Slow release fertilizers based on polyphosphate/montmorillonite nanocomposites for improving crop yield

2.1 Materials and reagents

Urea (CO(NH₂)₂, >99%) was supplied from Luxi Chemical Co., Ltd. (Shandong, China), potassium chloride powder (KCl, >98%) was purchased from Qinghai Salt Lake Potash Fertilizer Co., Ltd. (Qinghai, China), montmorillonite (>96%) was purchased from Lingshou County Ruixin Mineral Powder Factory (Hebei, China), and phosphoric acid (H₃PO₄, >85%) was purchased from Kaitong Chemical Industry Co., Ltd. (Tianjin China).

2.2 Fertilizer preparation

2.3 Characterization of the prepared fertilizers

The prepared fertilizers were characterized using an infrared absorption spectrometer (Nicolet IS 10, USA). The solid samples were ground into powder, pressed into KBr pellets, and scanned in the range of 400–4000 cm⁻¹. The morphologies of the different fertilizers were observed using scanning electron microscopy (SEM, SU8080, Hitachi Ltd., Japan) and a transmission electron microscope (TEM, JEM 2100F, JEOL Ltd., Japan). X-ray diffraction (XRD) analysis of the composites was performed using a Shimadzu XRD 6000 diffractometer (Bruker D8 Advance, Germany). The relative intensity was registered in a diffraction range (2θ) of 3–10°, using a Cu Kα incident beam (λ = 0.1546 nm). The scanning speed was 1° min⁻¹ and the voltage and current of the X-ray tube were 30 kV and 30 mA, respectively. Diffraction patterns were analyzed using JADE 6.0. The prepared fertilizer was analyzed by thermogravimetric analysis (TGA, TGA/DSC 3+, Mettler Ettler Toledo, Switzerland) under air flow condition with ramping rate of 10 °C min⁻¹. Surface areas were calculated using the BET equation. The surface-area (BET) and pore size (Dp) measurements of samples were done on gas adsorption instrument (ASAP 2460, Micromeritics, America) at 77 K.

The 3D internal microstructures of UM, UPUKM, and UPUKM after elution were observed using 3D X-ray Computed Tomography. Tomographic scans were performed using a synchrotron beamline for full-field microtomographic imaging. Samples were analyzed with a Nanotom S (Phoenix, Ltd., Germany), using a 100 kV 100 μA X-ray source, with a final image resolution of 2 μm/pixel. A filtered back-projection-based reconstruction software was used to reconstruct cross-sectional images from microtomographic projections to 3D images.

2.4 Particle strength determination

Particle strength or particle hardness was a measure of the resistance of granules to deformation or fracture under pressure. For the measurement of particle hardness, fertilizer samples were first screened with a standard sieve to obtain granules with a size of 3.50–4.50 mm. Then, the particle hardness was measured for hardness by applying an increasing compressive force on a single particle with a particle hardness tester (Yinhe Instrument Factory, Jiangyan, China). The load at which 20 granules of a fertilizer fractured was averaged and reported as the particle hardness (N) of the fertilizer.

2.5 Particle fluidity determination

The angle of repose is the slope measured at the base of a pile of uncompressed granular solids when runoff occurs (Dubey and Mailapalli 2019; Ferreira et al., 2021). It is an indicator of particle fluidity and roundness. A larger angle of repose indicates lower particle fluidity, and a smaller angle of repose indicates higher particle fluidity (Tian et al., 2019). A fertilizer sample of 100 mL was placed into the funnel of an angle of repose tester (FBS104, FURBS, China), and the accumulation height (h) of the fertilizer was measured. Angle of repose (ϕ) was calculated according to Eq. (1).

2.6 Optimization of UPUKM preparation by response surface methodology (RSM)

Response surface methodology is a good mathematical statistical method for process improvement and optimization with a small number of experiments (Danmaliki et al., 2017; Chelladurai et al., 2021). To investigate the effects of the raw material ratio, reaction temperature, and reaction time on the properties of the prepared UPUKM, a center composite design (CCD), a standard RSM design, was adopted for model development by using the Design-Expert V11.1.0 software.

