Airborne Particles in Swansea, UK: Their Collection and Characterization

Abstract

Urban air particulate matter (PM) has previously been associated with a variety of adverse health effects. It is now believed that the smallest particles, ultrafine or nanoparticles, are linked to the greatest health effects. The physicochemistry of these particles is likely to provide information regarding their toxicity. Therefore, the aim of this study was to further the understanding of the heterogeneous and changing particle concentrations in urban air, in conjunction with gaining an understanding of the physicochemistry of the particles. A Dekati electrical low-pressure impactor was used to collect the particles and real-time data in a busy traffic corridor in Swansea, Wales, over a period of 10 nonconsecutive weeks. Particle concentrations in the street canyon were analyzed and particle physicochemistries investigated using a variety of techniques. Particle number concentrations were found to vary both diurnally and from day to day in the traffic corridor. Of all particles, the nano to fine size fraction was consistently identified in the highest concentrations (maximum: 140,000 particles cm⁻³). Particle physicochemistry was found to vary as a function of size, with larger particles exhibiting a greater variety of morphologies (and consequently particle types) and associated metals.

Air pollution episodes have been noted since Roman times, with evidence of small-scale scientific atmospheric pollutant investigations as early as the 17th century (Kretzschmar, 2007). However, it took one-of-a-kind events such as the Meuse Valley fog in Belgium in 1930 (Nemery et al., 2001) and the Great London smog of 1952 (Whittaker et al., 2004; Davis et al., 2002; Elsom, 1987) to incontrovertibly link airborne particulate matter (PM) to adverse health effects. These events served as a wake-up call, leading to technological improvements, funding, and research (Donaldson, 2003). It is now postulated that the smallest particles, nano- or ultrafine particles, generally defined as particles with at least one dimension below 100 nm (Donaldson et al., 2001; Oberdörster et al., 2005), are being linked with the greatest adverse health effects in epidemiology studies, studies in vitro, and to a large extent, animal studies in vivo (Donaldson et al., 2001; Brown et al., 2001; Oberdörster et al., 2005). While this association is now well established, the actual causes of adverse health effects continue to be debated and are not well understood.

Over the range of particle sizes, it is nanoparticles that have consistently been found in the highest concentrations in urban air (Tuch et al., 2003; Ketzel et al., 2004; Mejìa et al., 2007). Concentrations in urban air have repeatedly been shown to reach levels of 10⁴ to 10⁵ particles cm^-3 (Kittelson et al., 2004) during peak traffic flow periods. This causes concern that at these high particle levels human clearance mechanisms cannot function efficiently at removing particles (Oberdörster, 1995), leading to particles remaining in contact with cell surfaces for longer periods of time. This persistent contact or “particle overload” has been highlighted as a potential contributing factor when assessing the toxicity of airborne particles. This issue is complicated by the variety of particles that populations are exposed to on a daily basis. Urban air particles are a complicated and heterogeneous mix (Donaldson et al., 2005), combining a wide range of particle characteristics such as size, morphology, surface reactivity, biopersistence, and chemistry in every sample. This emphasizes the importance of fully characterizing particulates in all investigations (Harrison & Yin, 2000).

This study used an interdisciplinary approach to investigate particle physicochemistry within a traffic canyon. Particles were collected using a Dekati electrical low-pressure impactor (ELPI) into 12 size fractions. The collection was completed at two locations: an urban air traffic canyon and a rural background location. Due to the small masses in each of the collected size fractions, they were then combined into three analyzable size fractions (Up to 615 nm, 616 nm–2.39 μm, and 2.4–10 μm). The three size fractions were physicochemically evaluated using tools including field emission scanning electron microscopy (FE‐SEM) and inductively coupled plasma–mass spectrometry (ICP-MS).

