Pyrolysis Treatment Enables Safe Application of Sewage Sludge in Horticulture: Tracking Potentially Toxic Elements Through the Biochar-Soil–Plant System in Tomato

Biochar production via pyrolysis is a promising option for treating sewage sludge (SS). Here we investigated the uptake of potentially toxic elements (PTEs) from SS and its biochar (450 °C) into cherry tomato (Lycopersicon esculentum L.) plants and their fruits in pot experiments (2, 5 and 10%), to assess the health risks associated with fruit consumption. We further studied the ability of the PTE-rich amendments to immobilize PTEs already present in soil through spiking of soil (pH 7.4) with Cd, Cu and Zn. Our findings suggest that tomato consumption is not a human health risk when grown in soil amended with SS and its biochar; in most treatments PTE uptake into fruits was similar to the control. Vegetative plant growth was highest in SS-amended soils. Yet 5 and 10% biochar-amended soils significantly increased tomato fruit yield relative to untreated SS application (400–500% higher than 10% SS-amended soil) and significantly decreased fruit uptake of Cu and Zn. Importantly, biochar also significantly immobilized PTEs present in contaminated (spiked) soil. As expected, available (DTPA-extractable) PTE contents were typically lower in biochar-amended soil and sequential extraction indicated that the PTEs were predominantly bound in more recalcitrant soil fractions (organic matter associated and Fe–Mn oxides-bound). We conclude that pyrolysis can be a key technology for treatment of PTE-rich SS and subsequent use of the produced biochar in urban, and potentially even PTE-contaminated, horticultural sites.


Introduction
The amount of sewage sludge (SS) from urban and industrial sources produced globally is increasing with the world's population and level of industrialization [1]. Sewage sludge contains organic matter (OM), macro-nutrients (N, P and K), and micro-nutrients, and therefore has been applied on agricultural land as soil amendment and fertilizer for decades [2,3]. Unfortunately, it also contains contaminants, including potentially toxic elements (PTEs), pathogens and organic compounds, such as polycyclic aromatic hydrocarbons [4]. Therefore, SS application can pose a health risk for humans and the environment [5] and it is necessary to treat SS before soil addition. In some countries, SS application to agriculture is now being restricted or forbidden altogether [6,7].
Pyrolysis is a technique that converts biomass under oxygen-limited condition into a material called biochar. The process reduces the volume of the material, sterilizes it and decomposes or removes organic contaminants effectively [4,[8][9][10]. Consequently, the application of SS biochar to agricultural land is of much less environmental concern than the application of untreated SS [11,12]. Conversion of SS into biochar has advantages over current practices, such as landfilling but also incineration from energy and environmental perspectives; SS biochar can increase soil fertility, improve water retention and immobilize contaminants in soil [3,10,13,14]. Modifications of SS, such as potassium doping, can produce biochar with nutrient-release properties beneficial for plant growth enabling agronomic use of SS biochar [15].
During pyrolysis, PTEs, such as cadmium (Cd), copper (Cu) and zinc (Zn) present in the biomass material are fixed and enriched in the biochar [16,17]. However, the availability, mobility, and leachability of PTEs are lower in biochar-amended soils compared to SS-amended soils [18]. Pyrolysis reduces the environmental risk of PTErich biochar as a soil amendment compared to unpyrolysed material because it is more efficient in the sorption of cations and therefore reduces PTE release [19,20]. Yue et al. [21] reported that typically, the total content of PTEs in soil increases following the application of SS biochar, yet a large proportion of the added PTEs are considered unavailable for plant uptake.
The large amounts of SS available in urban centers and high concentrations of nutrients in SS biochar make it ideal for use in urban horticultural systems and the ability of pyrolysis to immobilize PTEs in SS has been wellproven [11,15]. Tomato is the most important horticultural crop worldwide [22] and therefore, in the first part of this experimental series, we focus on assessing the effects of SS biochar on tomato growth and tomato fruits.
The soils in many urban centers are contaminated with PTEs [23,24] and it is well-known that biochars and nonbiochar organic materials, such as compost can sorb and immobilize PTEs in soil [25]. The ability to sorb PTEs and hence reduce PTE plant uptake would be an added benefit of SS biochar application in horticultural systems. This could make low-value, contaminated land close to human population centers usable for horticultural production. The key research questions for the second part of our experimental series are: (i) is PTE-rich biochar able to sorb additional PTEs from soil and if so (ii) does this happen without releasing biochar-inherent PTEs.
In this study, we investigate the effect of SS and SSderived biochar (applied at 2, 5 and 10%) on plant growth and PTEs uptake into vegetative organs and fruits of tomato plants and subsequent health risk associated with consumption of tomato fruits. We performed experiments in both (i) uncontaminated, low-PTE soil and (ii) soil spiked with Cd, Cu and Zn. Spiked soils were used to study sorption of PTEs present in soil by PTE-rich SS biochar and unpyrolysed SS. In addition, we assessed the DTPA-extractable PTE content in soil and their chemical fractionation in the rhizosphere and bulk soil. This aim was to track (i) inherent PTEs in SS biochar and (ii) PTEs present in contaminated soil through the biochar-soil-plant system.

