Agronomic quality of sludges
The main agronomic parameters are given in Table 2. All the studied DSS had significant OM and tN contents, which offers good prospects for agricultural use. The cation exchange capacity (CEC) was in the range expected from the high OM content. The pH values were standing between slightly acid 6.4 and alkaline 8.1. Such values were not expected to cause problems, particularly with regard to the mobility of trace metals (Yong and Phadungchewit, 1993). The tN content ranged between 3.0 and 4.4 %, a result similar to that of other studies (Stark and Clapp, 1980). The C/N ratio ranged between 5.2 and 7.6. Some authors consider that such low values may promote nitrogen leaching. C/N ratio should ideally be raised to around 20 by adding bulking agents having a high C/N ratio (Tisdale and Nelson, 1993). Others, however, using the same type of sludge, showed that, in Mediterranean areas, biomass production was better improved by using sludge-alone amendment rather than sludge added to bulk agent amendment (Bousselhadj et al. 2004).
Regarding available P and K values, significant variability was observed between DSS samples (Table 2). PO43- varied from 4.6 to 0.2 and K+ from 3.7 to 0.6 g.kg-1. These values are in the range of total P and K values found in other studies (Wen et al. 1997; Samars et al. 2008; Tavazzi et al. 2012), which confirms their useful practical use as fertilizer. The measured electrical conductivity (EC) is only indicative, because there is no standard value for this type of measurement, which, in addition, is quite rare in the literature. It gives an indication of the salinization hazard for soils sensitive to this problem. In our case, the values for a 1:10 water extract varied from 409 to 1844 µS.cm-1, which would correspond to 762 to 3432 µS.cm-1 for a 1:5 water extract (Perez-Espinosa et al. 2000). These values are lower or in the range of values observed elsewhere (Perez-Espinosa et al. 2000), therefore without specific salinization hazard.
Table 2 Agronomic parameters of the studied urban dewatered sewage sludges depending of the WWTP location (EC: electrical conductivity, tN: total N, CEC: cation exchange capacity), OM: organic matter
Total content of heavy metals
The total content of the studied heavy metals did not vary significantly between DSSs (Table 3). For metals for which there is a legal maximum value, all SSs values were below the permitted threshold, except for Ni. TDM and BDM samples had Ni content higher than the French and Chinese legal ceiling value and sample AZF had Ni content higher than Chinese ceiling value. TDM and BDM sample had a Cr content lower, but rather close to the Chinese limit value. High Ni and Cr content in DSSs were frequently observed in other studies (Juarez et al. 1987; Zufiaurre et al. 1998; Lasheen et Ammar, 2009). High content of heavy metals in WWTP sludges usually do not originate from domestic effluents, but from industries that discharge effluents in the sewer system or from entry of runoff stormwater into the sewer system (Sörme and Lagerkvist, 2002; Singh and Agrawal, 2008).
Cr and Ni compounds may come from metal surface treatment or damascening industries. Ni compounds may also come from ceramic manufacturing industries and Cr compounds from textiles and tanning industries (Islam et al. 2017). In our study, the high Ni and Cr content in DSSs must be related to the fact that in the studied area there is no strict quality control of effluents from small industries which discharge directly in the sewer network.
Table 3 Total content of heavy metals (mg.kg-1 of dry matter ± std) in sludge and ceiling values application to the land. Fr: France NF U44-041; US: U.S. EPA CRF 503.13; Ch: China GB4284-84. In red, values higher than one of the legal ceiling value
# Unavailable data.
No legal maximum value was found for Ag, Co and Ti. Ag content in the studied DSSs ranged from 1.1 to 7.4 mg.kg-1. These values can be considered to be low in comparison with the few data available in the literature. A study from USEPA (2009) considered 74 sludge samples representative of U.S. WWTPs and found that Ag content was ranging from 2 to 195 mg.kg-1 (20 on average), 50% of the samples having a value higher than 13. Pradas del Real et al. (2016) reported a value equal to 14 mg.kg-1 for a sludge from Switzerland.
