Diffusive samplers for ambient air quality monitoring

Abstract

Air pollution remains a matter of concern for governments, regulators and the general public. Diffusive, or passive, samplers provide a cost-effective, versatile and spatially dense method to sample air pollution that does not require the significant investment needed to operate traditional air quality monitoring instrumentation. This Technical Brief describes the theory and operation of diffusive samplers and how their sampling rates may be calibrated. It also discusses considerations for the deployment of diffusive samplers in outdoor environments and provides examples of their practical use in sampling different air pollutants.

---

Introduction

Diffusive samplers – also called passive samplers – are commonly used in environmental and occupational hygiene measurement to sample a variety of organic and inorganic compounds in matrices such as air, water and soil. In particular, they are regularly employed to sample ambient air, where they assist in determining the average concentration of a number of trace, and potentially harmful, pollutant gases over the period that they sample for – often called the exposure period – typically ranging from a week to a month. Diffusive samplers offer a cheaper, less obtrusive and more spatially dense alternative to continuous monitoring methods at fixed monitoring locations that necessitate expensive analysers requiring power, costly environmentally controlled housings, regular human intervention and frequent maintenance (they also offer similar advantages for workplace air monitoring, particularly for personal sampling, where the need for sampling pumps is eliminated). However, unlike continuous monitoring methods that actively sample the air and then analyse its composition in situ and in real time, diffusive samplers acquire their air sample passively by diffusion and then, in a separate step, the diffusive sampler is recovered and its collected sample analysed, usually at a central laboratory.

Theory of operation

Diffusive samplers exist in many different designs, including tube (both axial and radial) and badge types. The most common axial tube format consists of a steel, acrylic or quartz cylindrical tube containing a sorbent. One end of the tube is open, and the other end is closed. At the closed end there is a sorbent, which the analyte(s) of interest either chemically reacts with, or physically adsorbs onto, depending on the mechanism of operation. The sorbent material is chosen so that, as far as possible, only the desired pollutant is sampled. Following the required exposure period, the concentration of the analyte is determined using a suitable quantification method. This generally involves solvent extraction of the captured sample followed by spectrophotometry or ion chromatography, or thermal desorption of the sample followed by gas chromatography. Diffusive samplers therefore allow determination of the mass of the target analyte(s) collected, which can then be used to determine the average concentration of the analyte in the sampled air over the exposure period.¹ The operation of diffusive samplers may be described by Fick’s first law of diffusion and the amount of analyte, M (in mol), collected over a time, t (s), by the diffusive sampler is given by:where U is the diffusion transport rate (mol s⁻¹), often referred to as the sampling rate or uptake rate, D is the molecular diffusion coefficient of the analyte under consideration (m² s⁻¹), A is the cross section of the diffusion path (equal to the cross section of the diffusive sampler) (m²), L is the total length of the diffusion path (equal to the length of the diffusive sample up to the sorbent) (m) and c₀ is analyte concentration in the sampled air (mol m⁻³). It is assumed that the concentration of the analyte of interest decreases along the length of the diffusive sampler as the analyte is sampled and is zero at the sorbent. This is demonstrated diagrammatically in Fig. 1.

Fig. 1 Operation of a tube-type diffusion sampler of length L and area A, with the open end at the bottom and the closed end at the top where the sorbent is placed. The analyte of interest, at a concentration c₀ immediately outside the tube, diffuses into the open end of the tube, then along its length, and is captured by the sorbent at the closed end of the tube. A concentration gradient is established such that the concentration of the analyte decreases across the length of the tube, reaching zero at the sorbent.

