*Article* **Technologies for Fertilizers and Management Strategies of N-Fertilization in Coffee Cropping Systems to Reduce Ammonia Losses by Volatilization**

**Tainah Freitas 1, Lucas Bartelega 2, César Santos 2, Mateus Portes Dutra 2, Leonardo Fernandes Sarkis 2, Rubens José Guimarães 1, Anderson William Dominghetti 3, Pauliana Cristina Zito 1, Tales Jesus Fernandes <sup>4</sup> and Douglas Guelfi 2,\***


**Abstract:** The aim of this study was to quantify NH3-N losses from conventional, stabilized, slowrelease, and controlled-release N fertilizers in a coffee field. The N fertilizers analyzed were prilled urea, prilled urea dissolved in water, ammonium sulfate (AS), ammonium nitrate (AN), urea + Cu + B, urea + adhesive + CaCO3, and urea + NBPT (all with three split applications), as well as blended N fertilizer, urea + elastic resin, urea-formaldehyde, and urea + polyurethane (all applied only once). NH3-N losses (mean of two crop seasons) were statistically higher for urea + adhesive + CaCO3 (27.9% of applied N) in comparison with the other treatments. Loss from prilled urea (23.7%) was less than from urea + adhesive + CaCO3. Losses from urea + NBPT (14.5%) and urea + Cu + B (13.5%) were similar and lower than those from prilled urea. Urea dissolved in water (4.2%) had even lower losses than those treatments, and the lowest losses were observed for AS (0.6%) and AN (0.5%). For the single application fertilizers, higher losses occurred for urea + elastic resin (5.8%), blended N fertilizer (5.5%), and urea + polyurethane (5.2%); and urea-formaldehyde had a lower loss (0.5%). Except for urea + adhesive + CaCO3, all N-fertilizer technologies reduced NH3-N losses compared to prilled urea.

**Keywords:** N-fertilizers; NH3 emission; urease inhibitors; slow- and controlled-release N-fertilizers; *Coffea arabica*; sustainable agriculture

#### **1. Introduction**

Brazil is the largest coffee (*Coffea arabica* L.) producer and exporter worldwide, and the constant search for better beverage quality and sustainability is essential in different coffee production systems. The application of nitrogen (N) fertilizers is imperative to achieve an adequate yield of this high-value crop. Nitrogen is the nutrient most extracted by the coffee plant and the nutrient of second greatest export by coffee beans [1,2].

Some studies using 15N have shown that coffee plants take up less than 25% of the N fertilizer when applied as conventional urea [3,4]. The dynamic transformation of N forms in the soil and the varying pathways of N losses in coffee growing areas result in low N fertilizer use efficiency (NUE) [4]. Ammonia (NH3-N) volatilization is the primary N loss in coffee production areas in Brazil [2,5,6], particularly when conventional urea is applied on the soil surface with plant residues and without fertilizer incorporation [7]. This is a common practice in systems of perennial crops such as coffee.

In 2017, the amount of N fertilizers used in the world was estimated at 150 Tg N per year [8], and may reach 260 Tg N per year in 2050 [9]. About 50% of global N fertilizer production is represented by urea [10,11]. The NH3-N losses from urea can be intensified

**Citation:** Freitas, T.; Bartelega, L.; Santos, C.; Dutra, M.P.; Sarkis, L.F.; Guimarães, R.J.; Dominghetti, A.W.; Zito, P.C.; Fernandes, T.J.; Guelfi, D. Technologies for Fertilizers and Management Strategies of N-Fertilization in Coffee Cropping Systems to Reduce Ammonia Losses by Volatilization. *Plants* **2022**, *11*, 3323. https://doi.org/10.3390/ plants11233323

Academic Editor: Przemysław Barłóg

Received: 26 September 2022 Accepted: 20 October 2022 Published: 1 December 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

when specific soil properties are combined with climatic conditions favorable to this loss. Such properties and conditions include the application of urea on moist soil followed by an absence of rainfall, increased soil and air temperature, increased N doses, application of N on soils with low cation exchange capacity, and alkaline soil pH [12,13]. NH3-N losses can exceed 50% of the applied dose, considering N applications in multiple crops [14,15]. The equivalent of one in three applications of N fertilizer is lost by volatilization in coffee production systems using conventional urea as an N source [2,5,16].

NH3-N losses not only reduce NUE and cause agronomic damage, but also lead to environmental problems. These problems include air pollution, due to the acidifying nature of NH3 [17], and greenhouse gas emissions to the atmosphere. NH3 gas is an indirect source of nitrous oxide (N2O), which has a global warming potential 310 times greater than carbon dioxide (CO2) [18]. It is estimated that 1.4% of the total volatilized N is converted to or lost as N2O [19].

The 4R principles (right nutrient source, right rate, right time, and right place) guide various management practices to minimize nutrient losses and the C footprint and increase N retention in the soil [20]. The development and proper use of enhanced-efficiency fertilizers (EEFs) may reduce these N losses [20–22]. Technological development of fertilizers is currently one of the strategies most investigated for improving NUE [23–25].

The fertilizer technology market had a compound annual growth rate estimated at 12% from 2014 to 2020 [26]. In addition, some European countries, including Germany and the Netherlands, have already adopted measures banning the application of conventional urea without incorporation and encouraging the use of some technologically enhanced fertilizer.

Enhanced efficiency fertilizers are in four main categories: stabilized fertilizers, slowrelease fertilizers, controlled-release fertilizers, and their blends [27]. Stabilized fertilizers can inhibit some stages of N transformation in the soil through additives such as urease or nitrification inhibitors (e.g., N-(n-butyl) thiophosphoric triamide—NBPT) [28,29]. Some chemical compounds, such as boric acid, and metallic ions, such as copper (Cu), may also function as urease inhibitors when used in adequate concentrations [28,30].