Soil, Sewage Sludge, and Biochar Sampling and Their Properties
An agricultural soil (within depth 0-30 cm) was sampled from the Azandarian region located in the Hamedan province, Iran. The SS sample was collected from the municipal wastewater treatment plants in the Sanandaj province, Iran. Soil and SS samples were collected in 2018. The biochar sample was prepared from SS through slow pyrolysis at a temperature of 450 °C for 7 h.
The soil, biochar and SS samples were analyzed by methods suggested by Rowell [26]. The results are summarized in Table 1. Electrical conductivity (EC) and pH were determined in deionized water with a soil-solution ratio of 1:5. The content of OM was measured by the dichromate oxidation procedure. Calcium carbonates equivalent (CCE) was determined by the acid neutralization method. Determination of cation exchange capacity (CEC) was done by a neutral 1.0 M NH 4 OAc saturation method. The surface structure of SS and biochar was analyzed by JSM-840A scanning electron microscopy (SEM). Fourier transform infrared (FTIR) spectrum of SS and biochar was used to determine the chemical functional groups by a Perkin-Elmer FTIR spectrometer 17,259 using KBr disks.

Greenhouse Experiment
The pot experiment was designed for untreated, control (C) and spiked soils (S). For preparing the spiked soils, the PTE amounts equivalent to 10, 250, and 400 mg kg −1 of Cd, Cu, and Zn, respectively were added to the soil by using chloride salts. These contents were selected based on the average values of regulatory standards of PTEs in agricultural soil reported by He et al. [27]. The non-spiked (control soil) and spiked soils were incubated at field moisture for three weeks. Finally, the SS and biochar were separately mixed with the soils at rate of 0%, 2%, 5% and 10% on a dry weight basis. Cherry tomato (Lycopersicon esculentum L.) seeds were sown in peat moss and perlite combination placed in multiple-cell containers. After emergence about 14 days after sowing, one cherry tomato was planted in each pot (containing 2 kg of soil). The experiment was conducted in three replicates in a glass house at temperatures of 20-30 °C. The period of experiment was three months (from December 2018 to March 2019). The pots were irrigated with deionized water approximately up to field capacity. Rhizosphere and bulk soil samples were separated by pulling plants out of the soil and gently shaking them by hand at the end of the greenhouse experiment. The soil that remained on the roots was regarded as the rhizosphere soil, and the soil that separated from the roots was considered as the bulk soil [28]. Then, rhizosphere or bulk samples were air-dried and used for analysis.

Plant Analysis
The fresh weight of each part (root, shoot, and fruit) was measured. Then, samples were placed in an oven at 70 °C for 48 h and the dry weight of each portion was taken. The PTE content in plants was measured via the HNO 3 /H 2 O 2 wet digestion method given by Cao et al. [29]. Firstly, 0.3 g of the plant samples was placed in a Teflon container. Then, 5 ml of concentrated HNO 3 was added and the samples were left overnight. The next day, 2 ml H 2 O 2 was added and samples were placed in a microwave digester until the mixture became clear. The aliquots were collected in the bottle diluted to a final volume of 25 ml with deionized water. At the end of the digestion period, PTE content in roots, shoots, and fruits was analyzed by atomic absorption spectrophotometry (AAS).

