By-products of the thermoelectric industry and applied mixtures

The experiment was realized in the Laboratory of Environmental Edafology of the Department of Soils and Natural Resources, Faculty of Agronomy, Universidad de Concepción, Chillán.

Treatments consisted in one control (soil without any addition) and the following mixtures of synthetic gypsum, coal ash and urea: M₁: 50% synthetic gypsum, 50% coal ash, M₂: 50% synthetic gypsum, 35% coal ash, 15% urea, M₃: 45% synthetic gypsum; 40 coal ash, 15% urea, M₄: 65% synthetic gypsum, 20% coal ash, 15% urea, M₅: 55% synthetic gypsum, 30% coal ash, 15% urea (Table 1).

Table 1 Amount of synthetic gypsum, coal ash, and urea used to make different mixtures applied to an Alfisol before incubation trials. 

Synthetic gypsum and coal ash were provided by thermoelectric industry Guacolda located at Huasco (28°28.08’33” S, 71°13’08.18” W), Chile. Chemical characterization of FGD gypsum and coal ash (Table 2) was performed according to the methodology described by ^{Sadzawka et al. (2006)}. Total N and OM were measured by digestion, while the contents of P, K, Mg and Ca were determined by atomic absorption spectrometry (Perkin-Elmer spectrometer, model 1100B, Phoenix, Arizona, USA).

Table 2 Chemical characterization of synthetic gypsum and coal ash used to make different mixtures applied as treatments. 

nd: Not determined.

Soil used for incubation trials

The soil corresponds to order Alfisol (Fine, halloysitic, mesic, Aquic Palexeralfs) (^{Stolpe, 2006} ), samples were collected at a 0-20 cm depth from San Ignacio (36°46’51.01” S, 72°00’29.62” W). The management and preparation of the soil samples prior the establishment of the experiment, as well as the determination of the bulk density of the soil were performed based on ^{Sandoval et al. (2012)}.

The chemical soil characterization (Table 3) was performed according to the methodology described by ^{Sadzawka et al. (2006)}. Water pH and OM were measured, while concentrations of macronutrients (P-Olsen, K, Ca, Mg, Na and S) and micronutrients (Fe, Mn, Zn, Cu and B) were also determined. The soil has a clay loam texture and a bulk density of 0.85 g cm^-3.

Table 3 Initial chemical characterization of the Alfisol used in this experiment to determine N mineralization. 

MO: Organic matter, CE: electrical conductivity.

Assessment of net N mineralization

The effect of the different mixtures was investigated by a completely randomized experimental design with three replicates per treatment. Treatments consisted of a dose equivalent to 2 t ha^-1 each mixture (M₁, M₂, M₃, M₄, M₅), in addition to a control (soil without amendment). Eighteen experimental units consisting of 200 g of previously sieved soil were prepared in polypropylene bags. Once treatments were applied, experimental units were placed under controlled conditions in an incubator (D-3162, Uetze-Hänigsen, Germany) maintaining soil moisture at 80% field capacity and 25 °C. Five evaluation times corresponding to 0, 2, 4, 6 and 8 wk were used.

The concentrations of NO₃ ^- and NH₄ ⁺ were determined by colorimetry with salicylic acid and Nessler reagent, respectively (^{Keeney and Nelson, 1982}) and the entire mineral N by the sum of both N forms for each incubation period. Net mineral N concentration (N_MN) was calculated by the difference between total inorganic N concentrations of each evaluated time minus the total concentration determined at time 0.

Potential N mineralization and mineralization rate were determined using the non-linear method of average squares. This regression analysis assumes that N mineralization is a first-order reaction (^{Hirzel et al., 2010}), using the following model:

^{_{Nmt = N0 [1-℮(-k} t _)]}

where ^_Nmt is the mineral N accumulated at a specific time (mg kg^-1), ^_N0 is the potentially mineralizable N (mg kg^-1), k is the mineralization rate and finally t is the incubation time (wk).