The slow-release or chemically modified fertilizers are products of condensation of urea with aldehydes (e.g., formaldehyde and acetaldehyde). Controlled-release fertilizers are those with coatings that control the release of nutrients by diffusion or by a physical barrier (e.g., sulfur, wax, and polymer) [24,31]. In addition to these technologies, combining enhanced efficiency fertilizers and conventional N sources gives rise to another category of fertilizers, what are known as blends. N fertilizer blends are produced from the physical mixture of different fertilizer technologies with conventional N sources (stabilized, slow-release, or controlled-release fertilizers). Combining these N sources has many advantages, including a reduction in production costs compared to separate application of slowor controlled-release fertilizers, the optimization of dynamics between nutrient release and plant nutrient uptake, and a reduction in NH3-N losses compared to conventional urea [32,33].

In addition to the different technologies available on the market, diverse N fertilization application strategies can be used, such as the mechanical incorporation of fertilizers into the soil. However, incorporation cannot be used in some cropping systems. In coffee fields, for example, mechanical incorporation of fertilizers may damage the root system, whose greatest activity occurs in the first 0.30 m of the soil [34]. Thus, urea dissolved in water, applied via a jet directed to the soil, can be a promising alternative, since urea would be automatically incorporated.

In this study, we tested the hypothesis that enhanced efficiency N fertilizers and other fertilizers, such as ammonium nitrate and sulfate and prilled urea diluted in water, are options more suitable than conventional urea to reduce NH3-N losses in coffee production systems. We chose the main technologies available on the market to perform this study. Thus, the objective of this study was to quantify NH3-N losses by volatilization from conventional, stabilized, slow-release, and controlled-release N fertilizers, as well as from

a fertilizer blend, that were applied for two crop seasons on a coffee growing area in the production stage.

#### **2. Results**

#### *2.1. Weather Conditions*

The accumulated precipitation in the fertilization period of the 1st year was: 337 mm, 289 mm, and 69 mm at the 1st, 2nd, and 3rd split fertilizations, respectively, totaling 694 mm. In the first seven days after each fertilizer application, these rainfall values corresponded to 78 mm (23%), 171 mm (59%), and 19 mm (27%) (Figure S1). In the 2nd year, the rainfall accumulated during the fertilization period was: 153 mm, 124 mm, and 178 mm at the 1st, 2nd, and 3rd split fertilizations, respectively, totaling 455 mm. In the first seven days after each fertilizer application, these rainfall values corresponded to 102 mm (67%), 64 mm (52%), and 40 mm (22%) (Figure S3). The mean annual rainfall over the two years of assessment was 1243.3 mm.

The relative air humidity was higher than the critical relative humidity of urea (75%) for most of the period after fertilization in the two crop seasons. The mean temperatures in the same period were 21.2 and 22.6 ◦C in the first and second crop seasons, respectively. The minimum temperatures were 19 and 18 ◦C, and the maximum temperatures were 28 and 25 ◦C. Between the years 2015 and 2017, January had the highest mean temperature (30 ◦C), and June had the lowest mean temperature of (14 ◦C). The potential evapotranspiration (ETP) was around 899 to 873 mm per year [35].

#### *2.2. Daily and Accumulated N-NH3 Losses*

In this study, the results were divided into two topics for a better understanding of the treatments and for a fair comparison among the N-sources. The first topic includes the results of fertilizers applied in three split applications, and the second topic describes fertilizers applied in a single application.

#### 2.2.1. Fertilizers Applied in Three Split Applications

The daily and accumulated losses of N-NH3 of the three fertilizer applications in each crop year were influenced (*p* ≤ 0.05) by the technologies for N-fertilizers. Except for urea+adhesive+CaCO3, all the technologies for N-fertilizers, ammonium nitrate, ammonium sulfate and stabilized fertilizers reduced N-NH3 losses compared to prilled urea. For the 2015/2016 crop season, the maximum losses or peaks of daily NH3 volatilization for prilled urea occurred 1.3 days after application (~5.8 kg N ha−1). For ammonium nitrate and ammonium sulfate, the maximum loss occurred 6.3 and 4.5 days after application, with values of 0.03 and 0.01 kg N ha−1, respectively. For fertilizers stabilized with Cu + B and NBPT, the maximum loss occurred at 4.9 and 2.6 days after application, with 1.0 and 2.5 kg N ha<sup>−</sup>1, respectively. Lastly, for urea dissolved in water and urea+adhesive+CaCO3, the maximum loss occurred at 1 and 1.15 days after application, with 1.5 and 1 kg N ha−<sup>1</sup> (Table 1). In the 2015/2016 crop season, the mean accumulated losses in the first seven days were 9.2, 6.7, and 13% of the applied N for the first, second, and third split applications, respectively (Figure S1). In the 2016/2017 crop season, the maximum loss for prilled urea occurred two days after application, with a value of 7.4 kg N ha<sup>−</sup>1. For ammonium nitrate and sulfate, the maximum loss was at 13 and 5 days after application, with values similar to those of the first crop season (lower than 0.2 kg N ha<sup>−</sup>1). Urea + Cu + B and Urea + NBPT showed maximum losses at 3.7 and 4.3 days, with values of 4.0 and 3.2 kg N ha−1, respectively. Urea dissolved in water and urea + adhesive + CaCO3 had maximum losses at 1.5 and 1.9 days after N fertilization, with values of 1.6 and 10.3 kg N ha−<sup>1</sup> (Table 2). In the 2016/2017 crop season, the mean general accumulated losses in the first seven days were 16, 8.5, and 11.7% for the first, second, and third split applications, respectively (Figure S3).


**Table 1.** Regression parameters adjusted for the accumulated and maximum daily losses of N-NH3 from conventional and stabilized N fertilizers in the 2015/2016 crop season.

α: Asymptotic value (percentage of estimated maximum accumulated loss); b: Day when the maximum ammonia loss occurs; k: relative index; MDL (maximum daily loss of ammonia) and NBPT: N-(n butyl) thiophosphoric triamide.

