Analytical grade sulfuric acid (98%), hydrogen peroxide (50%), Nessler polyvinyl alcohol, Kjeldahl Nitrogen indicator and 12N potassium hydroxide were obtained from Hach. Lactophenol blue was obtained from Merck. Glucose Sabouraud agar (4%) was obtained from Becton Dickinson GmbH. Diammonium phosphate (DAP) with a composition of 18% nitrogen, 46% phosphorus and 0% potassium was obtained from Delcorp.

Residual cellulose/polyvinyl alcohol hydrogel was manufactured at Universidad EIA (Table 1). Sieved and washed sand (2 mm) and Andisol (Fulvudand), were provided by SOBIOTECH (Table 2) and characterized by Grupo Interdisciplinario de Estudios Moleculares-GIEM at Universidad de Antioquia and Laboratorio de Suelos at Universidad Nacional de Colombia, in Medellin, respectively. Zea mays seeds (Poaceae) 96 % pure and 80 % Germ were kept away from light, at a relative humidity of 30 % and a temperature of 4 °C, to inhibit germination. The seeds’ surfaces were disinfected with water and 70 % alcohol for 2 min. Then, they were immersed in 0.05 % sodium hypochlorite for 5 min, under constant stirring. After disinfection, the seeds were washed with sterile water and laid over wet sterile paper sheets for two days to complete germination. Germinated seeds were placed in plastic cylindrical holders (11 cm x 9 cm) with 200 g andisol type soil and thinned after 15 days to limit the number of plants to one per holder.

Table 1

Source: The authors

Table 2

Source: The authors

DR/200 UV/Vis spectrophotometer (Hach). 400-900nm. Tungsten halogen lamp. Spectral band: 7nm. Wavelength: ± 2 nm. Resolution: 1nm. Photometric linearity: ± 0.002 A. Photometric reproducibility: ± 0.005 A.

Digesdahl Digestor (Hach). Model 23130-20. Temperature range: 100 °C - 400 °C. Temperature control ± 15 °C. Flushing capacity: 11.5 L/min at a water flow rate of 6.5 L/min. Minimum pressure: 51.7 kPa.

Phenom Scanning electron microscope. Optical zoom: 20x -135x. SEM zoom 80x - 130000x. Sample size 32 nm. Acceleration voltage: 5 kV - 10 kV.

Olympus optical microscope CX21. 10x/18 ocular. Binocular tube inclined at 30°. Interpupilar distance range: 45 mm - 75 mm. Rotatable quadruple lense holder. 4x, 10x, 40x and 100x lenses. Halogen lamp 6V/20W. Fixed mechanical slide with cable movement: 120 x 132mm. Range of displacement: 76 mm(X) by 30 mm(Y).

Fig. 1 shows that, from day five on, hydrogel loses up to 10% of its structure. This reduction in mass is related to the erosion of residual cellulose and polyvinyl alcohol chains that failed to crosslink during hydrogel formation; this behavior indicates low disintegration during the evaluation period.

Figure 1 Source: The authors..

3.2. Release of NH _⁴⁺ by leaching

Fig. 2 shows the release profiles for NH₄ ⁺ (diammonium phosphate and hydrogel) from day 0 to day 60 of the evaluation period.

Figure 2 Source: The authors

During the first five days of evaluation, diammonium phosphate released 43 ± 2 % and 34 ± 2 % of NH₄ ⁺ in soil and sand matrices, respectively. The high solubility of diammonium phosphate in water (57.5 g / 100 mL at 10ºC) produced this rapid release (called Tigger effect). On day 30, NH₄ ⁺ release reached levels of 58 ± 2 % and 48 ± 1 %; these data are similar to those found on day 60 -this indicated that migration of NH₄ ⁺ from the fertilizer to the soil matrix had reached equilibrium. During the evaluation period, larger concentrations of NH₄ ⁺ were found in solutions obtained from soil (pH 4.8), than in those obtained from sand (pH 7.2); this was to be expected as pH was an essential factor in determining the solubility of nutrients in soil, where NH₄ ⁺, NO₃ ^-, CO₃ ^2- and SiO₃ ^2- displayed a larger displacement of oxidation equilibria (nitrification and denitrification), and also dissociation of NH₄ ⁺ ions (eq. 1 and 2) to the left side of the equation, increasing the concentration of NH₄ ⁺ ions in the soil solution and favoringleaching [11]. In addition, the soil exchange complex displays a lower pH, which increases protonation at the points of exchange; this translates into positive charging of the colloidal mycelia, avoiding adsorption of NH₄ ⁺ on the surface.

The different media did not exhibit statistical differences when hydrogel was used. Release of NH₄ ⁺ after days 5, 30 and 60 was 8 ± 2 %, 11 ± 1% and 12 ± 2%, respectively. The first stage of release from the hydrogel can be attributed to the erosion of residual cellulose and polyvinyl alcohol chains as mentioned before [12]; structural loss also plays a role. According to the European Standardization Committee [13], a fertilizer is called “a controlled release fertilizer” when:

It releases no more than 15% after 24 h.