Statistical analysis

The statistical analysis of data obtained in terms of N mineralization was performed with ANOVA, while the effect of means was analyzed by Tukey’s test with a confidence level of 95% (α = 0.05). Potentially mineralizable N (N₀) and mineralization rate (k) were determined using the Gauss-Newton method for nonlinear least squares. Data were analyzed using SAS software (SAS Institute, Cary, North Carolina, USA).

Nitrogen mineralization

The concentrations of net mineral N (N_MN) determined in the Alfisol had significant differences (p < 0.05) respect to control treatment (Table 4). Throughout the test, N_MN concentrations in the Control and M₁ treatments differed (p < 0.05) from the N_MN concentrations presented by treatments M₂, M₃, M₄ and M₅. Differences were only determined between the Control and M₁ treatments (p < 0.05) in the last evaluation. Increasing N_MN in soils without addition of organic amendments or N fertilizer has been reported by ^{Masunga et al. (2016)}. After 2 wk incubation, treatments M₂, M₃, M₄ and M₅ showed no differences (p > 0.05) respect to N_MN concentration. During week 4 N_MN concentration generated by treatments M₄ and M₅ did not show any significant differences (p > 0.05), but they differed (p < 0.05) from treatment M₂. The highest N_MN increase occurred at week 6, when treatment M₄ generated the highest N_MN concentration (p < 0.05). Regarding M₂, M₃ and M₅, these treatments did not present significant differences among them (p > 0.05). However, at week 8, treatment M₃ shown the highest concentration of N_MN (p < 0.05) with respect to all mixtures with urea in its formulation. Finally, at week 8 there were differences (p < 0.05) in N_MN concentration between treatments M₄ and M₅. The behavior of treatments M₂, M₃, M₄ and M₅ respond to the observations of ^{Nave et al. (2009)}, who found increments of N mineralization up to 62% after N supply in forest soils.

Table 4 Net concentrations of mineral N (N_MN) determined at each incubation time in an Alfisol. 

Mixture 1: 50% synthetic gypsum; 50% coal ash. Mixture 2: 50% synthetic gypsum; 35% coal ash; 15% urea. Mixture 3: 45% synthetic gypsum; 40% coal ash; 15% urea. Mixture 4: 65% synthetic gypsum; 20% coal ash; 15% urea. Mixture 5: 55% synthetic gypsum; 30% coal ash; 15% urea.

CV: Coefficient of variation, LSD: least significant difference by Tukey’s test.

Different letters in a same column means significant differences between treatments (p < 0.05).

Respect N_MN dynamics presented by the Alfisol (Figure 1), control treatment and M₁ generated a constant increase in N_MN concentration during all incubation time. With exception of Control and M₅ treatment, all treatments had a slight decrease in N_MN concentration during week 4. Later in week 6, an increase in N_MN was observed in all treatments, an increase that was maintained in treatments M₁, M₂ and M₃ until that of week 8 evaluations. Treatments Control, M₄ and M₅ presented a decrease in N_MN concentration in the last week of incubation generating differences between them (p < 0.05). The behavior of N_MN concentration observed in Figure 1 is similar to that reported by ^{Laos et al. (2000)}, ^{Mohanty et al. (2013)} and ^{San Martín et al. (2016)}, who observed a steady increase in N_MN concentration with mild lows between weeks. These results also were similar in all treatments to the results of ^{Masunga et al. (2016)} using different amendments used in agriculture from animal and green origin.

Figure 1 Net concentration of total mineral N determined at each incubation time in an Alfisol. 

Due to the null contribution of N from the coal ash and FGD gypsum, the increase in N_MN concentration is attributed to the mineralization of OM present in the soil and the mineralization of urea in treatments M₂, M₃, M₄ and M₅. A soil N concentration greater than 60 mg kg^-1 is considered high. Based on this, Alfisol showed high concentrations of N available throughout the incubation time after applying treatments M₂, M₃, M₄ and M₅, whereas M₁ only had a high concentration at week 8. In this study, N_MN concentrations were higher than those reported by ^{San Martín et al. (2016)}, when applying pelleted waste from the paper industry with and without addition of Ulva lactuca L. in an Entisol.