**Table 2.** Regression parameters adjusted for the accumulated and maximum daily losses of N-NH3 from conventional and stabilized N fertilizers in the 2016/2017 crop season.



**Table 2.** *Cont.*

α: Asymptotic value (percentage of estimated maximum accumulated loss); b: Day when the maximum ammonia loss occurs; k: relative index; MDL (maximum daily loss of ammonia) and NBPT: N-(n butyl) thiophosphoric triamide.

Regarding the accumulated N-NH3 losses in the 2015/2016 crop season, the mean value of losses was 10.6% of the applied N (average of the three split applications) (Table 3, Figure S5). For the treatments, the mean losses decreased as follows: Urea + adhesive + CaCO3 (25.5% of applied N) = prilled urea (23.2%) > urea + NBPT (13%) > urea + Cu + B (7.4%) > urea dissolved in water (4.5%) > ammonium sulfate (0.3%) = ammonium nitrate (0.2%). For the 2016/2017 crop season, the mean value was 13.7% (average of the three split applications). As for the treatments, the mean losses decreased in the following order: urea+adhesive+CaCO3 (30.3% of applied N) > prilled urea (24.2%) > urea + Cu + B (19.7%) > urea+NBPT (16%) > urea dissolved in water (4.5%) > ammonium sulfate (0.9%) = ammonium nitrate (0.8%) (Table 3, Figure S6).

**Table 3.** Mean accumulated losses of ammonia (% of applied N), for conventional and stabilized N fertilizers, in three fertilizations in the coffee plantation, during the 2015/2016 and 2016/2017 crop seasons.


NBPT: N-(n butyl) thiophosphoric triamide. Note: In each crop season, 300 kg N ha−<sup>1</sup> per year were split into three equal applications for conventional and stabilized N fertilizers, totaling 600 kg N ha−<sup>1</sup> for both crop seasons. Means followed by the same lowercase letter in the column do not differ by the Scott-Knott test (*p* ≤ 0.05). Mean of the six fertilization sources performed between November and February during both seasons \*\* (PCRDU) Percentage change decrease compared to Prilled Urea. \*\*\* Negative value indicates increased volatilization compared to prilled urea.

#### 2.2.2. Fertilizers Applied in a Single Application

In this section, the results of slow-release and controlled-release fertilizers and a blend will be presented. Here, fertilizers were applied in a single application, as they have the mechanism of gradual release of N to the soil. The results showed significant differences in N-NH3 losses by volatilization for both seasons. In the 2015/2016 crop season, the maximum loss occurred 35 days after the application for urea+elastic resin, with a mean value of 0.4 kg N ha−1. For the Blend N-fertilizer, at 24.7 days (0.19 kg N ha−1); ureaformaldehyde, at 7 days (0.08 kg N ha<sup>−</sup>1); urea+polyurethane, at 28.2 days (0.37 kg N ha−1). Regarding the 2016/2017 crop season, these N-sources behave similarly, with low values on the day of maximum loss. The urea+elastic resin treatment showed maximum loss at 40 days (~0.2 kg N ha<sup>−</sup>1); Blend N-fertilizer, at 9 days (~0.5 kg N ha−1); urea-formaldehyde, at 9 days (~0.04 kg N ha<sup>−</sup>1); and urea+polyurethane, at 31 days (0.22 kg N ha−1) (Table 4).

**Table 4.** Regression parameters adjusted for the accumulated and maximum daily losses of N-NH3 from slow-release and controlled-release N fertilizers in the 2015/2016 and 2016/2017 crop seasons.


α: Asymptotic value (percentage of estimated maximum accumulated loss); b: Day when the maximum ammonia loss occurs; k: relative index; MDL (maximum daily loss of ammonia) and NBPT: N-(n butyl) thiophosphoric triamide.

The losses accumulated by these sources in the 2015/2016 crop season were higher for the treatments: urea + polyurethane (6.4% of applied N) > urea + plastic resin (5.7%) = Blend N-fertilizer (4.6%) > urea-formaldehyde (0.6%) (Figure S2). In the 2016/2017 crop season, losses were higher for the Blend N-fertilizer (6.5% of applied N) = urea + plastic resin (5.9%) > urea + polyurethane (4%) > urea-formaldehyde (0.5%) (Table 5, Figure S4).

**Table 5.** Mean accumulated losses of ammonia (% of applied N), for slow-release and controlledrelease N fertilizers, in one single application in the coffee plantation, during the 2015/2016 and 2016/2017 crop seasons.


Note: In each crop season, 300 kg N ha−<sup>1</sup> per year were applied into one single application for slow and controlledrelease N fertilizers, totaling 600 kg N ha−<sup>1</sup> for both crop seasons. Means followed by the same lowercase letter in the column do not differ by the Scott-Knott test (*p* ≤ 0.05). \* Sum of precipitation during the evaluation periods, which were 208 and 235 days in the 2015/2016 and 2016/2017 crop seasons, respectively. \*\* Mean of the six fertilizations performed between November and February of each crop season/year. \*\*\* Mean of precipitation during the evaluation periods, which were 208 and 235 days in the 2015/2016 and 2016/2017 crop seasons.

#### *2.3. Summarizing Results of Ammonia Losses from N-Technologies*

Considering the way that the study was designed and conducted, it is not possible to compare the results of all technologies. However, a sequence of loss values presented by the sources can be established, considering the average of the two years of study. Thus, the decreasing order for the split treatments would be as follows: urea + adhesive + CaCO3 (27.9% of applied N = 84 kg N) > prilled urea (23.7% = 71 kg N) > urea + NBPT (14.5% = 43 kg N) = urea + Cu + B (13.5% = 40 kg N) > urea dissolved in water (4.2% = 12.6 kg N) > ammonium sulfate (0.6% = 1.8 kg N) = ammonium nitrate (0.5% = 1.5 kg of N). The decreasing order for the sources applied at a single time were: urea + elastic resin (5.8% = 17.4 kg N) = Blend Nfertilizer (5.5% = 16.6 kg N) = urea + polyurethane (5.2% = 15.6 kg N) > urea-formaldehyde (0.5% = 1.59 kg).

