Introduction

Industrial and urban activities, including fossil fuel burning, are inevitably associated with the release of potentially toxic elements (PTEs) to the atmosphere (Cusack et al., 2020; Ramírez et al., 2020; Zang et al., 2021). Emitted trace elements and nutrients from anthropogenic sources are eventually deposited in soil where they get progressively mixed with geogenic sources (Antoniadis et al., 2019; Kabata-Pendias, 2011; Rinklebe et al., 2019; Yuan et al., 2019), a process which affect their speciation and availability, and therefore their toxicity to humans, plants and animals (Huang et al., 2020; Korai et al., 2021; Shaheen et al., 2020).

Egypt is one of the most populated developing countries (United-Nations, 2019), and its capital (Cairo) is one of the world’s mega cities and one of the largest urban/industrial agglomerations in the world (Cheng et al., 2016; United-Nations, 2018). The transformation of Egypt from an agricultural to an industrial economy and the urban expansion that started in the 1950s, in addition to agricultural waste incineration (Mohamed et al., 2015), have resulted in an increase in the environmental pollution with PTEs. By the 1990s, the concentrations of PTEs in the atmosphere of major Egyptian cities, around motorways and in industrial areas, had reached deleterious levels (Hassanien & Horvath, 1999). The environmental protection policies adopted by the Egyptian government, e.g., the relocation of industrial areas outside cities and using natural gas to fuel power stations, taxis and public buses, have resulted in a decline in the atmospheric concentrations of some PTEs to safe levels (Hassan et al., 2013).

Soil is currently the main reservoir of PTEs in Egypt (Shetaya et al., 2018) and although the biogeochemical status of classical PTEs, e.g., Pb, Zn and Cu, has been investigated in numerous studies, little is known about the bioavailability and geochemical behavior of other soil PTEs, particularly the emerging PTEs (e.g., V, Co, Mo, Sb and Sn) and their potential effect on public health. Despite the successful attempts of Shaheen et al. (2019a; 2020) and Shaheen and Rinklebe (2018) and others to investigate the origins and mobility of various PTEs in the soils of Egypt, they have mainly focused on limited agricultural areas. Therefore, there is a pressing need to study the biogeochemical status of nonclassical and emerging PTEs, especially in the most populated and polluted industrial and urban areas in Egypt, e.g., Greater Cairo. This is a key to drafting future policymaking and remediation strategies and may also help to understand and ‘predict’ the biogeochemical behavior of PTEs in other arid and semiarid largely populated developing countries similar to Egypt.

The main aim of this work is, thus, to investigate the status of nine PTEs (V, Cr, Co, As, Se, Mo, Cd, Sn and Sb) in soils and plants and to assess their risks on the health of the Egyptian population using risk assessment models. This has been achieved by determining the total content of the studied PTEs in 116 surface soil samples and associated 72 plant samples collected from industrial, urban and agricultural locations, across Egypt, which have been subjected to various degrees of aerial depositions from different emission sources. Chemical fractionation (of soil) and statistical analysis approaches were used to understand the origins and mobility of the studied PTEs in relation to soil properties and level of pollution. Risk assessment indices were also used to evaluate the level of soil pollution and to assess the carcinogenic and noncarcinogenic risks posed by these PTEs on the health of the residents (adults and children) of the studied areas.

Materials and methods

Soil and plant sampling

A total of 116 soils were sampled from 11 sites including industrial, urban and agricultural locations that have been exposed to historical aerial depositions to various degrees and from a variety of sources across Egypt (Table 1). Soil samples were surface soils (0–20 cm depth) randomly collected across transects within 100 m of main roads. A total of 72 soil-associated leafy and fruiting plants commonly consumed by local residents and animals were collected from some of the soil sampling locations (Table 1). A sampling map showing soils and plants sampling locations is included in the supplementary materials (Fig. A1; Appendix A).

Table 1 Soil and plant samples locations, codes, and description. Letters I, U and G in the sample codes refer to the industrial, urban or agricultural nature of the sampling location, respectively
Full size table

Soils and plants characterization and elemental analysis

Collected soil samples were air-dried, sieved to < 2 mm particle size, homogenized and stored in clean polyethylene bags at room temperature. Approximately, 25 g of each soil was ground to a fine powder in an agate ball mill (Retsch, Model PM400). To measure soil pH, 5 g of the < 2 mm sieved soils were suspended for 30 min in 12.5 mL ultra-pure water (18.3 MΩ cm−1) and pH was measured after 5 min using a combined glass electrode (Ag/AgCl; PHE 1004, Hanna pH-209) (Atkinson et al., 2011). Soil organic carbon (SOC) was extracted by suspending c. 2 g samples of the < 2 mm soils in 0.1 M NaOH solution and shaking for 24 h (You et al., 2006). SOC concentrations in the centrifuged and filtered extracts were measured using a Shimadzu total organic carbon analyzer (TOC-VCPH). Soil organic matter content (SOM) was then estimated by multiplying SOC by the van Bemmelen's factor of 1.72 (Grybos et al., 2007).

Total soil concentrations of V, Cr, Co, As, Se, Mo, Cd, Sn and Sb in addition to Fe, Al and Mn were determined by digesting 0.2 g of finely ground soils with a mixture of trace analysis grade (TAG) 2.5 mL hydrofluoric acid (40%), 2 mL HNO3, 1 mL HClO4 and 2.5 mL H2O, using a Teflon-coated graphite block digester (Model A3, Analysco Ltd, UK) (Marzouk et al., 2013). For plants, edible parts (leaves or fruits of leafy and fruiting plants, respectively) were washed carefully with DI water before oven-drying at 60 °C and milling (Retsch ZM 200 centrifugal mill). Subsamples (c. 0.2 g) were microwave digested (in duplicate) with 6 mL TAG HNO3 for 15 min at 1400 W (Anton Parr, UK) and syringe filtered (0.22 µm) (Nabulo et al., 2011). Multi-element analyses of the diluted soil and plant digestates were undertaken by ICP-MS (section A.1; Appendix A).

