Introduction

Phosphorus (P) is an indispensable element for plant growth, and the application of P fertilizer plays a crucial role in promoting the development of modern agriculture. The majority of P is sourced from phosphate rock, with its economic mining estimated to last approximately 100 years1. To ensure sustained access to P for food security, it is imperative to adopt essential sustainability strategies such as recycling of P from P-enriched wastes into agricultural land. Animal manure, in particular, holds significant potential as a fertilizer due to its high P content (up to about 3% dry matter)2. The P quantity from animal manure was estimated to reach 150 million t3. However, animal manure often contains pharmaceuticals, pathogens, and heavy metals that can pose risks to ecosystem and enter the food chain. Therefore, there is a need to establish safe and efficient methods for utilizing P in animal manures.

Among various treatment techniques for animal manure, pyrolysis has emerged as an effective alternative approach. Pyrolysis not only dramatically reduces the volume by up to 50%4, but also effectively eliminates pathogens and organic contaminant5 while immobilizing inorganic elements present in the feedstock6. Moreover, this process generates high value-added production such as bio-oil, bio-gas, and biochar from animal manure pyrolysis. The advantage lies in the enrichment of P within biochar due to its low volatility at temperatures below700 ℃7. During pyrolysis, the majority of organic phosphates (OP) are transformed into inorganic phosphates (IP), with some water-soluble IP being converted into less labile forms like hydroxyapatite and oxyapatite at higher temperature8. However, the P availability within biochar is much lower than that found within feedstock, making it unsuitable as a direct source of P-fertilizer9. Nevertheless, a temperature threshold above or equal to700 °C during pyrolysis is considered as an effective approach for reducing the potential risk of heavy metal in biochar obtained from animal manure10,11,12. Therefore, there is a need to resolve this contradiction.

Thermal treatment of P-rich waste with alkali additives has been proved to be a promising method to increase available P content. Calcination of sewage sludge ash blended with alkali additives (Na2CO3, Na2SO4, or K2SO4) at the temperature between 800 and 1000 °C converts poorly plant available P from whitlockite (Ca3-x(Mg, Fe)x(PO4)2) and AlPO4 into highly plant available calcium alkali phosphates ((Ca, Mg)(Na, K)PO4) via Rhenania reaction (Eqs. 1–3)13,14,15,16. In contrast, K-doping (2% and 5% potassium acetate) increases P availability in biochar derived from sewage sludge pyrolyzed at 700℃ through the formation of K2HPO4, however, its mechanism is unknown17. Our earlier study found that co-pyrolysis of K-rich agricultural biomass waste (corn stalk) and P-additives (Ca(H2PO4)2 or NH4H2PO4) at 500℃ could form some potassium phosphates (KH2PO4) or (Ca, Mg)(Na, K)PO418. These results suggest that the reaction production of P and K depends on the precursor types and thermal treatment temperature. However, Buss et al. (2022) found that the enhancing effect of K-doping on water extractable Cr and Cu content was found in biochar derived from sewage sludge at high temperature (600℃, 700℃, 800℃ and 900℃)19. Therefore, optimal process conditions are necessary for producing excellent biochar.

$$Ca_{3} \left( {PO_{4} } \right)_{2} \, + \,Na_{2} CO_{3} \, + \,SiO_{2} \, \to \,2CaNaPO_{4} \, + \,CaSiO_{3} \, + \,CO_{2}$$
(1)
$${\text{Ca}}_{{3}} \left( {{\text{PO}}_{{4}} } \right)_{{2}} + {\text{2Na}}_{{2}} {\text{SO}}_{{4}} + {\text{Al}}_{{2}} {\text{O}}_{{3}} + {\text{SiO}}_{{2}} \to {\text{2CaNaPO}}_{{4}} + {\text{2NaAlSiO}}_{{4}} + {\text{CaSiO}}_{{3}} + {\text{SO}}_{{2}}$$
(2)
$${\text{Ca}}_{{3}} \left( {{\text{PO}}_{{4}} } \right)_{{2}} \, + \,{\text{2K}}_{{2}} {\text{SO}}_{{4}} \, + \,{\text{Al}}_{{2}} {\text{O}}_{{3}} \, + \,{\text{SiO}}_{{2}} \, \to \,{\text{2CaKPO}}_{{4}} \, + \,{\text{2KAlSiO}}_{{4}} \, + \,{\text{CaSiO}}_{{3}} \, + \,{\text{SO}}_{{2}}$$
(3)