Theoretically, the sampling rate may be calculated using eqn (1), provided the dimensions of the diffusive sampler are known, along with literature values for the diffusion coefficient, D. However, this approach is complicated because D is temperature (T) and pressure (p) dependant and knowledge of this dependency is not always well established. The analyte concentration, c₀, may also be pressure and temperature dependent. In theory and, from the ideal gas law, c₀ ∝ T⁻¹p. This suggests an overall dependency of . However, rarely is this temperature dependence ideal, not least because the atmosphere does not behave as an ideal gas, plus there are other possible confounding factors, such as varying literature values for D. Moreover, effects caused by the small relative changes in temperature and pressure across the long sampling times involved are generally minor in comparison with analyte concentration variations over the same period and the overall measurement uncertainty. Furthermore, eqn (1) does not accommodate effects such as analyte losses to other parts of the sampler, the collection efficiency of the sorbent not being 100%, and other non-ideal behaviour associated with surface effects close to the sorbent, and edge effects near to the inlet of the sampler. Diffusive samplers may also show sensitivity to humidity or cross-sensitivity to other ambient gases because the sorbent used is rarely perfectly selective. It is usually the case that sampling rates calculated from first principles introduce biases into measurements² and therefore an experimental determination of the sampling rate is recommended. This is a time-consuming process, although it usually only requires determination of the sampling rate for a particular design of diffusive sampler – often known as ‘type testing’ – and not for each individual sampler. The latter approach would in any case be impractical since most diffusive samplers are single use. Experimental determination of the sampling rate involves either repeated exposures of diffusive samplers to controlled atmospheres in the laboratory over a range of relevant analyte concentrations simultaneously with reference analytical instrumentation,² or co-location of the samplers with reference instrumentation in the field for sufficiently long periods of time. A plot of analyte amount collected against the product of analyte concentration multiplied by the exposure time gives a gradient that is nominally equal to in eqn (1), but is in effect an empirically determined constant, k (m³ s⁻¹), describing the operation of that particular sampler type, such that:

M = Ut = kc₀t (2)

The dependence of k on temperature may also be determined experimentally under controlled conditions so that the results of real samples may be appropriately corrected for the average temperature measured locally during sample collection. To improve consistency of measurement, the calculation of results is often standardised. For example, nitrogen dioxide (NO₂) measured using Palmes tubes (see below) in the UK assumes a relationship of D ∝ T^1.81 with literature values of D, known at 273 K, being corrected to the mean UK temperature of 284 K.³

Aside from ‘axial’ diffusive samplers where gases diffuse parallel to the cylinder axis, other designs also exist, and these provide different benefits. One of these is a ‘radial’ diffusive sampler where gases diffuse parallel to the cylinder radius. These are still cylindrical in design but have a smaller diffusion path and larger cross section when compared with axial samplers,⁴ as shown in Fig. 2. This allows for greater sampling rates and therefore shorter exposure periods are possible to sample a measurable quantity of analyte.

Fig. 2 Diagram of a radial sampler where the sorbent is placed inside a porous cartridge in the centre, surrounded by a microporous sintered cylindrical tube, which allows diffusion to take place through it.

Sampling in outdoor environments

When diffusive samplers are used in outdoor air environments, there are further considerations. Diffusive samplers are always exposed vertically with their open end facing downwards to avoid direct interference from rain, atmospheric deposition and bird fouling. Some samplers may also have meshes placed over their open end, designed to stop the undue influence of wind on the sampling rate whilst not adversely affecting the cross section of the diffusion path. To provide further protection from wind, rain and direct sun, diffusive samplers are often housed in ‘bird box’ type enclosures with an open floor. These housings also discourage undue attention from passers-by when in public settings! Samplers are always sited far enough away from large surfaces, such as walls and buildings, so that their sampling rate is not affected, and they sample a representative ambient composition, and not stagnant air. Lampposts, road sign poles and metal mesh fences are usually good locations.