Geochemical fractions of soil PTEs

The chemical fractions of V, Cr, Co, As, Se, Mo, Cd, Sn and Sb in selected 58 soil samples collected from all 11 locations were extracted according to a sequential extraction protocol (SEP) adapted from Li and Thornton (2001) as summarized in Table 2. Soil samples were prepared in duplicate, centrifuged after each extraction step and the supernatant solution filtered and assayed for V, Cr, Co, As, Se, Mo, Cd, Sn and Sb concentrations by ICP-MS (Sect. A.1; Appendix A). Carry-over was accounted for gravimetrically. To minimize the inevitable accumulative error in the residual phase (F5), the F5 concentrations (mg kg−1 soil) were calculated by subtracting the summation of the corresponding element concentrations in the first four steps (F1–F4) from the total soil concentrations measured in the HF–HClO4–HNO3 acid soil digestates (Shetaya et al., 2019b).

Table 2 Sequential extraction procedure, as
Full size table

Soil contamination indices

Soil geo-accumulation (I geo) index

The geo-accumulation index (Igeo; Eq. 1) was used to quantify the degree of contamination of each TE in comparison to its crustal concentration.

$$I_{{{\text{geo}}}} = \log_2\left( {\frac{{C_{{{\text{soil}}}} }}{{1.5 C_{{{\text{bkg}}}} }}} \right)$$
(1)

where Csoil is the measured concentration of an element in soil (mg kg−1) and Cbkg is the average crustal level of that element. Crustal average concentrations (displayed in Table 3 (Emsley, 2011)) were used here due to the lack of documented local background soil concentrations for some of the emerging contaminants (Co, Mo, Sb and Sn) and in order to have a uniform scale for comparison between all of the studied PTEs (Birch, 2017; Kabata-Pendias, 2011).

Table 3 Concentration (mg kg−1 soil) of PTEs in the studied soils.
Full size table

The Igeo index has seven degrees of contamination ranging from uncontaminated to extremely contaminated: (i) uncontaminated (Igeo ≤ 0), (ii) uncontaminated to moderately contaminated (Igeo 0 − 1), (iii) moderately contaminated (Igeo 1 − 2), (iv) moderately to heavily contaminated (Igeo 2 − 3), (v) heavily contaminated (Igeo 3 − 4), (vi) heavily to extremely contaminated (Igeo 4 − 5) and (vii) extremely contaminated (Igeo ≥ 5) (Antoniadis et al., 2017b; Li et al., 2014).

Pollution load index (PLI)

The PLI is used to determine the level of soil contamination with PTEs and is defined as the nth root of the multiplications of the contamination factor of elements (Islam et al., 2015; Shaheen et al., 2018). Individual elements contamination factors (CF) are calculated by Eq. 2.

$${\text{CF}} = \frac{{C_{{{\text{soil}}}} }}{{C_{{{\text{bkg}}}} }}$$
(2)

where Csoil is total element concentration in the soils; Cbkg is the background reference element concentration in uncontaminated soils (Shaheen et al., 2017) (used here as the average crustal level of any given element as explained in Sect. “Soil geo-accumulation (Igeo) index”).

The PLI for each soil can then be calculated using Eq. 3.

$${\text{PLI }} = \, \left( {Cf_{e1} . Cf_{e2} \ldots . \, Cf_{en} } \right)^{1/n}$$
(3)

where Cfe1, Cfe2 and Cfen are the CF of elements 1, 2, …, n, and n is the number of measured elements.

PTEs in plants

A proportion of the PTEs concentrations measured in the plants extracts is likely originating from extraneous contamination of the plants tissues with re-suspended soil particles. Washing plants with DI water or even with stronger reagents, e.g., EDTA, has been reported to be not entirely effective in eliminating soil contamination due to the embedment of submicron soil dust within plant tissue (Notten et al., 2008). The concentration of vanadium in plants tissues can be a good indicator of soil contamination since most of soil V is locked as immobile V(III) within ferric oxide crystals (Schwertmann & Pfab, 1996). Therefore, most of soil vanadium is inaccessible for systematic uptake through plants roots. The proportion of other PTEs in plants originating from extraneous soil contamination (PTEcont) can thus be calculated from Eq. 4 (Joy et al., 2015).

$${\text{PTE}}_{{{\text{cont}}}} = \frac{{V_{{{\text{plant}}}} . {\text{PTE}}_{{{\text{soil}}}} }}{{V_{{{\text{soil}}}} . {\text{PTE}}_{{{\text{plant}}}} }}$$
(4)

where Vplant and Vsoil are the concentrations (mg kg−1; dry weight basis) of V in plants and associated soils, respectively. PTEplant and PTEsoil are the concentrations (mg kg−1; dry weight basis) of any other PTE in plants and associated soils, respectively.

The relative enrichment of PTEs in plants through systematic uptake (via roots) can be estimated by calculating the bio-concentration factor (Eq. 5) which represent the soil to plant transfer factor of any given element (Antoniadis et al., 2021; Zeng et al., 2018).

$${\text{BCF = }} \frac{{C_{{{\text{plant}}}} }}{{C_{{{\text{soil}}}} }}$$
(5)

where Cplant and Csoil are the concentrations (mg kg−1; dry weight basis) of PTEs in plants and associated soils, respectively.

