Biowaste materials and peat for OMFs were collected from industries across Italy. Biowaste materials were selected based on their similarities to peat C_org content, C_org stability based on the alkyl C/O alkyl C ratio and the humification degree, low heavy metal contents, and processing facilities. The GC was produced by combining vegetable waste with pruning residues composted for 6 months and was manufactured by Roffino S.R.L. The VC was produced from bovine manure that was composted for 6 months, which then undergoes a vermicomposting process, and was manufactured by Conitalo S.p.A. ACEA Pinorelese Industriale S.p.A. manufactured the MSWC by mechanically separating the organic fraction of municipal household waste and stabilizing it through integrated anaerobic digestion and aerobic composting (Mainero, 2008). Details regarding MSWC production can be found in the study conducted by Moretti et al. (2020). Organic materials were fully characterized in the study of Sitzmann et al. (2023).

The OMFs were manufactured by SCAM S.p.A. (Modena, Italy). The OMF manufacturing process involved mixing the organic material with ammonium sulfate, urea, diammonium phosphate, and water in a granulator. The water dissolves the mineral N, whereby impregnating the organic material with the dissolved fertilizer, forming a cohesive dough. This cohesive mass acts as a coating for the diammonium phosphate that acts as the granule nucleus. After creating the OMF granules, they are dried at 60°C until they reach ~4% moisture content. The mineral N fertilizer is added slightly over the target concentration of OMF N, accounting for volatilization of a partial N concentration during the OMF drying. Finally, the granules are sieved, resulting in OMFs with a diameter of 2–5 mm (60% <3 mm). For more details, see Sitzmann et al. (2024a,b).

A pot experiment was conducted in the summer of 2022 on the campus of the Department of Agricultural, Forest, and Food Sciences of the University of Turin in Grugliasco (NW, Italy) under semi-controlled conditions in an open tunnel greenhouse with a North-South orientation. The greenhouse was covered with a transparent cloth to guarantee proper illumination for plant development while preventing rainwater from entering. During the growing period, the minimum and maximum average temperatures recorded were 18 and 33°C, respectively. A tomato (Solanum lycopersicum L.) cultivar H1301, suitable for industrial processing, was used as a test plant. Twenty-day-old tomato plants grown in a nursery were transplanted, one per pot, into pre-fertilized pots on 25th May.

The experiment was a randomized block design separated by benches with four replicates for each treatment (the total number of pots was 36). However, one block was discarded due to pest development during the 2nd month of the tomato growth cycle. Each pot was 20 cm in diameter at the top and slightly thinner at the bottom, and it was placed on a plate that was used to provide sub-irrigation. Each pot was filled with 12 kg of air-dried soil sieved to 5 mm at a density of 1.40 g cm⁻³, resulting in a soil depth of 30 cm. The soil used in the experiment has a loam texture, a neutral pH in water (7.2), 8.4 g kg⁻¹ organic C, and 1.1 g kg⁻¹ total N. The total carbonate content was 4.4 g kg⁻¹. The exchangeable K was low (0.1 meq g⁻¹), while the exchangeable Ca and Mg were 7.1 and 0.9 meq g⁻¹, respectively, giving a low cation exchange capacity of 110. The Olsen P-value was 13.5 mg kg⁻¹.

To simulate fertilizer incorporation, fertilization was carried out by removing soil of ~2 cm depth, then dispersing the fertilizers homogeneously, and later covering again with the previously removed soil. Except for N₀P₀, each treatment received 81 mg N kg⁻¹ soil (equivalent to 170 kg N ha⁻¹ for 15 cm soil depth) and 18 mg P kg⁻¹ soil in pre-planting (Table 1). OMFs with a C_org-N-P₂O₅ concentration of 7.5-20-10 and 7.5-10-5 kg 100 kg⁻¹ OMF received ~4.9 and 9.8 g OMF pot⁻¹, respectively. For mineral control, MF_NP, urea, and ammonium sulfate were combined so that 75% of the N was supplied in the form of ammonium and 25% in the form of urea to emulate the N forms of OMF. Supplemental fertilization was carried out using with potassium chloride (KCl 60%), adding 79 mg K kg⁻¹ soil to all pots to avoid potential K deficiencies.

During the 1st month, pots were watered daily with 1 L of water and then with 2 L daily until harvesting. Water was added to the bottom plate to avoid nutrient leaching.

Tomato plants received lambda-cyhalothrin (KarateZeon^®, Seneffe, Belgium; 17th June), deltamethrin (Decis^® Evo, Villefranche, France; 1st July), and copper oxide (Fenicrit^®, Bologna, Italy; 6th June and 4th July) as insecticides and fungicides.

