Source Identification and Pollution Factors of Elements in PM2.5 Samples Obtained in Akure, Ondo State, Nigeria

Air quality deterioration is a big concern all over the world. It affects both humans and animals in a variety of ways. Air pollution must be quantified to monitor its negative impacts and stakeholders to proffer mitigation options. On this note, our research team has taken measures in Nigeria to monitor and measure PM2.5, its associated elements, as well as local meteorological parameters during sampling period. In addition, enrichment, contamination index, and principal component analysis of associated elements were studied. For ten months (January–October 2018), PM2.5 samples were collected at three locations in Ondo State, Nigeria with coordinates (the Federal University of Technology, Akure (FUTA)—low-density residential (LDR), Oba Ile (high-density residential (HDR), and Museum (IND, High density traffic and commercial area) (January–October). The Contamination Factor (CF), Enrichment Factor (EF), and Pollution Load Index (PLI), as well as Principal Component Analysis (PCA) of the samples were determined. Average PM2.5 mass concentrations [101 µg/m3 (FUTA); 120 µg/m3 (Oba Ile); and 176 µg/m3 (Museum)] were found to be higher than the WHO (10 µg/m3) and USEPA (15 µg/m3) normal limits. The concentrations of the most prominent elements were in the order; K > Na > Ca > P > Al, while the trace elements displayed Cu > Fe > Zn > Ni > Cr > Mn > Pb. Oba-Ile had the highest metal enrichment in the order: Ti > Mn > Zn > Cu > Ni > Cr. The PCA resolved four factors vis-à-vis vehicular activities, biomass burning, and soil dust as the major PM2.5 emission sources. The findings could be useful in the current drive to develop national air quality guideline for PM2.5.


Introduction
Urban air pollution is typically caused by a complex mix of gaseous and particulate air pollutants such as nitrogen dioxide (NO 2 ), sulfur dioxide (SO 2 ), fine particulate matter (PM 2.5 ), and ground-level ozone (defined as aerodynamic diameter particulate matter (PM 2.5 ) (Abulude et al. 2021). Atmospheric particulate matter has been linked to negative effects on human health, visibility, and atmosphere, as well as habitats and physical property degradation (UNEP/WHO 1994). The scale of atmospheric aerosols affects human health and the climate. As a result, studies of size-segregated atmospheric particulate matter particularly fine (PM 2.5 ) fractions are often required to gain a better understanding of its sources as well as aids in proffering abatement options. Different sources (natural and anthropogenic) and atmospheric residence periods can exist for PM 2.5 fraction; however, PM 2.5 sources are majorly linked to anthropogenic activities such as industrial, biomass and petroleum components combustions, waste incineration, vehicular emissions, and so on (Ezeh et al. 2015;Suriano et al. 2015;Penza etal. 2018;Suriano 2020).
Continuous exposure to PM 2.5 exposure has a slew of negative health implications. In exposed populations, there have been increase in morbidity and mortality rates resulting from infections, cardiovascular, chronic pulmonary, and neurodegenerative diseases, as well as a decrease in life expectancy (Kim et al. 2015). PM 2.5 can be breathed into the deepest recesses of the lungs due to its vast surface area, small size, and the fact that it contains a significant amount of toxic properties. It settles on the alveoli and also makes its way into the bloodstream, posing a health risk to humans (Zhang et al. 2016). Fe, Co, Ni, Cu, Zn, V, Cr, Mn, As, Pb, and Cd can increase reactive oxygen species production (ROS) whose excess can overpower the body's antioxidant defense mechanism, leading to oxidative stress, inflammation, and disease infections (Sameenoi et al. 2012;Charrier and Anastasio 2015;Park et al. 2018).
Trace elements are efficient atmospheric air pollution tracers (Querol et al. 2007) and are commonly used as input in receptor modeling studies to help in the interpretation of particular emission sources. The Principal Component Factor (PCA), is one of the versatile modeling approaches which efficiently uses elemental concentration data and classifies potential source contributions. The characterization and source detection of trace elements in PM 2.5 are critical in air pollution prevention and control. These methods have attracted a lot of attention from researchers in recent times, because PMs Study could reveal information regarding emission sources and strength.