METHODS

Site Details

Particle collection was completed at two localities: an urban air site, and a rural control site. Neath Road in Swansea, Wales, United Kingdom, was the urban collection site. Neath Road is a main commuter traffic route into Swansea City, and a recognized traffic hotspot (Figure 1). The area was designated an air quality management area (AQMA) based upon its pollutant concentrations. Traffic levels are high (approximately 18,000 vehicles per day) due to the road forming a main commuter zone between Swansea and Neath. Swansea is also an old industrial port city, which has been undergoing a process of urban renewal for a number of years. The locality was therefore expected to consist of a cocktail of particle types that were contributed by the main sources: urban, industrial, and marine. Sampling was completed over a period of 10 nonconsecutive weeks during one season (thus reducing the impact of seasonal-related meteorological differences) between 05/12/07 and 28/02/08, resulting in both particle collections and real-time particle data.

FIGURE 1. Location map showing the Neath Road, Swansea sampling site (black circle) in relation to surrounding features.

The traffic corridor is orientated north-northeast to south-southwest, with the predominant wind direction in a similar trajectory (northeast–southwest). Small-scale industrial sites are located city-wide; however, the predominant wind direction (blowing straight from the sea and onto the site) reduces the impact of local industry. Port Talbot to the southeast represents the most substantial industrial area in the vicinity, potentially contributing particles dependent upon the wind direction.

Brecon, the rural control site, is located approximately 42 km northeast of Swansea. Sampling lasted for a period of 3 wk, producing only a 1-wk usable sample due to an atypical dust storm (correlated to an event originating from the Sahara) and a neighbor's bonfire. While achieving the 1-wk usable particulate sample, a local mains power failure resulted in no real-time data collection.

Instrumentation

Particles were collected using a Dekati electrical low-pressure impactor (ELPI). The ELPI is an inertial-based cascade impactor, which accumulates both real-time particle data and particle collections onto substrates. It divides particle data into 12 size fractions, from 7 nm to 10 μm, 3 of which are within the “nano” size range, and particle collections from 30 nm to 10 μm. ELPI cutoff diameters (Keskinen et al., 1992) and particle concentration profiles (Zervas & Dorlhène., 2006), have been confirmed in previous studies. A flow rate of 30 L/min was maintained using a Sogevac Leybold vacuum pump. The ELPI stages were loaded with 25-mm aluminium foil substrates. Substrates were weighed using a microbalance (Sartorius Micro SC-2) pre- and post-sampling to determine the particulate mass. Substrates were not coated with grease (as recommended by the manufacturers) in order to reduce contamination during subsequent ICP-MS analyses (Fujitani et al., 2006). The equipment setup on-site included the collection head, Teflon tubing, ELPI, pump, and laptop for equipment control and data collection.

Statistical Testing

Graphing and statistical testing were completed using Microsoft Excel, with SPSS (version 16) used for nonparametric particle analysis including Spearman's rank correlation coefficient.

Particle Characterization

Analytical Electron Microscopy

In preparation for field emission scanning electron microscopy (FE-SEM), the aluminum foil substrates were cut into sections. Approximately one-eighth of each collection foil was used for analysis. Epoxy resin (Araldite) was used to attach the foil substrate sections to 12.5-mm aluminium SEM stubs (Agar Scientific). Samples were then coated with gold using a sputter coater (Bio-Rad SC500). Samples were imaged using a Philips XL30 FE-SEM. A range of working conditions in secondary electron mode was utilized to maximize image quality, including a working distance of 5–10 mm, accelerating voltages 5–20 kV, spot size 4, and a gold foil aperture.