Owing to their advantages over expensive instrumentation for continuous ambient air monitoring, diffusion tubes have found a variety of applications in sampling air quality pollutant gases where dense spatial resolution is an advantage, but high time resolution is not a consideration. Most diffusion samplers are exposed for periods of between a week and a month, in order to acquire enough analyte so that a fit-for-purpose overall measurement uncertainty may be achieved – usually less than 20% (relative) at the 95% confidence level. Given that most air quality legislation targets the maximum allowable annual average concentration of pollutant gases, diffusive samplers producing long-term averages still find significant use in assessing compliance with regulation and population exposure. It is also the case that certain diffusive samplers can provide measurements that meet the data quality objectives for fixed reference measurements for both EU and UK air quality legislation. Since these diffusive samplers are also cheap to use and easy to deploy, they are also very useful to identify areas of high pollutant concentration, where it might be important in future to site fixed air quality monitoring stations with more expensive instrumentation. The long-term averages produced by diffusive samplers mean they are not suitable for identifying short-term pollution events, providing air quality warnings, measuring maximum peak concentrations, or elucidating atmospheric chemical processes.

Practical examples of air pollution monitoring

Diffusive samplers for NO₂ were first developed by Edward Palmes⁵ and were intended to measure the average NO₂ concentrations that miners were exposed to during their shift in underground mines. Often referred to as Palmes diffusion tubes (PDTs), they have since been widely deployed to measure NO₂ concentrations in urban air environments. PDTs consist of an acrylic tube with a capped end (top) and an open end (bottom). A stainless-steel grid, coated in the sorbent compound, triethanolamine (TEA), is placed at the capped end of the sampler. Once NO₂ reaches the TEA-coated grid, it reacts to form nitrite (NO₂⁻) and is irreversibly trapped. Following exposure, the mass of the collected nitrite is determined by an analytical procedure, known as the Griess test, in which a purple azo dye consisting of a solution of naphthyl-1-ethylene diamine dihydrochloride (NEDD), sulfanilamide and orthophosphoric acid is added to the PDT. The nitrite dissolves in the dye and the absorbance of the resulting solution at 540 nm is then measured with a UV-visible spectrophotometer calibrated with azo dye solutions to which a range of known concentrations of nitrite have been added. Nowadays, NO_x (NO_x = NO + NO₂) diffusive samplers are also becoming more common. They are similar in design to the Palmes NO₂ sampler, but they additionally contain an oxidant that converts nitrogen monoxide (NO) into NO_2. This allows both NO and NO₂ (i.e., total NO_x) to be captured by the sorbent. Therefore, by co-exposing NO₂ and NO_x samplers, the concentration of NO can also be determined.

Diffusive samplers are also regularly used for volatile organic compounds (VOCs) and ammonia (NH₃). Ammonia diffusive samplers are generally similar in design to the NO₂ samplers. Ammonia that is collected can be extracted using deionised water and measured as ammonium ions, NH₄⁺, using ion chromatography or an ammonia flow injection system. VOC samplers are generally axial in design but are packed with sorbent along their entire length. The sorbent, however, is not selective for just one species and therefore many VOCs are captured by physisorption. Following exposure, the VOCs are removed from the sorbent by either thermal desorption (with the additional analytical considerations that brings^6,7) or solvent extraction, and analysed using gas chromatography (GC). The sorbent may then be reused.

 

This Technical Brief was prepared by Richard J.C. Brown, Valerio Ferracci and Nicholas A. Martin (National Physical Laboratory, UK) on behalf of the Analytical Methods Committee and was approved by the AMC on 17 April 2025.

References

1. R. H. Brown, J. Environ. Monit., 2000, 2, 1–9.
2. N. A. Martin, V. Ferracci and N. Cassidy, et al. , Atmos. Environ., 2019, 199, 453–462.
3. Department for Environment Food & Rural Affairs, Practical Guidance: NO₂ Diffusion Tubes for LAQM, https://laqm.defra.gov.uk/air-quality/air-quality-assessment/practical-guidance/, accessed April 2025.
4. O. Motta, R. Cucciniello, R. La Femina, C. Pironti and A. Proto, Talanta, 2018, 190, 199–203.
5. E. D. Palmes, A. F. Gunnison, J. DiMattio and C. Tomczyk, Am. Ind. Hyg. Assoc., 1976, 37, 570–577.
6. I. Pengelly, Anal. Methods, 2020, 12, 3425–3428.
7. I. Pengelly, Anal. Methods, 2021, 13, 2345–2348.

---