The aforementioned reactions suggest that a molar ratio of alkali metal/P=1 is sufficient to produce (Ca, Mg)(Na, K)PO4, which exhibits poor water solubility but good solubility in both alkaline and neutral ammonium citrate20. However, previous study reported that Na/P molar ratios of 1.3–2.7 were necessary to achieve high alkaline ammonium citrate-soluble P13,14. Si is mainly present as silicates and SiO2 in agricultural and industrial wastes21,22, thus, the side reactions of alkali additives with Si may be responsible for the increased consumption of these additives. Previous research has shown that the extent of reaction between alkali additives and Si compounds depends on precursor types and pyrolysis temperature23,24. For instance, the reaction between Na or K-containing additives and silicates is thermodynamically favored over P15. In addition, KOH and K2CO3 are more efficient to enhance the availability of Si in rice straw biochar than CaO due to their higher base nature, leading to easier reactivity react with SiO2 (Eqs. 4–5)24. Therefore, it can be inferred that there is an interaction between P and Si when limited amounts of K were added. To the best of our knowledge, no prior studies have explored the effect of different types of K on P speciation in livestock manure derived-biochar.

$${\text{SiO}}_{{2}} \, + \,{\text{KOH}}\, \to \,{\text{K}}_{{2}} {\text{SiO}}_{{3}} \, + \,{\text{H}}_{{2}} {\text{O}}$$
(4)
$${\text{SiO}}_{{2}} \, + \,{\text{K}}_{{2}} {\text{CO}}_{{3}} \, \to \,{\text{K}}_{{2}} {\text{SiO}}_{{3}} \, + \,{\text{CO}}_{{2}}$$
(5)

In this work, two inorganic (KOH and K2CO3) and two organic (C6H5K3O7 and CH3COOK) K compounds were selected as K sources. Swine manure may contain high concentrations of Cu and Zn because of feed additives11,12. The main aims of this study were (1) to characterize the properties of biochar pretreated by different K sources, (2) to assess the change of P and heavy metals speciation; and (3) to reveal the underlying mechanisms using XRD, solution31P-NMR, and FTIR.

Materials and methods

Feedstock and biochar preparation

The swine manure (SM) was obtained from a farm in Shijiazhuang City, China. Subsequently, the SM was dried at 105 °C for 24 h in an oven, ground using a crusher, and sieved through a 60 mesh sieve. Two inorganic potassium (K2CO3 and KOH) and two organic potassium (CH3COOK and C6H5K3O7) were chosen as the source of K. Impregnation of the SM with K was achieved by adding 20 g of dried SM into the prepared K solution in a beaker at a mixing ratio of 5% K (w/w), followed by stirring the mixture for 1 h at room temperature using a stirrer. Afterward, these mixed samples were directly dried at 105 °C until rtheir water content reached below 5%. The resulting dried samples were subjected to produce biochar in a tube reactor under N2 atmosphere (99.999%, flow rate: 300 mL min-1) at a heating rate of 10 ℃ min-1 and held at the desired temperature of 700 °C for 2 h.. The produced biochars were labled as follows: SMB-swine manure; SMB + K1-swine manure + K2CO3; SMB + K2-swine manure + KOH; SMB + K3-swine manure + CH3COOK; SMB + K4–swine manure + C6H5K3O7.

Thermogravimetric analysis (TGA)

The mass loss analysis of SM with or without P was investigated using a NETZSCH STA 449F3 thermogravimetric analyzer. To ensure consistency of all experimental parameters, about 10 mg of samples were heated from ambient temperature to 800 ℃ at a heating rate of 10 ℃ min-1 under an N2 atmosphere (99.999%, flow rate: 50 mL min-1).