Health risk assessment of PTEs in soil

Humans are exposed to PTEs in soil through direct ingestion, dermal absorption and inhalation of soil particles. The doses received (mg kg−1 day−1) from each exposure route were calculated from Eq. 6, 7 and 8 for ingestion, dermal and inhalation, respectively (Doabi et al., 2018; USEPA, 1989, 2004).

$$D_{{{\text{ing}}}} = \frac{{\text{C . IngR . EF . ED . CF}}}{{{\text{BW . AT}}}}$$
(6)
$$D_{\text{drm}} = \frac{{\text{C . AF . SA . ABS . EF . ED . CF}}}{{\text{BW . AT}}}$$
(7)
$$D_ {\text{inh}} = \frac{{\text{C . InhR . EF . ED . CF}}}{{\text{PEF . BW . AT}}}$$
(8)

where Ding, Ddrm and Dinh are the ingestion, dermal and inhalation doses (mg kg−1 day−1), respectively, C is the concentration of any given element (PTE) in soil (mg kg−1), which was calculated for each sampling location as the 95% upper confidence limit of the mean (95% UCL; Table A8 (Appendix A)) to yield a reasonable maximum exposure to that element (Doabi et al., 2018; Ferreira-Baptista & De Miguel, 2005). Terms values (with source references) and more detailed information can be found in section A.2 (Appendix A).

Noncarcinogenic risks were calculated for each PTE from any given exposure route as a hazard quotient (HQ; Eq. 9) (Doabi et al., 2018; USEPA, 1989).

$${\text{HQ}} = \frac{{D_{x} }}{{D_{{{\text{ref}}}} }}$$
(9)

where Dx is the dose received from any exposure route (Ding, Ddrm or Dinh) and Dref is the reference dose (the toxicity threshold value for a specific chemical). Terms values (with source references) and more detailed information can be found in Section (Appendix A).

The noncarcinogenic hazard index (HI) was then calculated for each PTE as the summation of the corresponding HQs of all exposure routes. The combined HI of all PTEs was calculated similarly by summation of individual HIs (Doabi et al., 2018; USEPA, 1989). A value of HQ or HI that is ≤ 1 indicates no significant risk of noncarcinogenic health issues for exposed population; a value of HQ or HI > 1 means high risk of noncarcinogenic health effects.

Carcinogenic hazard quotients (HQcar) for As and Cd (HQcar) were calculated using (Eq. 10) (Doabi et al., 2018; Ferreira-Baptista & De Miguel, 2005).

$${\text{HQ}}_{{{\text{car}}}} = D_{x} . {\text{SF}}$$
(10)

where Dx is the dose received of As or Cd (mg kg−1 day−1) from any exposure route (Ding, Ddrm or Dinh) and SF is the carcinogenicity slope factor (mg kg−1 day−1). Terms values can be found in section A.2 (Appendix A).

The carcinogenic risk (CR) was then estimated as the summation of the corresponding HQcar from different exposure routes (Doabi et al., 2018; Ferreira-Baptista & De Miguel, 2005). Carcinogenic risk (CR) indicates the probability of developing any type of cancer during a lifetime (70 years) of exposure to carcinogens. Carcinogenic risk (CR) values lower than 1 × 10–6 indicate negligible risk while those higher than 1 × 10–4 indicate a probable risk of developing cancer; CR between 1 × 10–6 and 1 × 10–4 indicates acceptable risk for population health and social stability (Doabi et al., 2018; Li et al., 2014). The meaning of a cancer risk of x is that there is a risk of additional x persons developing cancer due to the exposure to PTEs on top of the usual number of persons expected to be ailing with cancer. Thus, a CR of 1 × 10–6 means that one additional person on top of the usually expected one million cancer patients is in the risk of developing cancer due to the effect of PTEs (Antoniadis et al., 2019).

Quality assurance and statistical methods

All analyses were undertaken in duplicate and repeated if the differences were > 5%. Reported differences between data sets were always checked for significance using paired t-test or 2-sample t-test and p-values < 0.05 were considered significant. Quality assurance of the multi-element extraction procedures and the analyses of soil and plant digestates by ICP-MS were verified with reference soil (NIST 2711) and plant (NIST 1567a) materials. For all of the assayed PTEs, the recovery was 96 ± 4% in average. The quality of the sequential extraction was audited using 11 soil samples in which the residual (F5) phase was digested using HF-HClO4-HNO3 (Atkinson et al., 2011); a recovery of 95–110% was achieved across all the investigated PTEs. The limits of detection (LOD) for all the assayed elements were automatically calculated by the Qtegra software of the ICP-MS based on the analysis of 16 blanks (Marchetti et al., 2019). In risk assessment modeling, the 95% UCL (upper confidence limits) of the mean concentrations at each sampling site were calculated by 1-sample t-test using Minitab 17 software.

Results and discussion

Soil properties

Key soil parameters are displayed in Table A1 (Appendix A). All soils were alkaline and their average pH ranged from 7.8 to 9. Soil organic matter content was relatively low (< 1%) in most of the studied soils with the exception of locations CA-I and SH-I, which showed average organic matter contents of 1.78 and 2.72%, respectively.

Occurrence and contamination levels of PTEs in soil

The total content (mg kg−1 soil) of V, Cr, Co, As, Se, Mo, Cd, Sn and Sb in the investigated soils in comparison with crustal background abundances (Emsley, 2011) is displayed in Table 3. Selenium was highly elevated above its background levels in all sampling locations (~ 3–100 times its crustal abundance) and the highest concentration was found in the soils of the Pb smelter site (SH-I) (Table 3). The Igeo indices of Se were > 0 in all locations with a maximum of 4.7 in the Pb smelter site (SH-I, Table A2), indicating an overall moderate to heavy contamination of Se in the soils of Egypt. However, this could be due to the sedimentary origin of the Egyptian soils, which are known to be generally enriched in Se (Shaheen, 2009; Shaheen & Rinklebe, 2014, 2018; Shaheen et al., 2017, 2020).