Bioconcentration Factor
The bioconcentration factor (BCF) was calculated as the ratio of PTE content in plant tissues on dry weight basis (mg kg −1 ) at harvest and the content of total PTEs in soil (mg kg −1 ) [30]. This factor was determined for the root (BCF root ), shoot (BCF shoot ), and fruit (BCF fruit ) of the plant.

Human Health Risk Assessment
Estimated daily intake (EDI) (mg kg −1 day −1 ) and health risk index (HRI) of PTEs via cherry tomato consumption was calculated as follows [31,32]: where F IR is the daily vegetable consumption of 109.0 g per person per day [33], C is the PTE content in the fresh weight of cherry tomato fruits (mg kg −1 ), W is the average body weight of 70.7 kg adult and 32.9 kg children [32], ORD is the oral reference dose and the values of 0.001, 0.04 and 0.30 mg −1 kg −1 day −1 were considered for Cd, Cu, and Zn, respectively [34].
If the EDI of PTEs exceeds the ORD, the HRI becomes ≥ 1.0, and human is exposed to a level of contamination with potential adverse health effects [35].

Determination of Total and DTPA-Extractable PTEs
Total PTEs content in the soil, SS, and biochar was determined via the procedure proposed by Pietrzak and McPhail [36]. To determine the total PTEs content, 2 g of each sample was weighed into 50 ml tubes and 2.5 ml of 70% HNO 3 was added. The samples were left overnight to digest at room temperature. The next day, 7.5 ml of 70 wt.% HNO 3 was added to each tube and the mixtures were placed in a boiling water bath. Samples were digested at 100 °C for 8 h. After this time, samples were centrifuged for 3 min at 3500 rpm, and the supernatant was filtered into 50 ml volumetric flasks. The residual wet solid was rinsed with 10 ml and then with 5 ml of deionized water, centrifuged and the aliquots added into the 50 ml volumetric flask. Finally, PTE concentrations were analyzed by AAS.
Analysis of DTPA-extractable PTEs was carried out by the method suggested by Lindsay and Norvell [37].

Sequential Extraction
The sequential extraction of PTEs was performed using the procedure suggested by Tessier et al. [38] which was modified by Jaradat et al. [39]. In this procedure, PTEs were classified into solution and exchangeable (F1; EXCH), bound to carbonate phase (F2; CARB), bound to iron and manganese oxides (F3; Fe-Mn OX), bound to organic matter and sulfides (F4; OM), and residual (F5; RES). The mobility factor (MF) of PTEs was determined for the samples by Eq. 3 [40]. The units of MF and F1 to F5 are % and mg kg −1 , respectively.
As a check on the values from the sequential method, the sum of the five fractions (F1 + F2 + F3 + F4 + F5 in mg kg −1 ) and the total content of PTEs (TC in mg kg −1 ) determined by concentrated HNO 3 was compared in the recovery rate (R in %) equation:

Quality Control of Data and Statistical Analysis
To validate the analytical method, certified reference salts (Merck, Darmstadt, Germany) for each PTE were used to spike the soil. All experiments were carried out in triplicate and a blank sample was applied for each measurement. The instruments were calibrated by standard solutions before each measurement. The detection limits of AAS for Cd, Cu, and Zn were 0.02, 0.03, and 0.01 mg l −1 . The average total PTEs recovery percentages of Cd, Cu, and Zn for spiked soils were 92.8, 95.0, and 93.7, respectively, after spiking the soils with the desired amounts of salts containing Cd, Cu and Zn. Statistical analyses using a significant level of p < 0.05 were carried out using SAS [41]. One-way ANOVAs followed by Duncan's multiple range post-hoc test were applied to identify significant differences among treatments for data on bioavailability, fractionations, and uptake of PTEs. The experimental design was completely randomized with three replicates. Correlations between PTE contents in plants and soils were determined using Pearson's correlation analysis.

