 months).^[86,87] See EN 14625^[84] for QA and QC.

• Nitric oxide-chemiluminescence-based monitor (NO-CL). This method uses chemiluminescence with NO as a reagent. The NO-CL O₃ analyzer detects O₃ based on the reaction of O₃ in sampled air with NO reactant gas forming excited nitrogen dioxide (NO₂*). The NO₂* emits a photon at 600 nm-2800 nm when it returns to its ground state. The emitted photon is then detected by a photomultiplier (PMT), and the PMT count is proportional to the O₃ numbers in the sampled air. In comparison to the conventional UV absorption technology, this method presents advantages including the avoidance of common interferences such as aromatic hydrocarbons, fine particles and mercury (Hg) vapor.^[88] This device operates on any full-scale range between 0-100 ppbv and 0-2000 ppbv with a precision of 0.5% of the reading. LOD is < 0.3 ppbv (e.g., Model T265, Teledyne API; San Diego, CA, USA).

For more information about the instruments that could be used, the US Environmental Protection Agency (EPA) provides a list of reference and equivalent methods for several pollutants.^[89]

• Molecular diffusion by passive sampling. Standards EN 14412 (specific to indoor air) and EN 13528-3 provide guidance for the selection, use and maintenance of passive samplers. Different cartridges can be used, including:

  1. Nitrite-coated filter (e.g., Ogawa), where the oxidation of nitrite to nitrate by O₃is quantified by IC following the OSHA method ID-214 for the analysis.^[90] However, this OSHA method uses a calibrated sampling pump and a two-piece polystyrene cassette containing two nitrite-impregnated glass fiber filters being the sampling device an active sampler.

  2. Silica gel coated with 4,4’-dipyridylethylene (e.g., Radiello®)^[91], where O₃ is quantified through the measurement of 4-pyridylaldehyde that is extracted with MBTH (3-methyl-2-benzothiazolinone hydrazone hydrochloride) to form the corresponding product, which is determined by UV-Vis spectrophotometer (Figure S9). Silica gel ensures the presence of water, necessary to complete ozonolysis reactions. Typical LOD is 2 µg m⁻³ for 7 days of exposure.

Passive cartridges typically have an uncertainty of 10% and are cheap and easy to handle, allowing simultaneous sampling at several locations. However, as referred above, the passive sampling method is only suitable for long-term monitoring, with the sampling time varying from one day to one week. This method was employed in some European studies, such as the SINPHONIE project^[24] and it has been recommended by WHO for sampling O₃ in different public setting for children in order to assess health exposure.^[92]

The 2B Technologies 714^[93] is a portable gas generator, and can generate accurate ppbv concentrations of NO₂, NO and O₃- very useful for field calibration. Accuracy is specified as ±2% concentration ±2 ppbv.

5.2.2. Hydrogen peroxide (H₂O₂)

Aqueous H₂O₂ is the active ingredient of several household non-bleach cleaning products. However, evidence from studies conducted in simulated and controlled indoor settings in the absence of any cleaning products suggested that reactions between O₃ and terpenes can also be an important source of indoor H₂O₂, at high indoor O₃ levels^[11,12,94] (Table 1). In addition, H₂O₂ may be produced by commercial air cleaners in indoor air. Air cleaners using reactive ions and/or reactive oxygen species (ROS) have become prevalent during the COVID-19 pandemic.^[95] Moreover, plasma air cleaners generate O₃ and other reactive oxygen species such as hydroxyl radicals, superoxides, and H₂O₂.^[96] Measurements aiming at the assessment of H₂O₂ concentrations in indoor air are scarce. For example, there are a few studies conducted under realistic and manipulated indoor environments, but there is still a lack of information about the background concentrations in real indoor settings. Considering yields and typical indoor scenarios, the maximum indoor H₂O₂ concentrations are projected to be comparable to the maximum outdoor concentrations. Exposure route is mainly by inhalation and the respective OSHA permissible exposure limit (PEL) is 1 ppmv as a 8 hour time weighted average.^[97]

The techniques of measuring gas-phase H₂O₂ by impinger and diffusion scrubber have been investigated.^[98] However, other methods to sample H₂O₂ in ambient air have been described by some authors.^[99,100] The most commonly reported sampling and analysis techniques are listed below.

• Diffusion scrubber of PTFE with a solution of Ti(IV)-4-(2-pyridylazo) resorcinol, in where the aliquot is analyzed by HPLC with UV-Vis detector at 508 nm. It requires a pump to force air into the scrubber. The method has a LOD of 9 pptv and 60 min of time resolution.^[101] The main advantages are the low-cost and the fact that interference by other inorganics, as O₃ or SO₂, is negligible.

• Monitor based on the chemiluminescence oxidation of luminol by H₂O₂. This system uses a photomultiplier tube (PMT) as detector with a LOD of 25 pptv, with wide linear range up to 100 ppbv and fast response for continuous monitoring. It is simple to handle and not expensive. Regarding interferences, at typical ambient levels NO₂ and SO₂ do not interfere with the reading; however, the O₃ equivalent interference is ∼<0.5%.^[102]

• Automatic analyzer based on wet chemical dual-enzyme. This liquid-based technique was developed by Lazarus et al.^[103]. The detection is based on the liquid phase reaction of peroxides with p-hydroxyphenylacetic acid (POPHA) catalyzed by peroxidase. This reaction produces a fluorescent dimer that can be excited at 326 nm and detected between 400 and 420 nm. The reaction is sensitive to all peroxides in the solution, but it is faster for H₂O₂. To separate the signal produced due to the presence of H₂O₂ from the other peroxides, two parallel channels are used.^[104] Automatic routine measurements require regular calibrations of the instrument. As liquid phase standards of H₂O₂ are unstable, they must be prepared fresh, directly before use and stored cool and dark even during the calibration process. For automatic calibrations the instrument can be equipped with an internal H₂O₂ permeation source (optional). The instrument requires a cooling box or fridge for storing some of the solutions, necessary for operation at a temperature of ∼4 °C. To ensure a maximum stripping efficiency, the stripping solution should also be kept cool and dark. The enzymatic reaction with peroxidase has only two known significant interferences to other substances: O₃ < 1: 3500 and NO < 1: 8000. The original dual enzyme technique and subsequent improvements using HPLC have been described elsewhere.^[99,100]

The methods for sampling and analysis of reactive oxygen species along with some characteristics are listed in Table 4.

Table 4. Methods for sampling and analysis of reactive oxygen species.

Pollutant	Kind of Instrument	Mode of operation^a	Standards	Analysis Technique	Brand (e.g.,)	Accuracy/precision	LOD	Operation range	Ref.
Reactive oxygen species
Ozone (O3)	Automatic analyzer	OnM	EN 14625 ISO 13964	Ultraviolet Photometry (UV-Pho)	Envea (O342e) Teledyne Ins.(model T400)	Precision 0.5% of reading above 100 ppbv	0.2– 0.4 ppbv	Min: 0–10 ppmv Max: 0–10 ppmv	^[^86^,^87^]
	Automatic analyzer	OnM	n.a	Nitric oxide- Chemiluminiscence (NO-CL)	Teledyne Ins (model T265)	Precision 0.5% of the reading	<0.3 ppbv	0–2000 ppbv	–
	Diffusive sampler. Nitrite-coated filter (DS-IC)	OffM	EN 14412	Ion Chromatography with conductivity/ absorbance detection	Ogawa	n.a	n.a	n.a	^[^90^,^105^]
	Diffusive sampler. Silica gel coated with 4,4′-dipyridylethylene (DS-Spec)	OffM	EN 14412	UV-Vis spectrophotometer	Radiello® Passam	Uncertainty 14.5 % over the whole sampling rate linearity range	2 µg m^-3 for 7 days of exposure	2–400 μg m^-3	^[^24^,^106^]
Hydrogen peroxide (H2O2)	Diffusion scrubber of PTFE with a solution of Ti(IV)-4-(2-pyridylazo) resorcinol (Denuder)	OffM	n.a	HPLC with spectrophotometric detector at 508 nm	n.a	Precision of standard solution between 1.2 % and 2.1%	9 pptv (sampling volume of 55 L)	n.a	^[^101^]
	Monitor	OnM	n.a	Chemiluminescence oxidation of luminol by H2O2 (Luminol-CL)	n.a	n.a	25 pptv	25pptv–100 ppbv	^[^102^]
	Automatic analyzer based on wet chemical dual-enzyme (WCDET)	OnM	n.a	Fluorescence spectroscopy	Aero-Laser (AL 2021)	n.a	100 pptv	100 pptv − 3 ppmv (gas)	^[^104^]

n.a: not available.

