A challenge associated with assessing the risk posed by emerging atmospheric pollutants is often the lack of historical data available for identifying and quantifying such pollutants so that trends in concentrations can be determined retrospectively. While opportunities do not exist for gaseous pollutants, archives of air filters could provide the records needed in the case of particulate matter, as long as the selected filter material is suitable for the analytical methodologies proposed. Thus, as emerging atmospheric pollutants are identified, historical filters could be data-mined to assess whether the specific pollutant of interest was present in the ambient air, and, if so, in what quantity, as well as how it varied temporally.

New Zealand’s Institute of Geological and Nuclear Sciences (www.gns.cri.nz) maintains a carefully logged, dated, and stored filter archive of air particulate matter samples from across New Zealand (National Air Particulate Speciation Database DOI: https://doi.org/10.21420/R58Q-FZ78). The database dates back over 20 years and consists of over 40,000 samples, including a long historical record captured at various sampling locations across the Auckland region (Talbot and Lehn, 2018). The filters that form the archive were originally sampled and analysed using non-destructive techniques (Ion Beam Analysis or X-ray fluorescence spectroscopy) to extract elemental compositional data for use in receptor modelling and apportionment of the sources of urban air particulate matter pollution (Davy et al., 2012; Davy et al., 2017a). The filters, primarily 47 mm diameter polytetrafluoroethylene (PTFE) or polycarbonate membrane filters, have since been stored individually in sealed petri dishes located in cabinets held at room temperature. Thus, the filters remain in-tact and suitable for reuse.

Over recent years there has been increased awareness that fibre-like particles might be present in the air across Auckland. This has come to attention due to work regarding silicas and, more recently, fibrous zeolites, with specific attention paid to erionite. Erionite is a Category 1 carcinogen (IARC, 1983), with exposure linked to malignant mesothelioma and other lung cancers in regions where it has been found in disturbed bedrock (Dogan et al., 2006; Carbone et al., 2011; Gualtieri et al., 2016) and where it has been found to be present on the near-surface dust and soil (Van Gosen et al., 2013; Harper et al., 2017; Beaucham et al., 2015).

Zeolites, including fibrous erionite, have been identified in several locations in the Miocene Waitemata Group sediments (Sameshima 1978), exposed across much of the Auckland urban and suburban regions. For example, in the upper Waitemata Harbour at Riverhead (Davidson and Black, 1994; Patel and Brook, 2021), fibrous erionite has been identified in riverbank exposures underneath a thin veneer of Quaternary sediments, as well as in rock outcrops in shore platforms in the intertidal zone. Erionite has also been identified in the Miocene Waitakere Group volcanics at Te Henga Quarry −9 km to the west of Henderson, in the west of Auckland (Brook et al., 2022). The quarry was operational until 2015 and used for aggregate minerals, as well as in a cutting at the nearby Waitakere Railway Station (Figure 1).

Thus, erionite, along with other fibrous silicas and zeolites, can be considered to be emerging contaminants in Auckland. Their importance in this specific context is especially relevant given that the risk of their disturbance has increased in recent years with Auckland’s rapid urban expansion across previous greenfield sites (Stats NZ, 2021). Moreover, there has been increased awareness recently about the risk posed by fibrous mineral forms including silicas (Otsuki et al., 2016; Sanchez et al., 2009) and zeolite minerals such as erionite (Brook et al., 2020; Patel and Brook, 2021). In this study, the GNS filter archive is used to investigate its potential to reconstruct the historical evolution of fibrous silicas and zeolites in the air across Auckland.

An important limitation is that there is currently limited literature on methods for identifying and quantification of erionite on filters collected from air sampling (Beaucham et al., 2015; Carbone et al., 2011), with no apparent standard methods followed. However, the literature suggests that filter measurement techniques encompassing microscopic analytical methodologies similar to those used for asbestos quantification (such as those provided by NIOSH 7400 and 7402 and ISO 13002) are potentially the most suitable (NIOSH et al., 1999). The instruments proposed include the scanning electron microscopy (SEM), an instrument often used in the analysis of air particulate matter (González et al., 2018; Li and Shao, 2009). SEM techniques have been used in numerous studies in both quantitative and qualitative analysis of filter samples, including in the analysis of particle size, morphology and composition to infer sources (Tripplett et al., 2013). In addition, when combined with energy-dispersive x-ray spectroscopy (EDS), this method provides information about both the morphology and the chemical composition of the particles present on the filter (Dogan, 2003).