The levels of Sb, Cd, Sn and As in most of the studied soils were slightly below or moderately above their respective average geogenic levels with a few exceptions (Table 3). Antimony concentrations were extremely elevated (~ 300 times its background levels) in the Pb smelter soils (SH-I) (Table 3) with an Igeo index of 7.2 (Table A2; Appendix A), suggesting severe Sb contamination in this location soils. Cadmium (Cd) concentrations in the soils of locations SH-I (Pb smelter) and SH-U (heavy traffic) were almost 13- and 20-fold its crustal abundance (Table 3), with an average Igeo of 2.3 and 1.1 (Table A2), respectively, suggesting moderate Cd contamination from anthropogenic sources. High levels of Sn (18.3 and 35.3 mg kg−1) were also found in the industrial locations SH-I (Pb smelter) and CA-I (waste treatment) soils, with Igeo indices of 2 and 1.3 (moderate contamination), respectively (Tables 3 and A2). Arsenic was also noticeably elevated (6.78 mg kg−1) in location SH-I (Pb smelter) (Table 3) with an Igeo index of 1.2 (Table A2), indicating low to moderate level of As contamination in this site.

The considerably high enrichment of Se, Sb, Cd, Sn and As in the Pb smelter (SH-I) soils is likely due to the nature of Pb smelting operations which is inevitably associated with the release of various elements (Ettler et al., 2012; Makhinova & Makhinov, 2020; Rieuwerts & Farago, 1996; Stafilov et al., 2010). In addition, the Pb production in the smelter included recycling scrap Pb (e.g., from old batteries) (Labib et al., 2003; Safar et al., 2014), which may explain the considerably higher soil concentrations of As, Se and Sb, as they are often alloyed with Pb to improve its physicochemical properties, e.g., rigidity, conductivity and corrosion resistance (Crompton, 2000; Lyon, 2010).

The soil concentrations of V (14–157 mg kg−1), Cr (24–123 mg kg−1), Co (3.2–23 mg kg−1) and Mo (0.26–2.88 mg kg−1) were generally below or marginally above their corresponding average crustal abundance in most of the studied locations. However, V and Cr showed slightly elevated levels in the soils of the urban and industrial locations SH-U (heavy traffic) and CA-I (waste treatment), respectively. Nevertheless, the geo-accumulation index (Igeo) was < 0 for V, Cr and Co in all of the studied locations (Table A2), indicating negligible anthropogenic input of these three PTEs. Similarly, although the soil contents of Mo in the urban/motorways locations IS-U and AM-U were almost twice its crustal level of 1.2 mg kg−1 (Table 3), the calculated Igeo in these two locations (− 0.2 and 0.1, respectively; Table A2) suggested trivial anthropogenic inputs. The fact that the range of these elements was indeed wide, while their Igeo values were < 0, shows that there is a difficulty in the use of correct background values. In some detailed studies, e.g., Reimann and Garrett (2005) and Reimann and de Caritat (2017), it was shown that the identification of appropriate background values is a complex task and that in many cases a range of background values per element instead of a single value may be used. However, in this work, given the complexity of this issue, we used the crustal abundance values, which helped us in keeping a uniform approach in all of the studied areas.

The pollution load indices (PLI; Eq. 3.), of all PTEs combined, ranged from 0.26 in the soils of the agricultural location IS-G to 3.6 in the Pb smelter (SH-I) soils (Fig. 1 and Table A3 (Appendix A)). PLI was higher than unity in the industrial and urban locations SH-I, CA-I (waste treatment) and SH-U (heavy traffic), indicating significant pollution with the studied PTEs in the soils of these three locations. In most of the studied soils, Se was the largest contributor to the total pollution load followed by Sb and Cd. However, in the soils of four locations that are characterized by heavy industrial and/or urban activities (SH-I, SH-U, CA-I and CA-U), the contribution of Sb, Sn and Cd to the pollution load exceeded that of Se (Table A3).

Fig. 1
figure 1

Box and whisker plots showing pollution load indices (PLI; Eq. 3) of all PTEs combined in all of the investigated sampling locations. The boxes represent the second and third quartile limits whereas the whiskers are extended to the furthest figure within 150% of boxes heights. Horizontal lines are medians, black circles are means and black diamonds are outliers. Letters G, I and U in the location codes refer to the agricultural, urban or industrial nature of the sampling site, respectively. The locations are arranged in a descending order according to their mean PLI values. PLI data are also displayed in Table A3 (Appendix A)

Full size image

Intercorrelations between PTEs and soil parameters

High Pearson correlation coefficients of r > 0.8 were only found between V, Co, Fe, Mn and Al (Table A4; Appendix A). This can be explained by the ubiquity of primary ferromagnesian and secondary phyllosilicate clay minerals (which are naturally rich in Fe, Mn and Al) in the Egyptian soils (Shaheen et al., 2020). These minerals usually encapsulate several PTEs and are a major source of geogenic Co and V, which supports the earlier conclusion that V and Co are mostly of geogenic origin. Chromium and Sn were also highly correlated with SOM (0.59 and 0.74, respectively), likely due to their high affinity toward organic matter (Gustafsson et al., 2014; Nakamaru & Uchida, 2008).