Biochar characterization

The C, H, and N contents were measured with an elemental analyzer (Vario EL II, Hanau, Germany). The pH and EC values were determined with a samples/water ratio of 1:20 (W/V) after 18 h equilibrium using a multifunction meter (DZS-706, Leici, China). Ash contents were determined by the weight loss after heating at 650 °C for 4 h under air atmosphere. The total K, Ca, and Mg contents of both SM and biochars were analyzed by using inductively coupled plasma-atomic emission spectrometry (ICP-OES, DV2100, Perkin Elmer, USA) after digesting a sample of 0.50 g with 20 mL mixture of HNO3: HClO4: HF (4:1:2, V:V:V).

Fourier Transform Infrared (FTIR) spectra of above samples were recorded with Nicolet iS50 spectrometer (Thermo Fisher Scientific, USA) in the region of 400–4000 cm-1. X-ray diffraction (XRD) patterns were obtained using a powder X-ray diffractometer (PANalytical X'Pert3 Powder diffractometer, The Netherlands) equipped with Cu Kα radiation (λ = 0.154 nm) in the 2θ range of 10–80° at 40 kV and 40 mA.

Analysis of P speciation and bioavailability

Sequential extraction methods based on the Hedley protocol and its modified versions have been extensively employed to assess the relative mobility and abundance of different P species. In order to characterize the mobility and solubility of P in SM and its thermal treatment products, the dried solids were subjected to sequential extraction following Hedley’s method25,26. Briefly, 250 mg of the solid sample was added to a 50 mL polypropylene centrifuge tube and sequentially extracted by 20 mL extraction solutions, including deionized water (readily soluble P), 0.5 M NaHCO3 (exchangeable P), 0.1 M NaOH (Fe/Al mineral adsorbed P), and 1.0 M HCl (insoluble phosphates) solution, with each step lasting 16 h by end-to-end shaking. The residual fraction and total P were measured after ashing at a temperature of 650 °C using NaOH. The bioavailability of P in samples was analyzed by using water, 1 M neutral ammonium citrate (NAC), 2% formic acid (FA), 0.5 M NaHCO3. Briefly, samples were shaken in an end-over-end shaker with a ratio solid: solution being maintained at w/v = 1:100 for duration of 16 h before being centrifuged and filtered. The concentration of P in extracts was determined using an inductively coupled plasma-optical emission spectrometer (ICP-OES, Optima 2000DV, Perkin Elmer, USA).

Solution-state 31P NMR samples were obtained by NaOH-EDTA extraction. For each feedstock and biochar, a mass of 0.5 g of sample was shaken in 10 mL solution containing a mixture of 0.25 M NaOH and 0.05 M EDTA at 20 °C for 16 h and then centrifuged at 4000 rpm for 30 min. The supernatant was filtrated, collected, and then freeze-dried for 48 h. Subsequently, 100 mg of freeze-dried samples were dissolved in 1 mL of 1 M NaOH with 10% D2O. Solution 31P NMR spectra were analyzed using a Bruker AV400M NMR spectrometer operating at 162 MHz. Spectra were recorded with a pulse width of 0.9 μs (15°), an acquisition time of 84 ms, and broadband proton decoupling. A pulse delay time of 5 s was employed, and a total of 300–2000 scans were collected to obtain reliable signals. Chemical shifts were determined relative to 85% H3PO4 solution (0 ppm).

Measurement of Si forms in samples

The dissolved and available Si was extracted using 0.02 mol L−1 CaCl2 and 1 mol L−1 HAc-NaAc (pH = 4.0) buffer, respevtively. The Si concentration in filtrate was detected using molybdenum blue colorimetry.

Analysis of heavy metals speciation and evaluation of risk assessment

The chemical speciation of heavy metals in the samples was determined using the BCR sequential extraction procedure, which classified the metals into four fractions: acid-soluble (F1), reducible (F2), oxidizable (F3), and residual (F4). The detection of heavy metals in filtrates from digestion and BCR sequential extraction was performed using an ICP-OES (Optima 2000DV, Perkin Elmer, USA).

Risk assessment code (RAC) was employed to assess the environment risk of Cu and Zn in SM and biochars. RAC is defined as the percentage of heavy metals in F1 fraction27.

$$RAC\, = \,F1/C_{n} *100$$

Where F1 is the content Cu or Zn in F1 fraction and Cn is the total content of corresponding heavy metals.

