Chemical Analyses and Geographical Origins of Residential Attic Dust in Central South Africa

doi:10.21203/rs.3.rs-3876029/v1

Particulate matter (PM) serves as widely used air pollution proxy indicator. Substantial supporting evidence links exposure to PM with adverse health effects. This study compared chemical and morphological properties of long-term accumulated particulate matter, and possible sources, from various locations in Bloemfontein, Kimberley and Vanderbijlpark residential areas. Samples were collected from the attics of houses built over forty years ago. Cluster sources were identified by analysing four decades of backward Long-Range Transport (LRT) clustering techniques. Particle morphology and elemental composition of samples were determined using scanning electron microscopy with electron dispersive spectroscopy, and mineral content by X-ray fluorescence, X-ray diffraction and electron probe microscopy techniques. Considering a period spanning over 4 decades, several PM origins were determined by a combination of LRT cluster and mineral analyses, pointing to various moving sources all over Southern Africa.

air pollution

dust particles

trace elements

minerals

Globally, the number of premature deaths due to exposure to long-term air pollution is increasing every year and therefor became a matter of profound concern (Wang H et al., 2012). Apart from causing ill health in many, pollution also drives climate change, and thus the consequent international outcry in favor of alternative green and sustainable technologies.

Particulate matter is but one part of air pollution in general, howbeit a major component. These particles include suspended particulate matter, inhalable larger particles with a diameter of 10 µm – designated PM₁₀, finer PM_2.5 particles, ultrafine particles with a diameter of 100 nm or less; and lastly, soot. Long-term dust may include several of these categories, which may be the consequence of soil and meteorological conditions both at the location and far away, the time period over which it settled, as well as specific activities whereby it was generated (Sarpong SA et al., 2021). The number of research reports addressing atmospheric particulate matter are increasing, but tend to focus mostly on PM₁₀ and PM_2.5. Previous studies have identified PM sources such as biomass burning, road traffic and agricultural and industrial sectors, but limited information is available on PM resulting from mining activities (Worobiec A et al., 2011; Adeyemi A et al., 2021; van der Westhuizen D et al., 2023).

In general, elevated levels of chromium, nickel and manganese originating from rock soil and nickel smelters are linked to increased health problems, specifically lung and thyroid cancer. Building material degradation are found to be rich in copper and zinc, however more prevalent so in homes which experienced increased traffic emission pollution and wear and tear. Older homes with gardens and peeling paint makes a difference in the metals that are found in the dust; these houses have elevated levels of almost all metals, except for chromium (House Dust, 2023).

A study done in Maribor Slovenia found mining sediments consisting of rock fragments containing garnet, epidote, zircon and quartz, etc. (Mioč P et al., 1989; Šoster A et al., 2017; Mencin GE et al., 2019). The mountains surrounding Maribor contains minerals to which the origin of the sampled dust could be ascribed (Hinterlechner RA, 1971; Mioč P, 1978; Zupančič N, 1994). Another study in Slovania determined that air borne PM/dust entering attics are not influenced by indoor activities or weather conditions and therefore persistently accumulates over long periods of many years or even decades (Gaberšek M and Gosar M, 2021). PM deposits in snow, urban soil, attics, streets and residential dust were analyzed to determine particulate characteristics and chemical composition of the dust. In attic dust it found a combination of metals such as copper, manganese, zinc, iron and lead originating from ferromanganese plants (J Expo Sci Environ Epidemiol, 2016).

PM may often travel over long periods of time and even long distances and thus formally became known as long-range transport (LRT) clusters. A key tool used to determine the origin of these clusters is Backward Trajectory Modelling. In the current study, four LRT clusters of origin were determined, spanning a period of four decades in three provinces (Adeyemi A et al., 2021; Howlett‑Downing C et al., 2022).

2.1. Sampling

Dust in the attics of houses with tiled roofs were collected at 9 different sites in 3 different towns/cities, chosen for their proximity to main roads, infrastructure and industry. Tiled roof covering provides ample openings between tiles for fine dust to enter through over periods of many years, making them ideal sampling sites. The selected sites are shown in Table 1 and corresponding maps in Figs. 1 and 2, i.e. the Free State (FS) province capital, Bloemfontein – located in the central FS, Kimberley – lying on the FS-eastern border, and Vanderbijlpark in the north.

A Black & Decker 7.2V cordless wet and dry Dust Buster handheld vacuum cleaner was used, specifically for its gentle suction action. Some precautions were taken during sampling, eg. not collecting towards the sides of roofs, not in isolated areas, or where degrading wood and insulation fibers, etc. accumulated. After sampling, the dust samples were sieved through a small sieve to remove the larger particles from the smaller dust particles.

