Introduction

Globally, approximately 1.3 billion tons of food wastes were generated in 2023, accounting for 30–50% of total municipal solid waste1. Conventional disposal methods such as landfill and incineration posed substantial environmental risks, necessitating urgent advancements in sustainable resource recovery technologies2,3,4. While anaerobic digestion and aerobic composting currently dominate waste valorization strategies5, their widespread adoption was constrained by prolonged processing periods (20-60 days) and extensive land requirements6. For instance, microbial-mediated composting usually converts perishable organics into stabilized humic-like acid (HLA, see the Supporting Information Supplementary Table S1 for a complete list of abbreviations) and fulvic-like acid (FLA), yet suffered from inherently slow metabolic rates and poor directional control, contributing to relatively high carbon loss (i.e., 30–50%) and low water-soluble HLA (< 15%) and FLA (< 2%) yields7,8. Fulvic-like acids (FLA), known for their high water solubility, were highly mobile in the environment and primarily exhibited a stimulatory role9. In contrast, humic-like acids (HLA), which displayed the characteristic of alkali solubility and acid precipitation, demonstrated a greater capacity to bind with minerals and become deposited in the soil, thereby contributing more significantly to long-term carbon sequestration9. Consequently, developing an efficient and targeted humification process, particularly one that directed carbon transformation toward high FLA yields, would be of considerable importance.

In our recent study10, radicals generated in KOH/persulfate (PS) enabled waste milk rapid conversion into HLA/FLA via degradation-polymerization, achieving significant HLA and FLA yields of 18.9 and 25.5 wt% within 1 h, respectively. Nevertheless, high KOH and PS inputs (8 wt% each) remained a practical limitation. More importantly, the heterogeneous composition and high viscosity of food wastes may impede mass transfer and reaction kinetics compared to homogeneous milk11, thus the applicability of this approach required further systematic investigation.

Microwave-assisted advanced oxidation processes (MW-AOPs) had been investigated in wastewater remediation due to its unique advantages of uniform heating, enhanced mass transfer, and oxidant activation12,13. To explore its potential in chemical-saving and humification efficiency improvement, this work employed microwave and KOH co-activated PS (MW/KOH/PS) in the humification of waste potato, a model food waste, and compared with KOH/PS. The process efficiency, humification mechanism, product components, and HLA/FLA yields were systematically investigated as well as the product’s fertilizer efficacy. Furthermore, its efficacy towards practical food waste was verified using vegetable residue and cooked food waste. This work highlighted a time (minute-scale) and chemical-efficient (4 wt%, saving 75% chemicals) method for converting food wastes into FLA fertilizer, promoting their onsite and spontaneous recycling (Fig. 1).

Fig. 1: Minute-scale humification of waste potatoes and recycling of the fertilizer.
Fig. 1: Minute-scale humification of waste potatoes and recycling of the fertilizer.
Full size image

Microwave (MW)/KOH/persulfate (PS)-induced conversion from waste potato to FLA-rich fertilizer and application in the form of the slow-release fertilizer.

Results

Robust humification of WP induced by MW/KOH/PS and condition optimization

As shown in Fig. 2, during the humification of WP, the dynamic evolution of soluble fluorescent components was monitored using three-dimensional excitation-emission matrix (3D-EEM) fluorescence spectroscopy. The 3D-EEM spectrum was generally divided into five regions14,15: regions I and II were typically associated with simple aromatic proteins and amino acids, region IV referred to soluble microbial byproducts (SMP), regions III and V corresponded to humic-like acid (HLA) and fulvic-like acid (FLA).

Fig. 2: Three-dimensional excitation-emission matrix (3D-EEM) spectra of waste potatoes (WP) treated by MW/KOH/PS under different conditions.
Fig. 2: Three-dimensional excitation-emission matrix (3D-EEM) spectra of waste potatoes (WP) treated by MW/KOH/PS under different conditions.
Full size image

EEM spectra of (A) WP, and treated WP by (B) 2%KOH, (C) 2%KOH, (D) 2%KOH + 2%PS, (E) MW, (F) 2%PS/MW, (G) 2%KOH/MW, (H) 2%KOH/2%PS/MW, (I) 1%KOH/2%PS/MW, (J) 2%KOH/1%PS/MW, (K) 2%KOH/3%PS/MW, (L) 3%KOH/2%PS/MW for 10 min. EEM spectra of treated WP by 2%KOH/2%PS/MW for (M) 0 min, (N) 2 min, (O) 5 min, and (P) 15 min. Conditions: WP = 50 g, KOH or PS dosage = 0–3 wt% and microwave time = 0–15 min, ambient temperature = 22 ± 1 oC. Each experiment was performed three times independently with similar results (n = 3). Source data are provided as a Source Data file.

In WP (Fig. 2A), intense fluorescence signals were observed in regions I, II, and IV, suggesting that proteins/amino acids and SMP dominated its fluorescent components. Isolated treatments with KOH or MW (Fig. 2B, E) exhibited negligible alteration in fluorescence distribution compared to WP, indicating their insignificant effects used alone. In contrast, 2 wt% PS treatment (Fig. 2C) induced a marked reduction in fluorescence intensities across regions I, II, and IV, likely attributed to PS-driven oxidative degradation of proteins/amino acids and SMP16.

After treatment by 2%KOH/2%PS without MW (Fig. 2D) or MW/2%PS (Fig. 2F), florescence intensities in regions I, II and IV decreased, while those in regions III and V slightly elevated, implying their slight humification effects. This shift aligned with the generation of reactive radicals during KOH or MW activated PS processes, which may promote oxidative degradation and subsequent humification. Similarly, combined MW and 2 wt%KOH treatment (MW/KOH) (Fig. 2G) induced a slight fluorescence intensity increase in regions III and V, potentially reflecting base-thermal synergistic effects on humification17.