Five classification of RAC were showed as: (1) no risk, RAC < 1%; (2) low risk, 1% ≤ RAC < 10%; (3) medium risk, 10% ≤  RAC < 30%; (4) high risk, 30% ≤ RAC < 50%; (5) very high risk, 50% ≤ RAC.

Results and discussions

Thermogravimetric analysis

As depicted in Fig. 1, the thermal analysis of SM revealed three distinct peaks, corresponding to the loss of water (110 ℃), the decomposition of hemicellulose (295℃) and cellulose (330 ℃)28. Conversely, the third peak disappeared in K-mixed SM, indicating that the addition of K compounds facilitated the decomposition of cellulose, resulting in an overlapping zone during the decomposition of hemicelluloses and cellulose. This observation aligns with previous studies demonstrating that K could enhances the reactivity of cellulose and reduces the peak temperature of decomposition29,30.

Figure 1
Figure 1
Full size image

Thermogravimetric analysis of swine manure with or without K.

Biochar properties

The properties of SM and biochars are presented in Table 1. SM exhibited a slightly alkaline pH value of 7.58, and the pH of biochars increased to 10.87. After K addition, the pH of biochar increased 0.30, 0.37, 0.75, and 0.69 unit, respectively. Interestingly, the addition of CH3COOK and C6H5K3O7 led to the bigger increase in pH than KOH and K2CO3. As expected, the yield rate of biochar slightly increased after K addition, whereas the ash content obviously increased. Meanwhile, the elements including C, N, H, O and P, significantly or slightly decreased with the introduction of K, mainly due to the dilution effect. Notably, compared to SMB, lower H/C ratio in K-doped biochar indicated more aromatic structure. Although the total content of Ca and Mg slightly decreased, the exchangeable Ca and Mg content markedly increased after K compounds addition. It was expected that the total and available K content significantly increased.

Table 1 The properties of swine manure and biochar.
Full size table

Phosphorus availability and fractions

As depicted in Fig. 2, the water-extractable P content accounted for 33.19% of total P (TP) in SM, it drastically decreased to only 0.29% of TP after pyrolysis at 700 ℃, representing a remarkable 114-fold decrease. Other studies also reported dramatic declines in P availability following animal manure pyrolysis. For instance, Li et al. (2018) found that water extracted 49.5% of TP from raw poultry litter and 2.4–11.7% of TP in poultry litter-derived biochars pyrolyzed at 300–600 °C8. The Olsen solution (0.5 M NaHCO3, pH 8.5)-extractable P was comparable to the water-extractable P from SM (32.2%), whereas it performed better capacity in recovering P from SMB (4.3%). NAC solution extracted 56.0% of TP from SM, while SMB possessed the drastically lower fraction (24.6%). FA acted more efficiently than other extraction solutions in recovering P from SM (73.6%) and SMB (62.3%). Overall, the percentage of extractable P increased in the following order: water < NaHCO3 < NAC < FA, indicating that soluble P of SMB consists of stable and crystal Ca-P due to the complete dissolution of amorphous Ca-P in acidic ammonium oxalate7. Similar results were found by previous studies9,21. Comparing with SMB, doping of SM with K greatly increased the proportion of NaHCO3, NAC and FA extractable P to TP to 4.9-7.0%, 41.1-50.3% and 69.8-76.2%, respectively, whereas its effect on water extractable P was negligible (Fig. 2). Overall, the organic K addition showed better effect on the enhancement of bioavailable P than inorganic K irrespective of extraction solution.

Figure 2
Figure 2
Full size image

The percentage of available P in swine manure and biochar.