Table 1

Location of Sampling Sites
Suburb	City	Coordinates
Langenhoven Park	Bloemfontein -West	-29.103, 26.162
Universitas	Bloemfontein - West	-29.096, 26.170
Heidedal	Bloemfontein - East	-29.142, 26.261
Maselspoort	Bloemfontein - East	-29.034, 26.404
South Ridge	Kimberley - South	-28.771, 24.762
Monument Heights	Kimberley - South	-28.767, 24.760
Casandra	Kimberley - East	-28.741, 24.794
SE3 1	Vanderbijlpark - South	-26.741, 27.849
SE3 2	Vanderbijlpark - South	-26.736, 27.850

2.2. Scanning Electron Microscopy - Energy Dispersive Spectroscopy (SEM-EDS)

Elemental composition and particle morphology were determined using a Joel IT-200 SEM equipped with an energy dispersive X-ray spectroscope. Samples were prepared by mounting on carbon tap and coating with iridium (Ir), coating required for conductivity purposes. Imaging was performed at an accelerating voltage of 5 kV and zoom magnifications varied from 27 to 7500 times. EDS was performed at 15 kV accelerating voltage (Lee E et al., 2018).

2.3. X-Ray Fluorescence (XRF)

The major elements (namely Fe, Mn, Ti, Ca, K, S, P, Si, Al, Mg and Na) were analysed on a fusion bead containing 2 g of TSP sample and the flux lithium borate. A Panalytical Axios XRF spectrometer with a 4-kW rhodium (Rh) tube. H₂O loss was determined by measuring the mass loss after heating the sample to 110°C and the loss on ignition (LOI) after heating the sample to 800°C. The lower limit of detection for major oxides is 0.01 wt % (Willis F etal., 2014).

2.4. X-ray Diffraction (XRD)

Mineral analyses were carried out using a Panalytical Empyrean X-ray Diffractometer. The X-ray source is of copper (Cu) composition and the detector is an X celerator detector. The tube geometry configuration is a Bragg-Brentano geometry with a theta-theta goniometer configuration. The analysis is performed from 3.5° to 70° 2θ with tube settings of 45 kV and 40 mA. The software used is Data Collector v. 3.0c and Highscore v. 3.0e, equipped with the ICDD PDF-2 database for evaluation of diffractograms (Dinnebier RE and Friese K, 2003).

2.5. Electron Probe Micro Analysis (EPMA)

Small portions of each TSP sample were mixed with Clarofast to form an epoxy mount, which was then polished. These epoxies were then sputter coated using a Quorum Q150 and the minerals analysed using a Jeol JXA-iSP100 EPMA equipped with a tungsten filament with an energy dispersive X-ray spectrometer detector (Kulkarni KN et al., 2022).

2.6. Long-range Air mass Transport

Long-range transport clusters of air pollution source areas were determined using the geographical origin of air masses which travelled over Bloemfontein, Kimberley and Vanderbijlpark as proxies (van der Westhuizen D et al., 2023; Molnar P et al., 2017; Schwarz J et al., 2016; Tshehla C and Djolov G, 2018; Williams J et al., 2020). The Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) program was used to generate 72 h backward trajectories every day for the years 2022, 2012, 2022 and 1992 for the three different locations (Tshehla C and Djolov G, 2018; Williams J et al., 2020; Draxler RR, 2003). Global Reanalysis Meteorological Data from the National Oceanic and Atmospheric Administration Air Resource Laboratory (NOAA ARL) was employed, as used by the Global Centers for Environmental Prediction / National Centers for Atmospheric Research (NCEP/ NCAR) web server. 6 h intervals (00:00, 06:00, 12:00 and 18:00) were generated from analysis fields (2.5° x 2.5° resolution, 17 vertical levels, 10 000 m AGL), while wind fields were linearly interpolated between each interval (Tshehla C and Djolov G, 2018; Williams J et al., 2020; Wichmann J et al., 2014). Three different starting heights (250 m, 500 m and 750 m) with a fixed offset grid factor (250 m) were used to calculate 72 h daily backward trajectories; these trajectories were used as in cluster analyses of previous studies (Adeyemi A et al., 2021; van der Westhuizen D et al., 2023). The study period generated a total trajectory range from 4380 to 4392. These trajectories were used in the clustering analysis for each location; Bloemfontein, Kimberley and Vanderbijlpark. The clustering algorithm in the HYSPLIT program was derived from the distance between the associated mean cluster and trajectory endpoints (Adeyemi A et al., 2021; van der Westhuizen D et al., 2023). The optimal number of clusters was determined by using a total spatial variance (TSV) plot (HYSPLIT Tutorial, 2023).