To address the problem of air pollution, stakeholders would need to have adequate knowledge of air pollution by particulate matter and also on policies for abatement options. The best tool which is receptor modeling for PMs' source identification and quantification should be applied (Yu et al. 2013;Roy and Singh 2014;Ndamitso et al. 2016;Feng et al. 2007). The application of a multivariate technique to identify and assign sources to pollutants at the receptor locations irrespective of emission time, strength, and transport is known as receptor modeling. In most urban cities in the developing nations, there is very little or no published work on source apportionment and PM 2.5 and thus this study. Therefore, the objectives of this study are to determine PM 2.5 mass loads, investigate the relationship between PM 2.5 and local meteorology, and perform chemical compositions of PM 2.5 samples as well as the application of principal component analysis technique for source identification and quantification in Akure, South Western Nigeria.

Materials and Methods
Akure is the capital of Ondo State, South West and it has a population of about 421,100. Geographically, Akure lies on t 5°12′0″E, and 7°15′0″N, at a height of 353 m (GeoNames Geographical Database 2012). Akure is a fast-growing city that serves as a trading hub for cash crops such as cocoa, cola nuts, palm oil and kernels, bitter cola citrus fruits, and so on.
Three distinct sampling locations (National Museum and Monuments, Oba-Ile, and the Federal University of Technology-FUTA) were chosen (Table 1 and Fig. 1). These sites represents industrial (IND), high-density residential (HDR), and low-density residential (LDR) site classes. Samplings lasted for at least 8 h (8:00-16:00 h.) for ten (10) months (January-October) of 2018. GPS Map 76CSX (Garmin Ltd, Taiwan) was used to assess the sampling geographical coordinates. At the Observatory Unit of the Federal University of Technology in Akure, Ondo State, Nigeria, wind direction and speed, rain, and relative humidity (RH) were also monitored and measured to aid in the explanation of the influence of local meteorology on PM 2.5 aerosols.
PM was collected using a Schlumberger Model M250 "Gent" stacked filter unit sampler (Maenhaut et al. 1993). The samples were collected using Whatman nuclepore filters that had been pre-weighed and pre-conditioned, at flow rates ranging from 16 to 18 L min. Before and after exposure, the filters were conditioned for 24 h at about 25 °C and 50% constant humidity. The sampler was positioned at a height of around 1.7 m, so that air circulation around it was not obstructed. To prevent filter clogging and keep the flow rate within the sampler's prescribed limits, the effective sampling time was varied. This ensured accurate size fractionation and efficient selection. Detailed procedure on the use of Gent stacked filter unit sampler for PM 2.5 collection has been reported (Ezeh et al. 2012). The exposed filters were stored in a desiccator until chemical composition analysis using the XRF technique.
The EDXRF analyses of exposed filters were achieved via a portable facility (model PX 2CR power supply and amplifier for XR-100CR Si-pin Detector) at Centre for Energy Research and Development, Obafemi Awolowo University, Ile-Ife Nigeria. Sample irradiation was done by exposing the samples for bombardment by silver (Ag) anode X-ray fluorescence spectrometer at a voltage of 25 kV and current of 50 µA for 1200 counts in an external chamber setup. Characteristic X-ray of the sample was detected by the solidstate Si-pin detector system and spectrum acquisition was done using an Amptek model multi-channel analyzer, while elemental analysis was done using the thin target mode of the International Atomic Energy Agency (IAEA) Quantitative Analysis of X-ray Iterative Least (Q-Axil) square software. Prior to the analysis, certified Micrometer ® thin films (Table 2) were irradiated for calibration purposes.

Calculation
Principal component analysis (PCA) defined by James et al. (2015) was used to classify the sources of the pollutants Instrument reading (cm 2 ) × 1000 Filter area (cm 2 ) × Volume of air sucked (m 3 ) .
To identify the sources of PM 2.5 at the study areas, principal component analysis (PCA) with varimax rotation of the components was achieved using Minitab software. Varimax rotation was done to get a clearer pattern of the components without changing their relative positions. Detailed methodology for performing PCA can be found in Jolliffe et al. (2016) and De La Cruz et al. (2019).