Notably, the most pronounced transformation occurred under the combined MW/KOH/PS treatment (Fig. 2H), where fluorescence signals completely migrated from regions I, II, and IV to regions III and V. Fluorescence region integral (FRI) analysis (Supplementary Table S2, according to the method described in Supplementary Note S1) revealed that the cumulative fluorescence contribution of regions III and V (PIII+V) surged from 10.59% in WP (Fig. 2A) to 67.09% after MW/KOH/PS treatment (Fig. 2H). These findings proved the synergistic role of MW/KOH/PS in driving robust humification of WP, likely through synergistic mechanisms of thermal/KOH activation, radical oxidation, and base/heat hydrolysis18,19,20,21.

To optimize the humification efficiency of WP by MW/KOH/PS, the dosages of KOH and PS were evaluated within the range of 1-3 wt%. With the increase of their dosages, the fluorescence intensities (Fig. 2I–L and Supplementary Fig. S1A and B) and PIII+V (Supplementary Table S2) increased initially (from 1 to 2 wt%) and then decreased (from 2 to 3 wt%), achieving the maximum at 2%KOH/2%PS. Therein, the decrease trend at overdoses (above 2 wt%) of KOH and PS was consistent with our prior results in waste milk’s humification22, likely due to radical self-quenching at excessive oxidant conditions or accelerated decomposition of humification precursors/products.

Besides, the MW dose was optimized through adjusting the reaction time. As shown in Fig. 2M–P, Supplementary Figs. S1C, S2, the fluorescence intensity increased with MW exposure time and plateaued after 10 min, consistent with the similar PIII+V values of Fig. 2H, P in Supplementary Table S2. Therefore, the optimal humification conditions were chosen at dosages of 2%KOH/2%PS and microwave exposure time of 10 min through their synergistic effect. Besides, the EEM spectra of the product remain unchanged in 7 d (Supplementary Fig. S4), reflecting the high stability of the product.

Advantages of MW/KOH/PS humification compared to KOH/PS

The role of MW during WP’s humification was investigated on chemical consumption reduction and efficiency elevation. 3D-EEM spectra revealed that the florescence signals shifted from protein and SMP (Fig. 2A) to HLA/FLA regions after treated by KOH/PS at 2-8 wt% each for 60 min (Fig. 3A–C and Supplementary Fig. S1D), with stronger intensities at higher dosages (8 wt% each). Notably, treatments of 8%KOH/8%PS (in 60 min) resulted in comparable FRI distributions in HLA/FLA regions (65.7%) compared to MW/2%KOH/2%PS (67.1% in 10 min) shown in Supplementary Table S2, along with the slightly lower FLA content in the humified product (Fig. 3D). These results showed that the involvement of MW could substantially reduce the chemical input by 75% as well as shortening reaction duration by 83%.

Fig. 3: WP humification efficiency comparison between KOH/PS and MW/KOH/PS.
Fig. 3: WP humification efficiency comparison between KOH/PS and MW/KOH/PS.
Full size image

3D-EEM fluorescence spectra of WP treated by (A) 2%KOH/2%PS, (B) 6%KOH/6%PS and (C) 8%KOH/8%PS for 60 min without MW. D Fulvic-like acid (FLA) content of WP and the product after MW/2%KOH/2%PS and 8%KOH/8%PS treatment. E Temperature variation of the reaction system under different treatment conditions. F Total organic carbon (TOC) and total nitrogen (TN) loss under different treatment conditions. Conditions: WP = 50 g, KOH or PS dosage = 2–8 wt%, ambient temperature = 22 ± 1 oC. Each experiment was performed three times independently with similar results (n = 3). Error bars in (D, F) represent the standard deviations from triplicate tests. Data are presented as mean values  ±  s.d. Source data are provided as a Source Data file.

Besides, Fig. 3E showed that systems treated with 8%KOH/8%PS and MW/2%KOH/2%PS exhibited rapid temperature rise from ambient conditions (20 °C) to peak values of 80.3 °C (5 min) and 92.1 °C (10 min), respectively, significantly higher than that of 2%KOH/2%PS (38.4 oC). The infrared imaging results (Supplementary Fig. S5) revealed the even temperature distribution in the reaction system during microwaving. Specifically, in the KOH/PS system, radicals were initially generated through KOH-activated PS and induced exothermic humification reactions along with temperature rise, and the released heat would activate PS in turn and further contribute to radical generation and the humification16. In comparison, MW/KOH/PS showed more significant humification effect than heating (90 °C)/2%KOH/2%PS (Supplementary Fig. S3), indicating that PS activation in MW/KOH/PS was realized through microwave-specific non-thermal effects, such as the formation of localized hot spots and the lowering of activation energy23, besides exogenous heating and KOH24,25. These mechanisms accelerated radical formation synergistically and finally realized similar humification effects to 8%KOH/8%PS.

Critically, Fig. 3F showed that the carbon loss during the humification induced through MW/2%KOH/2%PS and 8%KOH/8%PS were 20.1% and 43.4%, respectively. Meanwhile, the nitrogen loss induced through MW/2%KOH/2%PS and 8%KOH/8%PS were 15.8 and 50.2%, respectively. That is, MW/2%KOH/2%PS could significantly reduce the carbon and nitrogen losses by 53.7 and 68.5% respectively, compared to 8%KOH/8%PS, likely attributable to abbreviated reaction duration and thus minimized oxidative decomposition. This highlighted its potential in enhancing carbon retention besides concurrent resource efficiency (chemical/time savings), positioning it as a sustainable strategy for rapid fertilization of food wastes26,27. Besides, as showed in Supplementary Table S3, the tail gas produced mainly included 220.7 mg CO2, 28.1 mg CO, 1.1 mg CH4, 8.8 mg volatile organic compounds (VOCs), and 3.8 mg NH3 per g WP (dry weight-based).