According to the previous experimental results (Fig. 1), the reaction of urea phosphate and urea will change the shape of fertilizer particles, so it is necessary to choose lower reaction temperature and longer reaction time. First, the optimum mass of urea phosphate (A) and potassium chloride (B) was evaluated under the reaction temperature of 90 °C and reaction time of 300 min. In order to obtain the fertilizer with more balanced nutrient elements, both the amounts of urea and montmorillonite were 1, and the masses of urea phosphate (A) and potassium chloride (B) were set to 0.5 and 1. The parameter range values in Table 2 are obtained by inputting the parameters into the design-expert V11.1.0 software. The masses of urea phosphate (A) and potassium chloride (B) were selected as independent variables (Table 2), and the fluidity and strength of fertilizer particles were taken as response parameters, with the matrix design shown in Table S2. In order to obtain the more energy-saving fertilizer preparation conditions, the reaction temperature was set to 80 °C and 120 ℃, and the reaction time was set to 120 min and 300 min, according to the previous experimental results (Fig. 1). The effects of reaction temperature (C) and time (D) on the fluidity and strength of fertilizer particles were analyzed using a 2² full factorial design (Tables 3 and S2). Finally, the results from the two RSM-CCD designs were combined to obtain the optimum process conditions for UPUKM preparation.

Close

Fig. 1

Particle hardness (a) and images (b) of the prepared fertilizers at different temperatures, FTIR spectra of fertilizer after preparation (c), and scheme for generation of polyphosphate via the polycondensation reaction between urea phosphate and urea (d). See Table 1 for details of the fertilizers.

Export to PPT

2.7 Evaluation of the nutrient release behavior of the prepared fertilizers

The nutrient release behavior of the prepared fertilizers was evaluated using a column leaching experiment. Rigid polyvinyl chloride columns with an inner diameter of 4 cm and a height of 30 cm were used. The columns were packed with a bottom layer of 15 cm quartz sand; a middle layer, which was the fertilizer sample; and an upper layer of 5 cm quartz sand (Fig. S1). Three grams of UPUKM and 2.25 g of conventional chemical fertilizer (CCF) were used so that the columns had the same nutrient contents. The CCF was a mixture of urea phosphate, urea, and potassium chloride. Filter papers were placed on top of the upper quartz sand layers to minimize surface erosion during rain simulation. A nylon mesh (80 mesh) was placed at the bottom of each column to prevent sand loss during the experiments. First, 76 mL water was added to saturate the column. Next, 408 mL more water was added. The amount of water used was based on the average seasonal rainfall in Taian City, Shandong Province (469 mm, National Meteorological Center of China, https://en.weather.com.cn/) and the cross-sectional area of the leaching column. The water flow rate was set to 34 mL h⁻¹ using a peristaltic pump (LabN1, Baoding Shenchen Pump Co., Ltd.), and the leachate was collected for analysis of nitrogen, phosphorus, and potassium analysis. Nitrogen was quantified using the Kjeldahl method. Total phosphorus content was determined using potassium persulfate digestion-molybdenum spectrophotometry. Total potassium was measured by flame emission spectrophotometry.