This study investigates whether a historical archive of air filters, using standard PTFE or polycarbonate membrane filters, can be used to identify and quantify the presence of an emerging contaminant, in this case erionite, using a combination of phase contrast microscopy (PCM) and SEM, coupled with EDS microscopic techniques. The result of the study highlights the value of retaining historical archives for the purpose of retrospective air pollution analysis.

For the initial screening process, three sampling sites were chosen to represent areas typically affected by substantially different airborne pollutants. These locations were Patamahoe (rural countryside location away from dwellings and busy roads), Queen Street (a central Auckland main shopping street), and Henderson (a residential/light industrial suburb in the West of Auckland, with sampling carried out at a site set alongside a major arterial road) (Figure 2). The Henderson site was established in 1997 as part of Auckland’s air quality monitoring network and is of particular interest, as it is located within 5 km of known locations of erionite-bearing rock strata (Patel et al., 2021).

Filter-based particulate matter samples were collected at each of the chosen sites on a 24-h time-integrated sampling basis using reference monitoring methods. Sampling for chemical speciation analysis began at Queen Street and Henderson from 2006 onwards, and for Patamahoe, sampling was conducted during 2010 only.

Particulate matter elemental compositional analysis [Na to U by accelerator-based Ion Beam Analysis (Barry et al., 2012)] and source apportionment data (receptor modelling by positive matrix factorization (Davy et al., 2017a; Sun et al., 2020) were used as an initial tool to identify the most suitable filters for reanalysis. The mass content of each element upon each filter was obtained from the historical data archives (National Air Particulate Speciation Database DOI: https://doi.org/10.21420/R58Q-FZ78), and these data were used to establish a subset of filters from each site with the highest soil/mineral content.

The criteria used were as follows; filters with high Al and Si content were most likely to contain substantial quantities of crustal matter particles since crustal matter is largely dominated by aluminosilicate minerals (Lide 1992). While any ambient air particulate sample could possibly contain fibrous mineral particles, only filters with high levels of aluminium (Al) and silica (Si) mass from compositional analysis and a corresponding crustal matter source component as evidenced by the source type (Soil for the Henderson and Patumahoe sites, Soil and Construction dusts for the Queen Street site) and corresponding source chemical profile (Supplementary Figures S2; Supplementary Table S1) were considered for further investigation to maximize the potential for identifying fibrous mineral entities. Due to the labour-intensive nature of filter screening, a total of 30 filters, 10 from each of the three chosen sites, were selected for further analysis based on the above criteria. Figure 3 is an example of source apportionment results for the filter from the Henderson site from 14 May 2011, a filter high in soil content based on the composition of Al and Si components (23% compared to the long-term average of 7%).

In the first instance, phase contrast microscopy was undertaken on all 30 of the filters, 10 from each of the three air quality sites. Two different microscopes were used to examine the filters, each offering a slightly different magnification quality: A Leica DM2700P Petrographic Microscope and a Leica MZ16 Binocular Microscope. In each case, the filters were scanned in their entirety to visually assess the presence and density of fibrous material cover. A careful record of the morphology (shape and visual dimensions) was made of the deposited material on each filter so the data could be referenced for later analyses. Both microscopes were network linked and enabled with digital imagery capabilities. Once fibrous materials had been located using the microscopes, the filters were set aside ready for SEM-EDS analysis.

A crucial part of this work is to test whether the fibrous nature of the PTFE filter would allow for identification of fibres on top of, or within, the filter itself. It is clear that most fibres are quite clearly identified as they are both larger than the fibres of the filter, and also generally sit at angles that are not consistent with the filter. An attempt to measure the typical woven fibre within the filter was made, however, the fibres were not consistent so no representative value could be obtained. However, it was noted that almost all woven fibres within the filter were less than 1 µm in length. It is, therefore, reasonable to assume that any fibres of interest less than 1 µm in length could be difficult to differentiate from the filter medium.