Supplementary Table S4 showed a preliminary techno-economic comparison between MW/2%KOH/2%PS and 8%KOH/8%PS. The total production cost for FLA fertilizer derived from per ton of WP was calculated by integrating costs of chemical consumption, electricity, equipment depreciation, and labor28,29. The analysis revealed that MW/2%KOH/2%PS incurred a total cost of 320 RMB/ton, whereas the conventional 8%KOH/8%PS process required 850 RMB/ton. Notably, the MW-integrated approach achieved a 62.4% reduction in overall costs, primarily attributed to the minimized chemical usage.

Humification product characterization

The components of the product were characterized comprehensively in Fig. 4. As shown in Fig. 4A, peaks of N-H (713 cm-1), aromatic C-O (1027 cm-1), C-N (1152 cm-1), carboxyl -OH (1578 cm-1), and aromatic C = C (1670 cm-1) became stronger after humification. These variations indicated the occurrence of aromatization, carboxylation, and amidation30,31. Meanwhile, the groups of aromatic C-O and C = C, amide N-H, and carboxyl C-O were consistently observed in standard FA (SFA), indicating the formation of FLA32. Similarly, the solid-state 13C NMR spectra in Fig. 4B showed that peaks belong to regions of oxygenated aliphatic carbon (50–100 ppm), aromatic carbon (100–160 ppm), and carboxyl C = O (160–220 ppm) become stronger after humification, align with the peaks in the spectra of SFA33. XPS C1s results in Fig. 4E showed that the C1s peak in WP could be deconvoluted into peaks located at 284.5, 285.5, 285.9 and 287.8 eV, corresponding to C-C, C-O, C = O, and Π-Π. After humification, Fig. 4F showed that peaks representing C-C/C = C, C = O, and Π-Π were strengthened, while C-O was weakened. Meanwhile, the peaks in the O1s spectra (Fig. 4G, H) also observed the strengthened C = O peak and weakened C-O peak in the product compared with WP. These results consistently reflected the occurrence of aromatization and carboxylation34,35.

Fig. 4: Components characterization of WP and Product under optimal humification conditions (MW/2%KOH/2%PS).
Fig. 4: Components characterization of WP and Product under optimal humification conditions (MW/2%KOH/2%PS).
Full size image

A Fourier Transform Infrared Spectrometer (FTIR) spectra and (B) Solid-state 13C Nuclear Magnetic Resonance (13C-NMR) spectra of SFA, Product and WP. C Thermo gravimetric analysis (TGA) spectra of WP and Product under nitrogen condition. D X-ray Diffraction (XRD) spectra of WP and Product. X-ray Photoelectron Spectroscopy (XPS) spectra of (E, F) C1s, (G, H) O1s for WP and Product. Conditions: WP = 50 g, KOH or PS dosage = 2 wt%, microwave time = 10 min, ambient temperature = 22 ± 1 oC. Each experiment was performed three times independently with similar results (n = 3). ND in Fig. 4H referred to not detected. Source data are provided as a Source Data file.

TGA spectra showed that the product possessed a higher thermal stability compared with WP (Fig. 4C), likely related to the conversion of perishable organics (proteins/amino acids) and SMP into HLA/FLA with higher thermal stability besides the introduction of K2SO4. Besides, the main contents in Supplementary Table S5 showed that a total of 10.5 g dry matter was contained in 50 g WP, corresponding to a moisture of 79%. Therein, contents of organic matter (OM) and dissolved organic carbon (DOC) accounted for 9.6 g and 0.6 g, respectively. Besides, WP (50 g) also contained 1.5 g K+, 0.6 g FLA, and 0.15 mg SO42-. After humification, OM content decreased to 8.7 g, corresponding to a loss of 9.4%. While DOC content increased to 6.5 g likely due to the formation of dissolving OM during humification. K+ and SO42- contents increased to 2.5 and 0.69 g (corresponding to the generation of 0.99 g K+ and 0.71 g SO42- from complete decomposition of 2%KOH and 2%PS) in the product respectively, indicating generated sulfur persisted predominantly as inorganic sulfate in the product. XRD spectra in Fig. 4D and XPS S2p spectra in Supplementary Fig. S7 confirmed the emerged peaks of K2SO4 in the product due to the decomposition of PS, consistent with the result in Fig. 4H. From a mechanistic perspective, sulfate radicals were known to react with organic compounds mainly through single-electron transfer or hydrogen abstraction pathways, rather than addition reactions36,37. This reaction behavior minimized the incorporation of sulfur into organic molecular structures, thus, no significant plant toxicity risk from sulfur species would be anticipated. Besides, K+ was an important inorganic fertilizer for plant, making the product like a kind of compound fertilizer38.

MW/KOH/PS-induced humification mechanism

Figure 5A showed the EPR spectra during WP’s humification. Both •OH and •SO4- radicals were detected at 2 min, with weakened signals at 6 min and disappeared at 10 min, indicating their possible involvement in the humification. With the addition of TBA as the •OH quencher to MW/KOH/PS humification system (Fig. 5B)25, the 3D-EEM spectrum showed slightly weakened signals in protein/amino acids regions compared with WP (Fig. 2A), with insignificant variation in HLA/FLA regions. This result indicated the key contribution of •OH in the degradation of protein/amino acids and HLA/FLA formation. Similarly, in the presence of EtOH, quenchers of both •OH and •SO4-39, the florescence peaks were almost unchanged, verifying the key role of •SO4- besides •OH in WP humification (Fig. 5C). Furthermore, Fig. 5D showed that the UV absorbance at 250–270 nm, representing the aromatization degree40, was significantly increased after humification, while the peaks were suppressed after addition of radical quenchers especially EtOH. Therefore, these quenchers inhibited the aromatization and humification, which were consistent with our previous observations in waste milk and glucose humification systems16.