2.8 Evaluation of fertilizer agronomic effectiveness

A pot experiment was conducted at the New Fertilizer Experimental Station of Shandong Agricultural University in Tai’an, Shandong Province, China (117°13′ E, 36°20′ N), from June to September 2019. The soil used was collected from Binzhou City, Shandong Province (117°92′ E, 37°90′ N). This coastal salinized fluvo-aquic soil is classified as a Parasalic Ochri-Aquic Cambosol according to Chinese soil taxonomy. The basic physical and chemical properties of the soil were as follows: organic matter 7.69 g kg⁻¹, total nitrogen 0.50 g kg⁻¹, total phosphorus 1.83 g kg⁻¹, total potassium 35.94 g kg⁻¹, available nitrogen 85.58 mg kg⁻¹, available phosphorus 13.95 mg kg⁻¹, available potassium 186.19 mg kg⁻¹, pH 8.72, and total salt content 0.21%. Pottery pots (30 cm in diameter and 36 cm in height) were used and maize (Zea mays L. cv. Zhengdan 958) were grown with one maize plant per pot. Three treatments were set up with four replicates: CK (without fertilizer), CCF (conventional chemical fertilizer with N-P₂O₅-K₂O of 21–14-20), and UPUKM (14–11-16). The CCF was composed of urea, diammonium phosphate, and potassium chloride. For the CCF and UPUKM treatments, each pot received 5.40 g N, 1.44 g P₂O₅, and 3.60 g K₂O at the beginning of the experiment.

At the jointing stage of maize, soil samples were collected from each pot for nutrient analysis. Soil NO₃^–-N and NH₄⁺-N were extracted with 0.01 mol L^-1 CaCl₂ (soil-solution ratio of 1:10) and determined with a continuous flow injection analyzer (AA3, Bran. Luebbe, Germany). Soil available potassium was extracted with 1 mol L^-1 ammonium acetate and quantified with a flame photometer. Soil available phosphorus was extracted with 0.5 mol L^-1 sodium bicarbonate and determined with an automatic discontinuous chemical analyzer (Smartchem200, USA).

2.9 Statistical analysis

All data were subjected to analysis of variance (ANOVA) using SPSS software (SPSS 22, USA) followed by a mean comparison using Duncan’s multiple range test (p < 0.05).

3.1 Preparation of fertilizer based on polyphosphate/montmorillonite nanocomposites

The particle shape of UM, UPM, KM and UPUKM was remained spherical at 30–140 °C, while the particle strength of fertilizer in UM, UPM and KM did not change significantly, and the particle strength was low (Fig. 1a, b). In UPUKM, the particle strength of the fertilizer gradually increased, reaching 63.2 N at 140 °C. The FTIR analysis showed that the peak at 913 cm⁻¹ in UPUKM was the tensile vibration absorption peak of P-O-P (Fig. 1c), indicating a polycondensation reaction occurred between urea phosphate and urea and polyphosphate was generated (Fig. 1d) (Gong, 2001; Wang et al., 2017; Naraginti et al., 2019), while no such peak was found in UM, UPM and KM. This also shows that the formation of polyphosphates increases the particle strength of fertilizers. The particle strength of fertilizer can be increased by forming polymer. Yamamoto et al. prepared a kind of urea/urea–formaldehyde/montmorillonite nanocomposite by extrusion and granulation and heating, which increased the mechanical strength of fertilizer based on urea/montmorillonite nanocomposite (Yamamoto et al., 2016). However, the fertilizer prepared by extrusion was cylindrical, which was not conducive to mixed application with other fertilizers and mechanized fertilization, and formaldehyde was added in the production process, which may pose a threat to human and environmental health (Wang et al., 2023). UPU and UPUM fertilizer granules deform at 80 °C and 100 °C respectively. Owing to the presence of UP and U, the characteristic peak (P-O-P) of polyphosphates was observed in the infrared spectra of UPU (895 cm⁻¹) and UPUM (911 cm⁻¹). The peaks of the P-O-P bond in UPUM and UPUKM shifted to higher wavenumbers than those in UPU, which may be due to the hydrogen bonds between polyphosphate and montmorillonite. The higher deformation temperature of UPUM than that of UPU could be attributed to the layered structure of montmorillonite in UPUM, which slows down the contact and reaction between urea phosphate and urea. On the other hand, it may be that potassium ions in the inter-layer space of montmorillonite in UPUKM might have been strongly adsorbed by the negative charges of the two adjacent layers, rendering montmorillonite a much more stable structure, slowing down the contact and reaction between urea phosphate and urea, and keeping the fertilizer particles spherical (Morrow et al., 2013). Simultaneously, TGA analysis was performed on different treatments of fertilizers, revealing a trend in thermal stability following the sequence UPUKM > UPUM > UPU. This trend was consistent with the observed phenomena during the reaction process and further showed that the addition of nanoscale montmorillonite could enhance the stability of the fertilizer. The addition of potassium chloride further improved the thermal stability of the fertilizers (Fig. S2). The characteristic peaks at 3457, 3364, 3427, and 3333 cm⁻¹ in the FTIR spectrum are attributed to the stretching vibration of –NH in urea phosphate and urea in FTIR (Fig. 1c). The peaks at 1034, 1038, and 1029 cm⁻¹ are related to Si-O-Si in montmorillonite. The peaks at 1080 and 1066 cm⁻¹ are due to the stretching vibration of P-O. The peaks at 1012 and 1003 cm⁻¹ correspond to the symmetrical stretching vibration of PO₂.