Particle Extraction

Particles were removed from the foil substrates for further physicochemical analysis using a novel freeze-drying technique. Onto each aluminum foil substrate, up to 1 ml of molecular biology grade water was pipetted. The foil and water were then frozen. Once fully frozen, the ice discs were peeled from the foils using ceramic tweezers. Samples were freeze-dried at –40ºC (model: Edwards Pirani 10) until no ice remained in the samples, a process taking varying lengths of time from overnight, to periods of 2 or 3 d depending upon sample size. Samples were combined into three size fractions (30–615 nm, 616 nm–2.39 μm, 2.4–10 μm) in order to provide samples sufficient for analysis, representing “nano–fine,” “fine,” and “fine–coarse” particle size fractions. The accuracy of the particle removal technique was assessed (Figure 2). Figure 2 compares the particle recovery efficiencies between the three size fractions. Percentage particle recovery is ascertained by weighing substrates before/after sampling to find total particle mass, and after extraction to find the particle mass that was removed from the substrate. Particle percent recovery therefore represents the mass percentage removed from the substrate using the extraction, in comparison with the original particle mass.

FIGURE 2. Particle mass extraction efficiency for the three analyzed size fractions (30–615 nm, 616 nm–2.39 μm, 2.4–10 μm). Error bars indicate the range of recovery efficiencies measured.

Particle removal using this technique is proven to be efficient (up to 98% particle recovery), removing the majority of the particle mass from the collection substrates. These removal efficiencies are comparable to (or more efficient than) those from other studies. Hartz et al. (2005) obtained a 60–85% mass recovery using a solvent-based extraction process. Jones et al. (2006) recovered 80% of particles with an initial wash of particles collected onto polyurethane foam (PUF) substrate. Further washing provides recoveries of up to 95%, comparable with this study.

Due to the high removal rates, particles removed using this methodology are considered to be representative of the particle sample as a whole. It is known that particle removal is most effective in the middle size fraction, a factor likely to be closely related to a larger initial mass and volume in this size fraction, combined with similar substrate adherence areas to the smallest and largest size fractions, reducing the relative percent of particles in contact with the substrate.

ICP-MS Analysis

Samples were digested for ICP-MS analysis using a CEM MDS-200 microwave system. Particle samples (n = 2) were washed into Teflon-coated composite vessels using 5 ml of 70% nitric acid. The samples were digested using an existing program developed for refractory carbon-based PM (Jones et al., 2006). The microwave program consists of a stepped increase in pressure to 80 psi for a period of 20 min, with a corresponding temperature rise to 180ºC. The program lasts for approximately 2.5 h, including warm-up and cool-down periods. Samples were then diluted to a level of 10 μg/ml (dependent upon their original weight) using deionized (>18 ΩM) H₂O. Raw data was corrected for blanks and controls accordingly.

RESULTS

Real-Time Particle Data

After processing the raw data using ELPIvi software, it is seen that throughout the daily cycle, on both weekdays (Figure 3, a–c) and Sundays (Figure 3, d–f), particle number concentrations are consistently highest in the smallest size fraction (D50% 7 nm). In this size fraction, particle number concentration peaked at 140,000 particles cm^-3.

FIGURE 3. Average daily particle concentration profile in Neath Road traffic corridor for weekdays and Sundays in three size fractions: (a) 7–615 nm, (b) 616 nm–2.39 μm, and (c) 2.4–10 μm.

During the weekday averages, there is a consistent daily concentration profile that is replicated in all three analyzed size fractions. The profile is characterized by a steep rise in particle numbers during the morning rush hour. Interestingly, while all three size fractions show this trend, particle numbers in the coarse size fraction (2.4–10μm; Figure 3c) do not begin to increase until 08:30 a.m., compared to a 06:00 a.m. rise identified in the two smaller size fractions. Similarly, the evening rush hour signal identified in the two smaller size fractions (7 nm–2.39 μm), which begins at 15:00 p.m., does not begin in the coarse size fraction until 17:00 p.m.

During weekdays, the “nighttime” particle concentrations (18:30–06:30) are significantly lower (95% confidence) than “daytime” particle concentrations (06:30–18:30) in the two smaller particle size fractions (7 nm–2.39 μm). When considering the coarse size fraction (2.4–10 μm), this statistical difference (95% confidence) is not identifiable.