In this study, a historical archive of filter data, originally collected for the purpose of compositional analysis for receptor modelling, was used to investigate the presence of FLP in filters. The elemental composition of particulate matter collected on filters was used to identify the main chemical fingerprint of crustal matter, and a sample of filters with elevated quantities of Si and Al were investigated further. Of the three sites investigated, based on scans of the filters using PCM, no fibres were found on any of the filters collected from either Patumahoe (rural background) or Queen Street (urban roadside). Therefore, these filters were not subjected to further examination via the more time-consuming and expensive SEM-EDS method. By contrast, the Henderson (suburban roadside/light industry) filters did reveal fibrous materials across the surface. The morphology of the structures was found to be largely fibrous with some long, thin fibres mixed in with some hexagonal shards (Figures 6, 7, 9).

The clustered nature of the components in the image in Figure 7 suggests that the long fibre might have impacted first, and then other fibres and particles impacted on top during the 24-h sampling period. Within the image and EDS spectra from Figure 7, the fragment numbered “2” closely resembles that of the zeolite erionite, as exemplified in Figure 8. Although in its captured form it has the morphological dimensions of a particle, it was likely fibrous when entering the atmosphere given the fibrous nature of erionite in the bedrock (Patel et al., 2021). There are obviously limitations to identifying individual fibres with SEM/EDS alone, and the results should be seen as indicative only. However, these findings do suggest the need for further investigation.

Another air pollutant of growing concern is respirable silica particles. These often fibrous materials are long-lasting in the atmosphere and easily inhaled given their small dimensions (Gualtieri, 2018; Waeselmann et al., 2020). The methods used for this paper show that it is possible to confirm the presence of these potentially hazardous particles and adequately describe the morphological and chemical composition of silica fibres and fibre-like structures, thus helping to identify them on archived filters. Whether smaller fibres found in the ultrafine range would be visible is unclear, perhaps unlikely given the PTFE fibrous mesh structure. However, such fibres are generally rare with mineral fibres usually found to be in the larger (5–20 µm) size range.

Finally, the method of scanning filters first for fibre-like structures was instructive and useful as a first step. However, it cannot be discounted that some particles on the filter may actually be fibres that had been crushed or fragmented into smaller pieces. Such fragments would also be informative about the historical airborne loading of potentially hazardous fibrous forms.

This paper presents the reanalysis of archived PTFE filters originally collected for the compositional analysis of Auckland’s ambient air particulate matter. Retrospective information on new air pollutants of interest, such as fibrous zeolites, was obtained, as demonstrated here by the identification on filters of previously unobserved fibrous materials.

The initial step of screening archived filters for analysis using those with the highest Si and Al was found to be somewhat successful. However, high soil content did not necessarily translate into fibres or fibre-like particles on all the chosen filters, with both mineral and non-mineral found in plentiful quantities within the historic Henderson samples, largely from the 2011 collection. A review of activities in the vicinity of the monitoring site during May 2011 will be undertaken with use of mapping, council records and press information (reporting fires, infrastructure construction, etc.).

The use of fibrous PTFE filters does increase the complexity of identifying fibres that are under a certain size, with fibres <1 µm being more likely to be embedded in the filter and, therefore, hard to distinguish from fibres from the filter itself. However, the use of PCM showed that the larger width of the fibre-like particles tended to allow them to sit on top rather than be embedded within the filter substrate. The thickness of the filters allowed only partial illumination from PCM, which could make the identification of smaller fibres difficult.

The subsequent use of SEM techniques for filters identified as interesting from PCM methods allowed for the identification of several fibrous mineral forms potentially including fragments of erionite and respirable silica. Given the limited number of fibres present, identification of the material or their origin was not possible with certainly. However, the presence of quarrying in the nearby vicinity, and the similarity of the EDS scan of zeolites within the filter to that of erionite found in Auckland’s bedrock, suggests that further investigation is warranted. These results open the door to the reanalysis of filters, collected initially for a different purpose, for emerging pollutants such as respirable silica, micro-plastics as well as mineral particles, to be identified using reasonably simple and quick methodologies. This study highlights the value in the retention and careful archiving of air quality filters for retrospective analysis for such a purpose.