#### **3. Discussion**

In this study, the weather conditions greatly influenced N-NH3 losses by volatilization, particularly rainfall and temperature. In both coffee crop seasons, most N-NH3 losses occurred in the first seven days for the N-fertilizers applied in three split applications. The rainfall in these first days was essential for incorporating fertilizers into the soil and reducing N-NH3 emissions. Such a pattern was evidenced in both seasons.

In 2015/2016, the accumulated rainfall in the first seven days (19 mm) of the third split application was the lowest. Such low rainfall led to an increase of 40 and 90% in N-NH3 losses compared to the first and second split applications, respectively. The same pattern was not observed for the 2016/2017 season. However, an important issue must be considered: in the first and second N-fertilization, the mean NH3 losses in the first seven days were 16% and 8.5%, respectively. Such lower NH3 loss of 8.5% can be due to the absence of precipitation in the first two days of the first split application, which favored the permanence of the NH4 <sup>+</sup> from the N-fertilizers for a longer time on the soil surface. Regarding the pattern observed in the third split application, the losses were significant until the 13th day after fertilizer application, which is due to the lack of rainfall in the first days. These results show that rainfall before or after fertilization affects N losses by volatilization, particularly in the first seven days after the fertilizer application. Considering the initial seven days as the most critical phase to lose ammonia after applying prilled urea, the use of technologies associated with urea is critical to reducing N losses by volatilization [36,37]. In this context, NH3 losses depend on the volume and intensity of the rainfall [38].

This relationship between precipitation and N-fertilizer incorporation into the soil becomes even more complex in coffee plantations, as the architecture of the coffee plant restricts the direct incidence of rainfall, thus limiting the incorporation of the N-fertilizer applied in the canopy projection. Plant residues on the soil surface also function as a barrier to fertilizer incorporation (Figure S13).

Regarding the efficiency of urea + NBPT in reducing N-NH3 losses by volatilization, it is possible that the NBPT inhibited the urease activity, which is responsible for urea hydrolysis [12,30]. The efficiency of the NBPT was evidenced by the delay of 1.3 days in daily ammonia volatilization peaks, the reduction in MDL by 63%, and the 38.8% reduction of the accumulated loss compared to prilled urea. Therefore, the NBPT effectively delayed the beginning of N-NH3 losses and reduced the accumulated losses compared to prilled urea. This delay possibly increased the chances of N-fertilizer incorporation by rainfall, which can reduce the losses of N by volatilization [39].

NBPT is currently the most used urease inhibitor worldwide [39]. A meta-analysis study reported that NBPT can reduce N-NH3 losses by 52% on average, compared to the ammonia losses of prilled urea [6]. Urease inhibitors are highly efficient in reducing N-NH3 losses by volatilization, but some aspects must be considered when NBPT is added to urea. These aspects include: its degradation under increased soil temperature [40], acidic soil pH [41], time and temperature of storage [12], and contact with phosphate fertilizers, which contain free acidity [42].

Urea + Cu + B reduced the N-NH3 losses by 68% and 18% compared to prilled urea in the 2015/2016 and 2016/2017 crop seasons. This efficiency in reducing losses is due to the potential for urease inhibition using Cu and B. The urease inhibition mechanism is due to the reaction of copper with the sulfhydryl groups of the urease enzyme, forming insoluble sulfites and inactivating the enzymatic action of urease [43,44]. Boric acid (H3BO3) can also inhibit urease activity but through a different inhibition mechanism. In this case, H3BO3 has a very similar structure to urea and functions as an analog substrate for ureases. Thus, it replaces almost perfectly the water molecules bound to Ni at the center of the reaction [45,46]. Urea treated with Cu and B is already commercialized in Brazil, and for this study, it was bought from the regional fertilizer market. Although Cu and B are potential urease inhibitors, the low concentrations found in some commercial products may not be enough to inhibit the enzyme [30]. Thus, proper concentrations must be evaluated in varying crops and cropping systems. Some issues related to the treatment process in the fertilizer industry still complicate the increase in the amounts of Cu and B added to urea, especially with the use of H3BO3, which has a low concentration of B (17%).

Ammonia losses in the treatment urea + adhesive + CaCO3 were 18% higher than prilled urea in two coffee crop seasons. In this treatment, calcium carbonate was used as an alternative to elemental sulfur to create a physical barrier around the urea granule. In the present study, this technology was inefficient due to its limited effect as a physical barrier for the urea granule. CaCO3 increased the porosity and the contact with water enhanced its dissolution. This characteristic was evidenced when the day of the maximum NH3 loss was anticipated as well as the increase in the maximum NH3 daily loss in relation to the prilled urea. In addition, CaCO3 increases the alkalization that occurs around the urea granule hindering the pH buffering capacity of the region where the urea is hydrolyzed. Such a physical barrier with CaCO3 in urea increases the N-NH3 losses. Furthermore, this concept was also verified by the NH3 accumulated losses from the da urea + Adhesive + CaCO3, which was higher than the prilled urea. Thus, we concluded that the alkaline (CaCO3) coating urea was inefficient to reduce the ammonia losses by volatilization.

There are two N-fertilizers widely used in Brazilian coffee plantations, namely ammonium nitrate and ammonium sulfate. In this study, the reduction of N-NH3 losses for these two sources was higher than 97% in both crop seasons. The irrelevant NH3 losses from these N-sources are related to their acidic-to-neutral reaction in soil, mainly at pH < 7 [47]. Another positive aspect is that these fertilizers do not depend on weather conditions at the moment of their application. Thus, both ammonium nitrate and sulfate can be smart options for N-fertilization in coffee crop systems.