Potentially mineralizable N

The N₀ and k determined for each treatment applied and their respective nonlinear equations, together with the mineral N concentration estimated by the first order nonlinear model are shown in Table 5 and Figure 2, respectively.

Table 5 Potentially mineralizable N (N₀), rate of mineralization (k) and non-linear model fitted on an Alfisol. 

Mixture 1: 50% synthetic gypsum; 50% coal ash. Mixture 2: 50% synthetic gypsum; 35% coal ash; 15% urea. Mixture 3: 45% synthetic gypsum; 40% coal ash; 15% urea. Mixture 4: 65% synthetic gypsum; 20% coal ash; 15% urea. Mixture 5: 55% synthetic gypsum; 30% coal ash; 15% urea.

Nm_t: Accumulated mineral N at a specific time (mg kg^-1).

Figure 2 Accumulated mineral N concentration for each evaluation time estimated by first order nonlinear model for each treatment applied in an Alfisol. 

Values of N₀ refer to the amount of organic N, which can be converted to inorganic forms soluble by microbial biomass activity (^{Ros et al., 2011}). In this respect M₁, which does not contain urea, presented the highest value of N₀. However, it is observed that, when applying the N₀ and k values determined in the nonlinear model, Control and M₁ treatments show the lowest concentration of potentially mineralizable N (Figure 2). On the other hand, mixtures with urea in its formulation, M₃ presented a N₀ of 523 mg N kg^-1 (Table 5). The values of the N₀ in all treatments applied to Alfisol were higher than those reported by ^{Hernández et al. (2002)}, ^{Mohanty et al. (2013)}, and ^{San Martín (2016)}. Because N₀ values are calculated using net mineralization values, according to ^{Guntiñas et al. (2012)} it must be considered that N₀ values will not depend only on the applied treatment, it also affects the type of use of the soil, moisture and temperature. Also must be considerate in future investigations that N₀ values determined under laboratory conditions could differ from those determined in field conditions, because urea fertilization decreased the decomposition of plant residues and SOM in soils with growing plants (^{Li et al., 2017}).

Regarding k, values determined in control treatment, M₁ and M₃ were lower than those reported by ^{Hernández et al. (2002)}, who determined k values of 0.6 and 1.14 when performing incubations of clay aggregate of 50 and 30 g kg^-1 soil of urban sludge, respectively. Likewise, the values k for the Control and M₁ treatment are lower than those determined by ^{San Martín et al. (2016)} in cellulose by-products without addition of U. lactuca. In mixtures with urea, k values of the treatments M₂, M₄ and M₅ are within the range of 0.42 and 0.77 determined by ^{San Martín et al. (2016)}, however the M₃ treatment presented a value lower than this range. According to this, k values determined in Control treatment respond to its natural capacity of N release from its OM content, this way our results for Control treatment are higher than those reported by ^{Dossa et al. (2009)} in unamended sandy arid soils with low OM content. Nevertheless, with little exception of M₁, all treatments shown k values higher than those reported by ^{Mungai and Motavalli (2006)} for green wastes. Moreover, in all our treatments without exclusion the results were higher than those reported by ^{Gil et al. (2011)} with k values of 0.002 and 0.009, when modelling N mineralization of bovine manure and sewage sludge compost in sandy clay loam soils.

Finally, a 32 wk projection of potentially mineralizable N is presented using the N₀ and k values determined based on the results of the net mineralization test (Figure 3). The response observed in treatments M₁ and M₃ (Figure 3) is attributed to the high value of N₀ and the low value of k determined by the Gauss-Newton method.

Figure 3 Estimated projection of accumulated mineral N concentration over 32 wk using a first-order nonlinear model for each treatment applied in an Alfisol.