Determining the morphology and elemental components of suspended dust/particulate matter by means of microscopy techniques is useful to aid analyses in particle size and mineralogical compositions. A total of seventeen elements were found in the current study (Lee E et al., 2018). The most common minerals that were identified were quartz and feldspars, making it difficult to determine the exact sources of these minerals. The reason is that these minerals are found in most rocks in South Africa, with feldspars in fact being the most common group of minerals within the continental crust, quartz being second.

Other prominent minerals that were found include calcite, ilmenite, mica, hematite and gypsum. The most common possible sources that were identified were sedimentary and iron ore. LRT combined with mineralogical analyses pointed to the Bushveld Complex, the Karoo-, Olifantshoek-, Cape- and Transvaal Supergroups, and the Hardveld Land Division (Botswana) being prominent sources in South Africa. As one of the most active mining countries in the world, the large mining industry in South Africa necessitate more research with regards to related PM pollution.

3.1. Scanning Electron Microscopy and Energy Dispersive Spectroscopy (SEM-EDS)

SEM-EDS is analytical technique that is ideally suitable for investigations of sample and material surfaces. Elemental composition and particle morphology were determined for each of the nine samples collected for the present study. All dust samples were kept in its original state during SEM-EDS analyses. Analyses therefor include results in its entirety, i.e. composed of both inorganic and possible organic constituents. The presence of Ir in the samples is due to coating of samples, as required for SEM analyses.

Table 1 lists precise sampling locations, i.e. suburbs and coordinates. In Bloemfontein four sites were identified and sampled, namely in house attics in Langenhoven Park (LHP) directly west of the main north-south national N1 highway, Universitas (U) directly east of the N1, Heidedal (H) that lies next to the city’s industrial area, and Maselspoort (M) that lies to the east of the city, but where the municipality’s water purification works is also located. The two sites west and east of the N1 highway were selected to investigate possible differences resulting from mainly westerly winds that blows in this area.

In Kimberley, three sites were identified and sampled, namely Casandra (C), South Ridge (SR) and Monument Heights (MH). Casandra (C) is located immediately north of the Ekapa mine and the provincial road (R64), west of the De Beers mine and northwest of the industrial area. South Ridge (SR) and Monument Heights (MH) are located between the N8 and N12 national highways and immediately north of the Transnet Railway Station and north-west of the Airport.

Vanderbijlpark was selected for being known to be a high pollution area. Two sites were sampled here, both being in the SE3 suburb, about 10 km south of the huge Arcelormittal steel and other engineering works.

The various particle morphology images and the elemental compositions for the four sites in and outside Bloemfontein, i.e. LHP. U, H & M, were taken at magnifications varying from 35 to 750 times, see Figures S1 – S8 in Supporting Information. The two sites west (LHP) and east (U) of the N1 highway were selected to investigate possible differences in view of mainly westerly winds that blow in the central Free State, with more PM resulting from heavy traffic expected downwind, i.e. in U. However, no significant differences between the elemental composition for all three sites, LHP, U and H located inside the city residential areas, were observed. Interestingly no Cl was found directly next to the municpal waterworks at M which is know for its use of ample amounts of Ca(OCl)₂ during the decontamination process.

Figures 3 and 4 show SEM-EDS elemental compositions for the LHP and M sites, with associated SEM images included. Only at Maselspoort (M), being located about 10 km outside, east of the city, one difference was observed in the elemental composition, i.e. Zn was also observed here.

The particle morphology images for the SR site in the mining city Kimberley were taken at magnifications ranging from 33 to 190 (SI: Figure S9), for the MH site at magnifications ranging from 35 to 230 (SI: Figure S11), and for the C site at magnifications ranging from 37 to 160 (SI: Figure S13). The graphical representations of the elemental compositions for the SR, MH and C site are shown in Figs. 5–7, while corresponding elemental compositions are shown in Figures S10, S12 and S14 respectively, in the supporting information. The same level of similarity was not observed for the Kimberley sites as in the Bloemfontein, eg. no sodium (Na) was detected in samples from the Kimberley sites, as opposed to both the Bloemfontein and Vanderbijlpark sites. The sites in Kimberley had key differences. The SR site did not contain Cl, but did contain Cu, which in turn was not observed at the other Kimberley sites. The MH site indicated Cl and uniquely also Ag, Cl was also observed in the C site.

Graphical representations of the elemental compositions for the SE3 1 & 2 sites in Vanderbijlpark are shown in Figs. 8 & 9, while corresponding elemental compositions are illustrated in Figures S16 & 17 in the Supporting Information. Particle morphology images for the SE3 1 and SE3 2 sites were taken at magnitudes ranging from 30 to 85 (SI: Figure S15) and magnifications ranging from 55 to 150 (SI: Figure S17) respectively. Of all nine samples the SE3 1 site is the only one that indicated Mn. Modelling of the origins of the various dust samples presented results shown in section 3.6 below.