Altogether, the application of urea dissolved in water by drench draws attention to the technologies used to mitigate prilled urea losses by volatilization. In the present scientific study, this treatment showed good efficiency in reducing N-NH3 losses by volatilization. The days of maximum loss occur very similarly to prilled urea applied on the soil surface. However, the losses in these days of maximum loss are, on average, 3.8 and 4.5 times lower than prilled urea for the 2015/2016 and 2016/2017 crop seasons, respectively. Moreover, the accumulated losses of urea dissolved in water were five and six times lower than those observed for prilled urea in the 2015/2016 and 2016/2017 seasons, respectively. Such reduced losses are due to the dissolution of urea in water, which percolates to subsurface layers in the soil carrying the urea molecules, thus reducing N-NH3 losses by volatilization. For this treatment, no additive was added to the conventional urea. However, urease or nitrification inhibitors can also be added to the urea solution [48], thus improving urea use efficiency, that is, the technologies available in the market can be associated with strategies that can further increase the N use efficiency.

Considering the management of coffee plantations in Brazil, the application of urea solution can be performed along with systemic insecticides. Such products are applied directly in the ground, under the projection of the coffee tree canopy. However, it is important to evaluate the compatibility between the products to be applied, as well as the spray volume used and its relationship with the urea concentration in the solution, related to the solubility product constant [16]. Another strategy would be to add micronutrients to the urea solution in order to standardize the distribution. In addition, increasing the concentration of B and Cu in the urea solution could inhibit urease activity and help mitigate ammonia losses.

In this study, it was not possible to compare the conventional, stabilized, slow-, and controlled-release fertilizers. However, the latter treatments showed interesting patterns when applied to coffee cropping systems. The basis of slow-release, controlled-release, and Blend N-fertilizers is urea, but the associated technologies lead to contrasting responses in N-NH3 losses. The average accumulated loss by those sources is lower than 6% of the applied N when averaging the two crop seasons.

This pattern observed for controlled-release fertilizers (urea + elastic resin, urea + polyurethane) is explained by the way N is released into the soil. The release of N in controlled-release fertilizers occurs by the diffusion of urea from inside the granule through the coating into the soil solution. This process starts with increasing steam pressure and water intake into the granule. Then, osmotic pressure inside the capsule increases and creates a diffusion gradient from the fertilizer to the soil solution [49]. The gradual release reduces the excess of N-mineral available in the soil solution, which is susceptible to volatilization, denitrification, and leaching. Controlled-release urea improves the synchronism between the N release from fertilizer granules and its absorption by the plants, thus reducing N losses and improving nitrogen use efficiency in coffee crop environments [2,50].

The chemical reactions in the Urea-Formaldehyde production process reduce the nitrogen solubility in water compared to conventional sources of N, owing to the formation of long and short polymerization chains. This reduced solubility has varying effects on the rates of N release over time. The methylene urea chains formed in the Urea-Formaldehyde production depend on the activity of microorganisms in a process similar to N mineralization in the soil. Such a process prevents all the N from being readily released into the soil and is thus subject to the transformations needed to produce NH3 [51–55]. From an agronomic perspective, the lower N-NH3 losses are due to the reduction of excessive mineral N in the soil solution, which is susceptible to N-losses. The same pattern was also observed in the controlled-release urea. In the present study, the release time of controlled-release or slow-release fertilizers such as urea-formaldehyde was not verified. However, these enhanced efficiency fertilizers should be further investigated regarding their potential for proper N supply for coffee crop systems.

The Blend N-fertilizer, a blend of urea stabilized with NBPT and urea coated with elemental sulfur and polymer, was also efficient in reducing N losses, which did not exceed 7% in both coffee crop seasons. In this blend, part of the urea is in the soluble form and is protected by the NBPT as a urease inhibitor. The Blend N-fertilizer improves the N provision to the coffee plants over time as it combines the fast release of the soluble urea mixed with NBPT and the controlled-release urea to provide nitrogen for a longer time. The N-NH3 losses were similar to the 100% coated treatments compared to blend N-fertilizers, thus demonstrating the efficiency of this technology to supply N to the coffee plant.

#### *Highlights of Economic View of N-Fertilizers Technologies*

From an economic perspective, here we present a short overview related to the prices of N-fertilizer technologies assessed in this scientific paper. Prilled urea has the lowest cost in the market, considering its increased N concentration and disregarding the high N-NH3 losses. In sequence, are fertilizers stabilized with NBPT, Cu, and B, which have similar market values, followed by urea + adhesive + CaCO3. The prices of Blend-N-fertilizer reduce as the proportion of the stabilized and conventional ureas increase in the physical mixture and their prices are higher than conventional and stabilized N-fertilizers. In addition, Blend-N-fertilizers have a lower cost compared to 100% of controlled-release urea.

In this context, urea + polyurethane and urea + plastic resin, which are controlledrelease or added-value fertilizers, have similar market prices. In addition, the prices of controlled-release fertilizers may vary according to the material used in the coating and

the thickness of the coating. Finally, urea-formaldehyde as well as controlled-release fertilizers require investments in industrial processing such as infrastructure with specific conditions to produce these added-value N-fertilizers. In some situations, in Brazil, ureaformaldehyde has been used as a blended form with conventional urea and/or ammonium sulfate reducing its price compared to pure urea-formaldehyde. In general, pure ureaformaldehyde has similar prices to pure controlled-release fertilizers.

From an agronomic/economic point of view, the decision on which N sources would be interesting for application in coffee plantations must consider the costs of the fertilizer application. Fertilization with conventional and stabilized fertilizers must be split into three or more applications. Slow- and controlled-release fertilizers can be applied in a single operation reducing the costs (labor, fuels, machine maintenance, and depreciation), time of mechanized operation on the farm, and soil compaction due to the reduction of N splits compared to conventional and stabilized fertilizers. On the other hand, the Blend N-fertilizer technology is more expensive than conventional and stabilized fertilizers, but almost always has a lower value compared to slow- and controlled-release fertilizers. Besides, Blend N-fertilizers provide better synchronism between the N release and its absorption by the plants.