Principle component analysis (performed with Minitab 17) was conducted on the total soil concentrations (mg kg−1) of all 9 PTEs, and total soil Mn, Al and Fe in all sampling locations. Scree plot (Fig. A2; Appendix A) suggests that the intercorrelations between all 12 elements can be dimensionally reduced to two major factors: PC1 and PC2 with Eigenvalues of 5.3 and 2.4, respectively, explaining 65% of the variance. Loading plot (Fig. 2a) shows that PC1 is mostly influenced by Mn, Al, Fe, V, Co and Cr, suggesting a common origin of these elements. Since Mn, Al and Fe generally reflect the bedrock lithology (Ekoa Bessa et al., 2020; Inácio et al., 2008; Shaheen et al., 2013); PC1 can thus be considered the ‘geogenic’ factor which again highlights the geogenic origin of V, Co and Cr in the studied soils. The second factor (PC2) was largely dominated by Se and Sb suggesting a significant anthropogenic influence on their loadings (Rodrigues et al., 2010), which is likely due to their very high levels in the Pb smelter location (SH-I) soils (Table 3). Arsenic, Sn and Cd were not inclined toward either PC1 or PC2, suggesting almost equal contribution in geogenic and anthropogenic loadings. The score plot (PC1 vs PC2; Fig. 2b) shows that soils from relatively contaminated locations, e.g., the industrial locations SH-I (Pb smelter) and CA-I (waste treatment), and the heavy urban and traffic locations CA-U and SH-U, within Greater Cairo, form discrete clusters representing combinations of anthropogenic and geogenic inputs that are characteristic to each location. Other soils from agriculture and lightly polluted urban locations were aligned closely to the ‘geogenic’ factor PC1.

Fig. 2
figure 2

Principle component (PC) analysis of the total soil concentrations (mg kg−1) of V, Cr, Co, As, Se, Mo, Cd, Sn, Sb, Mn, Al and Fe in all sampling locations by Minitab 17 software package. (A) Loadings plot of all elements and (B) Score plot of the two major PC factors (PC1 and PC2) showing sampling locations in different symbols and colors

Full size image

Geochemical fractions and potential mobility of PTEs

V, Cr, Co and As

The PTEs fractions are presented as percentage of the total content in Fig. 3 and as concentrations (mg kg−1 soil) in Table A5 (Appendix A). The residual fraction (F5) is mainly composed of minerals that contain PTEs within their crystal structure, and therefore, the F5-PTEs are considered permanently immobile (Li & Thornton, 2001; Shaheen & Rinklebe, 2014). On the other hand, the exchangeable, carbonate-, oxides- and organic-bound ‘non-residual’ fractions (F1–F4) are potentially mobile in response to changes in the equilibrium physicochemical conditions, e.g., pH, Eh and ionic composition of soil solution (Shaheen & Rinklebe, 2014).

Fig. 3
figure 3

Proportion (%) of a-Vanadium, b-Chromium, c-Cobalt, d-Arsenic, e-Selenium, f-Molybdenum, g-Cadmium, h-Tin and i-Antimony measured in each of the five sequential extraction phases: Exchangeable (F1), Bound to carbonates (F2), Bound to metal oxides (F3), bound to organic matter (F4) and Residual (F5). Industrial locations (HE-I (cement plant; CP), SH-I (lead smelter; LS) and CA-I (waste treatment; WT)) are displayed independently due to their different pollution sources while soils from urban (UB) or agriculture (AG) sites are combined together by location type. Full data set can be found in Table A5 (Appendix A)

Full size image

The F5 (residual) fraction was the largest geochemical fraction for V, Cr, Co and As accounting for more than 50% of total soil concentrations in most of the studied locations with only a few exceptions (Fig. 3a–d). The overall dominant F5 fractions of V, Cr, Co and As strongly suggest that they are mostly of geogenic rather than anthropogenic origins in the soils of Egypt (Shaheen & Rinklebe, 2018; Shaheen et al., 2017). The second and third largest fractions were the oxides (F3) and organic-bound (F4), respectively, constituting together an average of 28% of total soil pools of V, Cr, Co and As. Oxides (F3) and organic-bound (F4) phases are known to be the most important fractions for retaining PTEs in alkaline soils (Atkinson et al., 2011; Shaheen et al., 2017). The carbonate-bound fraction (F2) represented, in average, about 4% of total soil concentrations of V, Cr, Co and As, while the exchangeable fraction (F1) was almost nonexistent (< 1%) in all soils for these four elements. In alkaline soils, positively charged metal ions are typically capable of being sorbed onto negatively charged colloids in addition to lack of competition with H+ (Bradl, 2004; Shetaya et al., 2019a). Long-term contact of microelements with soil particles also leads to a progressive assimilation of PTEs to the non-exchangeable and less available soil pools (F2-F5); therefore, under these circumstances the exchangeable phase (F1) is usually a very minor fraction.

Slightly different patterns were observed in the soils around industrial locations, where anthropogenic inputs are likely a major source of PTEs. The non-residual V and As fractions (F1–F4) were higher than their corresponding residual fractions (F5) and accounted for 54% of total V in the cement plant (HE-I) location soils and 52% of total As in the Pb smelter (SH-I) soils. Raw materials used for the production of cement are naturally enriched in V (Shaheen et al., 2019b); in addition, high concentrations of mobile V species, e.g., VO43+, are found in the calcium silicate hydrate gel added to cement and responsible for its ultimate hardness (Achternbosch et al., 2003; Shaheen et al., 2019b) explaining the relatively higher non-residual (F1–F4) V content in the HE-I site. Ashes released from primary and secondary operations of Pb smelters are usually enriched in As and this leads to As contamination of the surrounding soils (Ettler et al., 2012; Temple et al., 1977). Arsenic in these ashes is usually released as As2O3 (Gu et al., 2019), which is transformed to the highly reactive As5+ in alkaline media such as our range of soils (Zvobgo et al., 2018). At relatively high anthropogenic inputs, as observed for As in the soils of the Pb smelter (SH-I) site (~ 4 times the background levels (Table 3)), the produced As5+ ions are expected to be mainly distributed among the non-residual (F1–F4) fractions, increasing the proportion of non-residual As in comparison with residual (F5) As. Chromium also was mostly (64% of total Cr) attached to the F4 phase (bound to organic matter) in the waste treatment location (CA-I). The high affinity of Cr toward organic matter is well documented (Gustafsson et al., 2014); therefore, the large F4-Cr fraction in the soils of CA-I location is likely due to the combination of the relatively high SOM (2.7%) and Cr (123 mg kg −1) contents in these soils.

