The composts studied in the experiment were prepared with a mixture of the solid fraction (SF) of pig slurry obtained from a piglets and sows farm located in southeast Spain and two different bulking materials: cereal straw and cotton gin waste. The SF was isolated using a screw-press system for solid-liquid separation (without flocculants). Two composting piles were prepared within the farm facilities: Pile 1, mixing the separated SF (SF1, solid fraction stored for one month) with cereal straw, as bulking agent, in a trapezoidal pile (3:2, v:v; 4.8 m³); and Pile 2, mixing SF (SF2, freshly separated) with cotton gin waste (2:1, v:v; 19.3 m³). The windrow system with aeration by mechanical turning was used for both piles. The bio-oxidative phase of Pile 1 lasted for 75 days (with 3 mechanical turning events in total) and that of Pile 2 for 120 days (with 5 mechanical turning events in total); the total composting times (including the maturation period) were 170 and 187 days, respectively. Compost samples were obtained by mixing seven sub-samples taken from seven representative sites of the corresponding pile, covering the whole profile (from the top to the bottom of the pile). Further details of the composting process can be found in Sáez et al. (2017).

The original solid fractions (SF1 and SF2) and the final mature composts (1 and 2) were characterized physico-chemically (Tables 1, 2). The moisture content was determined after drying at 105°C for 24 h and the organic matter (OM) concentration by loss on ignition at 550°C for 24 h. The electrical conductivity (EC) and pH were measured in 1:10 water extracts (w/v). The water-soluble organic carbon (C_w) was determined in 1:20 (w/v) water extracts, using an automatic analyzer for liquid samples (TOC-V CSN Analyzer, Shimadzu, Japan). The total organic carbon (TOC) and total nitrogen (TN) were determined in an automatic elemental microanalyzer (EuroVector Elemental Analyser, Milano, Italy). The humic-like fractions were extracted with 0.1 M NaOH (C_EX) from the composts and from the different particle size fractions, and from these the fulvic acid-like C (C_FA) was separated through acid precipitation of the humic acid-like C (C_HA); C_EX and C_FA were analyzed in an automatic carbon analyzer for liquid samples, and C_HA was calculated as the difference.

The solid fractions, final compost samples, and initial mixtures for composting were analyzed for their pseudo total heavy metal concentrations through microwave (ETHOS1, Milestone) assisted acid digestion (HNO₃/H₂O₂; 4:1, v/v) followed by measurement in an atomic absorption spectrophotometer (Thermo iCE 3000 series).

The mature composts were mechanically sieved at different particle sizes (mm): < 0.05; 0.05–0.5; 0.5–1.0; 1–2; and >2, using a vibratory sieve shaker (RETAC-3D, RETSCH). The different particle size fractions of the mature composts were analyzed for total Cu and Zn as indicated previously, and for soluble and exchangeable Cu and Zn through 0.1 M CaCl₂ extraction (1:10, w/v; 16 h) and measurement by atomic absorption spectrophotometry.

The potential phytotoxicity of the composts was evaluated through a direct acute toxicity plant growth test (ISO 15 799.1999), using different proportions of compost with an artificial soil (100/0, 50/50, 25/75, 12.5/87.5, 6/94, 3/97, and 0/100, w/w%). The artificial soil was prepared according to OECD 207.1984. Around 80 g of the corresponding mixtures (at 70% of their WHC) were placed in plastic seedbeds (10 replicates for each mixture). One seed of Zea mays L. was sown in each hole and the plants were left to grow for 2 weeks after 50% of the plants had emerged in the controls (artificial soil without compost), in a growth chamber with controlled temperature and relative humidity (11/13 h day/night cycles at 25/18°C and 58/70% relative humidity). For each treatment the number of emerged seedlings was recorded and the fresh weight of the plants (aboveground and roots) determined. The growth index for the fresh weight (g per aboveground of the plants) was expressed as a percentage with respect to the control (artificial soil without compost). The EC50 and LC50 (the compost concentration, in v/v%, at which plant fresh weight and seedling emergence, respectively, were reduced by 50%) were calculated by applying a linear regression analysis to the relationship between the logarithm of the percentage compost concentration and the percentage toxic effect on plant growth (fresh weight) and seedling emergence.

An indirect seed germination index (GI) test was performed using compost water extracts (1:10, w/v; 24 h) and Lepidium sativum L. (cress) seeds (Zucconi et al., 1981; de Bertoldi et al., 1983). Ten seeds were placed on filter paper in plastic Petri dishes, 2 ml of the water extracts were carefully added, the seeds were covered with a new layer of filter paper, and the dishes were closed. The Petri dishes (10 per extract and dilution, including deionized water controls) were wrapped in aluminum foil and incubated for 3 days at 21°C. The germination index (GI) was calculated from the percentage of seed germination (G) and root elongation (R), determined in comparison with the results for the controls, as: GI = [G × R]/100.