#### **4. Materials and Methods**

#### *4.1. Characterization of the Experimental Area*

The experiment was conducted in coffee plantations under field conditions for two crop seasons, 2015/2016 and 2016/2017, in Lavras, Minas Gerais (MG), Brazil (Figure 1). Lavras (910 m a.s.l., 21◦14 06" S 45◦00 00" W) is located in a traditional region of coffee production in Brazil, within the Campos das Vertentes geographical indication. According to Köppen's classification, the climate is Cwa, mesothermal with mild summers and dry winters. The mean annual precipitation is approximately 1472 mm, the mean annual temperature is 19.4 ◦C [56].

**Figure 1.** Location map of the experimental areas in Lavras, Minas Gerais, Brazil.

The coffee plantation in the production phase was planted with the "Catuaí Vermelho" cultivar, line 144. At the beginning of the experiment, the plantation was six years old. The spacing used in the planting was 3.7 m between rows and 0.7 m between plants, totaling 3861 plants ha<sup>−</sup>1.

The soil was classified as "Latossolo Vermelho Amarelo Distroférrico (LVdf)" according to the Brazilian System of Soil Classification [57], or Haplustox [58]. Before installing the experiment, soil samples were collected at the 0–0.2 m depth for soil texture [59] and soil chemical analyses. (Table 6) lists the result of the soil chemical analysis and texture.

**Table 6.** Chemical characterization and soil texture of the experimental area at the 0–20 cm depth, before the application of the treatments.


pH in water (1:2.5); P, K, and Cu extracted by Mehlich-1; B extracted by hot water; Ca2+, Mg2+, and Al3+ extracted by 1 M KCl; CEC = Cation Exchange Capacity at pH 7.0; OM = soil organic matter; BS = base saturation; Sand, silt, and clay = particle-size fractions.

#### *4.2. Experimental Design*

In this study, different sources of N-fertilizers were used, which were applied in a single application or split into three applications. Thus, two different group experiments were carried out in the same area, but the management practices (other than fertilization) remained similar. The experiments were designed as follows: Group 1) seven treatments, consisting of conventional and stabilized fertilizers and urea dissolved in water (management strategy) the experimental design in the field was randomized blocks with three repetitions, totaling 21 plots; and Group 2) four treatments, consisting of slow-, controlled-, and blend fertilizers, the experimental design in the field was randomized blocks with three repetitions, totaling 12 plots. For conventional and stabilized fertilizers, a dose of 300 kg ha−<sup>1</sup> was divided into three applications. For the other treatments (Group 2) a dose of 300 kg ha−<sup>1</sup> was applied in a single application. The treatments will be described in detail in the next topic. Each experimental unit consisted of 14 coffee plants. The ten central plants comprised a useful area for data collection.

#### *4.3. Characterization of the Fertilizers*

The fertilizers used in this study were chosen based on technologies used in Brazilian coffee production systems. They were divided into four groups and characterized according to the type of technology used. We photographed all fertilizers with a Canon camera, SL3 DSLR model, and an Olympus microscope, SZ60 Japan model. Fertilizers classified as controlled-release were characterized by scanning electron microscopy (SEM) and energy dispersion X-ray spectroscopy (EDS).

The first group included the conventional fertilizers: (1) prilled urea (45% N), (2) Ammonium nitrate (31% N), and (3) Ammonium sulfate (21% N and 24% S-SO4 2-). Another treatment containing prilled urea (45% N) diluted in water at a concentration of 50 g L−<sup>1</sup> was also added, aiming to reduce N-NH3 losses and to completely dissolve urea: (4) urea dissolved in water.

Another group amongst the technologies used in this study was the stabilized fertilizers, which have additives that can inhibit or delay some process of N transformation in the soil: (5) urea treated with Cu and B (44% N; 0.4% B as boric acid and 0.15% Cu as copper sulfate) and (6) urea treated with NBPT (45% N). This group of fertilizers consists of urease inhibitors (NBPT, NPPT, Cu, and B). The functioning of Cu and B as urease inhibitors depends on the concentrations added to the fertilizer. Besides being micronutrients, Cu and B have competitive and non-competitive urease inhibition capacities, respectively [60].

The group of controlled-release fertilizers was also included in this study: (7) urea coated with elastic resin (44% N and 43.8 μm of average coating thickness), (Figure S8) (8) urea coated with polyurethane (40% N and 56.4 μm of average coating thickness), (Figure S9).

The group of chemically modified or slow-release fertilizers was represented in this study by (9) urea-formaldehyde (26% N). This product results from the reaction of formaldehyde molecules (H2CO) with urea (NH2)2CO under controlled temperature and pressure. This reaction forms chains of C and N with different sizes and degrees of polymerization.

A treatment for physical protection of the urea granules was included in this study: (10) urea + adhesive + CaCO3. This treatment included a compound that agglutinates calcium carbonate (CaCO3), creating a physical barrier of adhesive and CaCO3. This barrier temporarily prevents contact between the soluble conventional urea and soil moisture.

Lastly, a fertilizer based on the physical mixture of technologies (blend) was added to the present study, constituting the treatment called: (11) Blend N-Fertilizer (39% N, 9% S0) (Figure S7). In this case, the release of N to the system occurs in different stages, as this blend is a mixture of conventional urea with a controlled release fertilizer (urea coated with elemental sulfur (S0) + polymer), measuring 67.5 μm of coating thickness and a stabilized fertilizer (urea treated with NBPT, most of the times). Therefore, the blend aims at the synchronization of the release of N by the fertilizer and its absorption by the plant, which reduces N excess in the system and N-NH3 losses by volatilization.

For the other treatments, the granulometry of the fertilizers varied from 1 to 4 mm, as officially specified by the Brazilian legislation for granulated fertilizers. Further physical characteristics of the treatments can be found in figures (Figures S7–S12).