^a OnM: on-line monitoring; OffM: offline monitoring.

5.3. N-Containing compounds

5.3.1. Nitrogen dioxide (NO₂)

The primary indoor sources of NO₂ are unvented kitchen combustion appliances (e.g., gas stoves), kerosene heaters, tobacco smoke and candles.^[14,107,108] Outdoor NO₂ levels resulting mainly from natural and anthropogenic sources (road traffic) can also have an important contribution to the NO₂ concentrations found indoors.^[109] Airborne NO can be oxidized in air to form NO₂ by available oxidants (oxygen, O₃ and VOCs) and this oxidation velocity is rapid, unlike indoor air where the oxidation process is much slower.^[14] Typical residential indoor concentrations are about half of those found outdoors. The factors favoring high indoor NO₂ levels are poor ventilation, the small size of the indoor spaces containing sources of NO₂ and frequent use of gas stoves or other combustion-related equipment/products, where concentrations may even exceed outdoor levels.^[14]

Nitrogen dioxide is an oxidizing toxic gas with important health effects affecting mainly the respiratory system (bronchoconstriction, airway inflammation, increase of bronchial reactivity and decrease in immune defense having as consequence an increased susceptibility to respiratory infections).^[14] The previous indoor guideline concentrations for NO₂ were 40 µg m⁻³ (annual average) and 200 µg m⁻³ (1 hour).^[14] But, WHO recently set new limit values for NO₂ of 10 µg m⁻³ for annual average and 20 µg m⁻³ for 24 h-exposure.^[69] If these values are exceeded, adverse effects on people, especially affecting vulnerable groups as children and people with asthma are almost inevitable.

The measurement planning for NO₂ in indoor air is described in ISO 16000-15 Standard.^[110] The following methodologies have been employed for determining concentrations of this inorganic compound in the air:

• Chemiluminescence from electronically excited NO_2, which is formed from the reaction between NO and O₃ (Molybdenum catalyst or Photolytic converter can be used to reduce NO₂ to NO), according to the reference method EN 14211:2012^[111]. The standards ISO 7996^[112] and ASTM D3824^[113] also provide information on continuous methods based on the principle of chemiluminescence. The system is an on-line and active monitor, with 1 minute resolution. Typical LOD is 50 pptv and accuracy of 10%, allowing different measurement ranges up to 1ppmv.^[114,115] There are several brands in the market, including Thermo Scientific – 42i, Teledyne, Eco-Physics CLD 770 (Figure S10). The measurement of NO₂ by this method is indirect and is calculated as the difference between NO_x and NO. This characteristic might be a source of error. Monitors using Molybdenum catalyst can be affected by positive interferences caused by photochemically formed NO_y species, while those using the photolytic converter may suffer from negative interference by photolysis of VOCs in the photolytic converter and consecutive peroxyradical reactions with NO. It requires periodic external calibration but it is easy to use, handle and maintain.^[116] See EN 14211^[111] for QA and QC of the analyzer.

• Oxidation of luminol by NO₂ which produces chemiluminescence that is measured by a PMT detector (Unisearch LMA 3D), being almost linearly correlated with the NO₂ concentration. This technique is based on the use of an on-line active monitor with 1 minute resolution and a LOD of 10 pptv. Positive interference by photo-chemical formation of O₃ and different peroxyacylnitrates (PAN) like species have been reported.^[117]

• Automatic analyzer based on cavity attenuated phase shift (CAPS) spectroscopy. It is also an on-line and active monitor with 1 minute resolution. It is a very reliable technique based on spectroscopy with LOD of 40 pptv and two operation ranges: minimum 0 pptv - 5 ppbv and maximum 0-1000 ppbv (Teledyne-T500U-CAPS and Aerodyne CAPS NO₂ Monitor). The monitor, which detects the optical absorption of NO₂ within a 20 nm bandpass centered at 440 nm, comprises a blue light emitting diode, an enclosed stainless steel measurement cell (26 cm length) incorporating a resonant optical cavity of near-confocal design and a vacuum photodiode detector. The main advantage of this method is that the measurement of NO₂ is direct, and thus less prone to suffer from interferences. Interferences due to glyoxal and methylglyoxal are reported but those can be removed by broadening the spectral analysis region above 455 nm.^[118,119]

• Laser induced fluorescence (LIF) spectroscopy. Another on-line and active method with a high sensitivity, characterized by a LOD of 80 pptv (10 sec). The main reported drawback is interference by water vapor, due to fluorescence quenching, which reduces the sensitivity of the method. In addition, the device needs periodic external calibration and is more expensive than chemiluminescence methods and other spectroscopic techniques.^[120] These devices are heavy (typically > 50 kg) and complex laser systems which can be a limitation for indoor use, although other instrument based on laser-induced fluorescence (e.g., GANDALF) have been developed with lightweight laser diodes (< 2 kg) and limits of detections similar to those of the best-performing larger instruments.^[121] Summing, this technique requires significant expertise to operate correctly.

• (Incoherent broadband) cavity-enhanced absorption spectroscopy (IBBCEAS). This is an automatic online and active optical system with 0.9 ppbv (5 sec) LOD and an overall uncertainty of 7.8%. For quantification of the compounds, it requires determination of the mirror reflectivity, which is described as very challenging.^[122]

• Diffusive sampling with triethanolamine. This passive sampling method is one of the methods for long-term monitoring recommended by ISO-16000-15:2008 standard.^[110] Analysis follows the Griess-Saltzman method according to EN 16339 standard.^[123] The LOD of the method depends on the sampler employed, but is in the order of 1 ppbv for a sampling exposure of 7 days (e.g., Radiello®; Figure S11). The typical accuracy is approximately 10%. The concentration of NO₂ in the sample is determined by IC or by spectrophotometry. It is a simple and well-established method that can also be used to provide information on personal exposure. It is preferred over chemiluminescence methods for long-term measurements due to its lower cost and the possibility of measuring concurrently in several locations. It is useful for long-term trends and exposure.^[66] Since NO₂ reacts in the presence of ultra-violet light, direct UV light should be avoided where diffusive samplers are placed.^[124,125] The same method can be used for different contaminants by selecting the proper reagent.

EN 16339 standard describes different types of diffusive samplers and provides information regarding QA and AC when using diffusive samplers.^[123] Yu et al.^[126] reviewed the different passive dosimeters for NO₂ in personal and indoor air sampling.

• Active sampling based on a fritted bubbler system according to the method ASTM D1607-91^[127]. This method covers the determination of NO₂ content in ambient air by the Griess-Saltzman reaction. It allows a maximum sampling time of 60 min at a flow rate of 0.4 L min⁻¹. The LOD of the method is 2 ppbv and the working range is from 0.002 ppmv up to 5 ppmv (4-10000 µg m⁻³). The analysis is also based on the Griess-Saltzman method. Nitrogen dioxide is absorbed in an azo-dye forming reagent. A red-violet color is produced within 15 min, the intensity of which is measured spectrophotometrically at 550 nm. This method allows to acquire concentration over shorter sampling times than the passive methods but a sampling pump is required for sampling. ISO 6768^[128] also specifies a modified Griess-Saltzman method for the determination of NO₂ using fritted bubblers system for sampling. The working range if from 3 to 2000 µg m⁻³ for sampling times between 10 min and 2 hours. For concentrations of NO₂ in excess of 500 µg m⁻³, as can occur in industrial atmospheres, other methods should be applied such as NIOSH 6014, but this method is not part of the objective of this review.

5.3.2. Nitric oxide (NO)

Nitric oxide levels in indoor air proceed from indoor sources such as unvented fuel burning appliances, heating systems and tobacco smoking but also from outdoor air. To this date, no ambient or indoor air guideline limit values have been set for this substance. The methods for measuring NO in the air are presented below.