The Hedley fractionation method categorizes the P species extracted by water, NaHCO3, NaOH and HCl to be readily soluble, exchangeable, Fe/Al mineral adsorbed and insoluble, respectively. In SM, HCl–extractable P accounted for 34% of TP, followed by water-extractable P (29%), NaOH-extractable P (18%), NaHCO3 -extractable P (13%), and residual P (5%) (Fig. 3). After pyrolysis at 700℃, the proportions of water, NaHCO3, NaOH-extractable P further decreased to 0.7%, 4%, and 10%, while the proportion of HCl- extractable P increased to 78%, indicating an increase in insoluble P content. In contrast, the proportion of residual P in biochar was slightly higher than that in SM (8% vs 5%). Doping of SM with 5% K only increased the fractions of NaHCO3 and NaOH- extractable P while simultaneously reducing the fraction of HCl extractable P, indicating a conversion from insoluble P into exchangeable and Fe/Al mineral adsorbed P26, and the organic K showed a better promoting effect compared to inorganic K.

Figure 3
Figure 3
Full size image

Fractionation of P in swine manure and biochars.

XRD

XRD analysis was conducted to reveal the P mineralogy in biochars. As depicted in Fig. 4, the peaks of CaCO3, SiO2, KCl and whitlockite (Ca7Mg2P6O24) were identified in SMB. Previous studies also reported the presence of crystalline Ca phosphate phase in biochar derived from animal manure31,32. Whitlockite and other Ca phosphate minerals exhibit high solubility for P under strong acid condition, which could explain the relative high percentage of HCl -extractable P observed in SMB. Additionally, it was observed that the introduction of K led to the formation of poorly water-soluble CaKPO4 that is readily soluble in weak acid and neutral ammonium citrate20. This accounts for a significant increase in NAC and FA-extractable P content for K-doped biochars. Furthermore, CaKPO4 was also formed in all thermally treated mixtures containing three common P compounds (CaHPO4, C12H10Ca2MgO14, and Ca3(PO4)2) in manure and four K compounds used in this study (Figure S1), indicating no discernible difference regarding its formation when various K compounds were added. This observation was not different from the previous study17, which might be due to the difference in P species between sewage sludge and swine manure. However, no SiO2 peak was observed in SMBK2, and lower intensity of SiO2 peaks were found in other K-doped biochars with decreasing order: SMBK1 < SMBK4 < SMBK3. This trend aligns with an increase in soluble Si contents in K-doped biochars (Table S1), suggesting their different reaction reactivity towards SiO2. The weaker reaction capacity of organic K with SiO2 resulted into a higher propensity for CaKPO4 formation in organic K doped SMB.

Figure 4
Figure 4
Full size image

XRD spectrum of biochar.

FTIR

As shown in Fig. 5, several peaks observed at 3270, 2917, 2849, 1636, 1419, and 1031 cm−1 in SM, which assigned to –OH, C-H, C = O, C-O, and C–O–C of stretching vibration, respectively, became weaker or shifted to lower wavenumber after pyrolysis due to decomposition of organic matter. In contrast, the intensity of peaks between 600 cm-1 and 500 cm-1 increased, which probably associated with inplane vibration of phosphate (PO43-) and asymmetric stretching of Si–O-Si33, reflecting the enrichment of SiO2 and phosphates.

Figure 5
Figure 5
Full size image

FTIR patterns of swine manure and biochars.

Two new peaks of 1466 and 1383 cm-1, which were ascribed to CO 2-3 , appeared in K-doped biochar, indicating the formation of K2CO3. The peak of 1031 cm-1 shifted to lower wavenumbers (996 cm-1), which could be attributed to the replacement of Si–O–Si bonds by Si–O–M bonds34, further confirming the formation of silicate, i.e. K2SiO3. The peaks at 591 cm-1 disappeared while the intensity of peaks of 546 cm-1 decreased. These alterations are directly linked with the transformation of Ca7Mg2P6O24 into CaKPO435.

Solution phase 31P NMR

To further investigate the chemical speciation of P in biochar, liquid 31P NMR spectroscopy was employed to analyze NaOH-EDTA extracts of SM, SMB, SMBK1 and SMBK4 (Fig. 6). Successful extraction rates for P were determined as follows: 44.3% for SM, 32.5% for SMB, 50.5% for SMBK1, and 64.3% for SMBK4 (Table S2). Notably, some minor peaks corresponding to monoesters phytate (at 4.9, 4.6, and 4.4 ppm)36 were observed SM. However, almost exclusively orthophosphate (PO43- at approximately 6.0 ppm)37, accounted for the majority of P species in SMB, SMBK1 and SMBK4 (98.1–99.1%, Table S2), indicating the decomposition of organic phosphate into orthophosphate during biochar formation process. Interestingly, the peak position of orthophosphate was slightly lower in both SMBK1 (5.91 ppm) and SMBK4 (5.93 ppm) compared to that of SMB (6.04 ppm). Quillard et al. (2011) found that the substitution of some calcium sites in β-tricalcium phosphate (β-TCP, Ca3(PO4)2) by potassium and/or sodium cations could lead to the NMR chemical shifts38.