3.2 XRD analysis of prepared fertilizer

XRD is commonly employed to determine the interlayer spacing between minerals and nanocomposites (Vo et al., 2019; Zhang et al., 2020). The larger basal spacing in UM (d = 1.81) than that in montmorillonite may be due to the entry of urea into the interlayer of M (d = 1.27) (Fig. 2). The basal spacing of UPM was the same as that of the montmorillonite, but the reflection was weaker because of the partial destruction of the layered structure of the montmorillonite (Yadav and Asthana, 2002; Pacuła et al., 2014). UP, U, K and UPU had no characteristic diffraction peak in the range of 2θ = 3 ∼ 10°. The interlayer spacings of KM was 1.15 nm. This distance may be because potassium ions enter the montmorillonite lattices, strengthening the bonding of adjacent crystal layers and, thus, reducing interlayer spacing (Jiang et al., 2014; Moslemizadeh et al., 2019). The interlayer spacing of UPUM and UPUKM was 1.43 nm and 1.41 nm, respectively. The difference in the trend of interlayer spacing between UPUKM and UPUM may be attributed to the fixation of montmorillonite by potassium in UPUKM, which reduced the interlayer spacing of montmorillonite as discussed in previous sections (Huang et al., 2018; Wang 2020). Overall, the interlayer spacings of UPUM and UPUKM were larger than those of the original montmorillonite, indicating the formation of polyphosphate-montmorillonite nanocomposites in both UPUM and UPUKM.

Close

Fig. 2

XRD patterns of the raw materials (a) and the prepared fertilizers (b). d-spacing was calculated from the Bragg Law: 2d sinθ = nλ. See Table 1 for details of the fertilizers.

Export to PPT

3.3 Morphology analysis of prepared fertilizer

Further investigation of the nanostructure was performed using field emission scanning electron microscopy (SEM) and transmission electron microscopy (TEM). As shown in Fig. 3, the raw MMT clay exhibited a thin, flexible plate-like morphology (Fig. 3a1). However, in UPUKM, urea phosphate reacted with urea to form chain polyphosphate, and montmorillonite was dispersed in the polymer in the fertilizer (Fig. 3b1). Additional, structural information was obtained from the TEM micrographs of the as-prepared samples. An ordered crystal layer structure of montmorillonite was observed. In UPUKM, the silicate layer of montmorillonite was dispersed in the fertilizer particles, but the interlayer structure became disordered and exfoliated. The polyphosphate/montmorillonite nanocomposites with an exfoliated structure were more beneficial to the mechanical properties. The mechanical properties of the nanocomposites are improved due to the strong interaction between the nanoparticles and the polymer matrix and the reduction of the distance between the particles (Xie et al., 2022).

Close

Fig. 3

The SEM (a₁, b₁) and TEM (a₂, b₂) images of montmorillonite (a) and interior of UPUKM (b).