In contrast to the weekday data, Sunday particle number concentrations peak at 38,000 particles cm^-3 at 20:30 p.m. The smallest (7–615 nm) and largest (2.4–10 μm) measured size fractions do not show a significant difference in particle number concentrations between “daytime” and “nighttime” hours (95% confidence). In contrast, the middle size fraction does indicate number concentration variation between daytime and nighttime hours 95% confidence.

Averaged data across the week (Monday–Sunday; Figure 4) illustrate the daily particle concentration profile differences at Neath Road, Swansea. Outputs for Monday–Thursday are consistent in terms of profile shape and magnitude in the smallest size fraction (7 nm–615 nm). This profile pattern begins to break down on Friday and Saturday, and by Sunday, the original number concentration profile was broken down completely, with smaller magnitudes and a different profile shape, with a particle concentration low during the morning replacing the number concentration high identified in the weekday data.

FIGURE 4. Average weekly particle concentration profile for Neath Road, Swansea.

Fine (616 nm–2.39 μm) and Coarse (2.4–10 μm) particles do not have a weekly concentration distribution similar to that of the smallest size fraction. The consistency of the number concentration profile (Monday–Thursday) identified in the smallest size fraction is not repeated in these size fractions. Instead, concentration profiles are generally more poorly defined, with occasional time periods appearing to be synchronized with the finest size fraction. In both larger size fractions, particle concentrations are higher from 12:00 p.m. Saturday to 00:00 a.m. Sunday than on the Wednesday and Thursday, which contain some extreme particle concentration lows, for example, Thursday (14:30 p.m.)

FE-SEM

As shown in Figure 5, particle morphology, and consequently type, increased in variability as particle size increased. Particles in the smallest size fraction (30–615 nm) have a consistent morphology of spherical to sub-spherical particles. In the middle size fraction, a combination of agglomerated spherical/subspherical particles and more sheet-like platy grains dominate. The largest size fraction (2.4––10 μm) exhibits much greater particle variability, with a range of particle morphologies visible (Figure 5, e and f), agglomerated spherical/subspherical particles, platy grains, cubic morphologies, larger spherical particles, and large near-spherical particles with nodules.

FIGURE 5. FE-SEM images of particles in the three measured size fractions collected in Neath Road, Swansea. (a) Particles in the 30–615 nm size range. (b) Close-up view of the 30–615 nm particle size range. (c) Particles in the middle size fraction (616 nm–2.39 μm), at a large-scale view. (d) Closer view of particles in the middle size fraction. (e) Particles in the largest size fraction (2.4–10 μm). (f) Closer view largest size fraction.

ICP-MS

The ICP-MS elemental analysis confirmed that iron (Fe), zinc (Zn), and magnesium (Mg) were the most abundant trace elements in the particles (Figure 6). Element concentrations were found to vary with respect to particle size, but differently between elements; for example, Fe and Mg were found to increase in concentration with increasing particle size, compared to nickel (Ni) and lead (Pb), which had the highest elemental concentrations in the smallest size fraction.

FIGURE 6. ICP-MS elemental analysis of the three analyzed size fractions. Bars represent the three different analyzed size fractions (white = D50% 30–615 nm; light gray = D50% 616 nm–2.39 μm; dark gray= D50% 2.4–10 μm), with top graph showing elements in parts per million (ppm) concentrations and bottom graph showing elements in parts per billion (ppb) concentrations. Error bars represent one standard deviation either side of the mean.

In terms of average PM₁₀ concentration, elements were identified in the descending concentration order Fe > Zn > Mg > Ni > Cu > Cr > Ba > Mo > Pb > Mn > Ti > V > Zr > Co > Cd. Using Spearman's rank correlation coefficient, associations were identified between a number of elements, including Fe and Cu, Fe and Ba, Fe and Mn, Mg and Co, Ni and Ba, Cu and Ba, Cu and Mn, and Ba and Mn, to a .01 confidence level.