The treatments used in this study were applied at the 300 kg N ha−<sup>1</sup> dose per year. The application of the treatments prilled urea, ammonium nitrate, ammonium sulfate, stabilized (urea + NBPT and urea + Cu + B) and urea + adhesive + CaCO3 were split into three doses of 100 kg N ha−<sup>1</sup> into the two crop seasons of the experiment. Urea dissolved in water was applied via drench at a dose of 1.6 L m<sup>−</sup>1, totaling 16 L per plot, following the same criteria described for the split application. The slow-release, controlled-release, and blend fertilizers were applied at a single dose of 300 kg N ha−<sup>1</sup> per year. All fertilizers were applied as topdressing, superficially, and under the canopy projection of the coffee plants. The applications for the 2015/2016 season were conducted on 6 November 2015, 11 January, and 10 March 2016. The 2016/2017 season received the applications following the same interval. The treatments received the slow-release, controlled-release, and blend-N fertilizers on the same day as the first application of the conventional and stabilized fertilizers.

#### *4.4. Complementary Management of Soil Fertility*

Liming was performed 60 days before applying the N fertilizers in each treatment plot, aiming to increase soil base saturation to 60%. The dose of 2 t ha−<sup>1</sup> of agricultural lime (PRNT 100%) was used in both crop seasons. Maintenance fertilization was performed with potassium chloride (KCl—60% K2O) and simple superphosphate (SFS—20% P2O5) fertilizers, applied at doses of 300 kg K2O ha−<sup>1</sup> per year and 100 kg P2O5 ha−<sup>1</sup> per year, respectively, under the canopy projection of the coffee plants [61].

The micronutrients were applied via leaf fertilization along with phytosanitary control. These procedures were performed both during the formation period of the coffee plantation and over the experimental period. A total of 5 kg ha−<sup>1</sup> of a commercial product containing the following nutrients were applied: 6.0% zinc (ZnSO4), 3.0% boron (H3BO4), 2.0% manganese (MnSO4), 10.0% copper (Cu (OH)2), 10.0% sulfur, 1.0% magnesium (MgSO4) and 10.0% K2O (KCl). The spray volume applied was 300 L ha−1, totaling three applications per year in 45-day intervals between November and February each year.

#### *4.5. Quantification of N-NH3 Losses*

The losses of N-NH3 owing to the application of N fertilizers were quantified using the semi-open collector adapted by Lara Cabezas [62]. In the first year, three PVC bases (0.2 m height and 0.2 m diameter) were installed 30 days before the application of the fertilizers in each experimental plot, under the canopy projection of the coffee plants, and at a depth of 0.05 m into the soil (Figure 2). These PVC chambers were kept in the field during the two years of the experiment.

**Figure 2.** Illustration of the collectors used in the quantification of ammonia losses.

Collection chambers were made in PVC with a diameter similar to the bases. The chambers had lids that prevented water to enter, but allowed air circulation. They had 0.5 m height and specifications as described in (Figure 2). The amount of fertilizer corresponding to the dose applied per hectare was added within each base (0.2 linear meters). To calculate the dose of N, we considered the useful distance in linear meters of one hectare and the 3.7 m spacing between lines, totaling 2702.7 m. The dose of N was corrected for the equivalent base diameter (0.20 m). For the N-sources whose fertilization was split into three applications, 7.4 g of N was added to each base on the same day that the fertilization of the plots was performed. As for the treatments that received one single fertilizer application, 22.20 g of N was added to each base. The collection chamber was added to one of the bases, in all plots, immediately after applying fertilizers on the bases.

Two laminated foam discs with 0.02 g cm−<sup>3</sup> density, 0.2 m thickness, and the same diameter as the PVC tubes were placed inside each semi-open collector. The foam discs were soaked with 80 mL of phosphoric acid (H3PO4; 60 mL L−1) solution and glycerin (50 mL L−1). The lower disc was placed inside the chamber at a height of 0.35 m from the soil, and the upper disc at 0.2 m from the lower one (Figure 2). The lower foam disk aimed to capture the ammonia released by the treatments, as the upper disk aimed to avoid contamination of the lower disk by N-NH3 released from the rest of the fertilized line.

Foam disks were collected on the 1st, 2nd, 3rd, 4th, 5th, 7th, 9th, 12th, 15th, 19th, 24th, and 31st days after the application of fertilizers in the 2015/2016 crop season for conventional and stabilized fertilizers (Group 1). In 2016/2017, the collections were performed on the 1st, 2nd, 3rd, 4th, 5th, 7th, 9th, 12th, 14th, 17th, and 20th days, and until the 34th day after applying the treatments. The collections in the treatments with slow- and controlled-release fertilizers (Group 2) were performed on the same day as the conventional and stabilized fertilizers. However, they were extended until the 208th day of the first crop season and until the 235th of the second crop season.

After each sponge change, the chamber was rotated from one base to another to consider the influence of the spatial variability of ammonia emission. This rotation allows a greater influence on climatic variations, such as temperature and precipitation.

The solution in the sponges collected from the field was extracted by filtration in a Büchner funnel connected to a vacuum pump. The extraction was performed after ten sequential washes with 40 mL of deionized water each. The extracts were stored in a cold chamber for a maximum period of 5 days, and after that, they were analyzed. From the extract, 20 mL aliquots were taken to determine the N content by distillation by the Kjeldahl method [63]. The N content in the sample was calculated according to equation 1, adapted from Nogueira and Souza [64]: TN = [(Va − Vb) × F × 0.1 × 0.014 × 100]/P1, in which, TN: Total N concentration in the sample (%), Va: Volume of hydrochloric acid solution spent on the sample titration (mL), Vb: Volume of hydrochloric acid solution spent on the blank titration (mL), F: Correction factor for 0.01 M hydrochloric acid, P1: Sample mass (g).

The values obtained in the N content calculations referred to the area occupied by the base of the chambers installed in the field. These values were then extrapolated to the percentage of loss of N-NH3 per hectare. The accumulated losses in the assessment were calculated by adding the losses from the 1st to the 2nd day, then adding this value to the losses of the 3rd day, and so on until the last day of collection.