• Chemiluminescence with Molybdenum catalyst or Photolytic converter, according to the standard method EN 14211:2012^[111]. This monitor is the same system as the one used to measure NO₂ (Thermo Scientific 42i; Eco-Physics CLD 770), with the same LOD and accuracy. The technique is on-line and active and, unlike NO₂, the measurement of NO is direct and therefore more reliable. It needs periodic external calibrations and is a widely used method as it is easy to use and maintain.^[115]

• Some manufacturers (e.g., Teledyne) have developed automatic analyzers based on CAPS. This instrument combines direct NO₂ measurements with highly efficient gas phase titration (GPT) to convert and measure the NO gas component. The LOD is less than 10 pptv and the working range of 0 to 1000 ppbv and a precision of 0.5% of reading above 5 ppbv. Finally, ASTM D3608-19^[129] provides a test method that covers manual determination of the combined NO₂ and NO content (total NO_x) in the range from 4 to 10 000 μg m⁻³ (0.002 to 5 ppmv).

5.3.3. Ammonia (NH₃)

Nazaroff and Weschler conducted a comprehensive and thorough review in which they reported that the dominant basic species in indoor air is NH₃.^[7] Concentrations of NH₃ tend to be significantly higher in indoor air than in outdoor air, often by a factor of ten or more.^[130] Exposure to ammonia inside buildings occurs due to emissions from the use of consumer surface cleaning products, which typically contain low levels of ammonia (between 5 and 10%), and by tobacco smoke, cooking byproducts and concrete.^[7,131] Another indoor source is human emission thought breath, skin, flatulence, urine and feces. ^[7] It is a priority compound as set by the US Agency for Toxic Substances and Disease Registry.^[132] The indoor air quality standard for China (GB/T 18883-2002)^[41,133] stipulates an indoor limit value of 0.20 mg m⁻³ (1-h average) for NH₃ and the U.S. Occupational Safety and Health Administration and British Health and Safety Executive^[134,135] have set limits on NH₃ exposure of 25 parts per million (ppmv) over an 8 h period and 35 ppmv over a 15 min period.

Many methods have been used to measure NH₃ in ambient air including both off line and on-line techniques; in recent years these methods have been improved with a finer temporal resolution. Amongst these methods we can cite the denuder technique (simple and annular) with analysis in laboratory, AMANDA system (Ammonia Measurement by Annular Denuder sampling with on-line Analysis) and passive samplers. Some recent commercial advances are Chemical Ionization Mass Spectrometer (ClMS), Quantum Cascade Tunable Infrared Laser Differential Absorption Spectrometer (QC-TILDAS), Differential Optical Absorption Spectroscopy (DOAS), Monitor for AeRosols, GAses in ambient air (MARGA) and Proton Transfer Reaction Mass Spectrometer (PTR-MS).^[136–141] Obviously, some of these techniques cannot be used in indoor air due to space and operational limitations such as DOAS, although a compact set-up using the same principle and based on Cavity Enhanced Absorption Spectroscopy (CEAS) in the infrared region has being developed to detect NH₃.^[142]

Other techniques commonly used for measuring NH₃ in outdoor air that can be applied to indoor air are described below:

• Cavity ring-down spectroscopy (CRDS) (e.g., Picarro G2103, Figure S12). This on-line and active method has a LOD of 0.5 ppbv at 1 minute resolution.^[64,143] As described above, it uses a tunable laser light which limits which analytes can be measured. As an optical system, this technique is very sensitive and reliable, although more expensive than other spectroscopic techniques. This kind of automatic analyzer has been commonly used to measure NH₃ in indoor air.^{[64,130,143,144]}

• Chemiluminescence analyzer (e.g., ECO Physics CLD 855 CY). Two catalytic converters of different characteristics allow sequential detection of NO_x and NO_x−amines by converting them into NO at 375 ◦C and 650 ◦C, respectively. The concentration of ambient NH₃ is calculated from the difference between NOx and NO_x−amine (NH₃ = NO_x-amine - NO_x). The range of measurement of NH₃ analyzer varies from 0–5 ppbv to 0–5000 ppbv (accuracy ± 0.050 ppbv of all the ranges).^[145] Zhang et al.^[146] measured the variation of NH₃ in a university building in Beijing with a chemiluminiscence analyzer (EC9842B, Ecotech Pty Ltd).

Continuous automatic analyzers that employ spectroscopic techniques or other techniques are commercially available, but there is a lack of published robust data on calibration and procedures for reliable field measurements. Currently, there is no accepted reference method for NH₃.

• Diffusive sampling (DS), according to EN 17346^[140,147] standard. There are different types of passive samplers to collect NH₃ in air (tube type, filter type and radial type). The sorbent commonly used includes citric, phosphoric, phosphorous and tartaric acids and the analysis is carried out by means of different methods including IC, flow injection analysis with conductivity detector, and spectrophotometry.^[140] The radial type (e.g., Radiello®) is made of micropopous polyethylene and impregnated with phosphoric acid. Ammonium ion is quantified by visible spectrometry (635 nm) as indophenol. Particulate matter containing airborne ammonium salts do not cross the diffusive membrane of Radiello®. The LOD depends on the type of passive sampler. In the case of the Radiello® sampler, the LOD is 0.001 mg m⁻³ for 24-hour exposure. As passive method its advantage is the low cost of the samplers and that it is easily deployable. It is suitable for providing information for chronic exposure assessment purposes given the long sampling period.

5.3.4. Nitric acid (HNO₃)

In indoor air, nitric acid is formed by the reaction of nitrate radicals with saturated organic compounds by abstracting a hydrogen, although these reactions are slow, on the order of 10⁻¹⁷cm³ molecule^_1 s^_1. Dinitrogen pentoxide can also form nitric acid via its reaction with water, especially in indoor surfaces. Nitric acid formation in the daytime atmosphere also occurs by hydroxyl radical (OH·) reacting with NO₂. On the other hand, HNO₃ in the presence of gas-phase ammonia (NH₃) can react to produce ammonium nitrate (NH₄NO₃) that can then suffer dissociation. The measured concentrations of HNO₃ in indoor air are very low (< 0.5 ppbv), suggesting that HNO₃(g) is lost to indoor surfaces, and depending on the method, concentrations are often indistinguishable from zero.^[7]

The most common sampling methods for the collection of HNO₃ with selected references can be found in a publication of Trebs et al.^[148]. Filters, denuder and filters, diffusion denuder, diffusion scrubber, laser-photolysis fragment-fluorescence (LPFF) and chemical ionization mass spectrometer (CIMS) have been used to collect and measure HNO₃ in the atmosphere. Few studies have measured HNO₃ in indoor air and most of them have used filters and denuders as sampling method. Recently, Vichi et al.^[149] developed a novel multipollutant diffusive sampler for simultaneous sampling of three different pollutants (HNO₃, HONO and NO₂) using the Analyst® passive sampler.

Sampling and analytical methods:

• Annular denuder systems (ADS). The sample enters the system at a flow rate of 10 L min⁻¹ through a glass inlet-impactor to remove coarse particles. Then, the air passes through a series of denuders which selectively remove gases from the sample stream. A sodium carbonate coated denuder collects HNO₃ (also, SO₂ and HONO). NH₃ can also be trapped using a citric acid-coated denuder. The denuders are extracted and analyzed by IC. The LOD based on sensitivity of IC analysis (24 h sampling at 10 L min⁻¹) is 0.07 ppbv for HNO₃.^[150] More information about this well-established and -known technique to collect these trace gases represented by denuder sequence-based systems can be found in Benner et al.^[151], Genfa et al.^[152] and Liang and Waldman^[153]. The drawback of this technique is that it is time consuming and labor intensive.