Properties of Soil, Sewage Sludge and Biochar
The biochar yield of SS pyrolysed at 450 °C was approximately 70%. Conversion of SS to biochar increased the pH from 6.4 to 6.6, and decreased CEC, OM, and EC from  [16,42,43]. The surface of SS was mostly compact and the pore structure was negligible demonstrated by SEM, whereas biochar had a porous surface (Fig. 5). The analyses of FTIR identified a peak at 3341 cm −1 and peaks at 2925 and 2851 cm −1 in the SS sample that can be assigned to hydroxyl groups and aliphatic structures, respectively (Fig. 6). These peaks decreased in the biochar sample, as also observed by Hossain et al. [42] and Song et al. [44] who suggested the decomposition of aliphatic C-H bonds at 450 °C. The peaks Root Shoot Fruit Fig. 1 Root, shoot and fruit dry weight (g pot −1 ) of tomato plants grown in non-spiked (a) and spiked (b) soil. Columns marked with the same letter (a-c for root, j-m for shoot and v-z for fruit) are not significantly different (Duncan's multiple range test) at the p < 0.05 level. Error bars represent standard deviations (n = 3). C control soil, SS sewage sludge, B sewage sludge biochar. *Marks the treatments that were so severely affected that no plant biomass could be measured showed at 1658 and 1548 cm −1 were caused by C=O and C=C functional groups, which also decreased after pyrolysis. Peaks at 1030 and 875 cm −1 were less affected by the pyrolysis treatment and are likely Si-O (quartz)/C-O bonds [39] and CaCO 3 , respectively (Fig. 6).

Plant Growth
The dry weight of roots and shoots increased with application of both unpyrolysed SS and biochar (in non-spiked soil; Fig. 1a). Maximum root and shoot dry weights were observed in the 2 and 5% SS-amended soils, followed by the 10% biochar-amended soil. The highest application rate of SS stunted plant growth, while the highest application rate of biochar enhanced plant growth. These results are in line with data on tomato published by Waqas et al. [45].
The highest fruit dry weights were observed in 2% SSamended, and 5 and 10% biochar-amended soils with values 2.4, 3.0 and 3.6 times higher than the control (Fig. 1a). Hossain et al. [46] reported that in their study fruit dry weight for biochar-amended soils was 64% and 21% higher than the control and SS-amended soil, respectively.
Plants grown in the soils that received extra PTEs (spiked with Cd, Cu and Zn) were severely negatively affected (Fig. 1b). Application of SS in addition to spiking soil with PTEs completely inhibited plant growth. Biochar addition, on the other hand, resulted in plant growth similar to the control. In the 10% biochar-amended soil root and shoot dry weight was even significantly increased compared to the control. Remarkably, the 10% biochar amendment was the only treatment in the spiked soil that demonstrated fruit yield (Fig. 1b).

Potentially Toxic Elements Contents in Plant Tissue
As expected, the roots demonstrate the highest bioconcentration factors (BCFs) due to an exclusion strategy that constraints transport of PTEs from the root towards shoot and fruit [47]. In the non-spiked soils, Cd content in the aboveground plant tissue only changed marginally with SS and biochar applications (Fig. 2a). Yet, Cd contents in the root were significantly higher in the treatments compared to the control soil ( Fig. 2a; Table 2). Furthermore, the contents of Cd in the shoot in 10% SS-amended soil, and in the fruit in 2 and 10% biochar-amended soils were significantly higher than those in the control and the other non-spiked treatments. Importantly, the addition of biochar to spiked soil reduced the content of Cd in the root and shoot (Fig. 2b) and BCF root and BCF shoot ( Table 2).
Biochar application to non-spiked soils reduced the contents of Cu in shoot and fruit compared to the SS treatments considerably (Fig. 2c). While 2% and 5% biochar increased the content of Cu in the root tissue compared to all other samples, the contents in the 10% biochar treatment were the same as in the control. The 10% biochar-amended soil also demonstrated the lowest values for BCF root , BCF shoot and BCF fruit of all the treatments (Table 2). In the spiked soil, biochar did not change the content of Cu in the shoots, but the root content increased (Fig. 2d).
The shoot and root Zn contents did not demonstrate a consistent response with SS and biochar application in non-spiked soil, yet both tended to increase compared to the control (Fig. 2e; Table 2). The Zn contents in the fruit were significantly greater in the presence of SS compared to biochar (Fig. 2e) and the values of BCF fruit decreased significantly in 5 and 10% biochar-amended treatments relative to the control ( Table 2). In the spiked treatments, 5% biochar also significantly decreased the root and shoot content compared to the control (Fig. 2f). Additionally, the values of BCF root decreased significantly following the addition of biochar at the rates of 5 and 10% and the values of BCF shoot decreased significantly following the 5% application of biochar ( Table 2).
The EDI of PTEs through cherry tomato consumption was in the order of Zn > Cu > Cd. Despite the observed EDI order, the HRI of Cd for both adults and children were greater than for Cu and Zn (Table 3), because EDI values of Cd exceeded recommended oral reference doses (ORD). In a study by Pajević et al. [35], tomato was one of the top three vegetables with elevated levels of Cd in the fruit and therefore also possessed high HRIs. Yet, in our study, all HRI values of Cd, Cu and Zn via cherry tomato consumption were below 1, indicating levels of contamination that do not pose a human health risk (Table 3). These values are in line with observations reported by Jalali and Karimi Mojahed [32] who also demonstrated that all HRI values of Cd, Cu and Zn in tomato do not pose a risk for adults and children.