The IBM SPSS Statistics 22.0 software was used for the statistical analysis. The standard deviation of the means was calculated for the chemical composition of the compost. A one-way ANOVA was performed to determine the effects of the different particle size fractions on the chemical parameters for each compost; and data were also subject to two-way ANOVA (General Linear Model) using particle size fraction and compost as independent factors. Differences between means were determined using Tukey's test. Before the statistical analysis, the data were tested for normality using the Kolmogorov–Smirnov test. Pearson's correlation coefficients between the chemical parameters and the TE concentrations of the different fractions and composts were also determined.

The solid fractions of slurry obtained by screw press separation on the farm showed elevated concentrations of Zn and Cu (Table 1). The concentrations of Cu in SF1 and SF2 were similar to those reported in other pig slurry separation studies (Popovic et al., 2012), but the concentrations of Zn were higher in the present study (>3,000 mg kg⁻¹ dw). This was probably because the pig slurry studied here was mainly produced by piglets, and this slurry has a higher concentration of Zn than that from fattening pigs (Moral et al., 2008), since Zn is provided as a feed additive (ZnO) to piglets in order to avoid digestion problems and to improve nutrients assimilation (Carlson et al., 1999; Woodworth et al., 1999). Also, Zn and Cu are excreted at relatively higher rates than other feed constituents (Shepard and Sapinelli, 2012), and over 90% of Cu and Zn has been reported to be retained in the solid fraction of pig slurry after solid-liquid separation using different technologies (Riaño and García-González, 2014).

The total concentrations of Cu and Zn increased in both piles during composting (Table 2) due to the degradation of organic matter (OM) and the consequent loss of weight in the piles (Ko et al., 2008; Smith, 2009). According to Sáez et al. (2017), the increase is most evident in the earlier stages of composting, when the rate of OM degradation is highest, in agreement with previous observations (Hsu and Lo, 2000; Ko et al., 2008). The Zn concentrations were greater than those found in compost prepared with pig slurry from fattening pigs (Santos et al., 2016), confirming its origin from the piglet diet. Therefore, the total concentrations of Cu and, especially, Zn were very high in both composts; the values for Zn were well above the limit established for compost in Spain and for fertilizer products in Europe (Ministerio de la Presidencia, 2017; EU, 2019).

In the present experiment, the concentrations of the metals (Cu and Zn) in soluble and exchangeable forms (0.1 M CaCl₂-extractable) were very low in comparison with the total values, showing low percentages of extractability, and both tended to decrease from the solid fraction to the mature compost, especially in compost 2 (Table 2). The values found for Cu were lower than those reported by Alvarenga et al. (2015) for compost from MSWs with lower total-Cu concentrations (180 mg kg⁻¹), giving an extractable/total Cu concentration ratio similar to that in the initial untransformed wastes (solid fraction and initial mixtures) in the present experiment. However, Alvarenga et al. (2015) found values of 0.1 M CaCl₂-extractable Zn in compost from MSW and from agricultural wastes similar to those found here, although their composts had lower total-Zn concentrations. This indicates that the Cu and Zn in the pig slurry composts have very low solubility, compared to other common composted materials. Decreases in heavy metal availability in different organic residual materials, especially MSW, have been found during the composting process, including the maturation period (Smith, 2009). Amir et al. (2005), using a sequential extraction, found that <2% of the total heavy metals were easily available (extracted in dilute KNO₃) in mature sewage sludge compost, and that the Cu and Zn were mainly associated with organic and carbonate fractions. The OM stabilization and humification processes, which occur during composting, could facilitate heavy metal (especially Cu) interaction with ligands of greater complexing strength, rendering them less bioavailable (Smith, 2009). Humic substances have been suggested as the main sites of metal sorption in compost (Song and Greenway, 2004), which is greater for Cu than for Zn. Then, the lower concentration of humic-like substances (NaOH-extractable C and C_HA; Table 1) found in compost 1 (prepared with cereal straw), with respect to compost 2 (prepared with cotton gin waste), could be responsible for the slightly greater proportion of available Cu and Zn in compost 1 (Table 2). Therefore, the bulking agent used had a significant effect on the solubility of metals, which may have affected both the degradation of the OM during composting and the OM humification process (Sáez et al., 2017). Also, the lower value of pH found in compost 1, with respect to compost 2, could have contributed to the higher solubility of Cu and Zn in the former. So, in addition to the feedstock for composting, the physico-chemical characteristics of the final compost affected the solubility and availability of Zn and Cu.

The particle size distribution of both mature composts followed a similar pattern: the coarsest fraction (>2 mm) was the most abundant and the finest fraction (<0.05 mm) the least abundant, with a tendency for the abundance to increase with the particle size (Tables 3, 4). The OM and TOC concentrations (as well as C_HA and C_FA) decreased with the particle size, which indicates that the inorganic components dominated in the smallest particles with the highest TN concentrations (Table 5).