• Diffusive sampler. Different passive samplers have been developed to collect exclusively HNO₃ or several pollutants simultaneously, including HONO and NO₂. The former device consists of a filter pack sampler assembly composed of one nylon membrane filter, one supported PTFE filter, a Petri dish, two PTFE rings and one polycarbonate ring. The nylon filter, which samples HNO₃(g) by a sorption mechanism, is housed inside the Petri dish between the two PTFE rings. Nylon filters are extracted in 1.0 mM KOH and analyzed by IC. The LOD is 5 pptv over a 30-day sampling period or 121 pptv over a 24-hour sampling period. This device allows to quantify ultra-trace measurements of HNO₃(g). Passive nylon filter sampling has the advantages of low cost, low maintenance, reduced field-personnel training, and the potential to obtain large spatial scale measurements.^[154]

• Vichi et al.^[149] developed a novel multipollutant diffusive sampler for simultaneous sampling of HNO₃, HONO and NO₂ collected at separate sampling stages. The design of the Analyst® passive sampler was used and modified to collect HNO₃, on the first filter and NO₂ and HONO in the successive absorbing pads. Filters are extracted by adding a solution of sodium bicarbonate (NaHCO₃) and carbonate (Na₂CO₃) and the analysis is carried out by means of IC equipped with an anion exchange column (AS12A column) (Figure S13). The lowest detectable concentration for the multipollutant samplers after one month exposure is about 1.4 μg m⁻³ for NO₂, 0.2 μg m⁻³ (∼0.08 ppbv) for HNO₃, and 0.5 μg m⁻³ for HONO.

In general, passive samplers for collection HNO₃ are prepared in the laboratory from commercial components.

• Chemical ionization mass spectrometry (CIMS). This CIMS apparatus for ground-based measurements of HNO₃ was developed, built and tested in the laboratory of MPI-K, Heidelberg (Germany).^[155,156] The main components of the instrument are a heated PFA Teflon sampling system, stainless-steel flow-tube reactor (FR) with a tubular inlay made of Teflon, a gas-discharge ion source, and a quadrupole mass spectrometer. In the sampling line and FR a pressure of about 50 mbar is maintained by an oil-free 30m³ h⁻¹ Scroll pump. Due to the pressure gradient between the FR and the atmosphere (usually around 790 mbar), ambient air is drawn at a flow rate of about (12.4 ± 0.5) slm (= L min⁻¹ at standard temperature and pressure) through a 1.35 ± 0.05 mm diameter heated PFA orifice and then through the 1/2 inch (outer diameter, 9.3 mm ID) heated PFA Teflon sampling tube into the left end of the flow-tube reactor. In situ calibrations with standards are required.

• Ambient ion monitor–ion chromatography (AIM-IC) system consists of ambient ion monitor (e. G model 9000D, University Research Glassware Corp.) for collection of inorganic gases (NH₃, HONO, HNO₃, and SO₂) and particles into aqueous solution and two IC systems (model ICS-2000, Dionex Corp.) for the analyses of anions and cations (NH₄⁺, NO₃⁻, and SO₄ ²⁻). Ambient air enters the AIM-IC system and gas phase species are collected by a parallel-plate wet denuder. After 1 h of sampling, the samples are injected onto two IC systems for simultaneous anion and cation analyses. A Dynacalibrator calibration gas generator (model 450; VICI Metronics, Poulsbo, WA, USA) with Dynacal permeation tubes (VICI Metronics) are used in conjunction with zero-air cylinders to generate known concentrations of HNO₃ gas standards. The LOD for HNO₃ is 14 ng m⁻³ .^[155,157]

• Long path liquid absorption photometer (LOPAP). The instrument is a monitor for the sensitive detection of HNO₃, which is sampled in a stripping coil. HNO₃ is detected in a long path absorption tube after conversion into an absorbing dye. Two channels are used for correction of interferences. The detection limit is 5-30 pptv for a sampling time of 6-2 min.^[158]

5.3.5. Isocyanic acid (HNCO)

Isocyanic acid has been recognized as a gas-phase acid in the outdoor atmosphere since 2008 and has recently been identified in ambient air at potentially concerning concentrations for human health.^[159] However, no study directly links inhalation exposure of HNCO to adverse health effects. The sources of HNCO include fossil fuel combustion, biomass burning, secondary photochemical production from amines and amides, cigarette smoke and combustion of materials in the built environment. More recently, experiments have recently demonstrated that HNCO is generated from gas phase oxidation of nicotine.^[9] In a residential kitchen in Toronto (Ontario), concentrations of HNCO were up to four times higher than outdoor concentrations.^[159]

The detection and measurement of HNCO in ambient air have been carried out using the following methods:

• Sample derivatization with impingers and analysis by liquid chromatography (LC) and electrospray mass spectrometry (MS) monitoring positive ions.^[160]

• Sample derivatization by means of denuders or impinger + filter and analysis by liquid chromatography mass spectrometry (LC-MS) according to ISO 17734-1:2013.^[161] This standard gives general guidance for the sampling and analysis of airborne isocyanates in workplace air.^[162,163]

• Selective hydrolysis followed by the detection of NH₃ with an ammonia-sensitive electrode.^[164]

• Fourier-transform infrared spectroscopy (FTIR).^[165] It is based on the Lambert-Beer Law and needs high long paths to achieve suitable detection limits, which makes it unfeasible for use indoors.

The above methods are able to detect gas phase HNCO but they are neither fast nor adequately sensitive to measure real-time concentrations.^[159]

• Chemical ionization mass spectrometer (CIMS). This technique allows the selective detection of organic acids by using acetate as the reagent ion in order to accomplish soft ionization of acidic species through proton abstraction in the gas phase so that, for example, HNCO is detected as its conjugate anion, NCO^-.^[166] This method is sensitive, selective and fast with LOD of 0.005 ppbv. CIMS does not require labor-intensive collection or preparation steps.^[159,167] Acetate CIMS has a negligible humidity dependence. An iodide reagent ion CIMS technique has also been developed for HNCO.^[168]

• Proton-Transfer Reaction Mass Spectrometer (PTR-MS) (Figure S14). This technique provides real-time data with high sensitivity and selective of airborne components in gaseous or vapor phase. The commercial instrument (e. g Ionicon Analytik GmbH) uses soft ionization (meaning little or no fragmentation of the analyte ion) through proton transfer from hydronium ions (H₃O⁺) produced from water vapor plasma and can detect HNCO mixing ratios from low ppbv up to ppmv.^[163,169] The HNCO signal depends on humidity which can complicate the quantification of the signal. One of the drawbacks of HNCO ambient measurements by mass spectrometry is that HNCO is not commercially available because it polymerizes at high concentrations.^[159] Pure isocyanic acid can be made by thermal decomposition of cyanuric acid or urea.^[162,163]

The methods for sampling and analysis of nitrogen-containing compounds along with the main characteristics are listed in Table 5.

Table 5. Methods for sampling and analysis of N-containing compounds.

Pollutant	Kind of Instrument	Mode of operation^a	Standards	Analysis Technique	Brand (e.g.,)	Accuracy/precision	LOD	Operation range	Ref.
N-containing compounds
Nitrogen dioxide (NO2)	Automatic Analyzer	OnM	EN 14211 ISO 7996 ASTM D3824-20 ISO 16000-15	Chemiluminescence (Molybdenum catalyst or Photolytic converter) (CL-Mo CL-Pho)	Thermo Scientific, Model 42i, Teledyne, Eco-Physics	Accuracy 10% Precision ± 0.4 ppbv (500 ppbv range)	50 pptv	50 pptv–1 ppmv	^[^114^,^115^]
	Automatic Analyzer	OnM	n.a	Luminol-Chemiluminescence (Luminol- Cl)	Unisearch  Associates, Inc. (LMA 3 D	n.a	<10 pptv	10 pptv–4 ppmv	^[^117^]
	Automatic Analyzer	OnM	n.a	Cavity Attenuated Phase Shift Spectroscopy (CAPS)	Teledyne-CAPS (model T500U)	Precision 0.5% of reading above 5 ppbv	50 pptv	Min:0–5 ppbv Max: 0– 1000 ppbv	^[^118^,^119^]
	Laser	OnM	n.a	Laser-induced fluorescence spectroscopy (LIF)	n.a	Accuracy ± 0.5 %	80 pptv (10 sec)	n.a	^[^120^]
	Automatic Analyzer	OnM	n.a	(Incoherent broadband) Cavity-enhanced absorption spectroscopy ((IBB) CEAS)	n.a	Uncertainty 7.8 %	0.9 ppbv (5 sec)	n.a	^[^122^]
	Difussive sampler with Triethanolamine (DS-TEA)	OffM	ISO 16000-15	Griess-Saltzman method, Spectrophotometry/ion chromatography	Radiello®, Ogawa, IVL, Gradko (Palmes)	Uncertainty 11.9 %	It depends on the sampler. LOQ 1 ppbv for 7 days of exposure (Radiello)	3–130 µg m^-3	^[^66^,^124^,^125^]
	Active sampler with fritted bubblers system (AS-FSB)	OffM	ISO 6768 ASTM D1607-91	Griess-Saltzman method. Spectrophotometry	–	–	It depends on the sampler	3–2000 μg m^-3 (ISO 6768) 4- 10000 μg m^-3 (ASTM 1607)	–
Nitric oxide (NO)	Automatic Analyzer	OnM	EN 14211	Chemiluminescence (Molybdenum catalyst or Photolytic converter) (CL-Mo/CL-Pho)	Thermo Scientific, Model 42i	Precision ±0.4 ppbv (500 ppbv range)	40 pptv (60 sec averaging time)	50 pptv–100 ppmv	^[^115^]
	Automatic Analyzer	OnM	n.a	Cavity Attenuated Phase Shift Spectroscopy (CAPS)	Teledyne-CAPS (model N500)	Precision 0.5% of reading above 5 ppb	< 10 pptv	Min:0–5 ppbv Max: 0–1000 ppbv	–