In summary, the elements found in the elemental analysis for each sample were oxygen, carbon, sodium, magnesium, aluminium, silicon, sulphur, potassium, calcium, titanium and iron. Contrary to the rest, the Kimberley site indicated no Na, Cl was found at LHP, H, MH, C, SE3 1 and SE3 2, Mn was found only at SE3 1 and Cu at SR, MH, SE3 1 and SE3 2, Zn was only found at M and Ag only at MH.

3.2. X-ray Fluorescence (XRF)

XRF studies found the main constituents of all the dust samples to be SiO₂, Al₂O₃, Fe₂O₃ and CaO, see Fig. 10. The Maselspoort sample was the only one that would not fuse, with consequence that the XRF analysis could not be done.

3.3. X-ray Diffraction (XRD)

XRD results (Fig. 11, Table S2 and Figure S19 – S27) show that the samples are mostly composed of the silicate component; quartz, plagioclase and/or k-feldspar. Figure 11 illustrates this minerals distribution, with quartz together with plagioclase and k-feldspar representing 55–89% of the samples. The Bloemfontein samples (LHP, U, H and M) are also more complex when compared to the Vanderbijlpark (SE3 1 and SE3 2) and Kimberley (SR, MH and C) samples. After the silicate mineral component, the next most prominent component is the carbonates. These XRD results correlate well with XRF in that quartz (SiO₂ content – 100 wt.%) and plagioclase (NaAlSi3O8 – CaAl2Si2O8, SiO₂ content: 67.39–44.40 wt.%) are the most prominent constituent. Another correlation between the XRF and XRD results is the Loss on Ignition (LOI) (range from 12.39–30.81 wt.%) of the XRF results (Table S3). These losses are indicative of the decarbonization of carbonates during the heating process.

3.4. Electron Probe Micro Analysis (EPMA)

The intention of completing the analyses for the current study by means of the EPMA technique was to determine the minerals that may be present at concentrations too low to be detected by XRD. Large amounts of quartz and silicate minerals (such as plagioclase and pyroxene) are present in all samples, which is expected. The samples from Vanderbijl Park have Fe-minerals present in addition to the silicates. These Fe-minerals are in the form pyrite, magnetite/hematite and Fe-metal. The samples from the Kimberley area contain small quantities of chromite, magnetite/hematite, ilmenite and Fe-metal, see Fig. 12. The samples from the Bloemfontein area, contain minerals similar to those found in the Kimberley area, with Maselspoort containing Zn-minerals and barite. The analysis of these minerals can be found in the supporting information (Table S3).

In addition to the chemical composition of these minerals, it is observed that the grains viewed under the EPMA are angular in shape, not well rounded, which is expected from wind-borne material. Water-carrier particles are typically more rounded.

3.5 Combined XRD, EPMA and HYSPLIT

South Africa is a mineral rich country with an extended mining industry that is spread across the country. Combining Long-Range Transport (LRT) clusters with minerals identified in the samples of the present study, as well as mining sources, provided a means to help identify the origins of the dust samples (Pawloski GA, 1985).

Table 2 shows the different minerals found in the samples (excluding the common minerals; quartz, plagioclase and k-feldspar), the most possible source, and the source positions from the HYSPLIT maps. The minerals that are present are generally common, but the sources mentioned are the most prominent and correspond to the source of the HYSPLIT data. The quartz and feldspars found in the samples are abundant earth crust minerals, thus making it difficult to attribute its specific origins (Amulele GM et al., 2022). Much of the South African landscape is covered by the Karoo Supergroup, which consists of layers of sedimentary rocks which are capped by basalts. Minerals such as quartz, plagioclase, clay minerals, pyroxene, k-feldspar, gypsum and mica are present in the Karoo Supergroup (Johnson MR et al., 2006). Amygdales in the basalts may be the source of stilbite (Johnson MR et al., 2006). The Cape- (Thamm AG & Johnson MR, 2006) and Olifantshoek Supergroups (Moen HFG, 2006) also consist of sedimentary rocks (some of which are metamorphosed).

The groups in the Transvaal Supergroup that are of interest in this study are the Banded Iron Formation (BIF) rocks and the limestones, which may be the sources of the Fe minerals, hematite and magnetite, and the carbonates respectively (Eriksson PG et al., 2006).

The Bushveld Complex (BC) is South Africa's source of platinum, which is found in ultramafic rocks and chromitite layers. These rocks consist mainly of pyroxenes, plagioclases, k-feldspars and chromites. The BC may be the source of the chromites, ilmenite, rutile, magnetite and olivine. Ilmenite and rutile (and anatase) are also common in the heavy mineral sands found and mined along the coast of South Africa (Cawthorn RB et al., 2006).