DISCUSSION

Particle Data Analysis

Throughout the 24-h sampling period shown in Figure 3, the highest particle concentrations were found in the smallest particle size fraction, particles 7 nm–615 nm. These findings reinforce studies in Brisbane (Mejìa et al., 2007) where peak particle concentrations were below 30 nm (82–90% of particles). A study in two German cities, Erfurt and Leipzig (Tuch et al., 2003), found the highest particle concentrations in the 10–20 nm size range, while an urban air study focused upon Copenhagen (Ketzel et al., 2004) and lasting several years placed the particle concentration maximum between 20 and 30 nm. This particle concentration maximum is attributed to the traffic contribution at these urban sites (Mejìa et al., 2007; Ketzel et al., 2004; Shi et al., 1999). The findings within the Swansea traffic corridor are therefore comparable to those in other locations, and the concentration maximum, combined with what is known about the street canyon, confirms that while the input of particles from other sources (for example, industrial and marine) contributes to the particle totals, vehicles are the dominant sources of particles at Neath Road in Swansea.

Particle concentrations throughout the day in the traffic corridor are high (mean: 52,000 particles cm⁻³) when compared against some urban areas sampled in similar studies. The German two-city study (Tuch et al., 2003) found a particle concentration maximum of 40,000 particles cm⁻³, while the Copenhagen study (Ketzel et al., 2004) detected an average of only 7700 particles cm⁻³ during a 3-mo investigation period. A study completed in Rouen, France (Gouriou et al., 2004), using an ELPI found average particle concentrations below 50,000 particles cm⁻³; if particular external factors were combined, concentrations in the range of 10⁶ particles cm⁻³ were sometimes obtained. This distribution is similar to the situation in the Swansea traffic corridor. While the mean averages at 52,000 particles cm^-3, specific events happening over time scales as short as seconds are influencing and dramatically increasing the particle concentrations observed in the traffic corridor at particular times, leading to concentration peaks of up to 140,000 particles cm⁻³ in the nano–fine size fraction. A Three European City study (Ruuskanen et al., 2001) obtained similar results, with an Erfurt peak at 188,000 particles cm⁻³ during the morning rush hour.

On weekdays, days dominated by traffic, all size fractions are identified as having a traffic-responsive profile. That is, it is possible to identify a morning and afternoon rush hour signal. The coarse size fraction was found to have a later rush hour peak (both morning and afternoon). This finding may be explained by the rapid sensitivity of nanoparticles to vehicle exhaust particles, as previously identified by Rodriguez et al. (2007), in a study carried out in Milan, Barcelona, and London. Nanoparticles were found to vary extremely quickly and significantly in response to traffic, a finding reinforced in a study of urban air particle concentrations in Helsinki (Buzorius et al., 1998), where individual vehicles were shown to affect the observed particle concentrations.

A number of studies demonstrated that particle concentrations are higher during the day, and linked to the vehicular particle source and its predominance during daytime hours (Buzorius et al., 1998; Laasko et al., 2003), as seen in the Swansea traffic corridor. Rodríguez et al. (2007) investigated further to determine that the difference between daytime and nighttime concentrations is much more pronounced in the nano–fine range—a result also found in this study on weekdays. On days not dominated by traffic sources (Sunday), this nanoparticle day–night variation was not significant, reinforcing traffic as a source of the smallest particles. This continuity between day- and nighttime particle number concentrations on Sundays might also be contributed to by the lack of industry and other related sources of particles on the weekend.

The morning rush hour peak was identified in this study, a finding also seen in a study at Marylebone Road (Harrison & Jones, 2005). A daily pattern, with nanoparticle peaks between 8 a.m. and 9 a.m. and between 4 p.m. and 5 p.m. identified in the German Two City study (Tuch et al., 2003) correlates with the nanoparticle morning and afternoon rush hour peaks identified in Swansea on weekdays. A link between nanoparticle concentrations and solar radiation was previously identified (Shi et al., 2001), perhaps explaining the sustained nanoparticle numbers observed at Neath Road between morning and afternoon rush-hour peaks.