n.a: not available.

^a OnM: on-line monitoring; OffM: offline monitoring

Table 5. Continued

Pollutant	Kind of Instrument	Mode of operation^a	Standards	Analysis Technique	Brand (e.g.,)	Accuracy/precision	LOD	Operation range	Ref.
N-containing compounds
Ammonia (NH3)	Automatic Analyzer	OnM	n.a	Cavity ring-down spectroscopy (CRDS)	Picarro G-2103	Accuracy 0.5 ppbv ± 5% of reading Precision of 0.10 ppbv + 0.1% of reading for 30 sec of data.	0.5 ppbv	Guaranteed range 0 − 500 ppbv Operational range 0 − 10 ppmv	^[^64^,^130^,^143^]
	Automatic Analyzer	OnM	n.a	Chemiluminiscence (CL)	ECO physics nCLD 855 CY; Teledyne-API Model T201; Ecotech Pty Ltd model EC9842B	accuracy ± 0.050 ppbv of all the ranges	n.a	0 − 5 ppbv; 0–5000 ppbv	^[^145^,^146^]
	Difussive sampler (DS)	OffM	EN 17346	Ion chromatography/ Flow injection analysis with conductivity detector/ spectrophotometry	Radiello®	Depends on the sampler expanded uncertainty (6.5 % − 29% for 0.001 mg m^-3 )	0.0002–0.0015 mg m^-3 3 for 4 sampling weeks 0.001 mg m^-3 (24 h exposure, Radiello®)	0.0005–0.150 mg m^-3 (depending on the kind of sampler)	^[^147^]
Nitric acid (HNO3)	Annular Denuder Systems (ADS)	OffM	n.a	Ion chromatography	n.a	Precision 2.6 %	0.07 ppbv (24 h sampling at 10 L min^-1)	n.a	^[^150^]
	Difussive sampler (DS)	OffM	n.a	Ion chromatography	n.a	Method precision was ± 3% across 5 calibrations; Accuracy 94 % (upper range standards)	5 pptv (30 day sampling period) or 121 pptv (24 h sampling period)	n.a	^[^154^]
	Continuous automated system	OnM	n.a	Chemical ionization mass spectrometry (Cl-MS)	n.a	n.a	< 50 pptv	n.a	^[^156^]
	Ambient ion monitor	OnM	n.a	ion chromatography	AIM model 9000 D, University Research Glassware Corp, IC (model ICS-2000, Dionex Corp.,)	n.a	14 ng m^-3	n.a	^[^155^,^157^]
	Automatic Analyzer	OnM	n.a	Long path liquid absorption photometer (LOPAP)	n.a	Precision 2-3%, accuracy 10%	5-30 pptv	5 pptv to 400 ppbv	^[^158^]
Isocyanic Acid (HNCO)	Continuous automated system	OnM	n.a	Negative ion proton transfer chemical ionization mass spectrometry (NI-PT-CIMS)	n.a	n.a	0.005 ppbv.	n.a	^[^159^,^167^]
	Continuous automated system	OnM	n.a	Proton transfer reaction-mass spectrometry (PTR-MS)	Ionicon Analytik GmbH)	n.a	2.5 ppbv	low ppbv (2.5 ppbv) up to 1000s of ppbv	^[^162^,^163^]

n.a: not available.

^a OnM: on-line monitoring; OffM: offline monitoring.

5.4. S-Containing compounds

5.4.1. Sulfur dioxide (SO₂)

Coal combustion used in cooking and heating has been reported to emit large amounts of SO₂, contributing to elevated indoor levels of this compound.^[170] Although not included in the WHO guidelines for indoor air quality, the limit values for ambient air of 20 µg m⁻³ 24-hour mean and 500 µg m⁻³ 10-minute mean have been established ^[171], while the European Environment Agency^[172] sets 350 µg m⁻³ 1-hour mean and 125 µg m⁻³ 1-day mean as limit values for the protection of health.

• Diffusive samplers using triethanolamine. Different types of samplers are used, e.g., Palmes^[173] and Ogawa^[170]. Samplers are analyzed by IC. The LOD is in the order of 4–8 µg m⁻³ for 1 day of exposition, with accuracy of 15%. A particle filter can be additionally used to collect the samples. Dynamic range: 0.1–200 µg m⁻³. This technique is appropriate for long-term measurements.

• UV Fluorescence SO₂ Analyzer (e.g., Teledyne T100), following the reference method established in EN 14212:2012^[174] and ISO 10498: 2004^[175]. This monitor has a LOD of 0.4 ppbv and different working ranges from 50 ppbv to 20000 ppbv, with precision of 0.5%. This widely used for indoor and outdoor applications^[64] is very accurate and includes temperature and pressure compensation and also adaptive signal filtering to optimize response.

• A new analyzer has been recently developed for in situ detection of SO₂. The PITSA (Portable In-siTu Sulfur dioxide Analyzer) prototype is based on non-dispersive ultraviolet (NDUV) spectroscopy. According to some authors^[176], it could be a promising alternative to presently used compact and low-cost SO₂ monitoring techniques, over which it has a series of advantages, including an inherent calibration, fast response times (< 2 s to reach 90% of the applied concentration), measurement range spanning about 5 orders of magnitude and small, well-known cross sensitivities to other gases. Compactness, cost-efficiency and detection limit (< 1 ppmv, few ppbv under favorable conditions) are comparable to other presently used in-situ instruments.

5.4.2. Hydrogen sulfide (H₂S)

Hydrogen sulfide is an air pollutant causing malodor and commonly emitted by both natural sources (thermal and cold sulfur-rich springs, wetlands, manure, coal pits, and gas and oil reservoirs, volcanic and hydrothermal systems) and anthropogenic activity (sewage treatment plants, concentrated animal feeding operations, oil and gas refineries, geothermal power plants, paper mills and vehicular traffic).^[177] Indoor sources of H₂S vary by region and include: emissions from groundwater containing sulfide, vapor intrusion associated with muck, septic tanks, wastewater mains, or landfills in the vicinity of the residence, conversion of sulfate by sulfate-reducing bacteria in the water distribution systems, the chemical and/or biological reduction of sulfate in the water heater and infiltration from outdoor air where high levels of H₂S exist.^[178] People also produce reduced sulfur compounds by a variety of mechanisms (production in the gut and oral cavity) as a result of microbial metabolism. Methane thiol, dimethylsulfide and H₂S have been reported to be the dominant gases causing bad breath.^[45] According to WHO, the air quality guideline^[42] for H₂S is 150 µg m⁻³ (average value in 24 h), although it should not exceed 7 µg m⁻³ (average value in 30 min) to avoid significant odor annoyance due to its typical rotten-egg smell (detectable at 0.7- 42 µg m⁻³ depending on individual sensitivity^[179]).