Export to PPT

3.4 Optimum UPUKM preparation process

The UPUKM preparation process was analyzed using ANOVA to identify the effect of the raw material mass ratio on the angle of repose (Y₁) and hardness (Y₂) of the fertilizer (Table 4). Two polynomial models with statistical significance were generated using a combination of the parameter estimates.

The regression equations were significant (p < 0.0001), the lack of fit was not significant (P > 0.05), and R² and Adj R² were>90% (Table 4), indicating that the models could explain most of the changes in the response variables.

The obtained surface obtained based on the quadratic models are presented in Fig. 5. When the mass of potassium chloride was constant, the particle hardness of the fertilizer increased, and the fluidity gradually decreased with an increase in the mass of urea phosphate. This finding was expected because with more polyphosphate generated due to the increase of urea phosphate, the particle hardness of the fertilizer increased. Simultaneously, with an increase in urea phosphate, as discussed previously, the reaction between urea phosphate and urea increases in intensity, and the fertilizer system becomes unstable, resulting in the deformation of fertilizer particles and an increase in the angle of repose or a decrease in particle fluidity. The optimal raw material mass ratio was found to be 1:1:0.93:1 (urea phosphate: urea: potassium chloride: montmorillonite), at which point the particle fluidity and hardness of the prepared fertilizer were maximized (see Table 5).

The effects of reaction temperature (C) and reaction time (D) on the hardness and angle of repose of the fertilizer particles are expressed by Eqs. (4) and (5):

The response surfaces of the response variables are presented in Fig. 6. Temperature strongly influenced on the hardness and fluidity of the fertilizer particles. The hardness and fluidity of the fertilizer particles first increased and then decreased with increasing reaction temperature, indicating that when the reaction temperature was too high, the reaction between urea phosphate and urea was so violent that the fertilizer particles were deformed, the fertilizer angle of repose increased, and the fluidity decreased.

The optimum process conditions identified by response surface analysis (Figs. 4 and 5) were reaction temperature, 103 °C; reaction time, 237 min; and mass ratio, 1:1:0.93:1 (urea phosphate: urea: potassium chloride: montmorillonite). Under these optimum conditions, the particle hardness and angle of repose of the prepared fertilizer were 64.75 ± 0.48 N and 30.84 ± 0.95°, respectively, which were close to the software-predicted values of 62.54 N and 31.21°, respectively.

Close

Fig. 4

Response surface plots showing effects of raw material mass ratio on particle hardness (a) and angle of repose (b) of prepared fertilizer. The two raw materials are urea phosphate (UP) and potassium chloride (KCl).

Export to PPT

Close

Fig. 5

Response surface plots showing effects of reaction temperature and time on particle hardness (a) and angle of repose (b) of the prepared fertilizer.

Export to PPT

Close

Fig. 6

Cumulative release of nitrogen (a), phosphorus (b), and potassium (c) from prepared fertilizers. CCF is a conventional chemical fertilizer with N-P₂O₅-K₂O of 21–14-20. See Table 1 for details of UPUK and UPUKM.

Export to PPT

3.5 Nutrient release behavior of the prepared fertilizers

N, P, and K were rapidly released from the CCF in the first three cycles (Fig. 6). For CCF, 100% of total N, 93% of total P, and 100% of total K were released and leached from the column three times, with up to 68% N, 72% P, and 60% K leached the first time. P was released and leached from the column five times.

Nutrient release from UPUKM was much slower: 11.60%–15.67% and 33.59%–55.59% of the three nutrients were released and leached from the column in the first and second cycles, respectively. After five leaching cycles, nutrient release from UPUKM became even slower than that aforementioned, and after 12 leaching cycles, approximately 92%-94% of the three nutrients were released. Because UP and urea enter the interlayer of montmorillonite, polyphosphates are formed via in situ polymerization. Nutrients in the fertilizer were hindered by the interlayer of montmorillonite, which reduced nutrient leaching. The slow release of potassium from UPUKM may be because the potassium ions were fixed in the montmorillonite lattice, forming a structure similar to that of mica, which can be used as a slow-release potash fertilizer (Said et al., 2018).