Fig. 5: Humification mechanism elucidation under optimum conditions (MW/2%KOH/2%PS).
Fig. 5: Humification mechanism elucidation under optimum conditions (MW/2%KOH/2%PS).
Full size image

A Electron paramagnetic resonance (EPR) spectra of •OH and •SO4− during the humification of WP. B, C 3D-EEM spectra of MW/2%KOH/2%PS -treated WP in the presence of tert-butanol (TBA) and ethanol anhydrous (EtOH). D UV-vis absorbance spectra of treated WP in different systems (E) PS decomposition rate of WP during humification. F Gel permeation chromatography (GPC) spectra of WP and Product. G FTIR spectra of treated WP at different intervals (0, 2, 6 and 10 min). H Synchronous and (I) asynchronous FTIR two-dimensional correlation spectra (2D-COS) of treated WP during humification. Conditions: WP = 50 g, KOH or PS dosage = 2 wt%, microwave time = 10 min, ambient temperature = 22 ± 1oC, TBA = 10 mol/L, EtOH = 10 mol/L. Each experiment was performed three times independently with similar results (n = 3). Error bars in (E) represent the standard deviations from triplicate tests. Data are presented as mean values  ±  s.d. Source data are provided as a Source Data file.

Figure 5E showed an initial (0–4 min) fast then (4–10 min) slow PS decomposition trend with time, reaching a 100% PS consumption in 10 min, which was consistent with the time-dependent variation of radical signals and humification process. Besides, the GPC results in Fig. 5F showed that the molecular distribution of dissolved organics in WP was mainly between 300-7370 Da, with the most distributed around 1141 Da. While after humification, that of the product increased to a range of 300–15000 Da, with the most distributed around 1405 Da, indicating that radical-induced polymerization likely occurred, which was considered as a characteristic reaction of humification39,41. Besides, the product’s soluble components derived from the alkali dissolution and acid precipitation method showed a relatively narrow molecular weight distribution range of 300-2700 Da in Supplementary Fig. S8. Since the general molecular weights of HA and FA were distributed in the ranges of 10000–100000 and 600–1500 Da respectively42,43, the GPC distribution in the product and Fig. S8 suggested the formation of HLA and FLA.

As shown in Fig. 5G-I, the FTIR spectra variation of treated WP were analyzed within 10 min through 2D-FTIR-COS. According to synchronous 2DCOS in Fig. 5H, overall structural variation mainly distributed at 1578, 1400, 1100 and 1020 cm−1, corresponding to aromatic C = C, carboxyl O-H, aromatic C-O and C-N, respectively. These peaks consistently showed a gradually strengthened trend (Fig. 5G). Asynchronous maps in Fig. 5I showed that the groups variation followed the order of aromatic C-O > carboxyl O-H > aromatic C = C > C-N (Supplementary Table S6). Collectively, this humification was mainly induced through radical-involved aromatization, carboxylation, and amidation (i.e., characteristic Mallard reactions)44.

Molecular transformation of DOM components during WP humification

A comprehensive molecular-level profile of the humification product was obtained through Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR-MS). After treated by MW/2%KOH/2%PS, significant compositional changes occurred to WP, including the relative enrichment of lignin substances (18% to 26%) and condensed aromatic structures (8% to 14%) (Supplementary Fig. S9). To better illustrate the transformation of dissolved organic matter (DOM) during humification, an improved van Krevelen (VK) diagram was constructed (Fig. 6A). The result showed that the transformation was generally towards higher O/C and lower H/C ratios, related to the oxidation and condensation reactions during humification45. Meanwhile, the molecular transformation of DOM was investigated through a linkage analysis of 30 reaction types, presenting the dominant reactions of oxygen addition, dealkylation, and Maillard reactions (Fig. 6B). These intermediates underwent aromatization and polymerization with amino acid-like structures via Maillard-type condensation. The resulting unsaturated aromatic compounds exhibited key physicochemical properties similar to those of humic substances.

Fig. 6: DOM composition variation during WP humification.
Fig. 6: DOM composition variation during WP humification.
Full size image

(A) The Van Krevelen diagram. (B) Radar plot presenting the number of possible precursor-product pairs during WP humification. Source data are provided as a Source Data file.

Based on the above analyses and compositional results of WP (Supplementary Table S7) and Product (Supplementary Table S8) based on Py-GCMS, a possible humification mechanism was proposed for •SO4-/•OH induced formation of FLA from WP (Supplementary Fig. S10). WP’s organic components (mainly amylum, protein and lipids) underwent •SO4-/•OH/microwave mediated degradation into small molecules (e.g., sugar, amino acids and aliphatic acid). These small molecules may undergo dehydration, cyclization, carboxylation and Maillard reaction to form aromatic polymers. These intermediates may be reacted by Hydrogen atom transfer (HAT)/Radical adduct formation (RAF)/Single electron transfer (SET)46, further generating organic radicals (•R) that polymerized into FLA.

Preparation and nutrient release behavior of product-based slow-release fertilizer

As shown in Fig. 7A, the product was conveniently granulated into FLA and K+ slow-release fertilizer after mixing with attapulgite (ATP, (Mg, Al)4(Si)8(O, OH, H2O)26·nH2O) at a weight ratio of 1:1. Figure 7B demonstrated that the SRF granules remained intact throughout a 30-day water immersion period, indicating a good water resistance. Concurrently, the aqueous solution gradually transitioned from colorless to brown. This color change was consistent with the progressively intensified fluorescence signals in the HLA/FLA regions observed over 30 days (Fig. 7B), suggesting the slow release of HLA/FLA. The release profiles of FLA and K⁺, presented in Fig. 7C, D respectively, confirmed their time-dependent release behavior, reaching release ratios of 98.15% and 97.7% after 30 days. Theoretically, K+ had higher solubility, smaller size and thus higher migration performance compared with FLA. However, in this work, FLA and K+ showed the similar release rate probably because of: (1) higher electrostatic force between negatively-charged ATP surface and K+ compared with commonly negatively-charged FLA; (2) access of K+ into the crystal structure of ATP through ion exchange effect with Ca2+/Al3+, and the formed ion bonds were stronger than the hydrogen bonds between FLA and -OH on ATP surface.