Se and Mo

Selenium and molybdenum showed highly variable proportions of the residual fraction (F5) depending on the sampling locations (Fig. 3e, f and Table A5). For example, the F5-Mo was as high as 92% in the soils of agriculture locations while the F5-Se was as low as 10% in the Pb smelter (SH-I) soils. However, both of Se and Mo had noticeably large (up to 55%) bound to organic fractions (F4) and small (< 9%) bound to oxides (F3) fractions in most of the studied soils. This highlights the strong association of non-residual Se and Mo with soil organic matter (Qin et al., 2012; Wichard et al., 2009). Both Se and Mo showed relatively higher exchangeable fractions in all sampling locations than the other studied PTEs (up to 22% for Se and 12% for Mo) except Cd. This suggests that a significant fraction of Se and Mo is immediately available in most of the studied soils, which is especially true for Se in urban and agriculture (nonindustrial) soils. The residual fraction (F5) of Se in the Pb smelter site (SH-I), at only 10% of total Se, was the lowest among all the studied PTEs and locations (except Cd in the cement plant site; HE-I). At such a very high concentration of total Se in the SH-I site (4.74 mg kg−1; ~ 100 time its crustal abundance), likely released by the Pb smelter, it was not unexpected for Se to be found mostly in the non-residual (F1–F4) phases and consequently the proportional contribution of its residual fraction (F5) was rather diminished.

Cd

Among all the studied PTEs, Cd showed the lowest proportion of residual phase (F5) ranging from 7% of total Cd in location HE-I (cement plant) to 35% in the urban soils (Fig. 3 g and Table A5). In the non-residual phases, the largest fractions were F2 and F3 (bound to carbonates and metal oxides, respectively). This was followed by the exchangeable (F1) and bound to organics (F4) geochemical fractions, a finding suggesting that soil organic matter plays only a minor role in the retention of Cd (Li & Thornton, 2001). The Cd exchangeable (F1) fraction was one of largest (4–13% of total Cd) in comparison with the other studied PTEs. Our results thus highlight the high environmental mobility of Cd and its great affinity toward non-residual fractions compared to other metals, e.g., Pb and Zn and Cu (Lair et al., 2008; Marzouk et al., 2013; Rinklebe & Shaheen, 2014), even in alkaline and organic-poor soils, and at relatively low anthropogenic inputs like the case in our range of soils.

Sn and Sb

The residual fractions (F5) of Sn and Sb, ranging from ~ 100% for Sn in agriculture soils to 62% for Sb in the soils of the cement plant location (HE-I), were considerably high in comparison with the residual fractions of all the other studied PTEs with only few exceptions (Fig. 3 h, i and Table A5). This may at the first instance indicates that Sn and Sb are mostly of geogenic origin in the soils of Egypt. However, this cannot simply be the case for Sn in the waste treatment (CA-I) and the Pb smelter (SH-I) locations because of the very high total Sn soil concentration (38 and 5 times its crustal abundance, respectively), which must have been mostly anthropogenic. Similarly for Sb, the residual fraction in the SH-I site (Pb smelter) soils was 67% of total Sb, despite the extremely high level of total soil Sb (~ 300 times its crustal abundance), which likely has been mostly released with the Pb smelting ashes. Dominant residual fractions of Sn and Sb in soils with high anthropogenic inputs (e.g., industrial, mining and smelting activities) have been previously reported and was attributed to the capacity of soil to scavenge and immobilize large quantities of Sb and Sn over relatively short time spans (Elbana et al., 2014; Wilson et al., 2010).

PTEs in leafy and fruiting plants (leaves and fruits)

In the urban locations (CA-U, SH-U, CI-U and IS-U), the order of occurrence of the studied PTEs in the sampled plants was Cr > Mo > V > Cd > As > Co > Sn > Se > Sb in leaves and Mo > V > Cr > Co > Cd > As > Sn > Se > Sb in fruits (Table 4). In the locations IS-G, SK-G and SZ-G (agriculture soils), the occurrence order was Mo > Cr > V > Cd > Co > As > Sn > Se > Sb in leaves and Mo > Cr > V > Co > Sn > Cd > As > Se > Sb in fruits (Table 4). With the exception of Cd and Mo, the above orders of occurrence resemble the corresponding soil and background (crustal) orders of occurrence (Table  3). This may indicate plant contamination with soil via foliar deposition of re-suspended soil particles followed by embedment into plant tissues in forms that resist washing (Notten et al., 2008). To examine this possibility, the ratio of plant V over soil V was used as a reference to differentiate between soil contamination and other plant enrichment pathways (e.g., systematic plant uptake) for the rest of the studied PTEs (Joy et al., 2015; Schwertmann & Pfab, 1996) (Eq. 4). It was found that the amount of plant PTEs originating from extraneous soil contamination ranged from ~ 50% (Cr in leaves) and ~ 45% (Co in fruits) to ~ 1% for Mo (leaves and fruits) (Fig. 4). It can also be noticed from Fig. 4 that most (> 95%) of the plants contents of Cd and Mo (both in leaves and fruits) were not originated from soil contamination.

Table 4 Concentration (mg kg−1) of PTEs in leaves and fruits (of leafy and fruiting plants, respectively).
Full size table
Fig. 4
figure 4

Percentage of PTEs in leaves and fruits (of leafy and fruiting plants, respectively) originating from extraneous soil contamination in comparison with the PTEs fractions originating from other sources (e.g., plant uptake or extraneous application of fertilizers), as calculated by Eq. 4. Values for each PTE were calculated as an average of all sampling locations

Full size image

This observation was supported by the calculation of the bio-concentration factors (BCF) values (Eq. 5), which showed significantly (P < 0.05) greater BCF values for Cd and Mo than all the other PTEs exhibiting a 4- to 300-fold increase (Fig. 5 and Table A6 (Appendix A)). This may indicate preferential systematic plant uptake of Cd and Mo relative to the other PTEs.