The enrichment factor (EF) of the element of interest is the ratio of the element's concentration to a reference value   (Gonzalez-Macias et al. 2006;Pandey et al. 2015). Fe was often chosen as reference factor, because it is assumed to have only source of crustal material. The enrichment factor (EF) is a tool for assessing whether the source of elements is from crust or human activity (Bhuiyan et al. 2010;Gowd et al. 2010). This factor was used to measure EF: where C i denotes the concentration of trace elements in the sample, C ref denotes the concentration of the reference element in the sample, B i denotes the background value of an element of interest, and B ref denotes the background value of the reference element in the study field. The results were categorized into seven classes based on the value of EF: If EF < 1 no enrichment, < 3 minor enrichment, 3-5 moderate enrichment, 5-10 moderate-to-severe enrichment, 10-25 severe enrichment, 26-50 very severe enrichment, and > 50 extremely severe enrichment. Contamination factor (CF) as defined by Dantu (2009) andBhuiyan et al. (2010) CFs of the elements of interest was used to calculate the overall contamination of samples by the elements (2010). These figures were calculated by dividing the concentration of the element of interest in the sample by the element's background concentration (Hakanson 1980) C i is the element i's concentration and B i is the element's geochemical background value. The pollution values are 0 = none, 1 = none-to-mild, 2 = moderate, 3 = moderate-toheavy, 4 = strongly contaminated, 5 = strong-to-very strong, and 6 = very strong in increasing order of contamination (Varol 2011).
The pollution load index (PLI) was created after Tomlinson et al. (1980) used Eq. (3) to compare pollution levels at different sites or at the same site over time. The estimation of the concentration factor obtained by dividing the measured concentration of an element with the background concentration of the same element in shale is used to calculate the PLI (Turekian and Wedepohl 1961)

Calculation of Air Quality Index
An index for any given pollutant is its concentration expressed as a percentage of the relevant standard (Abulude 2016), or (5) Air Quality Index = Pollutant concentration Pollutant standard level × 100.
One-way analysis of variance (ANOVA) with Duncan Multiple Range test at 95% confidence or p 0.05 was used to analyze the results. The correlation coefficients were determined based on the results. Minitab 16 Statistical Software was used to compute basic descriptions and PCA. Microsoft Excel was used to calculate the enrichment factor, contamination factor, and wind rose.

Results and Discussion
The average concentrations of PM 2.5 measured at the study site classes are presented in Fig. 2. During the sampling periods, mean concentrations of PM 2.5 ranged from 101 μg/m 3 (LDR) to 176 μg/m 3 (HDR), while the overall mean stood at 120 μg/m 3 . Compared with the results obtained elsewhere in some sub-Sahara mega-cities in Nigeria, the concentrations obtained in this study were lower than range of values reported for IND and HDR sites in Lagos (Ezeh et al. 2017). 150-606 μg/m 3 (industrial) and 110-460 μg/m 3 (high-density residential) reported by Ezeh et al. (2019) However, mean concentration of PM 2.5 obtained in this study were higher than mean values recorded in China. Acciai et al. (2017) and Huang et al. (2018) equally, higher than values (Gunchin et al. 2019), measured in Ulaanbaatar Mongolia, by 10%, Košetice Czech (Pokorná et al. 2018), by 10%, Terviso Italy (Squizzato et al. 2017), and by 15% and Athens Greece (Grivas et al. 2018) by 20%). Compared to PM 2.5 guidelines of the WHO and USEPA (Fig. 2), the results obtained in this study were more than three times of both reference values. This scenario is worrisome and could be dangerous for the inhabitants of the study areas.
Considering the US EPA's Air Quality Index (AQI) for reporting air quality (AirNow.gov 2018) deterioration by PM 2.5 ; Scale 0-50 is defined as a Good, 51-100 is Moderate, 101-150 is Unhealthy for sensitive groups, 151-200 indicated Unhealthy, 201-300 denoted Very Unhealthily, and 301 and above is classified as Hazardous. In this study, average PM 2.5 concentrations obtained at LDR and Oba Ile (IND) areas were categorized as Unhealthy for sensitive groups, while AQI for HDR receptor site is categorized as Unhealthy (Table 3). This index implies that members of sensitive groups may experience health effects.