Figure 6
Figure 6
Full size image

Solution phase 31P NMR spectrum of swine manure and biochar.

Heavy metals speciation

The chemical speciation distributions of heavy metals in the SM and biochar are illustrated in Fig. 7. The F1 + F2 fraction accounted for the predominant proportion of Cu (66%) and Zn (90%) in SM. After pyrolysis at 700℃, a noticeable reduction of the F1 and F2 fraction, accompanied by a corresponding increase in the F3 and F4 fractions, indicating that pyrolysis at high temperature could effectively immobilize heavy metals. Similar transformation patterns of Cu and Zn during the pyrolysis of livestock manure were reported by Meng et al.10 and Zeng et al.11.

Figure 7
Figure 7
Full size image

Chemical fractionation of Cu (a) and Zn (b) in swine manure and biochars.

The addition of all K compounds resulted in a decrease in the F1 + F2 fraction to less than 1% for Cu, while promoting the transformation of F3 to F4. The percentage of the F4 fraction of Cu increased from 47.6% (SMB) to 61.0%, 57.1%, 61.9%, and 57.9% for SMBK1, SMBK2, SMBK3 and SMBK4, respectively. In contrast to Cu, the addition of all K-compounds reduced the fractions of F1 and F3 for Zn but led to an increase in the fractions of F2 and F4 was observed. Our previous studies found that alkali and alkaline earth compounds including NaOH, CaO, CaCO3 and Ca(OH)2 obviously promoted the immobilization of Cu and Zn6,39. One the hand, the formation of K2CO3 led to the increase in pH of biochar, which could decrease the extraction efficiency of acetic acid, and thus led to the reduction of F1 fraction of Cu and Zn. On the other hand, K2SiO3 could react with Cu or Zn and other metals (e.g. Ca, Mg, Al, etc.) to form more stable silicate, which result into the increase of F4 fraction of Cu and Zn. However, Zn2(PO4)2 is one of the main compositions of Zn in biochar derived from swine manure40. Therefore, more soluble ZnxK1-xPO4 might be formed through a similar way of the transformation of Ca7Mg2P6O24 into CaKPO4, contributing to the rise of F2 fraction of Zn.

The values of RAC for Cu and Zn in SM and biochars are displayed in Table S3. Cu (RAC = 24.69%) and Zn (RAC = 29.62%) in SM showed a medium risk. SMB presented a low risk with an RAC of less than 10%. After the addition of K compounds, all biochars exhibited a lower risk (RAC of Cu and Zn < 1% and < 2%, respectively). Overall, there were no significant difference observed in the effect of four K compounds on heavy metals risk.

Conclusions

The present study demonstrates the different effects of four K compounds (KOH, K2CO3, CH3COOK, and C6H5K3O7) on P availability and heavy metals immobilization in biochars derived from swine manure, as well as their corresponding properties. K-doped biochar exhibit higher pH levels, whereas the addition of K compounds led to lower H/C, resulting from the catalysis activity of K. Furthermore, all K compounds facilitated the transformation Ca7Mg2P6O24 of into CaKPO4, thereby significantly enhancing P availability. Among them, CH3COOK, and C6H5K3O7 demonstrated superior performance compared to KOH and K2CO3 because of their weaker reaction capacity with SiO2. Additionally, the addition of K compounds increased the residual form of Cu and Zn while reducing their associated risk. In conclusion, co-pyrolysis of livestock manure with organic K compounds can yield a sustainable PK fertilizer with minimal heavy metals risk. These finding were obtained through laboratory experiment; therefore, future studies should investigate their performance as fertilizers and/or soil amendments in filed experiment.