In both composts, the highest Zn concentrations were found in the smallest particle size fraction (<0.05 mm), with the lowest TOC and C_HA concentrations, while, contrastingly, the concentrations of Cu were highest in the largest particle size fraction (>2 mm) that had the highest TOC, C_HA, and C_FA concentrations (Tables 3–5). In fact, Zn is added to the piglet diet as ZnO (Carlson et al., 1999), which is mainly excreted and recovered in the slurry (Moral et al., 2008), unaltered by the animal. This together with the low Zn(II) complexing capacity of the humic fraction (Hernández et al., 2006) and its ability to precipitate as Zn oxides at high pH values, suggest that the Zn in mature compost was mainly linked to hardly soluble inorganic forms. Total Zn concentration was correlated positively total-N (r = 0.555, p < 0.01 data of both compost), and negatively with OM and TOC (all at p < 0.001). Regarding extractable-Zn concentrations, negative correlations were found with C_HA, C_FA and C_EX (all at p < 0.001), all supporting the idea of Zn being scarcely retained in the organic fraction, and mainly linked with inorganic compounds. The highly significant correlations found between the Cu concentration and both TOC (r = 0.850 and 0.656 for composts 1 and 2, significant at p < 0.001 and 0.05, respectively; r = 0.756 at p < 0.001 for the data of both composts) and OM (r = 0.855 and 0.757 for composts 1 and 2, significant at p < 0.001 and 0.01, respectively; r = 0.692 at p < 0.001 for the data of both composts) reveal the suitability of this metal for retention by the OM. Also, the correlation between the concentration of extractable-Cu and that of C_HA and extractable-C was significant in both cases at p < 0.01 (r = −0.590 and −0.558, respectively), but the negative relationship indicates the retention of Cu by the humified OM produced during composting. Hernández et al. (2006) concluded that the stability constant of Cu complexes with humic acids are much larger than the corresponding values for Zn, even in pig slurry organic matter.

Due to the largest concentrations of Zn in the smallest particle size fraction, the removal of this fraction by mechanical screening of the composts could be seen as an option to reduce the concentration of this metal in marketable compost, as usually done for MSW compost. However, its elimination would remove only a small amount of the total metal content of the composts (0.28 and 0.64% of total Zn in Compost 1 and Compost 2, and 0.10 and 0.52% for Cu, respectively; Figure 1), maintaining virtually unaltered metal concentrations, since this fraction represented only a minor part of the compost (Tables 3, 4).

The GI increased for both piles throughout the composting process, from scarce seed germination in the initial mixtures (GI<1% in both composts) to 87.8 and 80.2% in final composts 1 and 2, respectively, which indicates the absence of phytotoxicity. Therefore, in spite of the relatively high total concentrations of Zn and Cu in the composts, the low solubility of the metals meant that they did not have any relevant toxic effects on L. sativum seed germination, underlining the relevance of metal availability to phytotoxicity.

In the growth test performed with Z. mays (Figure 2), there were significant decreases in aerial biomass when compost was present at ≥12.5% in the substrate, with root growth affected at 50% compost. Both composts stimulated the growth of maize seedlings when present at up to 3–6%, with respect to the control soil without compost; Compost 1 also stimulated root growth (Figure 2). Sáez et al. (2016) found that the presence of low rates of pig slurry compost (<20%) in a growing medium was beneficial for plant growth, as the compost provided essential plant nutrients (mainly N, P, and K). But, when high proportions of the composts were applied in the present experiment, Compost 1 was not able to maintain correct root growth, which resulted in a greater decrease in the growth index than for Compost 2.

The calculated values of EC₅₀ were 29% for Compost 1 and 40% for Compost 2, which indicates greater toxicity of Compost 1. In addition, the values of LC₅₀ (related to seedling emergence) indicated that Compost 1 applied at 62% can provoke a lethal effect on 50% of the seeds, while for Compost 2 no harmful effects were found.

Sáez et al. (2016) found that maize seed germination was delayed in growing media prepared with pig slurry compost mixed with coir or biochar at proportions >20% (v:v), with reduced plant growth; however, the percentage seed germination was reduced only at proportions >60%. In that study, the reduction of salinity and of the available Cu and Zn concentrations (so that they did not provoke plant metal stress), as well as the removal of the volatile organic compounds from the compost, were identified as the main benefits of the use of biochar in substrates prepared with pig slurry compost. In agreement with the present results, the authors concluded that the use of pig slurry compost as a component of growing media was only feasible at low ratios (<20%), to avoid phytotoxic effects.

All these results indicate that the presence of Zn and Cu at high concentrations in these composts did not negatively affect seed germination and plant growth when the composts were used at low rates, compatible with agricultural doses. But, the application of high proportions of both composts in the soil under real field conditions can cause the long-term accumulation of Cu and, mainly, Zn in the soil and changes in their solubility and availability leading to negative effects on plants.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.