Potentially Toxic Elements Extractability and Fractions
Overall, the addition of SS and biochar to non-spiked soils increased DTPA-extractable PTE contents as compared to the control soil (Fig. 3). This was confirmed in the fractionation assay where the RES PTE fraction decreased in   all treatments in favor of more extractable PTE fractions (Fig. 4). The DTPA-extractable PTE content increases with SS addition (non-spiked soils), clearly showing that SS adds available PTEs to soil. Overall, the DTPA-extractable contents of PTEs in SS amended soil (non-spiked soils) and untreated SS are significantly higher than in biochar and biochar-amended soil despite a higher content of total PTEs in biochar (Fig. 3; Table 1).
This result is less obvious in the fractionation assay (Fig. 4) where both amendments shift the proportions depending on type of PTE. Incorporation of SS and biochar into the non-spiked and spiked soils increased the CARB fraction of Cd compared to the control ( Fig. 4a and b). This could be attributed to the high content of CaCO 3 in SS (10.4%) and biochar (10.7%) ( Table 1) and the ability of CaCO 3 to decrease Cd activity in solution [48]. For all three PTEs, 10% SS and the biochar amendments increased PTE bound to OM. While the Fe-Mn oxide storage of Cu decreased with most amendment (Fig. 4c), it increased for Zn (Fig. 4e).
The DTPA-extractable PTE content in spiked soils increased more than 100-fold in comparison to non-spiked soils (Fig. 3). Importantly, the DTPA-extractable PTE content decreased with biochar application rate (Fig. 3b, d and  f). This is also the case for Cd and Zn in SS-amended soils, yet to a lesser extent (Fig. 3b and f). In the spiked soils, the added PTEs increased EXCH (Cd and Zn), CARB (Cu and Zn) and Fe-Mn OX fraction (Cd, Cu and Zn) over the non-spiked soils (Fig. 4). Distribution of PTEs in the OM fraction increased with amendments, in particular for Cu.
To compare the availability of PTEs, mobility factor (MF) based on data from the sequential extraction (see materials and methods) were calculated (Fig. 7). Cadmium showed the highest MF in all treatments, as was reported by Bogusz and Oleszczuk [49] who investigated the distribution of PTEs in SS-amended soil. The MF values of PTEs followed the order of Cd > Zn > Cu (Fig. 7). As expected, MF values in spiked treatments were greater than non-spiked treatments.
The DTPA-extractable contents of PTEs in the rhizosphere and bulk soil were comparable in most treatments (Fig. 3). Yet, in the non-spiked treatments, the DTPAextractable Cu content was significantly higher in the rhizosphere than that in the bulk soil sample at all rates of biochar applications (Fig. 3c). We correlated uptake of PTEs in aboveground plant tissue with DTPA-extractable contents and the sequential extraction assay (Table 4).
Unsurprisingly, the DTPA-extraction correlated best with all three PTEs in the non-spiked soil and less well with the artificial, spiked soils. Within the fractionation assay, Zn was predicted best, followed by Cu. Yet, for example the rhizosphere Cu content increased with biochar application rate, but the root tissue content dropped significantly at the highest biochar application rate to half of the value for 5% biochar (Fig. 2c). Similarly, the Zn tissue content was significantly higher in the 5% biochar-amended treatment and dropped in the 10% biochar-treatment (Fig. 2e), but the DTPA-extractable content consistently increased with biochar application rate (Fig. 3e). This clearly highlights that the root contents are weakly predicted by DTPA-extractions and that extractions as predictors for plant uptake only go so far. Direct measurements of contents in different tissues are the best way for determining plant uptake and for understand the underlying mechanisms.