Cobalt is of special interest because it can act on biogeochemical cycles, in particular with regard to the availability of nitrogen for plants (Perez-Espinosa et al. 2002). Here the Co content in sludges ranged from 9.5 to 27 mg.kg-1, which can be considered as high and even outstanding values compared to the scarce data available in the literature, that range from 1 to 13 mg.kg-1 (Grummitt, 1976; Perez-Espinosa et al. 2002; Pradas del Real et al. 2016; Malinowska and Jankowski, 2020). The higher value observed here (27 mg.kg-1, AZF sample) is, however, unlikely to have an effect on plant growth (Perez-Espinosa et al. 2002). A cumulative effect of DSS application over several years, however, is not to be ruled out, depending on the mobility of Co in the soil.
Titanium in sludges was more studied due to the increasing use in recent decades of TiO2 nanoparticles, especially in personal care products, and because it is used as a catalyst for photocatalytic degradation of organic pollutant in wastewater (Wielinsky et al. 2021). Data on bulk TiO2 content in DSSs, however, are also scarce. Values ranging from 305 to 1800 mg.kg-1 were measured in DSS issued from European WWTPs (Johnson et al. 2011; Pradas del Real et al. 2018; Wielinsky et al. 2021). Lazareva and Keller (2014) calculated, considering market studies, that the Ti content in sludges from New-York, London and Shanghai would range from 40 to 208 mg.kg-1. Here the Ti content in DSS ranges between 946 and 1092 mg.kg-1, values relatively high with regard to the literature. Titanium, mainly as TiO2, is widely used products from paints to personal care products (Braun, 1997). Poorly controlled house painting activity may be the cause of the high concentrations observed here.
Considering the total metal content, sludge from BKH and EST sewage treatment plants can be recommended as a fertilizer or as an organic soil amendment. For Ni, AZF sludge exceeds the Chinese standard, TDM and BDM sludges exceed both French and Chinese standards and therefore cannot be applied without costly prior treatment (Veeken et Hamelers, 1999; Wong et al. 2000). It would be, however, more relevant to trace the origin of nickel in the corresponding localities to avoid its transfer to the sewer network. Anyway, these data underline the importance of quality control of sludges with regard to heavy metals.
Statistical analysis (PCA, principal component analysis) (Fig. 2) was able to identify two groups of metals with differentiated behaviours. Thus, PCA shows good correlations between metals in the (Cu, Ni, Cr, Cd, Ti) group, as well as in the (Pb, Zn) group, these two groups not being correlated with each other. Co was anticorrelated with Pb and Zn. The two first components explained 83,6% of the variance. The position of the BKH and TDM points is explained by lower and higher values, respectively, for metals from the (Cu, Ni, Cr, Cd, Ti) group. The position of the EST point is explained by higher values for Pb and Zn and low value for Co; the opposite being observed for the AZF point.
Speciation of heavy metals
In Mediterranean soils where the sludges are spread, the fractions most likely to release bioavailable metals are the F-ex, F-ac fractions, as well as the F-oxi fraction. The latter corresponds to metals bound to sulphides, which are rapidly oxidized in a dry environment and, in the case of organic matter, of pH higher than 6.5 (Cherfouh et al. 2018).
The studied metals showed differentiated behaviours with, except for Ag, a great consistency between the WWTPs (Fig. 3, numerical values are given as Supplementary Material). For all metals, the exchangeable fraction (F-ex) was lower than 3.5 %.
- Ti, Cr and Ni were found mostly in the residual fraction (F-res), i.e. contained in minerals (more than 90% for Ti and Cr and more than 80% for Ni), therefore poorly bioavailable (Sims and Kline, 1991). Small amounts of these metals (<10%) were in the oxidable fraction (F-oxi), i.e. mainly bound to organic matter or sulphides and, for Ni, 8-12% were in the reducible fraction (F-red), i.e. mainly bound to Fe or Mn oxides, in agreement with other studies (Ahumada et al. 2004; Islam et al. 2017; Yang et al. 2017).
- The distribution of Co among the fractions looks like that of Ni, with a greater proportion in the F-oxi and F-red fractions. High values of the latter were observed for samples BDM and AZF. Cumulative content in the F-ex, F-ac, F-oxi fraction (max. 3.3 mg.kg-1 in the AZF sludge) were low enough, however, to rule out long-term risk with regard to plant growth.
- Zn an Pb were mostly in the F-red fraction, as observed elsewhere (Silveira et al. 2006; Chen et al. 2008; Zufiaurre et al. 1998), with, for Zn, a non-negligible proportion (4.2 to 15.9%) in the acid-soluble fraction (F-ac).