During the 2nd and 3rd months of tomato growth, phenology development, the number of stems, and leaf chlorophyll content were measured weekly. Phenological stage measurements were performed using the BBCH scale, focusing on the growth stage coding system for Solanaceae, provided by the Federal Biological Research Center for Agriculture and Forestry (Meier, 2001). The BBCH index was measured by considering the emergence of inflorescences, flowering, and fruit development, while the number of stems and fruits was counted manually. All shoots and branches with at least one visible node and a first genuine well-extended leaf were considered lateral stems. The measurements were taken on the same dates as the phenological stage survey. Leaf chlorophyll content was measured on a 6-mm² area using the SPAD-502 Portable Chlorophyll Meter (Konica Minolta Camera Co. Ltd., Tokyo, Japan). The tomato was harvested on 8th August, 75 days after the transplant. Tomato fruits were collected manually, and shoot mass was harvested at the soil level. The soil contained in the pot was divided into two sections (top and bottom) of 15 cm in height each, and from each section, roots were manually collected and hand-washed with the help of sieves. Dry matter (DM) biomass was measured separately for fruits, shoots, and top and bottom roots after drying samples at 60°C for 72 h. The total epigean biomass (fruit and shoot mass) was ground to 0.5 mm. Total N concentration in tissue was measured in an elemental analyzer (Thermo Scientific FlashEA 1112, Thermoquest). Similarly, total P concentration in tissue samples was analyzed through digestion with sulfuric and perchloric acid, followed by a spectrophotometer analysis using green malachite (Ohno and Zibilske, 1991). The epigean biomass N and P uptake was calculated as the product between the N and P concentration and the total epigean biomass.

The N use efficiency (NUE) and the P use efficiency (PUE) were used as indicators for the fertilizer's efficiency. The nutrient use efficiency was calculated as follows:

Data analysis was performed using R statistical software (version 4.2.2). For each crop, two sets of comparisons were made: (a) treatments N₀P₀, MF_NP, GC_{20 − 10}, MSWC_{20 − 10}, Peat_{20 − 10}, and VC_{20 − 10} were analyzed using treatments and block as independent values to determine differences between controls and organic materials in OMFs and (b) treatments MSWC_{20 − 10}, MSWC_{10 − 5}, Peat_{20 − 10}, and Peat_{10 − 5} were analyzed using C_org to mineral fertilizer ratios and block as independent variables. Values were tested for normality and homogeneity and transformed into a log or square scale when needed. A general linear model (GLM) with two or three factors was performed to compare treatments. The SPAD index and the number of stems were analyzed, including the day after transplanting, as independent variables and also independently for each sampling date. In contrast, biomass production, nutrient uptake, and efficiency were analyzed only once.

The chlorophyll content in leaves was analyzed based on the SPAD index, which is associated with the N content in tomato leaves (Wood et al., 1993) and fruits (Jiang et al., 2017). The SPAD index decreased by an average of 8% between 23rd June and 4th August (Figure 1A), indicating a lower chlorophyll concentration over time. Fertilized treatments had a significantly higher SPAD index overall than the control N₀P₀. However, when the data were analyzed for each day separately, there were no significant differences between treatments.

The chlorophyll content in leaves indicates no differences in N availability in soil between OMFs and mineral control, contradicting other reports that observed a slower N release into the N solution (Richards et al., 1993; Antille et al., 2013a, 2014; Florio et al., 2016). However, in previous reports from our study, we did not find significant differences in N release or NUE between OMFs and mineral controls in short-term ryegrass or maize pot experiments (Sitzmann et al., 2024a,b), where the interactions of the OMF organic fraction were mostly with P.

Morphological parameters such as the number of stems, or branch number, in tomatoes can be used as a reference for potential yield, as more stems can lead to increased branching and potentially more fruit-bearing sites, influencing overall yield and plant health (Ece and Darakci, 2009). In general, the number of shoots was significantly higher in Peat_{20 − 10}, the mineral control MF_NP, and MSWC_{20 − 10} than in the N₀P₀ control and VC_{20 − 10}, with intermediate values for GC_{20 − 10} (Figure 1B). Differences in the number of shoots occurred between the third, fourth, and fifth sampling dates, and the trend followed was higher values for Peat_{20 − 10} and MF_NP and lower values for N₀P₀ and VC.

The earlier entrance into the reproductive stage on week 7 for N₀P₀ was expected compared to the OMFs and the mineral control MF_NP (Figure 1C), as an earlier entrance to a reproductive phenological stage can indicate a nutritional deficiency in plants (Gaj et al., 2020). Peat_{20 − 10}, MSWC_{20 − 10}, and VC_{20 − 10} had the latest entrance to the reproductive stage (week 9). The results for VC_{20 − 10} are opposite to the number of shoots; however, as with Peat_{20 − 10} and MSWC_{20 − 10}, those treatments are correlated to a higher P uptake. In the formerly mentioned study, Sitzmann et al. (2024a) found that peat had the same initial P availability as the mineral control. At the same time, a vermicompost OMF—as well as the other biowaste materials—immobilized a fraction of the mineral P. This initial lower P availability could explain why VC_{20 − 10} initially had fewer stems than Peat_{20 − 10}, limiting their counting for the rest of the experiment.

The hypothesis of differences in growing parameters between OMFs and mineral control has not been confirmed. However, the results show that peat OMFs tend to outperform biowaste OMFs in terms of early growth, but no significant differences were observed after 75 days.