The difference in particle concentrations and distributions identified at the Neath Road collection site between weekdays and weekends was also noted in other studies (Buzorius et al., 1998) and was attributable to a reduction in commuter traffic and, to an extent, industrial processes during the weekends. These findings are not consistent for all studies (Mejìa et al., 2007), perhaps due to a reduced importance of commuter traffic-sourced particles in the study, and the dominance of other sources.

Identifiable in the Neath Road data are reduced particle number concentrations in the fine and coarse size fractions during Wednesday and Thursday, and increased particle number concentrations on Saturday and Sunday. If the smallest size fraction (7–615 nm) is taken to be representative of the particle number profile predominantly as a result of traffic, this finding reinforces that particles in the middle and largest size fractions are contributed to by a variety of sources other than traffic exhaust, perhaps road dust, marine particles, and industrial particles (Moreno et al., 2004).

The averaged weekly data at Neath Road traffic corridor identified variability in particle concentration signals for different days of the week, especially emphasized in the smallest size fraction, particles between 7 and 615 nm. Different particle signals were also identified in a study carried out in Milan, Barcelona, and London (Rodríguez et al., 2007), a finding explained by the importance of semivolatile compound condensation in urban areas. In contrast, a study at three sites within Birmingham, England (Shi et al., 1999), found that despite variable weather conditions, particle concentrations and distributions measured varied only negligibly. Day-to-day particle concentration and distribution variances at Neath Road may be assumed to be dependent upon traffic compositional, volume changes, or meteorological differences. Further studies are required to elucidate the relative contribution of each component.

Physicochemistry of Collected Particles

Carbonaceous material was found to be dominant in all size fractions, as identified from the FE-SEM imaging (nano-sized spherical to subspherical particles found singularly or in aggregates; Figure 5). Results from a characterization analysis of PM collected on the coast of Sicily (Rinaldi et al., 2007) agree with this finding, especially in the size range 50–140 nm. In this study, the smallest measured size fraction (7–615 nm) was also found to have the highest carbonaceous material of all the measured size fractions. These study findings are in agreement with others including those completed in Pasadena, CA (Hughes et al., 1998), Milwaukee, WI (Lough et al., 2005), and Belfast (urban), London (urban), and Harwell (rural) in the United Kingdom (Jones & Harrison, 2005). The large contribution of carbonaceous soot nanoparticles to the samples, whether as individual particles (or small groupings of particles) in the smallest size fraction or as larger agglomerates in the middle and largest size fraction, reaffirms traffic exhaust particles as the main particle source in the street canyon. The large contribution of traffic exhaust particles to total particle concentrations in urban settings is well documented (BéruBé et al., 2008).

Particles of cubic morphology, as recognized using FE-SEM imaging (Figure 5), are identified as marine-derived halite crystals (Jones & BéruBé, 2007), due to the proximity of the sampling site to the sea and the predominant wind direction (Figure 1). Those particles with perfect cubic morphology are likely to have grown in situ on the collection substrate, while more damaged particles are likely to have origins of either marine processes or road salting (Moreno et al., 2004). The combined factors of proximity to the sea and comparatively stable weather conditions suggest a predominance of marine-derived halite crystals. Large (coarse size fraction) spherical particles with nodules covering the surface are attributed to biogenic processes, confirmed by their behavior beneath the FE-SEM beam (BéruBé et al., 2008). FE-SEM imaging identified sheet-like particles in the largest size fraction. These particles (2.4–10 μm) are identified as mineralogical particles, perhaps derived from local or more distant areas of exposed crust and soil (BéruBé et al., 2008).

Due to the naturally variable wind directions encountered during a sampling period, the origin of industrial-generated perfect spherical particles may be local (within Swansea) or wind-blown from a distance (for example, Port Talbot to the southeast). Spherical particles are common in both urban and industrial air (Moreno et al., 2004).