Fig. 7: Slow-release fertilizer (SRF) preparation and nutrient release behavior.
Fig. 7: Slow-release fertilizer (SRF) preparation and nutrient release behavior.
Full size image

A Preparation flowchart of granular SRF using the Product and attapulgite (ATP), digital images of SRF in a water immersion system and time-dependent 3D-EEM spectra variation of the aqueous solution within 30 days. Release curve of (B) FLA and (C) K+ during 30 days water immersion of SRF. D FTIR spectra, (E) scanning electron microscopy (SEM) images, and (F) XRD spectra of SRF before and after slow-release. Experiments were performed three times independently with similar results. Error bars in (B, C) represent the standard deviations from triplicate tests (n = 3). Data are presented as mean values  ±  s.d. Source data are provided as a Source Data file.

FTIR spectra (Fig. 7D) revealed a significant weakening of peaks corresponding to amide N-H, aromatic C = C, C-O, aliphatic C-H, and O-H bonds after 30-day immersion, likely attributable to FLA release. SRF showed a dense surface morphology and lowered specific surface area in Fig. 7E and Supplementary Fig. S11D, compared to the “Blank” SRF (i.e., SRF particles without product loading) owning rod-like structure and relatively high specific surface area (Supplementary Fig. S11A, S11C), suggesting successful loading of the product in ATP with a nanonetwork structure47. However, post-immersion, the surface exhibited substantial rod-like structures, presumably resulting from the release of loaded contents from the ATP nanonetwork. Furthermore, XRD analysis (Fig. 7F) showed strong characteristic peaks of K2SO4 in the fresh SRF, which almost disappeared after 30-day release. This finding aligned with the significantly weakened distribution signals observed in the SEM-mapping results (Supplementary Fig. S3). In a word, these results confirmed the successful synthesis of SRF based on WP’s humified product. The synthesized SRF exhibited favorable slow-release properties for both FLA and K⁺ over a 30-day period.

Fertilization effect of WP-derived product in pot experiment

As shown in Fig. 8, the fertilization effect of the product on Shanghai Cabbage’ growth was compared with WP and K2SO4 through pot tests. After 21 d cultivation, the overall plant growth parameters (fresh weight, plant height, root length, germination rate, dry weight, chlorophyll content) showed an order of SRF > Product > K2SO4 > Untreated > WP (Fig. 8A-H). Therein, the fresh weight (actually the yield, Fig. 8D) of Product group was 5.2 g, 163.1%, 425.3% and 45.8% higher than the Untreated, WP, and K2SO4 group respectively. The decreased fresh weight of the WP group than Untreated group was likely owing to the negative effect of labile organic matter (proteins) in WP on plant growth. The positive effects of the Product indicated the effectiveness of the humification process in converting WP into a functional fertilizer. The better performance of the Product than K2SO4 demonstrated the significant contribution of HLA/FLA. Notably, SRF treatment further amplified the fertilization effect, resulting in better overall growth (including a 17.4% higher fresh weight) compared to the Product. Comparatively, ATP showed insignificant effect on the overall growth of plants compared to the Untreated group. Therefore, the slow release of FLA and K⁺ likely enhanced their absorption and utilization efficiency by plants48.

Fig. 8: Fertilization of product derived under optimal humification conditions (MW/2%KOH/2%PS).
Fig. 8: Fertilization of product derived under optimal humification conditions (MW/2%KOH/2%PS).
Full size image

A, B Digital images, (C) germination rate, (D) fresh weight, (E) dry weight, (F) plant length, (G) root length, and (H) chlorophyll content of Shanghai Cabbage with different treatments: Untreated, WP, ATP, K2SO4, Product, and SRF referred to 300 g soil alone, 300 g soil + 7.8 g WP, 300 g soil + 7.8 g ATP, 300 g soil + 0.87 g K2SO4, 300 g soil + 7.8 g Product, and 300 g soil + 15.6 g SRF respectively. Each pot contained 10 Shanghai Cabbage seeds. Each group was in triplicates and error bars represent the standard deviations. npot indicated the number of pots selected for analysis, and nplant referred to the number of plants selected in each group for analysis. Error bars in (C–E) represent the standard deviations from triplicate tests (npot = 3). Error bars in (F–H) represent the standard deviations from results of 12 independent plants in each group (nplant = 12). Data are presented as mean values  ±  s.d. in (I). Comparisons among results of different treatments were via one-way ANOVA analysis, and different letters refer to significant differences with Tukey’s test at p < 0.05 (p values are shown in the Source Data file). Source data are provided as a Source Data file.

Key soil indexes were also measured after harvest and shown in Supplementary Table S9. The soil organic matter (SOM) content showed an order of Product > SRF > WP Untreated. The higher SOM content in the Product group than WP indicated effective conversion of labile organic carbon into stable humic fractions during humification. Similarly, total and available nitrogen (TN and AN) consistently increased in WP and Product groups compared with Untreated likely due to their inherent high nitrogen contents in WP. Notably, total and available potassium (TK and AK) in Product group surged by 17.4% and 150.8% compared with WP, primarily due to external K+ supplementation during the humification process. The slightly lower values of these soil indexes in the SRF group compared to the Product group may be associated with the slow-release characteristics. Soil cation exchange capacity (CEC) values showed an order of SRF > Product > ATP > Untreated > WP > K2SO4, reflecting the positive effects of SRF, Product, and ATP treatments, likely attributed to the improved SOM and soil pH values compared to the Untreated (Supplementary Table S9).