Once NH₄ ⁺ finds its way into the substrate, water can wash it out via pore saturation, or it can remain available to plants and microorganisms [14]. Although the release profile in water does not represent the real span in which nutrients exit soil, it allows for comparison between commercial fertilizers and the hydrogel. In this way, we can test the capacity of our hydrogel to release NH₄ ⁺.

To analyze NH₄ ⁺ release data, we used the Korsmeyer-Peppas empirical model [15] (eq 3), for polyvinyl alcohol hydrogel controlled-release systems in soil [16,17].

In eq 3, is the fraction of drug or fertilizer used by time t, K is the constant that marks the speed of release and n is the release exponent.

To estimate n and K values, eq. 3 was linearized

Fig. 3 shows log (% NH₄ ⁺) versus time, Table 3 shows the statistical results of NH₄ ⁺ release data linearization. The data exhibits significant statistical correlation (p-value < 0.05) between log (NH₄ ⁺) and log (t), with a 95% confidence level; the correlation coefficients for the models used were larger than 0.99 -this indicates a strong correlation between variables.

Figure 3 Source: The authors..

Table 3

Source: The Authors.

From the equations shown in Table 3, the following equation was obtained:

Assuming :

Using the antilog function on both sides of the equation:

By representing as , then

Since k is constant then must also be a constant:

In this way, the leaching of NH₄ ⁺ in water adjusts to the Korsmeyer-Peppas model [18](eq. 3), which adequately describes the leaching process. This model explains the NH4+ release mechanisms for residual-cellulose/polyvinyl-alcohol hydrogel; the derived n and K values can be seen on Table 4.

Table 4

Source: The authors.

For polyvinyl alcoholhydrogels with different degrees of hydrolysis, a Fickian diffusion mechanism (n = 0.45) was proposed. For hydrogels with chitosan, xanthan and starch contents the mechanism proposed is not Fickian (0.45 < n < 0.89) as hydrogels disintegrated easily.

Hydrogel shows no significantly different n values for treatments in soil and sand, but values of n < 0.45 indicate a lower rate for water infiltration into the hydrogel than the rate at which its polymeric chains can relax. This phenomenon has been attributed to the high content of calcium carbonate in the hydrogel’s structure; even so, the process remains Fickian -such behavior is called “Less Fickian” [19,20].

Fig. 4 shows the control (left), diammonium phosphate-treated (center) and hydrogel-treated (right) samples used in the greenhouse tests; it is evident that plants in the right-hand side column are taller and display broader leaves.

Figure 4 Source: The authors.

Table 5 shows the aver age heights of the hydrogel-treated plants and the diammonium phosphate-treated plants; p > 0.05 suggests there are no significant differences in the treatments. However, both treatments are significantly different from the control group. The mass of the plants grown in hydrogel-fertilized soil was larger than that of diammonium phosphate-fertilized plants; these findings show the positive effects of hydrogel-fertilized soil on the bioindicator. The results shown confirm that nitrogen fertilizers can be optimized, averaging lower production costs and reducing environmental impact produced by traditional fertilization processes [21].

Table 5

Source: The authors.

Hydrogel produced the largest release of NH₄ ⁺. When hydrogel interacts with solid fractions in soil (organic and inorganic) it releases NH₄ ⁺, but the rate at which that release takes place is dependent on the degradation of the cellulose matrix. Degradation was verified on day 60 using scanning electron microscopy (SEM). Fig. 5b shows the presence of a microorganism growing on the surface of the hydrogel; given the characteristics of the material, microorganisms can exhibit cellular activity and contribute to its degradation.

Figure 5 Source: The authors.

Fig. 6 shows a filamentous fungus with cylindrical hyphae, spherical structures, ascomata (flask-like, some already broken due to the release of ascospores to the media) and conical-shaped porous hyphae (characteristic of the ascomycota) [22].

Figure 6 Source: The Authors.

After 60 days, we isolated (in 4 % Glucose Sabouraud agar) from hydrogel-fertilized soil, a fungus with the same characteristics as those found on the hydrogel’s surface. None of the tests performed revealed the presence of enterobacteria or E. coli (both regulated by Colombian legislation and the National Health Institute).

On the sixth day of isolation (Fig. 7a), an unlimited number of white cottony colonies was discovered throughout the Petri dish, all of which exhibited radial growth. On the backside of the plate, a diffused brown shade appeared (Fig. 7b). After ten days of isolation (Fig. 7c) the cottony colonies, once white, turned purplish, and then covered the entire agar, adhering to the walls. Meanwhile, the brown shade on the back of the plate turned pink (Fig. 7d).

Figure 7 Source: The Authors.

Table 6

Source: The Authors.

Fig. 8 shows images (at 40x) of the culture plates in lactophenol blue. On the images we observe the fungus’ mycelium, composed of thin septate hyaline hyphae (Fig. 8a), lateral phialides (Fig. 8b), oval and ellipsoidal microconidia and ascospores (Figs. 8c and 8d) [22]. The macroscopic and microscopic taxonomic characters identified in the fungal isolate match those of Fusarium spp.

Figure 8 Source: The Authors.