In Table 4, parameters with strong positive correlation are bolded. Although, there was a mild positive association between PM 2.5 mass concentrations and temperature (Table 4). This means that PM 2.5 concentration is a function of the ambient temperature. At the HDR, LDR, and IND sampling sites, wind speed (WS) has a significant positive association with PM 2.5 (r = 0.87, r = 0.61, and r = 0.35), with the HDR having the strongest connection. The HDR, LDR, and IND had r values of 0.59, 0.62, and 0.71 for ambient temperature and PM 2.5 values, respectively. With r values of 0.05 (FUTA), 0.28 (Museum), and 0.31 (FUTA), relative humidity (RH) had a weak relationship with PM 2.5 mass concentration at IND. It was also discovered that the meteorological parameters had relationships that effects PM 2.5 mass concentration. Rain (r = 0.67) was found to be related to wind speed. Wind also had positive (r = 0.50) relationship with humidity. These positive correlations suggest that local emission could account for air quality deterioration in Akure by PM 2.5 .
The summaries of wind directions in percentages and directions as the wind rose are shown in Fig. 3. Each month, the wind directions are denoted by the letters 'N' (north) to 'NNW (north-north-west).' The percentages represents the number of times each wind direction occurred during the sampling period. The wind blew majorly from the east in January, while in the following months, its directions were predominantly; ESE (February), SW (March), WSW (April), WSW (May), S (June), SSW (July), SW (August),    Fig. 3 Wind direction (July-October 2018) SW (September), and SW (October) (SW). These confirms that the wind direction during sampling were predominantly from the SW and WSW. Table 5 provides average local meteorology data measure during sampling period 3.71 mm, (Rainfall) 26.10 °C, (Temperature) 5.84 m/s (Wind speed), and 77.34% (Humidity). The variance ranged from 1.27 (rain) to 11.54 (humidity). Interestingly, rain had positive values, while others had negatives, according to the skewness results. Negative values were skewed to the left, while positive values were skewed to the right, implying that if the distribution of data is skewed to the left, the mean is less than the median, which is frequently less than the mode. If the data distribution is skewed to the right, the mode is frequently less than the median, which is less than the mean In addition, the monthly mean wind speed ranged from 4.00 to 7.63 m/s, while that of temperature and humidity ranged from 19.75 to 31.23 °C and 21 to 94.38% respectively. Rainfall ranged from 0 (dry season) to 3.5 mm (wet season).
The concentrations of major and trace elements in the PM 2.5 were calculated and are presented in Table 6. In addition, the results were compared to values reported for other cities in Nigeria and beyond. The overall summation of the elemental concentrations (Fig. 4) varied throughout the study sites, with the highest observed at the HDR site, followed by IND and LDR, respectively. The heavy human and vehicular traffic at the HDR location could led to the high concentrations of major and trace elements found there. Main and trace elements were classified into two groups based on their temporal and spatial variation; group 1 consists of Na, Mg, Al, K, Ca, Ti, V, P, Cr, Mn, Ni, Zn, and Fe, while group 2 consists of S, K, and Cu. Copper was found in high concentrations in the IND and HDR regions, while S and K were found in high concentrations in all the sites. The mean concentration of V (15-28 ng m −3 ), Mn (13-25 ng m −3 ), Ni (15-28 ng m −3 ), and Pb (ND) were all within the limits set by the World Health Organization (WHO; 1000, 150, 25, and 500 ng m −3 for V, Mn, Ni, and Pb, respectively), the European Air Quality Directive (EU Directive 2007/107/ EC; 20 and 500 ng m −3 for Ni and Pb, respectively), and the National Ambient Air Quality Standard (NAAQS) of China (500 ng m 3 for Pb) (Liu et al. 2020).