Supplementary Table S10 showed that the Ace and Chao values of the soil treated with Product, SRF, ATP, or K2SO4 were comparable to the Untreated group, indicating a well-maintained microbial richness. Importantly, both Product and SRF maintained high Shannon and low Simpson indices, suggesting stable community diversity. Conversely, the Ace/Chao and Shannon values of the soil treated with WP decreased significantly, revealing the potential inhibition of WP on community richness and diversity. Moreover, Fig. S13 showed that the relative abundance of Actinomycetota, a dominant phylum known for decomposing complex organic matter and enhancing nutrient availability49,50, was significantly increased in Product (14.9%) and SRF (15.4%) groups compared to Untreated (11.1%), ATP (9.7%), and K2SO4 (9.5%), indicating the positive effects of the product and SRF treatment on soil microbial community. Furthermore, Pseudomonadota, which often harbored soil-borne plant pathogens51, was markedly enriched in the WP group (52.6%) compared with Untreated (35.8%), suggesting the potential negative effects of WP on the soil microbial community.

Furthermore, Fig. S14 showed that SRF’s overall fertilization effect was more significant than that of commercial alternatives, including coal-derived slow-release fulvic acid (CFA) and coated potassium sulfate (CPF), as well as a blank SRF prepared without loading of product (details provided in Table S11). Considering that the costs of CFA (30 wt% FA content) and CPF (15 wt% K) were 2000 RMB and 1000 RMB per ton, SRF (estimated to be 800 RMB per ton according to Table 1, i.e., 320 RMB for humification and 480 RMB for slow-release fertilizer preparation) highlighted the promise of SRF as a high-performance and cost-effective slow-release fertilizer.

Table 1 Technical comparison between this work and conventional composting
Full size table

Besides, the biosafety of the product was monitored with earthworm as a kind of bioindicator (Supplementary Fig. S15 and Supplementary Note S5). After 7-d cultivation, the survival rate and weight gain rate in different groups followed an order of Product > Blank > WP, indicating that the Product with a good biosafety may favor the growth of earthworms. These findings validated the dual functionality of the humification product: (i) as a high-efficiency fertilizer through K+ and HLA/FLA-mediated plant growth promotion, and (ii) as a soil conditioner via carbon sequestration and nutrient reservoir construction.

MW/PS/KOH induced scaling-up humification of practical food wastes

The application feasibility of MW/KOH/PS was explored in scaling-up experiments (1 kg) using practical vegetable residue (VR) and cooked kitchen waste (CKW). Figure 9A showed that the color of the VR changed from green into brown after MW/KOH/PS treatment, likely a visual sign of humification. Meanwhile, Fig. 9B, C illustrated that the 3D-EEM spectra showed significant florescence shifts from regions I, II, and IV to region V, confirming the conversion of labile organic matter (proteins, amino acids, and SMP) into stabilized humic substances (HLA/FLA). After humification, a significant enhancement of UV absorbance at 250–270 nm appeared, suggesting the occurrence of aromatization (Fig. 9D). Besides, Fig. 9H showed a dramatic FLA content increase from 2.3 wt% (VR) and 2.6 wt% (CKW) to 32.3 wt% and 28.4 wt% after treatment, verifying the dramatic occurrence of humification. Notably, similar humification phenomenon was observed in both WP and CKW systems (Fig. 9D–H and Supplementary Fig. S16). Besides, Supplementary Figs. S17 and S18 showed that the growth of amaranth followed an order of VR-Product > CKW-Product > K2SO4 > Untreated > CKW > VR in field tests, indicating the fertilization of derived products from VR and CKW. These findings validated the universal applicability of MW/KOH/PS treatment in inducing scaling-up humification of diverse practical food wastes (i.e., VR and CKW).

Fig. 9: Scaling-up efficacy of MW/KOH/PS induced humification in practical food waste systems.
Fig. 9: Scaling-up efficacy of MW/KOH/PS induced humification in practical food waste systems.
Full size image

Digital images of (A) vegetable residue (VR, weight ratio of potatoes: spinach: tomatoes = 5:2.5:2.5) and (E) cooked kitchen waste (CKW) before and after humification. 3D-EEM spectra of (B, C) VR and (F, G) CKW before and after treatment. D UV-vis absorbance spectra and (H) FLA contents of VR and CKW before and after treatment. I Proposed food waste fertilization modes in household and park scales. Conditions: VR or CKW = 1000 g, KOH or PS dosage = 2 wt%, microwave time = 6 min, ambient temperature = 22 ± 1 oC, MW power = 3 kW. Each experiment was performed three times independently with similar results. Error bars in (H) represent the standard deviations from triplicate tests (n = 3). Data are presented as mean values  ±  s.d. Source data are provided as a Source Data file.

To be noted that based on the applied power, treatment time, and mass of material processed, in the small-scale (180 W, 50 g, 10 min) experiment, the specific energy inputs were calculated as 2160 kJ/kg. For the scaled-up trials using the industrial 40 kW microwave (3000 W, 1 kg, 6 min), the specific energy input was adjusted to 1080 kJ/kg. This adjustment was intentionally made after preliminary tests indicated that a lower energy density was sufficient to achieve similar or improved product characteristics under larger-scale processing conditions, likely due to more efficient energy transfer and improved penetration in the industrial system and less heat loss (smaller specific surface area). Therefore, although the absolute specific energy input differed, the scalability was considered successful as it maintained the desired product quality while improving process efficiency.