Due to the proximity of the Kimberley samples it is important to note that ilmenite, pyroxene and calcite are also found in kimberlites (Skinner EMW and Truswell JF, 2006). The Hardveld Land Division (Botswana) consists of sedimentary rocks, basalt and granite gneisses (Jones CR and Hepworth JV, 1973). This area is a source of dust in the Kimberley area.

Table 2

Minerals detected in samples, and possible sources according to the HYSPLIT data.
Area	Sample	Minor/trace minerals	Possible sources	HYSPLIT data – suggested sources
Bloemfontein	Langenhoven Park	Calcite	Sedimentary	Karoo supergroup, Bushveld complex, Olifantshoek supergroup, Cape supergroup
		Ilmenite	Heavy mineral sands
		Mica	Widespread
		Hematite	Sedimentary/Fe ore
		Magnetite/Hematite	Sedimentary/Fe ore
		Pyrolusite	Mn fields
	Universitas	Calcite	Sedimentary
		Ankerite	Sedimentary
		Ilmenite	Heavy mineral sands
		Mica	Widespread
		Hematite	Sedimentary/Fe ore
		Kaolinite	Widespread
		Gypsum	Sedimentary
		Monazite	Widespread
		Rutile	Heavy mineral sands
		Olivine	Bushveld complex/Kimberlite
	Heidedal	Calcite	Sedimentary
		Ilmenite	Heavy mineral sands
		Anatase	Heavy mineral sands/granites
		Goethite	Alteration of Fe minerals
		Mica	Widespread
		Hematite	Sedimentary/Fe ore
		Gypsum	Sedimentary
		Pyrite	Witwatersrand gold fields
	Maselspoort	Calcite	Sedimentary
		Hematite	Sedimentary/Fe ore
		Gypsum	Sedimentary
		Ilmenite	Heavy mineral sands
		Zincite	Bushveld complex
		Barite	Sedimentary/Bushveld complex
Kimberley	Southridge	Calcite	Sedimentary	Karoo supergroup, Bushveld complex, Transvaal supergroup, Cape supergroup, Hardveld land division (Botswana)
		Hematite	Sedimentary/Fe ore
		Chromite	Bushveld complex
		Ilmenite	Heavy mineral sands
	Monument Heights	Calcite	Sedimentary
		Magnetite	Bushveld complex/Fe ore
		Mica	Widespread
		Chromite	Bushveld complex
		Ilmenite	Heavy mineral sands
		Titanite	Pegmatite/skarn deposit
	Casandra	Calcite	Sedimentary
		Ankerite	Sedimentary
		Stilbite	Basalts
		Ilmenite	Heavy mineral sands
		Magnetite/Hematite	Sedimentary/Fe ore
		Pyroxene	Bushveld complex/dolerites
Vanderbijl Park	SE3 1	Ilmenite	Heavy mineral sands	Karoo supergroup, Bushveld complex, Olifantshoek supergroup, Transvaal supergroup, Cape supergroup
		Anatase	Heavy mineral sands/granites
		Hematite	Sedimentary/Fe ore
		Pyrite	Witwatersrand gold fields
		Zircon	Widespread
	SE3 2	Stilbite	Basalts
		Gypsum	Sedimentary
		Magnetite/Hematite	Sedimentary/Fe ore

3.6. Long – ranged transport clusters (HYSPLIT)

For the present studies the Free State province was selected due to the virtual absence of previous investigations related to PM pollution within the province. As area with highest population in the FS, being centrally located, lying within an extended agricultural land area, and situated at the intersection of main highway and railway lines, the capital, Bloemfontein, was firstly selected. For purposes of comparison, the diamond mining city of Kimberley that lies on the FS-eastern border and directly southeast of some of the world’s largest iron and manganese producing mines was also selected. Lastly, the traditionally heavy steel industry town of Vanderbijlpark in the north was also added. This town lies between the high pollution Gauteng province Witwatersrand area to its north and Sasolburg with its many fossil fuel-based chemical industries. With all these variables in mind, the aim was to also determine changes in air mass transport clusters over the passed four decades (1992, 2002, 2012 and 2022), at 10-year intervals. These changes are shown in Table 3 and Figs. 13–15.