The enrichment factor, or EF, is a tool for assessing the chemical structure of airborne particulate matter. It explains the relationship between the concentrations of an element (X) in the air and the concentration of a crustal element (such as Al, Ti, or Fe) in the typical continental crust (Hoffmann et al. 1972;Wedepohl 1971;Zoller et al.1983). The crustal composition (Taylor 1964) and Fe as the normalizing factor were used to measure EFs for each element. Since Fe is a crustal variable with less anthropogenic impact, it is a good option for a normalizing agent. Elements with an EF close to unity have strong natural source, while those with a high EF may be anthropogenic aerosols. In Fig. 5, the enrichment factor for PM 2.5 associated elements is shown and most elements were not enriched, since their EF values were less than five. For instance, respective EF values were 0.67-1.02 (Na), 0.08-0.14 (Al), 3.05-3.39 (S), 1.6-3.52 (K), 1 (Fe), 1.33-2.69 (Cr), and 1.23-2.69 (V) while Se and Pb had zero enrichment factors. Low enrichment factors could indicate that the pollutants could be from natural source, especially sea spray and soil dust components.
The pollution load index (PLI), which had values ranging from 0.1 to 1.2, was also used to categorize ecological risk. The mean PLI value (0.1) in this analysis was much smaller than the threshold (1) and suggested that there were no baseline contaminants or loads similar to background levels (Tomlinson et al. 1980). Because of the low PLI values, no drastic abatement measures are needed at the study areas; however, continued air quality deterioration by PM 2.5 . The findings showed that ecological risk assessment approaches can be used as decision-support mechanisms or instruments for determining the priority of air quality assessment studies.
For source apportionment and identification, PCA resolved five distinct factors with their corresponding components (Table 7). Factors 1,2,3,4,and 5 had 65.5,65.3,32.7,23.2,and 11.5% average variation, respectively. Factor 1 revealed significant loading for Mg, P, Ca, Cr, Fe, Ni, and Zn, indicating input from the burning of automotive fuels, wastes, and tobacco smoke in and around the museum district, due to the high volume of vehicular traffic and waste generation generated by the market within the area. Without a doubt, there are a lot of smokers in the city, and a lot of people trade their hobbies. This aggravates the danger to the environment. Suffocation and worsening of asthma and chronic bronchitis, as well as respiratory inflammation, eye and mouth pain, and premature death, are all possible health effects. P, Fe, Cu, and K, as well as their sources, which were mostly parented rock material, dust, soil, and fertilizer, were all present in VF2. However, these metals are also used in liming processes, which can result in respiratory problems, metallic dust poisoning, pulmonary fibrosis, lung cancer, dizziness, headaches, bronchitis, stomachaches, diarrhea, vomiting, loss of appetite, kidney and liver damage, and death (Wei et al. 2010). Cu, Ti, and Na loadings were highest in VF3, suggesting biomass combustion, diesel and residual oils, tobacco smoke, and steel non-ferrous alloys manufacturing. In the tannery industry, manganese is used in the form of salts [MnCO 3 and Mn (SO 4 ) 2 ] (Tariq et al. 2006). Ca sources from the PM were illustrated by VF4. Plants and animals, as well as the skeletons of animals, teeth, egg shells, coral, and many soils, are major sources of Ca in current PM. Calcium chloride is present in some water, especially seawater, at a concentration of 0.15%. Calcium does not occur in nature on its own. Calcium is mainly present in the form of limestone, gypsum, and other minerals. Significant loadings for S were found in VF5, indicating sources related to anthropogenic activities, specifically tire vulcanization. Many vulcanizers work along the streets in these neighborhoods, especially in the Oba Ile and Museum areas. Zinc is a vital trace element in living organisms, but it is insoluble and extremely rare in nature (Alloway 1990).

Conclusion
The PM 2.5 levels found in this study were higher than those recorded by the WHO, the World Health Organization, and the US Environmental Protection Agency, and some countries in Asia, Europe, and the United States. The PM is dangerous for sensitive groups, according to the AQI of the places. PM2.5's effect on trace and heavy element enrichment, source apportionment, and pollution load index was also investigated. In contrast to previous research on PM performed elsewhere, measured concentrations of macronutrients and heavy metals were higher or lower. Cr, Ni, Cu, Mn, and Zn were added to the PM. Pollution was minimal in the elements. Anthropogenic and non-anthropogenic origins of PM were identified through multivariate analysis. Metals in the air can be apportioned at the source, which can aid decision-makers in developing management strategies to minimize pollution. The climate should be continually and properly controlled in the locations under study.