Discussion
The application of SS at a low rate efficiently increased vegetative plant growth (shoot and root dry weight) of cherry tomato, whereas the addition of biochar at the high rate was more effective in increasing fruit dry weight. Higher concentrations of available nitrogen (N) and phosphorous (P) in SS relative to its biochar, stimulate vegetative plant growth [15,42]. Yet, in tomato the nutrient ratios seem more important for fruit yield [50] and therefore it is likely that the nutrient balance in our biochar was more favourable for fruit development in tomato. Remarkably, in the spiked soil, where Cd, Cu and Zn was added, 10% biochar application increased root and shoot growth compared to the control and enabled tomato fruit formation (no fruits in the control).
During pyrolysis, PTEs are concentrated in biochar, yet the PTEs are less extractable as confirmed in various studies [e.g. 15,51]. This resulted in lower uptake of PTEs in the aboveground plant part in our study, hence lower BCF and reduced EDI compared to unpyrolysed SS, as also reported by Waqas et al. [45]. We can conclude that the rate of uptake of PTEs into tomato fruit does not cause any health concerns, answering our first research question.
Compared to unpyrolysed SS, the PTEs within biochar were preferably bound in more recalcitrant form, in association with OM (Cd, Cu and Zn) and Fe-Mn oxides (Zn). The observed preference of Cu for OM and Zn for Fe-Mn oxides in this study agrees with Ignatowicz [52] and Huang et al. [53], who analyzed biochars by combining synchrotron X-ray spectroscopy/microscopy analysis and sequential chemical extraction. They report that Cu and Zn were desulfurized following the pyrolysis of SS, resulting in Cu-OM association and Zn association with Fe minerals. Yuan et al. [54] suggests that the pyrolysis process reduces DTPA-extractable PTEs because of its large specific surface area, pore structure, and the formation of organometallic complexes. In particular pyrolysis at relatively low pyrolysis temperature, such as the 450 °C performed here, enables PTE immobilization due to the retention of stable functional groups, while producing micropores and a rough surface area [44].
After soil spiking with Cd, Cu and Zn, biochar application reduced the extractable levels of all three PTEs. It actively immobilized Cd and Zn present in spiked soil by reducing the EXCH fraction, which bound to Fe-Mn OX and CARB instead reducing plant Cd uptake in all plant parts. This is remarkable considering the relatively high contents of PTEs already present in SS biochar and this demonstrates the value of SS biochar in remediation of PTE contaminated soil. This immobilization effect cannot be attributed to an increase in pH, which has been suggested as one of the main mechanisms of PTE immobilization by biochar, since the soil pH was higher than the biochar pH ( Table 1). The potential for PTE immobilization by biochar in alkaline soil was recently documented [55]. Yue et al. [21] pointed out that the precipitation of PTEs by inorganic components of biochar, including carbonates, oxides, and phosphates contributes to the increase of immobile fractions of PTEs in the biocharamended soil. The formation of strong metal-organic complexes decreases the mobility of PTEs considerably [52].
Biochar did not release any Cd, Cu and Zn in spiked soil, the opposite was the case, biochar was able to sorb additional PTEs. This answers our second research question: PTE-rich biochar can sorb additional PTEs and when doing so it does not release biochar-inherent PTE. It highlights that the sorption sites of SS biochar were not saturated with PTEs already. Documenting this phenomenon should also give higher confidence in biochar ability to retain its inherent PTEs on long-term when applied in non-contaminated soil. This capacity could enable the use of SS biochar on lowvalue, urban land for horticultural production.

Conclusion
Pyrolysis converts SS into a material with low risk for PTE release and stabilizes the carbon. Here we demonstrated that SS biochar can increase tomato yield compared to unpyrolysed SS despite resulting in lower vegetative plant growth. It can also immobilize PTEs in spiked soil without releasing biochar-inherent PTEs. This highlights the value of pyrolysis for treatment of (PTE-rich) SS and subsequent use of the produced biochar in horticultural systems. Establishment of horticultural production in urban centers using SS biochar could close the nutrient loop locally. The process also sequesters carbon that would otherwise be released back into the atmosphere, which further boosts pyrolysis as sustainable SS treatment option.