Using backward trajectories, changes in the origin of air masses over four decades could be determined for the three main locations, Bloemfontein, Kimberley and Vanderbijlpark. Four clusters of air mass origin were used for each specific year, as shown in Table 3. The clusters in 1992 for Bloemfontein were from Lesotho, NC, NW and AO. With the exception of NC, four decades later these clusters changed almost completely, namely to NC, FS, L and the Western Cape (WC). The clusters in 1992 for Kimberley were from the Atlantic Ocean (AO) and North-West (NW), Eastern Cape (EC) and Northern Cape (NC) provinces. Over four decades these clusters changed to Botswana, the NC, Free State (FS) and Limpopo (L) provinces. The clusters in 1992 for Vanderbijlpark (Gauteng province) were from NW, WC, L and Mapumalanga (MP) provinces. Forty years later these clusters changed to NC, FS, L and Kwa-Zulu Natal (KZN) provinces. It is thus seen that between 1992 and 2022, the directions of the air currents changed significantly, resulting in differences in source materials. The origins of the air masses show that the dominant movement in 2022 in Bloemfontein (averaging 51% over the four-decade period) and Kimberley (averaging 52% over the four-decade period) are in the north-east and south-west directions, whereas the prominent directions experienced in the Vanderbijlpark (averaging 52.5% over the four decade period) region are east-west.

Table 3 Long-ranged transport (LRT) clusters for each location in the years 1992, 2002, 2012 and 2022. Cluster relevant geographical areas are the neighboring countries of Lesotho and Botswana, the Atlantic Ocean (AO), and the North-West (NW), Eastern Cape (EC), Northern Cape (NC), Free State (FS), Limpopo (L), Western Cape (WC), KwaZulu-Natal (KZN), and Mpumalanga (MP) provinces of South Africa.

	Cluster	1992	2002	2012	2022
Bloemfontein	1	Lesotho (29 %)	NW (29 %)	FS (40 %)	NC (25 %)
	2	NW (37 %)	WC (20 %)	NC (27 %)	WC (31 %)
	3	NC (22 %)	FS (36 %)	L (20 %)	L (16 %)
	4	AO (12 %)	L (15 %)	EC (12 %)	FS (28 %)
Kimberley	1	NW (43 %)	AO (21 %)	NC (31 %)	NC (25 %)
	2	EC (17 %)	FS (32 %)	FS (27 %)	FS (28 %)
	3	NC (21 %)	L (17 %)	Botswana (25 %)	Botswana (34 %)
	4	AO (18 %)	Botswana (31 %)	WC (18 %)	L (13%)
Vanderbijlpark	1	MP (38%)	NW (37 %)	MP (40 %)	L (48 %)
	2	WC (13 %)	NC (12 %)	L (21 %)	FS (23 %)
	3	NW (18 %)	MP (38 %)	FS (34 %)	KZN (22 %)
	4	L (31 %)	KZN (14 %)	WC (5 %)	NC (7 %)

Particle morphology, mineral and elemental analysis pointed to the dust particulate matter essentially consisting of similar component, regardless of being of varied origins. The most common mineral was quartz, as expected from it being the most common mineral in the earth crust. Regardless having chosen sampling sites close to heavy industries, chlorine treatment water works, mines and heavy road traffic, typical pollutants from these potential sources were absent or hardly detected. This may be ascribed to the high volatility of pollutants typically associated with these sources, which otherwise did not settle in measurable quantities on ceilings of houses in residential areas.

From the Long-Range Transport cluster analyses it was noticed that clusters changed dramatically over the four decades for each of the three cities of this study. These clusters were also used to determine proposed source areas, concluding that the change in clustering extended the possible source areas.

This study assists towards improving the understanding of the combination of dust and the minerals and other PM pollutants it may carry. More specific studies may be conducted in future, such as focusing on specific minerals pollutants and determining the impact they have on the environment.

Author Contributions: Conceptualization and methodology, Karel G. von Eschwege; validation, Deidré van der Westhuizen and Karel G. von Eschwege; formal analysis, Deidré van der Westhuizen, Megan Welman- Purchase; investigation, Deidré van der Westhuizen and Megan Welman- Purchase; resources, Karel G. von Eschwege; data curation, Deidré van der Westhuizen and Karel G. von Eschwege; writing—original draft preparation, Deidré van der Westhuizen and Karel G. von Eschwege; writing—review and editing, Deidré van der Westhuizen, Janine Wichmann, Karel G. von Eschwege, and Megan Welman- Purchase;; visualization, Deidré van der Westhuizen, Megan Welman- Purchase and Karel G. von Eschwege; supervision, Karel G. von Eschwege and Janine Wichmann; project administration, Karel G. von Eschwege; funding acquisition, Karel G. von Eschwege. All authors have read and agreed to the published version of the manuscript.

Funding: This research was funded by the Central Research Fund of the University of the Free State, Bloemfontein.

Data Availability Statement: The data presented in this study can be made available from the corresponding author on request.

Conflicts of Interest: The authors declare no conflict of interest.