Adding mineral fertilizers alone or through OMFs increased the dry matter production in the shoots and top roots of the tomato plants compared to the control N₀P₀ (Table 2). Differences between treatments were observed only in the first 15 cm of the soil, which had, on average, six times higher root dry matter than the bottom 15 cm; the root DM was significantly higher in fertilized treatments than in the control N₀P₀. The shoot DM was significantly higher in fertilized treatments than in the N₀P₀ control; among the fertilized treatments, GC_{20 − 10} had a considerably lower DM than Peat_{20 − 10}, with other fertilized treatments being between these two. The highest total—berries, shoots, and roots—DM production was found in Peat_{20 − 10} and the mineral control, while GC_{20 − 10}, MSWC_{20 − 10}, and VC_{20 − 10} had intermediate values between the mineral control and N₀P₀.

Total plant N uptake at the end of the growing season was correlated to the SPAD index measured on the harvest day by a Pearson value of r = 0.5 (p-value <0.01), as all fertilized treatments were significantly higher than the control N₀P₀ (Table 3). OMFs did not present significant differences in NUE compared to the mineral control MF_NP.

As mentioned above, biowaste OMFs made with the same materials as the one used in this experiment had a lower content of plant-available P in the soil after 10 days of incubation than a peat OMF and mineral control, which caused a lower P uptake in maize plants after 30 days (Sitzmann et al., 2024a). The lower P availability caused by the biowaste OMF partially remained after 75 days, especially for GC_{20 − 10}, which demonstrated a lower P uptake and PUE than Peat_{20 − 10} at harvest (Table 3). Other biowaste materials mitigated the immobilization effects, showing intermediate values for P uptake, and did not differ from Peat_{20 − 10}. The organic material influence on the P availability of low-C_org OMFs has been associated with chemical interactions between both fractions rather than microbial immobilization (Sitzmann et al., 2024a). The Ca-P bonds in the OMF and the organic matrix pH have a higher influence on reducing P availability. However, as OMFs contained large fractions of ammonium sulfate and urea, it was expected that the nitrification process would reduce the pH in the fertilizer hotspot (Bouman et al., 1995; Dal Molin et al., 2020), and therefore, it could increase the P availability. However, this effect was not observed. The OMF granules were not concentrated in one unique hotspot, and hence, the buffer capacity of the soil, one with a high pH, could have readily adsorbed the H ions released during the nitrification, thereby reducing the impact of potential soil acidification on P availability.

Our second hypothesis considered that, despite differences during plant growth, after 75 days, tomato plants would have the same yield and nutrient use efficiency across OMF treatments. The second hypothesis has been confirmed for MSWC_{20 − 10} and VC_{20 − 10} but not for GC_{20 − 10}.

Differences in the mineral composition of the OMFs were expected to cause differences in N uptake between treatments. However, no differences were observed between OMF_{20 − 10} and OMF_{10 − 5} during the growth (Figure 2) and final yield (Table 2) of tomatoes. To fertilize at a ratio of 81 mg N kg⁻¹, OMF_{10 − 5} treatments used twice the amount of organic material. During the fertilization process, adding supplementary organic material will result in an increase of 4% N and 6% P for MSWC_{10 − 5} compared to MSWC_{20 − 10}. Furthermore, there will be an increase of 2% N and 2% P for Peat_{10 − 5} compared to Peat_{20 − 10} (Table 1). In addition, as urea is a richer source of N than ammonium sulfate, it will be in higher proportion in OMFs with 20% N (OMF_{20 − 10}). It can be considered that the OMF_{10 − 5} treatment, which contains a higher content of ammonium sulfate, received an additional S fertilization because the soil used in the experiment did not receive a supplementary S fertilization. A synergism between S and N was reported as S fertilization increases the NUE by promoting a higher N uptake (Jackson, 2000; Salvagiotti et al., 2009; Rietra et al., 2017).

However, a higher ratio of C_org to mineral nutrients could have promoted a higher degree of immobilization in soil, which would reduce the benefits of the additional organic N, humic C, and mineral S fertilization. Based on the characterization of the organic fraction discussed in the study conducted by Sitzmann et al. (2023), the immobilization would be more relevant for P than for N since pH and Ca values of MSWC (9.0, 47 mg g⁻¹ DM) and peat (8.0, 41 mg g⁻¹ DM) can enhance Ca-P bonds between the organic fraction and the mineral fertilizer and between the mineral fertilizer and the surrounding soil (Sica et al., 2023). The low soil P-values at the beginning of the experiment, associated with a lower release in OMF_{10 − 5} treatments, can be the reason for the lack of differences between the treatments, as P resulted in a limiting nutrient for a higher N and S uptake (Robertson and Vitousek, 2009; Vitousek et al., 2010).

Although the results showed no differences between using different concentrations of mineral fertilizers in OMFs, the results were inconclusive. Therefore, it is not possible to accept or reject the original hypothesis.