3.6 Fertilizer particle structures observed with XCT

To investigate the effect of urea phosphate and urea heating on the internal structure of fertilizer and the slow-release mechanism of fertilizer, UM, UPUKM, and eluted UPUKM fertilizer samples were scanned using XCT. XCT is a nondestructive method of scanning the internal structure so that the difference between each sample can be observed intuitively. X-ray computed tomography showed a clear core–shell structure in UM (Fig. 7a2_,3), which is a characteristic of agglomeration and granulation products. The instability of this core–shell structure results in its lower particle strength, which was consistent with the particle strength of UM in Fig. 1, which was only 10.2 N. There was no core–shell structure in UPUKM, and the structure was uniform (Fig. 7b2), with larger pores (Fig. 7b3_,4). This may be the result of the reaction of urea phosphate and urea to form long-chain polyphosphate, and simultaneously the polyphosphate enters the interlayer of montmorillonite and exfoliated its nanosheets to form an interlocking structure, causing the core–shell structure to disappear and the particle strength to increase (Zhu et al., 2019). The addition of nano-materials makes the matrix and nanoparticles maintain uniform dispersion and good adhesion, and significantly increases the mechanical properties of the composite (Domun et al., 2015; Heidari et al., 2021; Garg et al., 2022). The research showed that carbon nanotube reinforced polymeric composite doubly curved micro-shell panels (Huang et al., 2021). This has a good inspiration for our future work, for example, we can explore the effect of different nano-materials on the strength of fertilizer particles. After the leaching experiment, larger pores are visible on the montmorillonite skeleton of UPUKM (Fig. 7c4), which may have been left behind after the soluble matter was leached out. The montmorillonite skeleton still exists after UPUKM has been completely eluted (Fig. 7c1), indicating that the stability of the fertilizer in the soil was relatively high. Furthermore, the structure of the montmorillonite skeleton also hinders the diffusion of nutrients, thereby slowing down the release of nutrients in the fertilizer.

Close

Fig. 7

X-ray microtomography images of UM (a), UPUKM (b), and UPUKM after the leaching experiment (c): (1), (2) reconstructed 3D images, (3) 2D slice images, (4) magnified 2D slice images. See Table 1 for details of the fertilizers.

Export to PPT

3.7 Agronomic effectiveness of prepared fertilizer

The yield, plant height, and stem diameter of maize at the 12-leaf stage were 24.27%, 3.98%, and 10.13%, respectively, and were significantly higher in the UPUKM treatment than in the CCF treatment (Fig. 8). This finding demonstrates that the slow release and sustained supply of nutrients from the prepared composite fertilizer (UPUKM) was the key to increasing the crop yield.

Close

Fig. 8

Yield (a), plant height (b), stem diameter (c), available N (d), P (e), and K (f) at the jointing stage of maize in CK, CCF, and UPUKM. Soil of maize in CK, CCF, and UPUKM. Different letters above the bars indicate significant differences between the means by Duncan’s multiple range test (p < 0.05, n = 3). CK: without fertilizer; CCF: conventional chemical fertilizer with N-P₂O₅-K₂O of 21–14-20. See Table 1 for details of UPUKM (N-P₂O₅-K₂O of 14–11-16).

Export to PPT

At the jointing stage of maize, the soil nitrate nitrogen content in UPUKM was 28.02%, significantly higher than that in CCF, and available phosphorus was significantly higher (43.45%) (Fig. 8). Maize is a fast-growing crop that requires a high level of phosphorus at the beginning of its life cycle. In this study, the UPUKM treatment significantly increased the soil available phosphorus content and promoted the absorption and utilization of phosphorus by maize. There was no significant difference in the soil available potassium content between treatments. Therefore, the UPUKM in the formula could be used as a nanocomposite to slow the release of nutrient elements, prepare a good nutrient resource and thus improve the crop yield (Salimi et al., 2021).