• Passive samplers. Different passive samplers for collecting H₂S are commercially available. Radiello® cartridges are made of microporous polyethylene and impregnated with zinc acetate [Zn(CH₃COO)₂], then H₂S is chemisorbed by zinc acetate and transformed into stable zinc sulfide (ZnS) that is recovered by extraction with water. The sulfide reacts with N,N-dimethyl-p-phenylendiammonium ion in the presence of an oxidizing agent as ferric chloride in a strongly acid solution to yield methylene blue that is quantified by visible spectrometry at 665 nm.^[180] The LOD is 30 ppbv for 1 hour exposure or 1 ppbv for 24-hour exposure. However, it is worth mentioning that Venturi et al.^[177] conducted H₂S comparative measurements in air by passive samplers and a high-frequency analyzer (Thermo® 450i). The results showed that the concentrations of H₂S using Radiello® samplers were significantly higher than the average values measured by the Thermo® 450i analyzer during the Radiello® sampler exposure, especially when H₂S was <30 µg m⁻³.

Other passive samplers as those provided by Analyst® or IVL Swedish Environmental Research Institute are commercially available for collecting H₂S.^[8]

• Active sampling (Filter + Sorbent tube). The sampler consists of a Zefluor^TM PTFE prefilter (0.5 μm) and a sorbent tube with coconut shell charcoal, 400 mg/200 mg) according to the NIOSH method 6013^[181]. A flow rate range from 0.1 to 1.5 L min⁻¹ is recommended. The working range is from 0.6 to 14 ppmv (0.9 to 20 mg m⁻³) for a 20-L air sample. The upper limit of loading depends on the concentrations of H₂S and other substances in the air, including water vapor. High relative humidity (80%) increases the capacity of the sampler four-fold, relative to dry air. Sulfur dioxide is a positive interference, equivalent to H₂S by approximately twice the SO₂ concentration by weight. Methyl and ethyl mercaptans do not interfere. Later, the two sorbent sections are extracted separately, and the resulting sulfate ion is determined by means of IC. The LOD is 11 µg SO₄^2- per sample and the accuracy is ± 11%. The applicability of this method to indoor air is limited and it is only suitable for places where high concentrations of H₂S are expected.

• Portable H₂S gas meters. When concentrations of H₂S are expected to be high (hundreds of ppbvs) due to the proximity of a source, gas meters have been used to measure H₂S concentrations.^{[178,182,183]} Electrochemical sensors are usually used with these H₂S meters. The LOD is typically 5 ppbv with a range up to several ppbv. Jerome 631-X monitors H₂S using the change of resistance of a gold film to detect low levels of H₂S with a range from 0.003 to 50 ppmv in seconds and accuracy of ±0.003 ppmv at 0.05 ppmv.^[178,184]

• UV Fluorescence H₂S Analyzer (e.g., Teledyne T101). This kind of analyzer uses the proven UV fluorescence principle to measure H₂S at levels commonly required for ambient air monitoring. It is equipped with an internally mounted catalytic converter set at 315 °C to convert H₂S to SO₂. A switching mode alternately measures H₂S and SO₂ while showing both readings concurrently on the front display. This analyzer has a detection limit < 0.4 ppbv and a working range from 0-50 ppbv to 0-10 ppmv, with a precision of 0.5% of reading above 50 ppbv. It includes temperature and pressure compensation.^[185]

• Alternatively, non-dispersive ultraviolet (NDUV) spectroscopy analyzer (e.g., ABB Limas 11) and gas chromatography coupled with pulsed discharge helium ionization detector (GC-PDHID) (e.g., Agilent 7890B) have also been used to quantify H₂S. These instruments have been used to quantify H₂S reference gas mixtures^[185] but their use in not extended to analyze ambient air. ASTM D4323 - 15 Standard specifies a test method that covers the automatic continuous determination of H₂S in the atmosphere in the range from 1 ppbv to 3000 ppbv by rate of change of reflectance.^[186]

The methods for sampling and analysis of S-containing compounds along with their main characteristics are compiled in Table 6.

Table 6. Instrumentation and measurement techniques for the determination of S-containing compounds in indoor air.

Pollutant	Kind of Instrument	Mode of operation	Standards	Analysis Technique	Brand (e.g.,)	Accuracy/precision	LOD	Operation range	Ref.
S-containing compounds
Sulfur dioxide (SO2)	Automatic Analyzer	OnM	EN 14212 ISO 10498	Ultraviolet Fluorescence (UV-Flu)	Teledyne T100	Precision 0.5% of reading above 50 ppbv	0.4 ppbv	Min:0–50 ppbv; Max: 0 − 20000 ppbv	^[^64^]
	Diffusive sampler with triethanolamine (DS-TEA)	OffM	n.a	Ion chromatography	IVL (Palmes) Ogawa	Accuracy 15%	4 (IVL) − 8 (Ogawa) µg m^-3 (1 day)	0.1–200 μg m^-3	^[^170^,^173^]
Hydrogen sulfide (H2S)	Diffusive sampler (DS)	OffM	n.a	UV/ Visible Spectrophotometry for 24 h exposure.	Radiello® Analyst® IVL	Uncertainty 8.7 % over the whole exposure range	30 ppbv for 1 h exposure or 1 ppbv	–	^[^8^,^177^,^180^]
	Active sampler with-Filter + Sorbent tube. (AS)	OffM	NIOSH 6013	Ion chromatography- conductivity	SKC	Accuracy 11.8 %	11 µg per sampler (0.55 mg m^-3 in 20 L)	0.6–14 ppmv (0.9–20 mg m^-3 ) for a 20 L air sample	–
	Portable H2S gas meters (PGM)	OnM	n.a	Gold film sensor technology	Jerome 631-X	Accuracy ±0.003 ppmv at 0.05 ppmv, precision 5%	<0.003 ppmv	0.003–50 ppmv	^[^178^,^187^]
	Automatic Analyzer	OnM	n.a	UV fluorescence (UV-FLuo)	Teledyne T101	Precision 0.5 % of reading above 50 ppbv	< 0.4 ppbv	0–50 ppbv to 0–10 ppmv	^[^185^]

n.a: not available.

^aOnM: on line monitoring; OffM: offline monitoring.

6. Selection criteria for the sampling sites

If many buildings or indoor spaces are candidates for a study, it is important to define criteria for the selection of the sampling sites. Buildings or indoor spaces to select for the assessment plan should be as representative as possible of the entire group of candidate buildings/indoor spaces and or representative of the exposure to which population to study is exposed. In general, the definition of the criteria for selecting sampling sites is related to the objective of the measurement.

As an example, in studies assessing pollutants in different schools, the criteria have been established for selecting at least three classrooms per school building. In addition, the recommendation is that classrooms should be selected based on how representative they are of the school building and how routinely and continuously they are used during school hours; also, their location on different floors, orientation, how long they have been used, are other factors.^[24,92] Building recruitment usually occurs on a voluntary basis, requiring the consent of the respective building manager(s). In the European OFFICAIR project where modern offices were evaluated in different countries, the selection of buildings was carried out by requiring the following criteria: new or recently retrofitted building (less than 10 years), building operating with usual activities at least one or two years before starting the study, and building with no major renovation before starting the study.^[54] In homes, for instance, it is common to have to choose which rooms people spend more time, typically living rooms and bedrooms^[27], or kitchens if the interest is to characterize the chemistry in the cooking activity. Logically, long-term exposure to emission sources such as building materials, can be better characterized in the room(s) where people spend a higher percentage of their time.^[47] But in reality it is sometimes difficult to identify the precise location for the installation of samplers/equipment before starting the works in the field, because in some cases that relevant information is missing.