Adeyemi A, Molnar P, Boman J and Wichmann J (2021), Source apportionment of fine atmospheric particles using positive matrix factorization in Pretoria, South Africa., Environ Monit Assess Vol 193, pp 716. https://doi.org/10.1007/s10661-021-09483-3
Amulele GM, Lanati AW and Clark SM (2022), The electrical conductivity of albite feldspar: Implications for oceanic lower crustal sequences and subduction zones., American Mineralogist. Vol 107 (4), pp 614–624. https://doi.org/10.2138/am-2021-7836.
Cawthorn RB, Eales HV, Walraven F, Uken R and Watkeys MK (2006), The Bushveld Complex in: Johnson MR, Anhaeusser CR and Thomas RJ (Eds) The Geology of South Africa, Geological Society of South Africa., Johannesburg/Council for Geoscience, Pretoria., pp 261-282.
Dinnebier RE and Friese K (2003), Modern XRD methods in mineralogy, Introduction to the Mineralogical Sciences: Encyclopedia of Life Support Systems (EOLSS). pp 1-57.
Draxler RR, Rolph GD (2003), HYSPLIT (Hybrid Single-Particle Lagrangian Integrated Trajectory) Model. https://www.ready.noaa.gov/HYSPLIT.php.
Eriksson PG, Altermann W and Hartzer FJ (2006), The Transvaal Supergroup and its precursor Complex in: Johnson MR, Anhaeusser CR and Thomas RJ (Eds) The Geology of South Africa, Geological Society of South Africa., Johannesburg/Council for Geoscience, Pretoria., pp 237-260.
Gaberšek M and Gosar M (2021), Towards a holistic approach to the geochemistry of solid inorganic particles in the urban environment., Science of the Total Environment Vol 763, pp 144214.
Hinterlechner RA (1971), Pohorske metamorfne kamenine., Geologija Vol 14, pp 187–226. http://www.geologija-revija.si/dokument.aspx?id=210.
House Dust, https://theconversation.com/house-dust-from -35-countries-reveals-our-global-toxic-contaminant-exposure-and-health-risk-172499 (Accessed 8 November 2023).
Howlett‑Downing C, Boman J, Molnár P, Shirinde J, Wichmann J (2022), PM_2.5 Chemical Composition and Geographical Origin of Air Masses in Pretoria, South Africa., Water Air Soil Pollut Vol 233, pp 271. https://doi.org/10.1007/s11270-022-05746-y.
HYSPLIT Tutorial 2023, https://www.ready.noaa.gov/documents/Tutorial_2021/html/traj_clus.html (accessed April 2023).
J Expo Sci Environ Epidemiol (2016). Author manuscript; available in PMC 2016 January 01.
Johnson MR, van Vuuren CJ, Visser JNJ, Cole DI, Wickens HV, Christie ADM, Roberts DL and Brandl G (2006), Sedimentary rocks of the Karoo Supergroup Complex in: Johnson MR, Anhaeusser CR and Thomas RJ (Eds) The Geology of South Africa, Geological Society of South Africa., Johannesburg/Council for Geoscience, Pretoria., pp 461-500.
Jones CR and Hepworth JV (1973), Geological Map of Botswana. Director of Geological Survey and Mines with the authority of the Ministry of Commerce, Industry and Water Affairs, Gaborone.
Kulkarni KN, Tripathi A, Varshney A, Chandra J, Kumar S, Sangal S, Paul D & Mondal K (2022) Electron Probe Micro-Analyzer: An Equipment for Accurate and Precise Micro-Composition Analysis., Electron Microscopy in Science and Engineering., pp 71–93.
Lee E, Terblans JJ & Swart HC (2018), The effect of pH on the luminescence properties of Y2O3: Bi phosphor powders synthesised using co-precipitation., Vacuum Vol 157, pp 237 – 242.
Maps, https://www.sa-venues.com/maps/default.htm (accessed November 2023).
Mencin GE, Jamšek RP, Trajanova M, Gale L, Bavec M, Anselmetti FS, Šmuc A (2019), Provenance and morphostratigraphy of the Pliocene-Quaternary sediments, in the Celje and Drava-Ptuj Basins (eastern Slovenia)., Geologija Vol 62, pp 189–218.https://doi.org/10.5474/geologija.2019.009
Mioč P, (1978), Basic Geological Map of SFR Yugoslavia 1:100.000, Explanatory Text for Slovenj Gradec Sheet., Zvezni geološki zavod, Beograd, pp 74.
Mioč P, Žnidarčič M (1989), Basic Geological Map of SFR Yugoslavia 1:100.000, Explanatory Text for Maribor and Leibnitz Sheet., Zvezni geološki zavod, Beograd, pp 60.
Moen HFG (2006), The Olifants Supergroup Complex in: Johnson MR, Anhaeusser CR and Thomas RJ (Eds) The Geology of South Africa, Geological Society of South Africa., Johannesburg/Council for Geoscience, Pretoria., pp 319-324.