Fig. 5
figure 5

Average bio-concentration factors (BCF; logarithmic scale) (Eq. 5) of toxic and potentially toxic PTEs in fruits and leaves (of fruiting and leafy plants, respectively). Full data set is displayed in Table A6 (Appendix A)

Full size image

The exhibited high mobility of Cd and its tendency to be weakly bound by soil (Lair et al., 2008; Marzouk et al., 2013), in addition to its dominant non-residual fractions in the studied soils (Fig. 3 and Table A5), may have caused its relative enrichment in the sampled leafy and fruiting plants. Nabulo et al. (2011) demonstrated that the plant uptake of Cd is strongly related to its soil availability. Moreover, the similarity of the ionic radius and valence of Cd+2 and Ca+2 (Kim et al., 2002) facilitates the plant uptake of Cd through active transport mechanisms and via Ca channels, as opposed to other PTEs which are mostly taken up by plants through diffusion and permeation (Das et al., 1997). The apparently higher BCF values of Cd in leaves than in fruits (Fig. 5 and Table A6) can be attributed to the faster growing nature of leafy plants and the consequent enhanced accumulation of systematically transferred pollutants (Barraza et al., 2021; Zheng et al., 2007). Also, a reasonable explanation is that Cd, as a typical trace metal, is not mobilized within plant after uptake in aerial biomass; thus, translocation from early-developed vegetative plant parts, i.e., leaves, toward fruits is not to be expected (Antoniadis et al., 2017a).

In highly acidic soils, Mo is strongly fixed to soil and is inaccessible to plants, while in soils with pH > 5, Mo is released to soil solution as oxoanion molybdate (MoO42–) which is soluble and readily available (Kaiser et al., 2005). Since our soils are alkaline (pH 7.9–9), this may explain the relatively high mobility of Mo and its uptake by leaves and fruits (Shaheen et al., 2017). Molybdenum is an essential micronutrient for plants that is required for atmospheric symbiotic nitrogen fixation (Wichard et al., 2009) and in Egypt, it is a common practice to deliver it to plants in the form of foliar spray to increase the yield of the crops (Elrys et al., 2018). This may have also contributed to the observed enrichment of Mo in the studied plant samples. There was no significant difference (P > 0.05) between the BCF values of Mo in leaves and fruits.

The conclusions about the uptake pathways of Cd and Mo are supported by the higher correlation of plant-Cd with total soil Cd (r = 0.93 for fruits and 0.71 for leaves) than the respective Mo correlations (r = 0.6 for fruits and 0.51 for leaves) (Table A7; Appendix A). This highlights the systematic uptake (via roots) as the major enrichment mechanism of Cd and the foliar incorporation (of sprayed fertilizers) as the predominant route of Mo uptake in our range of leafy and fruiting plants.

Risk assessment of soil PTEs

Ecological risks

Due to the lack of local standard values for emerging nonclassical TEs (e.g., Co, Mo, Sn and Sb) in soil, we have used here the standard values of Finland (MEF; Table 3), which are internationally accepted (Tóth et al., 2016; Van der Voet, 2013). It was found that Sb concentrations exceeded the ‘Higher Guidelines’ values of 50 mg kg−1 in the SH-I (Pb smelter) site, which means that Sb in this location is an ecological and health hazard for all land use activities. Vanadium exceeded the ‘Lower Guidelines’ value (150 mg kg−1) in the urban location (SH-U) indicating that the soils of this location presents an ecological risk if the lands are used for activities other than industry and transportation. Arsenic, cadmium and antimony exceeded the ‘Threshold’ values of 5, 1 and 2 mg kg−1, respectively, in locations SH-I (As and Cd), SH-U (Cd) and CA-U (Sb), indicating that these areas need further assessment for potential health or ecological risks associated with As, Cd and Sb contamination.

Noncarcinogenic health risks

The individual hazard index (HI) values for all soil PTEs, from all three exposure routes combined, were significantly (P < 0.05; paired t-test) higher for children than for adults except for Se, Cd, Sn and Sb (Fig. 6A and Table A8 (Appendix A)). The higher HI for children is likely due to the children higher ratio of contaminant to body weight and their different physiology and psychology (Doabi et al., 2018; Li et al., 2014; Shaheen et al., 2020). Nevertheless, the HI values for children and adults of individual PTEs were mostly below 1 (Fig. 6 A and Table A8), indicating negligible risk of noncarcinogenic health risk (from individual soil PTEs). The only exception was Sb in location SH-I (Pb smelter), which showed a HI of 3.2 for children, indicating very high noncarcinogenic risk to the health of local children deriving mostly from Sb (Table A8; Appendix A).

Fig. 6
figure 6

Noncarcinogenic health indices (HI) for children and adults from all three exposure routes combined (ingestion, dermal and inhalation) displayed for: (A) individual PTEs in all sampling location as a box and whisker plots; the boxes represent the second and third quartile limits whereas the whiskers are extended to the furthest figure within 150% of boxes heights, medians are shown as horizontal lines and outliers are not displayed, and (B) combined HI for all PTEs by sampling location. HI ≤ 1 indicates no significant risk of noncarcinogenic health risks and HI > 1 means high risk of noncarcinogenic health effects. PTEs (A) and sampling locations (B) are arranged in a descending order according to their Children HI values. Full data set is presented in Table A8 (Appendix A)

Full size image

The total HI values (of all PTEs combined) were < 1 for adults in all sampling locations (Fig. 6B), indicating trivial noncarcinogenic health risks from the investigated soil PTEs. However, the total HI exceeded the safe noncarcinogenic threshold of HI = 1 for children in the urban location SH-U, and in the industrial locations SH-I and CA-I (Fig. 6b and Table A8 (Appendix A)), indicating significant risk to the health of the local children in these locations.