Figure 9I showed that the MW/PS/KOH process enabled on-site fertilizer production from food waste at both household and park scales. The integrated procedure comprised three sequential stages: pretreatment, humification, and fertilizer recycling. During pretreatment, food waste underwent sorting (manually or via magnetic separation) and pulping (using household or industrial-grade mechanical pulpers), converting raw waste into biodegradable slurry (particle diameter < 1 mm, impurity content < 3%). Subsequently, the humification of slurry could be realized in a microwave oven (household or industrial scale) after KOH and PS addition within 10 min. The derived humified product could be directly applied as fertilizer for adjacent green spaces (e.g., residential gardens or urban parks). This process demonstrated advantages of ultra-low chemical consumption, equipment versatility, rapid treatment efficiency, and closed-loop resource utilization. The designed application modes favored the efficient recycling of food waste, alleviating the potential pollution from storage and long-term composting. To be noted, the reagents used in this process required careful handling, particularly in household applications. Therefore, the development of mild chemical substitutes was a significant part of our future perspective.

Discussion

MW/KOH/PS treatment enabled rapid conversion of food wastes (i.e., waste potatoes, vegetable residues, and cooked food wastes) into fertilizer rich in FLA (15 wt%) and K⁺ (7.6 wt%). The overall operation cost in this study, including materials, labor, electricity, equipment depreciation, maintenance, plant depreciation and package was 250 RMB, close to conventional composting calculated in Supplementary Table S12. The involvement of microwave could significantly reduce chemical consumption by 75%, and shorten the reaction period from 60 to 10 min, thus reducing carbon and nitrogen loss by 20.1 and 15.8%, respectively. This offered distinct advantages in both cost-effectiveness and carbon/nitrogen retention. The inherent flexibility and low cost of microwave equipment (listed in Supplementary Table S12) further facilitated its significant potential for the efficient humification of food waste.

In contrast to traditional composting (Table 1), this technology featured drastically shortened duration (10 min vs. 20–60 d), higher FLA yield (10.9–15% vs. 1-2%), lower carbon loss (21% vs. 30–60%), lower nitrogen loss (15.8% vs. 20–50%), ultrahigh sterilization rate (> 99%) (Supplementary Fig. S19 and Supplementary Table S13), and elimination of C/N ratio adjustment. These advantages supported “generate-and-recycle within hours” for food waste across diverse scales, from households to commercial entities. Consequently, this approach saved substantial transportation and storage costs and effectively mitigated the risk of secondary pollution during processing, exhibiting high application potential in food waste recycling.

Methods

Reagents and materials

All reagents used in this study were of analytical grade. Potassium persulfate (K2S2O8, purity ≥ 99.5%), potassium hydroxide (KOH, purity ≥ 95%), hydrochloric acid (HCl, 36.0–38.0%), anhydrous ethanol (EtOH, C2H6O, 99.5%), tert-butanol (TBA, C4H10O, ≥ 99%) and 5,5-Dimethyl-1-pyrroline-N-oxide (DMPO) were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). ATP and aminosilicone oil (ASO) were supplied by Fufeng Biotechnology Co., Ltd. (Inner Mongolia, China). A microwave oven (Midea PM20A1, 800 W, 2450 MHz) and an industrial microwave oven (LY-40KW, 5800*1060*1700 mm, Shanghai Longyu Microwave Equipment Co., Ltd.) were employed for microwave-involved processes. Shanghai cabbage seeds (Fast-growing 605) were obtained from Shanghai Vegetable Seed Company. Cooked food waste (diameter<0.5 mm, 80% moisture content) was obtained from China Everbright Environment Group Limited. Adult earthworms (Eisenia fetida) were provided by Ruizhou Fishing Tackle Co., Ltd. (Jiangsu Province, China). Potatoes (Longdan No.1) was provided by Chunzhengsheng Co., Ltd. (Shanxi Province, China). Deionized water was used for all experiments, except for pot test where tap water was applied.

Humification of WP, vegetable residues and cooked food wastes

Prior to humification, waste potatoes and vegetable residues (weight ratio of potatoes: spinach: tomatoes = 5:2.5:2.5) were pulped into slurry using a crusher. The moisture contents of WP and VR were around 89.5% and 90.2%, respectively.

In small scale test, solid KOH and PS at dosages of 0–4 g were added simultaneously to 50 g WP slurry with stirring (50 g). Then the mixture was subjected to the microwave oven (160 W, 0–15 min) to achieve humification. The stated microwave power (e.g., 180 W) referred to the average output power, which was achieved by the oven’s inherent pulsed operation mode at its maximum setting (800 W). The reported irradiation time denoted the total processing time. The real-time temperature of the reaction system during microwave heating was monitored using a microwave-resistant thermometer equipped with a wireless display (DT-200PRO, Data Test Co., China). Besides, infrared imaging device (SKDUTC, SD-1401) was applied for revealing the temperature distribution during reaction. The humification procedure of WP under water bath conditions was provided in Supplementary Note S2.

To be noted, sulfuric acid solution (1 mol/L) was used to adjust the pH of the humified product from 11.0 ± 0.1 to 7.8 ± 0.2 for fertilization evaluation in the slow-release test and subsequent pot/field experiments.

In scaling-up test, solid KOH and PS with 20 g each were added to 1 kg slurry of WP, VR or CK under stirring (50 g). Then the mixture was placed individually on the conveyor belt of the industrial microwave oven and treated at a power of 3 kW for 6 min.

Radical identification

Radicals (•OH and •SO4-) produced during the process were determined on an electron paramagnetic resonance (EPR) spectrometer using DMPO as the spin-trapping agent52,53. In addition, quenching experiments were carried out by adding 6 mol/L EtOH and TBA (higher than 10 times of KOH and PS dosages) to 50 g WP before KOH and PS to obtain the contributions of those radicals.

Extraction and quantitative analyses of HLA and FLA in product

The quantitative analysis of HLA/FLA was conducted according to the alkali dissolution and acid precipitation method detailed in BS ISO 19822-20182843. First, the product was adjusted to a pH of approximately 10.5 and then shaken at 100 g for 1 h. Subsequently, the product was centrifuged at 4000 × g for 20 min at 4 °C. The insoluble substances (precipitate 1) were separated from the soluble humic acid (SHA) in the supernatant. Next, the pH of the supernatant was adjusted to 1 with hydrochloric acid and stand overnight at room temperature. Then, the supernatant (FLA) was separated from precipitate 2 (HLA) after centrifugation at 4000 g for 20 min. Finally, the product (around 50 g), precipitate 1, and precipitate 2 were weighed after freeze drying. The yields of HLA and FLA could be calculated according to formulas (1)-(4).

