Urea-functionalized amorphous calcium phosphate nanofertilizers: optimizing the synthetic strategy towards environmental sustainability and manufacturing costs

Nanosized fertilizers are the new frontier of nanotechnology towards a sustainable agriculture. Here, an efficient N-nanofertilizer is obtained by post-synthetic modification (PSM) of nitrate-doped amorphous calcium phosphate (ACP) nanoparticles (NPs) with urea. The unwasteful PSM protocol leads to N-payloads as large as 8.1 w/w%, is well replicated by using inexpensive technical-grade reagents for cost-effective up-scaling and moderately favours urea release slowdown. Using the PSM approach, the N amount is ca. 3 times larger than that obtained in an equivalent one-pot synthesis where urea and nitrate are jointly added during the NPs preparation. In vivo tests on cucumber plants in hydroponic conditions show that N-doped ACP NPs, with half absolute N-content than in conventional urea treatment, promote the formation of an equivalent amount of root and shoot biomass, without nitrogen depletion. The high nitrogen use efficiency (up to 69%) and a cost-effective preparation method support the sustainable real usage of N-doped ACP as a nanofertilizer.

2000 1500 1000 500 Wavenumber (cm -1 ) Absorbance (a. u.) U20-ACP-2x U20-ACP-2.5x U20-ACP-3.75x U20-ACP Figure S2. FTIR spectra (a) and XRD patterns (b) of the urea-doped materials prepared with technical grade reagents and nominal urea content of 20, 15 and 13.5 w%: TG-U20-NACP (green), TG-U15-NACP (magenta) and TG-U13.5-NACP (blue) materials. All spectra show the typical bands related to ACP, as well as the characteristics bands of nitrate ions (1384 cm -1 ) and urea (1668 cm -1 and 1465 cm -1 ). Only TG-U13.5-NACP is entirely amorphous, the other two samples showing co-precipitation of crystalline urea. Figure S3. XRPD patterns of the ACP materials, post-synthetically doped with nitrate ions: nitrate content: 1 % (blue line), 2 % (orange) 5 % (green) and 10 % (brown). The XRPD pattern of pure KNO3 is shown in black. Characteristics Bragg peaks of KNO3 can be barely appreciated when a 2 % of nitrate is added, being more evident at the 5 and 10 % levels.  Figure S4. SAXS data and relative fit for the NACP-1x sample, measured prior to urea postsynthetic functionalization. The log-log plot shows the intensity vs. Q dependence, where Q = 4sin/, 2 being the scattering angle,  the wavelength of the X-ray radiation used (here,  = 1.5418 Å). Data were fitted by a population of spherical nanoparticles following a number-based lognormal distribution of diameters with average diameter of 6.8 nm and relative polydispersity of 63%. Figure S5. Urea and calcium release profiles from Nano-U-ACP under suspension in water at room temperature (Urea, red circles, Calcium ions, blue circles). Taking into account that, in the presence of some complexing citrate generated by nanoparticle dissolution, the concentration the "truly free" Ca 2+ ions values may be somewhat underestimated, a slightly larger overall cumulative release of phosphate is favored.

S1. Assessment of the phase composition of technical grade reagents by XRPD:
Before any synthesis using technical grade (TG) reagents was conducted, their purity was assessed (Step 3 of Scheme 2 in the main text) by XRPD on data measured in the 5-85° angular range (see Figures S7, S8 and S9). For calcium chloride, due to its elevated hygroscopicity, XRPD data were collected in a shorter range (13-36º 5 Results are provided in the insets of Figure S8. The TG reagent "calcium nitrate" contained a complex mixture of hydrates and its purity was gravimetrically determined. 2.510 g were heated in an oven from room temperature to 300 °C with a 10 º min -1 ramp and kept at 300 °C for 2 h, until anhydrous Ca(NO3) was obtained, whose phase purity was certified by XRPD (ICSD code: 52351). The water content was then estimated from the measured weight loss.