- Cu was mainly distributed between F-red, F-oxi and F-res fractions, with a F-ac fraction ranging from 1.7 to 7.3%. Compared to other metals, Cu is well known to be mainly associated with the oxidizable fraction, which is confirmed by our results with best affinity of this fraction comparatively to Cr, Zn, Ni, Pb and Co. This behaviour is similar to that observed by other authors (Chen et al. 2008; Walter et al. 2006, Zufiaurre et al. 1998).
- Cd was the only metal having a low or under detection limit residual fraction (F-res). Most Cd was distributed between the F-ac and the F-red fractions, with a significant part (10.3 to 15.8%) in the F-oxi fraction. Similar distribution has already been described by in sludge from in middle-south region of China (Chen et al. 2008). The low Cd contents in the fractions were low enough to rule out a long-term risk.
- Ag was the only metal whose distribution was quite different from one sludge to another.
The specific behavior of each metal with respect to each fraction results in similarities between stations in the distribution of metals, as shown in figure 4. There were, however, significant differences between WWTPs, which are best highlighted by statistical analysis (Fig. 5). The BKH and EST sludges were characterized by lower values for most of the metal fractions F-red, F-oxi and F-res, and proportionally higher values for the exchangeable fractions F-ex and F-ac. In contrast, the AZF, BDM and TDM sludges were characterized by higher values for F-res, F-oxi and F-red fractions of metals. AZF and sludges were separated from the TDM sludge mainly due to a higher Co values in the F-oxi fraction.
The potential risk highlighted above due to high Ni and, to a lesser extent, Cr, can be put into perspective by the speciation data. More than 80% of Ni and Cr were found in the F-res fraction, the least likely to be made bioavailable by a change in acidity or redox potential in the soil. The pH usually greater than 6 of the soils of the region where these sludges can be spread is not favourable to the solubilization of Ni- or Cr-bearing metallic particles or minerals (Smith, 1994). Here the sum of the F-wat, F-ac and F-oxi Ni fractions, all capable of rapid release of bioavailable Ni, ranged from 6 (BDM) to 13 (TDM) mg.kg-1. Considering that sludge application is around 20 t.ha-1, neglecting leaching and plant uptake, and using 1.5 kg.dm3 as soil bulk density, it would take 230 (TDM) to 500 (BDM) years to accumulate in the topsoil 10 cm the bioavailable Ni content above which toxic effects can be observed (Kumari et al. 2018). The speciation of Ni can be very different in other places. In a Slovenian sludge studied by Scancar et al. (2000), more than 81% of the high Ni content (621 mg.kg-1) were contained in potentially labile fractions. In contrast to Ni, while the total Cu content was below the authorized limits, the sum of the F-wat, F-ac and F-oxi fractions, all capable of rapid release of bioavailable Cu, ranged from 26 (BKH) to 102 (TDM) mg.kg-1. Considering the same hypothesis than for Ni, it would take 3 (TDM) to 11 (BKH) years to accumulate in the topsoil 10 cm the bioavailable Cu content above which inhibition of nitrification can be observed (3.8 mg.kg-1 soil) (Cela and Sumner, 2002). These considerations underline the insufficiency of the regulations which only consider the total metal contents and not the conditions of their bioavailability.
Whatever the agronomic value of the sludges, the speciation of the metals in sludges makes it possible to identify the sites on which the search for sources of contamination would be necessary. Analysis of the treated wastewater gives information on the contamination at the moment of sampling, while the sludge gives information on the contamination integrated on the entire deposition time. Here the sources must be searched within the sewage network of the AZF, TDM and BDM WWTPs. The analysis showed that the contaminations concerned several metals simultaneously, the higher values of total Ni being accompanied by higher values for most F-red, F-oxi and F-res fractions of Cd, Co, Cr, Cu and Ti. Possible sources can be industries or cottage industries related to painting (Lokhande et al. 2011; Tesfalem and Abdrie, 2017), metalworking and metal plating (Quin et al. 2018), tannery, textile dyeing (Imtiazuddin and Mumtaz, 2013; Uma et al. 2016). Here we identified metal plating industry in the BDM area and metalworking activities in the TDM area. An exhaustive identification of the sources is, however, an activity in its own right which must be carried out in collaboration with the competent regulatory services.