7. Criteria for sampler location

If possible, ISO 16000-1 recommends placing samplers/equipment in the center of the room.^[47] In addition, the sampler should be placed 1–2 m from the wall and 1–1.5 m from the floor to be representative of the typical breathing zone of the occupants. In some situations, due to specific circumstances, the sampler can be at specific locations. For example, when measuring NO₂ in the indoor air generated from the use of a cooking appliance, a concentration gradient is generated due to thermal movement of air resulting in lower NO₂ concentrations below the height of the gas cooker than those observed above it (ISO 16000-1^[47]). It is then recommended to sample in various areas, at different heights. On the other hand, if the monitored space is large (e.g., dining rooms, big offices, auditorium, gyms) or if a non-homogeneous distribution of the indoor air is expected, it is recommended to collect comprehensive information and subdivide the area in order to accurately define a relevant number of sampling locations. For the specific case of monitoring CO₂, a single sampling point for rooms of 50 m² is reported to be adequate, but for larger rooms more sampling points are recommended in order to determine all concentration gradients, especially if the efficiency of ventilation is being investigated. In this case, a separation of 1.5–2 m from people is necessary in order to avoid the direct influence of the air breathed by people nearby.^[61] In fact, for achieving concentrations of pollutants representative of the air quality of the whole room, the recommendation of placing samplers at a distance higher than 1.5–2 m from declared pollution sources has to be considered. In addition, sampling locations near to heating and ventilation systems (or other sources of heat) should be avoided. For an indoor environment served by mechanical ventilation, it is recommended to take into account that the sampled volume during 1 h should be less than 10% of the air volume introduced by the respective ventilation system. If the ventilation rate cannot be measured (or this information is not available), the sampled volume per hour should be less than 10% of the volume of the studied room.^[47,48]

8. Timing, duration and frequency of sampling

The concentration level of a pollutant in indoor air is likely to present significant temporal variations. Factors including the age of the building, activities conducted inside the building, the season and the time of day when sampling can influence the result of indoor air measurements. Therefore, it is paramount to consider the possibility of temporal variation of pollutant concentration in the study design. Important information to take into account when choosing the timing of sampling is data related to identify relevant patterns of ventilation, pollution sources, building use and occupants activities, temperature and relative humidity.^[47] For example, if a pollutant source with a continuous emission is known to exist indoors, variations on the amount of fresh air introduced in the room (ventilation) will govern temporal fluctuations in the indoor concentration of the emitted pollutant(s).

As indoor concentrations and their variation throughout the time may also be influenced by the level of outdoor air pollution (e.g., O₃ and NO₂), it can be necessary to collect information about the emission from outdoor sources. Emissions occurring in the outdoor air typically reach the indoor air with a certain time lag, depending on the existing air exchange rate.^[49] It is recommended to monitor the ambient outdoor air, near the ventilation inlet.^[51]

If there is a continuous indoor source of a pollutant, fluctuations in ventilation rates will vary its concentration. For short-term monitoring, it is important to conduct the sampling under representative ventilation conditions, in accordance with typical use and maintenance of the indoor spaces.

The term "duration of sampling" refers to the time period over which the sampler collects a sample, addressing the questions of sampling period. The duration of sampling must be established depending on the objective of the measurement, the nature of the indoor pollutants, typical exposure duration, the potential health effects of the monitored pollutants (acute/short-term or chronic/long-term), the emission characteristics of the sources (and other factors influencing the concentration levels) and the detection and quantification limits of the analytical method. As referred above, short-term measurements should be used for compounds causing acute effects and are usually carried out through active sampling, while long-term measurements are appropriate for compounds causing chronic effects on health and are carried out generally by means of passive sampling. However, continuous monitoring with a direct-reading instrument can also be used also if long-term averages are required (as in the case of CO₂) or if the identification of temporal fluctuation patterns or of peak concentration periods are part of the plan. As shown in Table 2, long-term sampling is usually used to determine the levels of pollutants under normal condition of occupancy or real-life conditions while short-term sampling is used for collecting pollutants emitted from temporary sources (such as during cleaning activities) or temporary exposures (e.g., users of indoor swimming pools that typically uses the facilities during 1 to 2 hours^[188]). More information about duration of sampling is available in ISO 16000-1.^[47]

As an example, for studies aiming the assessment of health risks to hazardous chemicals in indoor air in public settings for children (schools, kindergartens, and daycare centres) using passive samplers, two different approaches can be considered for the sampling duration: i) 24-h measurements for five school days (24 h/5 days); ii) measurements during periods when children are present indoors (5-8 hours every day of a school week; 5-8 h/5 days).^[92] The first approach can result in an overestimation of exposure for certain pollutants which may accumulate when windows are closed or mechanical ventilation is switched off, but can also underestimate the exposure to other specific pollutants that are generated during the period of occupation of the rooms. The second approach requires that samplers be capped at the end of each school day and uncapped the next morning. In this last case, the LOD is higher and this strategy requires additional resources if technical staff must cap and manually uncap the samplers every day. For that, the first option using passive sampling from Monday morning to Friday afternoon has been preferable for measuring inorganic compounds such as NO₂ or O₃.^[24,92] However, for other compounds such as CO that can cause acute health effects or using CO₂ to calculate ventilation rates it is recommended to continuously measure using monitors or dataloggers. This would also apply to studies with a monitoring strategy for indoor gaseous pollutants in office building, with sampling from Monday morning to Friday afternoon by means of passive sampling to collect inorganic compounds such as NO₂ and O₃.^[189]

The sampling frequency is defined as the number of samples taken over a given time interval (e.g., one year). As occupant activities, ventilation characteristics and pollutant dynamics may vary from day to day and/or show a seasonal variability, no single sample taken at one particular time can give a reliable indication of the overall exposure.^[49] Significant seasonal variability in concentrations is recognized for some indoor air pollutants. Typically natural ventilation is poorer during the cold season, due to thermal constrains related to opening windows; this may lead to increased concentrations of indoor pollutants. In turn, peaks of pollutant concentrations related to either higher outdoor temperature or low air exchange rate, or a higher infiltration of outdoor air pollution can also be observed during summer.^[47]

Lastly, the duration of a sampling campaign and the number of samples collected should be adjusted to realistic time, costs, and technical staff-related limitations.

9. Concurrent sampling in the outdoor environment

The presence of traffic, industrial and agricultural-related air pollution sources in the vicinity of a building can adversely affect indoor air quality due to either air infiltration or air intake from the mechanical ventilation system. Therefore, pollutant concentrations in the outdoor air surrounding the building should be included in calculations and analyses of the indoor air quality in order to investigate the contribution of emissions from outdoor sources. It is common to calculate the indoor to outdoor (I/O) concentration ratio or concentration difference (I-O) rather than the indoor concentration only. Measurements of the outdoor air should be, whenever possible, conducted simultaneous to indoor air sampling. In addition, it should be performed at the same level or floor of the indoor sampling site when natural ventilation or air infiltration are important because vertical concentration gradients in outdoor air pollutants may occur. Samplers should be located close to the building but at one meter away from the wall. If the building has a heating, ventilation and air conditioning (HVAC) system, sampling should be done near to the air intake.^[47]

10. Quality assurance and quality control (QA/QC) program

The Data Quality Objectives (DQO) for each pollutant should be clearly defined at the beginning of the project. Quality Assurance and Quality Control (QA/QC) procedures are used to ensure you achieve your DQOs. They refer to activities incorporated into the design of an indoor air assessment plan with the aim of ensuring and monitoring the quality of the results obtained. A section on QA/QC is usually included in each standard (international, European, etc.) for the compound of interest. Minimum criteria for QA/QC practices are defined in each standard. These minimum DQOs should be adequate for most applications, but can be optimized by the researcher, depending on the needs of the specific study.

QA refers to management activities involving planning, implementing, documenting, assessing and reporting that assure that data are of known and documented quality while QC refers to technical activities that measure whether and how well the goals established in the quality assurance component were met. The QA/QC program covers mainly the sampling and the analytical method.

The general QA/QC procedures briefly described below are very important to obtain analytical result of quality, reliability and consistency and to avoid the need for repetition of sampling and analysis.

10.1. Sampling quality control

The parameters regarding sampling quality control define how to conduct replicate samples, laboratory and field blanks, and series sampling. Table 7 summarizes quality control parameters, their description and purpose, and the frequency of QA/QC samples collected during the study.

Table 7. Quality control parameters that must be included in indoor air sampling studies.