Wherein mSHA, mproduct, mFLA, and mHLA represented the dry weights of SHA, product, FLA and HLA, respectively. m1 and m2 represented the dry weight of precipitate 1 and the weight of K2SO4 calculated according to the stoichiometric method under the condition that PS has completely decomposed.

$${{{\rm{m}}}}_{{{\rm{SHA}}}}={{{\rm{m}}}}_{{{\rm{product}}}}-{{{\rm{m}}}}_{1}-{{{\rm{m}}}}_{2}$$
(1)
$${{{\rm{m}}}}_{{{\rm{FLA}}}}={{{\rm{m}}}}_{{{\rm{SHA}}}}-{{{\rm{m}}}}_{{{\rm{HLA}}}}$$
(2)
$${\rm{HLA}}\%=\frac{{{\rm{m}}}_{{\rm{HLA}}}}{{{\rm{m}}}_{{\rm{product}}}}\times 100\%$$
(3)
$${\rm{FLA}}\%=\frac{{{\rm{m}}}_{{\rm{FLA}}}}{{{\rm{m}}}_{{\rm{product}}}}\times 100\%$$
(4)

Fabrication of SRF and nutrient release behavior

The humification product derived from 50 g of WP under optimized conditions (1 g PS, 1 g KOH, 160 W MW irradiation for 10 min) was mixed with 50 g of ATP. The mixture was then granulated to produce spherical fertilizer particles approximately 5 mm in diameter. Subsequently, these particles were soaked in 25 mL of ASO for 5 min. SRF was obtained after air-drying.

For slow-release test, SRF (30 g) was immersed in 1 L of deionized water. Solution aliquots (5 mL) were collected at given intervals to determine the FLA content in the aqueous system. The released FLA content was determined using UV-vis spectrophotometry at its maximum absorbing wavelengths of 270 nm.

Pot experiment with Shanghai cabbage and biosafety evaluation

To evaluate the fertilization of WP’s humification product and SRF, pot experiment was conducted with five treatments in triplicate: (A) 300 g soil (Untreated), (B) 300 g soil + 7.8 g WP (WP), (C) 300 g soil + 0.87 g K2SO4 (K2SO4, corresponding to the K+ dosage in 7.8 g Product), (D) 300 g soil + 7.8 g Product (Product), and (E) 300 g soil + 15.6 g SRF (SRF). The soil was collected from the field (121◦213 E, 31◦059 N) in campus of Donghua University (Shanghai, China) with fundamental properties provided in Supplementary Table S14. In each pot (top width and length of 9.8 cm, bottom width and length of 7.5 cm, height of 10.5 cm), 10 Shanghai cabbage seeds were sprayed evenly in soil at depth of around 1 cm (schematic diagram of the pot provided in Supplementary Fig. S12), and the pots were put in an incubator (humidity of 60%, 22 oC). The plants were exposed to LED lighting (photosynthetic photon flux density of 400 µmol/m²/s, 16 h light/8 h dark) and irrigated with 15 mL of tap water daily. After harvest in 14 days, plant parameters and rhizosphere soil were collected for analysis.

Besides, earthworm was used as a bioindicator to evaluate the biological safety of the Product and SRF. All the experiments were carried out with three replicates for each treatment. The detailed procedure was provided in Supplementary Note S3.

Characterizations

After filtration through a 0.45 μm membrane, the components of WP, VR and CKW during humification were preliminarily analyzed by a 3D-EEM fluorescence spectroscopy (F-7000, Hitachi High Technologies Co., Japan) and UV-vis spectrometry (UV-1900i, SHIMADZU Co., Ltd. of Japan). TOC and DOC were measured using a TOC analyzer (Muti N/C 210, Germany). Molecular weight distribution was measured using a GPC (LC-20A HPLC, SHIMADZU Co., Ltd. of Japan). K+ was measured with an inductively coupled plasma optical emission spectroscopy (ICP-MS, iCAP™TQ, Thermo Fisher). Radicals were identified via an EPR spectrometer (Magnettech MS5000, Germany). The photosynthetic photon flux density of the LED light for the pot experiment was measured via a calibrated quantum sensor (LI-COR LI-190R, or equivalent model).

The structural components of the freeze-dried samples were analyzed through the X-ray Photoelectron Spectroscopy (ESCALAB 250, Thermo-VG Scientific Co., USA), Fourier Transform Infrared spectroscopy (NEXUS-670, Nicolet Co., USA), X-ray Diffraction (D/max-2550VB/PC, Rigaku Co., Japan), and solid state 13C Nuclear Magnetic Resonance (Bruker AVANCE III 600 M). Thermal stability was analyzed with a Thermal Gravimetric Analyzer (TGA, TG-209F1, NETZSCH Germany). The soil indexes, including SOM, TN, TP, TK, AN, AP, and AK were measured according to the procedures previously reported36. Soil microbial community in the pot experiment was analyzed with the procedure provided in Supplementary Note S4. The surface morphology of the SRF was measured with Scanning Electron Microscopy (SU8010, Hitachi Japan).

Statistical analysis

Experiments were conducted at least in triplicate throughout the study, with data presented as mean ± standard deviation. Data analysis was conducted with Microsoft Excel (Microsoft CoWPoration) and Origin Pro 2021 (OriginLab Co, Northampton, MA). Statistical analysis was conducted using SPSS (Inc., Chicago, IL) based on one-way analysis of variance (ANOVA).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.