S2. Optimization of synthesis and post-synthetic N-doping of ACP nanoparticles prepared with technical grade reagents and tap water.
Once the best conditions for the preparation of Nitrogen-doped ACP nanomaterials have been determined, we studied the feasibility of carrying out the same procedure by using technical grade reagents. As these inexpensive technical grade reagents may have suffered uncontrolled storage conditions and can contain non-negligible amounts of materials with a different composition than the nominal one, we firstly assessed their phase purity, or their phase composition by XRPD (Figures S7-S9). The chemical formula and composition calculated for each starting material are shown in Table S3. The purity of urea, sodium citrate dihydrate and sodium carbonate compounds was easily confirmed ( Figure S7). On the other hand, different hydrated forms or presence of impurities were identified in the other salts. The quantification S15 of the different phases ( Figure S8), present in the potassium phosphate dibasic, potassium nitrate and calcium chloride raw materials, provided the chemical compositions illustrated in Table S3. Finally, as the XRPD pattern of calcium nitrate showed a complex mixture of different hydrates, its water content was determined gravimetrically, after applying a thermal treatment leading a pure anhydrous crystalline phase ( Figure S9).
After the correct compositions of the different technical grade reagents were determined, these raw ingredients were used to prepare new doped-ACP materials by means of the postsynthetic doping approach (labelled as TG-Ua-NACP powders, see Scheme 1). Upon adopting the optimized conditions from the analytical grade reagents tests (U20-NACP-2x), TG-U20-NACP exhibited a small co-precipitation of crystalline urea. Reducing the concentration of urea to 13.5 wt% enabled to obtain an entirely amorphous material, both nitrate ions (visible in the IR spectrum by the characteristic peak at 1384 cm -1 ) and urea molecules (IR bands falling at 1668 cm -1 and 1465 cm -1 , Figure S2). The whole N content determined by elemental analysis for TG-U13.5-NACP was of 6.8 wt%, with a mere 1.3 wt% deviation from the reference U20-NACP-2x material.
Considering that all urea employed during the preparation (13.5%), is incorporated/adsorbed in/on the nanoparticles, the content of nitrate can be estimated around 2.2 wt%, a value not too distant from that reported by us for the one-pot doping approach (2.5 wt%). a Nevertheless, the efficiency of nitrate doping (calculated as the relation between the mass of nitrate in the material and total mass of nitrate added) is only of 3.42 wt %, being 96.58 % of added nitrate lost during the washing process.
Prompted by these promising results with urea and with the aim to increase the efficiency of nitrate ions doping, new attempts were also performed to incorporate NO3within ACP by means of a post-synthetic functionalization process. However, XRPD of the nitrate-doped hybrid materials shows that the usage of a mass fraction of potassium nitrate higher than 2% leads to the undesired co-precipitation of crystalline KNO3 ( Figure S3). These findings highlight the higher affinity of urea than nitrate, toward ACP nanoparticle decoration.

S3. SAXS experiment
A small aliquot of the NACP-1x sample (used as the amorphous calcium phosphate material for the post-synthetic urea doping) was analyzed by Small Angle X-ray Scattering (SAXS) for particle size and shape determination. The SAXS measurement was performed at Aarhus University using an in-house SAXS instrument 6 equipped with a rotating Cu anode source, side-by-side Montel multilayer mirrors for beam monochromatization and focusing, and a S16 Vantec 500 (Bruker AXS) detector. The collimation system consists of two-pinholes, where the one close to the sample is a scatterless pinhole with edges of Ge crystal. [7,8] The powdered sample was dispersed on acetate Scotch tape and mounted in the beam in the integrated vacuum of the SAXS instrument. A piece of the same tape was measured and subtracted as background.
The SUPERSAXS program package (in-house developed -C.L.P. Oliveira and J.S. Pedersen, unpublished) was used for the data reduction.

S4. DSE modelling of SAXS data
The NACP-1x SAXS signal was analyzed through a total scattering method based on the Debye scattering equation (DSE) using the DebUsSy Suite 9 (code availability. The DebUsSy program suite (freely available at http://debussy.sourceforge.net) performs calculations using atomistic models of nanoparticles (NPs) and a bottom-up approach to generate populations of NPs of different shapes and increasing sizes. DSE enables the computation of the total X-ray scattering pattern of randomly oriented and non-interacting particles from the set of interatomic distances calculated from the atomistic model. [9][10][11] Firstly, a population of NPs of spherical shape with diameters up to 20 nm was built. Pseudo-multiplicities vs equi-spaced pair distances were calculated and encoded in databases, according to a Gaussian sampling of interatomic distances to speed up calculations. 12 The calculated pattern is then optimized (monitoring the goodness of fit -GoF = square root of reduced  2 ) against the experimental data using a simplex minimization algorithm by adjusting the size parameters, namely the (number-based) average diameter and standard deviation of a univariate lognormal distribution function. This modelling approach, typically applied to Wide Angle X-ray Total Scattering data of nanocrystalline materials, can be effectively applied in the small angle region, provided that interparticle interactions can be reasonably neglected. [13][14][15] In the specific case of NACP-1x data, the DSE analysis was performed by assuming that i) in the Q range typical of SAXS experiments, the total scattering pattern is not affected by atomic-scale features; ii) the atomistic model of NPs with the crystal structure of hydroxyapatite (hexagonal crystal structure P63/m) 16 can reasonably account for the average electron density of ACP, as a suitable structural model for ACP does not exist.
The DSE best fit (GoF =3.158, Figure S4) suggests that primary particles of the NACP-1x sample exhibit spherical shape with average diameter of 6.8 nm and a size dispersion of 63% (lognormal number-based distribution). The corresponding mass-based values are 17.8 nm (average diameter) and 53% (size dispersion).