Among the investigated PTEs, Cr had the highest contribution to HI, from all three exposure routes combined, contributing an average of ~ 41% of HI (in all soils for children and adults) (Fig. 6A and Table A8); this was followed by V (~ 23%), As (~ 22%) and Sb (~ 11%). The total contribution of Co, Mo and Cd combined was only ~ 4% in average.

Ingestion was by far the most important route of exposure for noncarcinogenic risks, recording an average contribution (for all PTEs, except Se and Sn, and in all locations) of ~ 88% for children and ~ 81% for adults. This was followed by the dermal (~ 9% for children and ~ 13% for adults) and then the inhalation (~ 2% for children and ~ 8% for adults) exposure routes (Table A8). Data of Se and Sn were excluded due the unavailability of dermal and inhalation reference doses.

Carcinogenic health risks

Carcinogenic risks (CR) were only estimated for As and Cd due to the lack of carcinogenicity slope factors for other PTEs. Despite the obviously higher CR values of Cd than As in locations SH-I (Pb smelter) and SH-U (urban), the CR values of each of As, Cd and the combined As + Cd CR in all sampling locations were between 1 × 10–6 and 1 × 10–4 (Fig. 7 and Table A8). This suggests an overall tolerable/acceptable probability of cancer development due to the exposure to soil As and Cd during the lifetime of the exposed population of between one additional person per 10,000 existing patients to one per 1,000,000 patients (Doabi et al., 2018; Li et al., 2014). However, some studies suggest that CR values between 1 × 10–5 and 1 × 10–4 are ‘relatively’ unacceptable and further actions may be needed (Li et al., 2002; Yang et al., 2018). This was true for Cd (alone) and the combined (As + Cd) CRs in SH-I (Pb smelter) and SH-U (heavy urban) locations (Fig. 7 and Table A8). Among all three routes of exposure, ingestion was the dominant route for carcinogenic risks, similar to the noncarcinogenic risks, with a relative contribution of > 99% (Table A8).

Fig. 7
figure 7

Carcinogenic health risks (CR) for As, Cd and both combined from all three exposure routes (ingestion, dermal and inhalation) displayed by sampling locations. CR < 1 × 10–6 = negligible risk, CR > 1 × 10–4 = probable risk of developing cancer and 1 × 10–6 < CR < 1 × 10–4 = acceptable or tolerable risk. Sampling locations are arranged in a descending order according to their total CR (As + Cd). Full data set is presented in Table A8 (Appendix A)

Full size image

Uncertainties in health risk assessment

Health risk assessment modeling usually includes some degree of uncertainty; for example, the chemical form, toxicity level and bioavailability of each element play an important role in determining its associated risks (Ferreira-Baptista & De Miguel, 2005; Oomen et al., 2002; Huang et al., 2020). Another source of uncertainty is the local parameters, e.g., average body weight, average life expectancy due to genetic causes and life style (e.g., time spent outdoors/indoors and level of activity). Since there were no available local Egyptian parameters from reliable sources, in this work we have, as much as possible, used regional parameters from countries that are relatively similar to Egypt in the level of development, population size and climatic zone, e.g., Iran and Nigeria (Doabi et al., 2018; Isinkaye, 2018; Taati et al., 2020). For some other input parameters, we had no option but to use the values listed in the USEPA data sets (USEPA, 1989, 2004), which are the most comprehensive available risk assessment databases. Nevertheless, the results of this study identified the soil PTEs and exposure routes that represent the highest degrees of risk to the health of the Egyptian population, in addition to the locations where soil pollution with PTEs poses a significant risk to the health of the local residents. Therefore, despite the inevitable uncertainties, our results can be very useful for policymakers and researchers (in Egypt and similar countries), as it will guide them to focus their remediation, legislative and further in-depth and site-specific scientific investigation efforts on the ‘hotspots’ areas, and on the pollutants and exposure pathways of the most priority.

Conclusions

Our results demonstrate that V, Cr and Co in the Egyptian soils are mostly geogenic in origin, whereas As, Se, Mo, Cd, Sn and Sb have considerable anthropogenic input. The largest exchangeable and non-residual pools were measured for Se, Mo and Cd, indicating their relatively high mobility in the terrestrial environment of Egypt. For Se, this was mainly due to its large anthropogenic inputs (~ 3 – 100 times the background levels), a load that is mostly distributed among non-residual soil pools. The increased mobility of Mo and Cd is likely due to their physicochemical characteristics, i.e., Mo tendency to be readily available in alkaline soils as MoO42–, and the intrinsic affinity of Cd toward non-residual fractions. Both leafy and fruiting plants were highly enriched with Mo and Cd, which was likely due to systematic uptake (via roots) for Cd and foliar incorporation of externally applied fertilizers for Mo. Our model investigating the TE concentrations entering plant as foliar impurities ruled out the possibility that the enrichment of plants with Mo and Cd resulted from extraneous soil contamination (via, e.g., deposition of re-suspended soil dust on plants surfaces), because the proportion of extraneous soil Mo and Cd was < 5%. Overall, there was no wide scale contamination with PTEs in the terrestrial environment of Egypt. However, in some urban and industrial areas within Greater Cairo, the levels of V, As, Cd and Sb were elevated beyond safe levels for the ecosystem. Although noncarcinogenic and carcinogenic health risks from individual soil PTEs were deemed acceptable in all of the studied locations, there was significant noncarcinogenic risk to the health of local children living in urban and industrial zones within Cairo from all the PTEs combined.

The findings of this work may also help in understanding the mobility, bioavailability and health risks of the studied PTEs in the developing and largely populated countries in the more arid zones of the world similar to Egypt and the parameters likely to control PTEs fate under these circumstances. Also, the summarized findings in this study may help to develop innovative and applicable methods for management of PTEs-contaminated soils.