Quality control parameter	Description	Frequency	Purpose
Replicate samples	Two or more samples collected simultaneously.	At least 10 % replicate samples per each batch of samples.	To assess repeatability of sampling method.
Laboratory blank	A sampler that is not handled in the field and is stored in the laboratory.	At least one laboratory blank per batch of samples.	To assess contamination from storage in the laboratory.
Field blank	A sampler which accompanies sampler to field and back to laboratory.	At least one field blank per batch of samples.	To assess contamination from transportation of samplers to and from the field.
Series sampling (with active sampling)	Two or more cartridges attached in series.	One series sampling set-up per sampling location with highest expected concentrations.	To detect “breakthrough” and loss of sample.
Backdiffusion (with passive sampling)	Backdiffusion happens when the adsorbed mass of analyte is far from the maximum amount allowed by the adsorbing medium capacity.	It is only necessary to follow the instruction given by manufactures regarding sampling time and concentration of pollutants.	To improve confidence in measured concentrations.
Pump flows (active sampling)	Verification of pump flows by use of calibrated flowmeters. The flow rate should be within a range of ± 5%.	At the start and end of each sampling period.	The start and stop flow rates are averaged to obtain an average flow rate.

10.2. Storage stability and sample handling

The requirements for the holding time and the sample preservation method, as recommended in the respective reference sampling methods used, when available, should be followed in order to maintain the sample stability. The analyte loss, gain or degradation could lead to inaccurate results. After each day of sampling, samples must be shipped to the laboratory for analysis. Samples are only stable for a certain period of time and must be analyzed within this period. Also, they may require transportation in portable fridges to avoid their degradation due to high temperatures. Understandably, this section is only applicable to offline monitoring with passive and active samplers employing cartridges, and is not relevant for continuous monitoring analyzers.

10.3. Uptake rate for passive samplers

ISO standards, when available, such as for NO₂^[110], describe the uptake rate that is recommended, depending on the type of passive sampler used. On the other hand, for samplers that are commercially available, the manufacturer defines the uptake rate for sampling the compound of interest. Note that, since the uptake rate can be temperature dependent, it is always necessary to measure the temperature during the period of sampling. This allows to correct for the sampling temperature when calculating concentration of air pollutant(s) according to the manufacturer instructions.

10.4. Analytical quality control

Many analytical methods are complex and usually involve multi-step procedures. Several analytical parameters such as precision, accuracy, linearity, LOD and LOQ must be calculated to ensure the application of the selected method during the analysis of samples. Table 8 summarizes the parameters for the analytical quality control and how to calculate them. More information about analytical parameters is available in a recent WHO report on sampling methods.^[92] For more information about precision and accuracy measurements, see the ISO 5725 series of standards.^[193]

Table 8. Parameters for the analytical quality control.

Quality control parameter	Description	Sampling	Analysis	Calculation
Precision^a	Degree of agreement between independent measurement obtained in well-defined conditions (ISO 5725-3:1994^[^190^]). Precision measures the variability associated with repetitive analyses of the same sample.	Data variability may be evaluated by taking field duplicates or multiple air samples at a location.	Data variability may be evaluated with Certified Reference Materials (CRM) and/or laboratory-fortified blanks with analytes at a known concentration (3–5 samples are recommended). Different levels of concentration can be used.	$\mathit{RSD}\left(\mathrm{\%}\right)=\frac{SD}{\overline{X}}x100$ SD: standard deviation $\overline{X}:$ average value
Accuracy^a	It is the closeness of a measurement to the true value (ISO 5725-1:1994^[^191^]). Accuracy can be evaluated with the calculation of desorption efficiency or recovery.	–	Accuracy may be valuated with CRM and/or laboratory-fortified blanks with analytes at a known concentration (3– 5 samples are recommended). Different levels of concentration can be used. Values for acceptable accuracy are define in the international Standards or must be defined if they are not available.	$DE=\mathrm{ }\frac{\overline{X}}{X_{\mathit{ref}}}$ DE: Desorption efficiency $\overline{X}$= average amount recovered $X_{\mathrm{r}ef}$= Spike amount
Linearity (Instrument calibration)	Calibration involves analyzing a series of standards which contain the analytes of interest at several different concentrations.	–	The analytical system must be initially calibrated using a minimum of 5 standard concentrations in the range of expected concentrations in the samples in order to determine the linear response of the target analytes. Continuing calibrations are mid-level calibration standards containing all of the compounds of interest designed to assess the accuracy of an instrument. They are usually run on an on-going (at least daily) basis. It is recommended that the lowest instrument calibration standard be run at the LOD.	By means of a plot of instrument response versus standard concentration. The equation of the plot or the response factor can be used to calculate the concentrations.
Limit of detection (LOD)	It is the smallest measure that can be detected with reasonable certainty for a given analytical procedure.^[^192^] It presents different definitions	–	It can be calculated: from the blank samples, from the signal to noise ratio and from the calibration curve at low concentrations. Laboratories should provide information about the method used to estimate LOD and LOQ. It is recommended that a LOD study be performed twice a year to verify that the values obtained are below the lowest standard used for reporting or for each batch of samplers in case of passive or active samplers using adsorbents. They are usually expressed in units of µg m^-3 for both outdoor and indoor air	$\mathit{LOD}\mathrm{ }or\mathrm{ }\mathit{LOQ}=\frac{k\mathrm{ }x\mathrm{ }SD}{b}$ k: is a factor of 3 (for LOD) and 10 (for LOQ) SD:is the standard deviation of the blank signal, the standard deviation of the ordinate intercept or residual standard deviation of the linear regression b: the slope of the linear regression
Limit of quantification (LOQ)	It is the lower limit of precise quantitative measures.

^a The samples are extracted and analyzed following the same procedure used for field samples

11. Conclusion

The most studied inorganic pollutants in indoor air are carbon dioxide (CO₂), carbon monoxide (CO), nitrogen oxides (NO_x) and ozone (O₃). CO, NO_x and O₃ are mainly outdoor pollutants while CO₂ is mainly, in the absence of combustion processes, a metabolite emitted by occupants. Other inorganic compounds that can be present in indoor air although at lower concentration are NH₃, HNO₃, HNCO, SO₂, H₂S and H₂O₂.

Determining the concentration of pollutants in indoor air is critical for many goals such as assessing exposure and health risk, characterization of indoor sources and contribution of outdoor sources, determining temporal and spatial trends in indoor air quality and assessment of compliance with indoor air quality guidelines. Despite the large number of studies about monitoring indoor air quality, our understanding of the composition of indoor air as well as the chemical transformation driven by oxidants is limited. Redressing this imbalance, the field of indoor air chemistry is growing rapidly.

Nevertheless, studies on determining indoor air concentrations and characterizing emissions and related chemistry are limited for some of the inorganic pollutants explored in this publication, especially for H₂O₂. Reactions between O₃ and terpenes can be an important source of indoor H₂O₂ at high indoor O₃. Terpenes are used as odorants in a variety of consumer products and are also among the active ingredients used in many cleaning products. Aqueous H₂O₂ is the active ingredient in several household non-bleach cleaning products and in some air cleaners. It can also be an important indoor oxidant under certain conditions and is directly linked to the hydroxyl (OH) and hydroperoxyl (HO₂) radical budget. In spite of this importance, oxidation processes, emissions and chemistry from H₂O₂ are poorly characterized in indoor air.

In summary, this report describes the different methods for sampling and analysis of a list of relevant inorganic pollutants, reviewing their advantages and drawbacks. In general, passive sampling is the preferable method to assess the long-term exposure and health risk because they are cheap and easy to use, allowing the deployment of several units in different rooms, although laboratory analysis is required. Active sampling, and especially continuous monitoring, provides information with high temporal resolution not achieved with passive samplers; however, some of these techniques are expensive and bulky and as consequence are not suitable for monitoring multiple places simultaneously. Some of the instruments described in this report have been used only in outdoor air but all, due to their characteristics, can be potentially deployed in indoor environments.

Many factors must be considered when planning a monitoring campaign including limitations due to cost, expert staff, noise and size of instruments, amongst others. The newest monitoring systems, based on reliable low-cost sensors would provide much improved spatial and temporal resolution data in indoor environments, broadening our knowledge on the field of indoor air chemistry. Nevertheless, further studies are needed to properly validate these innovative technologies by comparison to the existing reference and equivalence methods.

Acknowledgements

Thanks are due to the European Cooperation in Science and Technology (COST) through the financial support to COST Action INDoor AIR POLLution NETwork (INDAIRPOLLNET), CA17136.

Disclosure statement

No potential conflict of interest was reported by the authors.