Molnar P, Tang L, Sjoberg K & Wichmann J (2017), Long-range transport clusters and positive matrix factorization source apportionment for investigating transboundary PM_2.5in Gothenburg. Sweden., Environmental Science: Processes and Impacts Vol 19, pp 1270–1277. https://doi.org/10.1039/C7EM00122C.
Pawloski GA (1985), Quantitative determination of mineral content of geological samples by X-ray diffraction., American Mineralogist. Vol 70 (7-8), pp 663–667.
Sarpong SA, Donkoh RF, Konnuba JK, Ohene-Agyei C and Lee Y (2021), Analysis of PM2.5, PM10, and Total Suspended Particle Exposure in the Tema Metropolitan Area of Ghana., Atmosphere Vol 12, pp 700. https://doi.org/10.3390/atmos12060700.
Schwarz J, Cusack M, Karban J, Chalupničkova E, Havranek V, Smolik J & Ždimal V (2016), PM_2.5chemical composition at a rural background site in Central Europe, including correlation and air mass back trajectory analysis., Atmospheric Research Vol 176, pp 108–120. https://doi.org/10.1016/j.atmosres.2016.02.017.
Skinner EMW & Truswell JF (2006), Kimberlites Complex in: Johnson MR, Anhaeusser CR and Thomas RJ (Eds) The Geology of South Africa, Geological Society of South Africa., Johannesburg/Council for Geoscience, Pretoria., pp 651-660.
Šoster A, Zavašnik J, Ravnjak M, Herlec U (2017.), REE-bearing minerals in Drava river sediments, Slovenia, and their potential origin., Geologija Vol 60, pp 257–266. https://doi.org/10.5474/geologija.2017.018.
South African map, https://southafricamap360.com/south-africa-map (accessed November 2023).
Thamm AG & Johnson MR (2006), The Cape Supergroup Complex in: Johnson MR, Anhaeusser CR and Thomas RJ (Eds) The Geology of South Africa, Geological Society of South Africa., Johannesburg/Council for Geoscience, Pretoria., pp 443-460.
Tshehla C & Djolov G (2018), Source profiling, source apportionment and cluster transport analysis to identify the sources of PM and the origin of air masses to an industrialised rural area in Limpopo., Clean Air Journal Vol 28, pp 54–66. https://doi.org/10.17159/2410-972x/2018/v28n2a18.
van der Westhuizen D, Howlett-Downing C, Molnár P, Boman J, Wichmann J and von Eschwege KG (2023), Source apportionment of fine atmospheric particles in Bloemfontein, South Africa, using positive matrix factorization., Environmental Monitoring and Assessment, preprint. https://doi.org/10.21203/rs.3.rs-3265323/v1.
Wang H, Dwyer-Lindgren L, Lofgren KT, Rajaratnam JK, Marcus JR, Levin-Rector A, Levitz CE, Lopez AD, Murray CJL (2012), Age specific and sex-specific mortality in 187 countries, 1970–2010: A systematic analysis for the global burden of disease study 2010., Lancet Vol 380, pp 2071–2094.
Wichmann J, Sjoberg K, Tang L, Haeger-Eugensson M, Rosengren A, Andersson EM, et al. (2014), The effect of secondary inorganic aerosols, soot and the geographical origin of air mass on acute myocardial infarction hospitalisations in Gothenburg, Sweden during 1985–2010: A case crossover study., Environmental Health Vol 13, pp 1–16.
Williams J, Petrik L & Wichmann J (2020), PM_2.5 chemical composition and geographical origin of air masses in Cape Town, South Africa., Air Quality, Atmosphere and Health, pp 1–12. https://doi.org/10.1007/s11869-020-00947-y.
Willis F, Feather C & Turner K (2014), Guidelines for XRD analysis, James Willis Consultants CC, Cape Town., pp 544.
Worobiec A, Potgieter-Vermaak S.S., Berghmans P, Winkler H, Burger R and Van Grieken R (2011), Air Particulate Emissions in Developing Countries: A Case Study in South Africa., Analytical Letters Vol 44, pp 1907–1924. https://doi.org/10.1080/00032719.2010.539734.
Zupančič N (1994), Petrografske značilnosti in klasifikacija pohorskih magmatskih kamnin., Rudarsko metalurški zbornik Vol 41, pp 101–112.

No competing interests reported.

ArticleSADustPMSI.docx

Chemical Analyses and Geographical Origins of Residential Attic Dust in Central South Africa

Status:

Version 1

Abstract

Figures

1. Introduction

2. Materials and Methods