The metals identified in the particle samples (ICP-MS analysis) were found to rise in variety with increasing particle size, as also found in the Milwaukee study (Lough et al., 2005). PM₁₀ was seen to contain more metals than PM_2.5, perhaps due to the greater variety of contributing sources to the larger size fractions; including crustal, traffic, biological, and technogenic-type sources. In another study, investigation of analytical SEM images identified that particles under 1 μm predominantly consist of traffic-derived soot (Baulig et al., 2004). Lin et al. (2005) demonstrated a more bimodal distribution of elements within particulate samples, for example, a peak in the nano-size fraction and a peak in the particle size range 3.2–5.6 μm in a study conducted in southern Taiwan.

Iron was found to be the most abundant metal in the particles, in agreement with results from other physicochemical analysis studies (Hughes et al., 1998; Lough et al., 2005; Baulig et al., 2004). Some elements identified by ICP-MS analysis were identified as partly arising from diesel emissions, for example Fe, Ca, Si, Mg, and Mn (Wang et al., 2003); a number of these elements are also associated with crustal components, for example, Fe, Ca, Si, and Mg (Lough et al., 2005). This highlights the fact that source apportionment is extremely complicated, with different studies identifying different tracers for the same source, and different sources for the same tracer or combination of tracers.

The elemental concentrations identified in this study (ICP-MS analysis; Figure 6) are much lower than in London 1950s PM samples (Whittaker et al., 2004). Comparisons include 157 ppm Fe concentration at Neath road and 19,294 μg/g London 1955 sample, and 1.3 ppm Mn concentration at Neath Road and 508 μg/g from the London 1955 samples. Additionally, in a paper by Shao et al. (2007), outdoor Beijing particulate matter was collected and analyzed by ICP-MS. Levels of 17 ppm Mn in the Beijing air can be compared with 39 ppb (Neath Road). Therefore, total metal concentrations of PM from urban Swansea air are lower than concentrations identified in historic studies (Whittaker et al., 2004) and in rapidly developing countries (Shao et al., 2007). This finding is to be expected (Donaldson, 2003) due to improved legislation and current British technological requirements, and more local factors including meteorological conditions, road usage, and the prevalence of local polluting industries.

Metal concentration order at the Neath Road collection site (Fe > Zn > Mg > Ni > Cu > Cr > Ba > Mo > Pb > Mn > Ti > V > Zr > Co > Cd) can be compared to sequences in the literature for urban locations (Whittaker et al., 2004: Fe > Pb > Cu > Mn > V > As > Co; Chandra Mouli et al., 2006: Fe > Mn > Ni > Cu > Pb > Co; da Silva et al., 2008: Cu > Pb > Ni, Sb > Ce). The difference between the concentration orders of metals at different sites highlights the importance of local factors, including geography, meteorology, and variability of sources and source compositions. Correlations were identified between some of the metals analyzed by ICP-MS. These correlations may indicate the same or similar elemental sources, for example, correlation between Ba and Ni may be associated with road exhaust emissions (Dongarrà et al., 2007).

SUMMARY AND CONCLUSIONS

Particulate matter within the Neath Road street canyon, Swansea, Wales, was studied for particle concentration variations and particle physicochemical properties. The particle concentrations within the traffic corridor were found to be consistently highest in the smallest size fraction, with particle concentrations and daily patterns comparable to previous studies in this field. Evening and weekend concentrations of particles were significantly lower than for daytime particles, highlighting the role of traffic exhausts as a primary and influential provider of the smallest (and most abundant) particles.

Generally, with increasing particle size, particle morphology and type rose in variability, with particles in the nanoparticle range being dominated by traffic exhaust particles. The associated metal content rose in both amount and variety of types with increasing particle size. The ICP-MS analyses generally added to and reinforced results from the FE‐SEM and were useful in providing bulk elemental analysis.

Notes

This research was funded by the National Environment Research Council (NERC).