#### *4.6. Statistical Analysis*

The treatments were submitted to a non-linear regression analysis using a logistic model (equation 2) to evaluate the ammonia loss by volatilization: Yi = [α/1 + ek (b − daai)] + Ei, in which, Yi is the i-th observation of the accumulated loss of N-NH3 in %, being i = 1, 2, ... , n; daai is the i-th day after the application of the treatment; α is the asymptotic value that can be interpreted as the estimation of the maximum accumulated loss of N-NH3; *b* is the abscissa of the inflection point and indicates the day of the maximum loss by volatilization; k is the value that represents the precocity index, and the higher its value, the lower the time needed to reach the maximum loss by volatilization (*α*); Ei is the error associated to the i-th observation, which is assumed to be independent and equally distributed according to a zero average standard and constant variance, E ~ N (0, I σ2).

This model is already used to estimate plant growth but has recently been applied to estimate the N-NH3 accumulated loss [6,65,66].

To estimate the maximum daily loss (day when the highest N-NH3 loss occurred), that is, to determine the inflection point of the curve, the following equation was used: MDL = k × (α/4), in which, k is a relative index used to obtain to a maximum daily loss of ammonia (MDL), and α is the asymptotic value that can be interpreted as the maximum amount of accumulated N-NH3 loss. The "nlme" package was used in the modeling of the N-NH3 losses data, using the R 3.3.1 software [67].

Normality and homoscedasticity of the data were verified by the Shapiro-Wilk and Bartlett tests, respectively. Then, an analysis of variance was performed to test the influence of the N sources on the N-NH3 losses by volatilization. The significance of the differences was evaluated at *p* < 0.05. After validating the statistical model, the mean values were grouped by the Scott-Knott algorithm using the R 3.3.1 software [67].

#### **5. Conclusions**

Nitrogen fertilizers such as conventional urea can be used to improve nutrient use efficiency in coffee production environments by using technologies such as urease inhibitors and polymer coatings. Altogether, conventional urea had ammonia losses equal to 24% of N applied to promote lower N-use efficiency during two coffee seasons. Calcium carbonate as a physical coating around the urea granules performed poorly compared to all the other N-fertilizer technologies with ammonia volatilization losses 18% greater than conventional urea. Urea dissolved in water is an interesting N-fertilization management strategy for coffee farmers as the ammonia losses were only 4.2% of the applied N. Urea stabilized with N-(n-butyl) thiophosphoric triamide (NBPT) is a useful industrial innovative technology to mitigate ammonia losses because urease inhibitor as additive reduces ammonia losses by 39%. Slow- and controlled-release urea and Blend N-fertilizer are interesting addedvalue N-fertilizers to improve coffee crop nutrition over time because they can be applied in a single mechanized operation with ammonia losses lower than 7% of the applied N. Conventional N-fertilizers such as ammonium nitrate and ammonium sulfate showed negligible ammonia losses demonstrating its potential as interesting choices in comparison with conventional urea to mitigate ammonia emissions. Also, they can be applied regardless of soil humidity and climate conditions. In summary, in this scientific paper, we presented some highlights of cutting-edge technologies as a plan for the efficient use of N-fertilizers in coffee crop production environments. However, our research group is engaged in

similar studies in the coffee crop, where not only aspects related to ammonia loss are being evaluated, but also emission of the other GHG, soil enzyme activity, and aspects related to plant nutrition, thus allowing better understanding of the N cycle for the coffee plant.

**Supplementary Materials:** The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/plants11233323/s1, Figure S1: Daily N-NH3 losses by volatilization in the first, second, and third split application (a, b, and c) of conventional and stabilized N fertilizers. Rainfall, average temperature, and relative air humidity after splitting the N fertilization in the 2015/2016 crop season (d); Figure S2: Daily (a) and accumulated (b) N-NH3 losses by volatilization in slow- and controlled-release N fertilizers. Rainfall, average temperature, and relative air humidity after splitting the N fertilization in the 2015/2016 crop season (c); Figure S3: Daily N-NH3 losses by volatilization in the first, second, and third split application (a, b, and c) of conventional and stabilized N fertilizers. Rainfall, average temperature, and relative air humidity after splitting the N fertilization in the 2016/2017 crop season (d); Figure S4: Daily (a) and accumulated (b) N-NH3 losses by volatilization in slow- and controlled-release N fertilizers. Rainfall, average temperature, and relative air humidity after splitting the N fertilization in the 2016/2017 crop season (c); Figure S5: Accumulated N-NH3 losses by volatilization in the first, second, and third split application (a, b, and c) of conventional and stabilized N fertilizers in the 2015/2016 crop season; Figure S6: Accumulated N-NH3 losses by volatilization in the first, second, and third split application (a, b, and c) of conventional and stabilized N fertilizers in the 2016/2017 crop season; Figure S7: Thickness of surface coatings by MEV images (a) and elemental composition of surface coatings (b) of the controlled release source: urea coated with elemental sulfur (S0) + polymer (Blend N—fertilizer); Figure S8: Thickness of surface coatings by MEV images (a) and elemental composition of surface coatings (b) of the controlled release source: urea coated with plastic resin; Figure S9: Thickness of surface coatings by MEV images (a) and elemental composition of surface coatings (b) of the controlled release source: urea coated with polyurethane; Figure S10: Images of the conventional N fertilizers: urea (a), ammonium nitrate (b), and ammonium sulfate (c); Figure S11: Images of the stabilized N fertilizers: urea treated with NBPT (a) and urea treated with Cu and B; Figure S12: Images of the fertilizers: Slow-release: Urea-formaldehyde (a) and Physical barrier: Urea + adhesive + CaCO3 (b); Figure S13: Dry leaves in the canopy projection of the coffee plant hindering fertilizer incorporation.

**Author Contributions:** Conceptualization, D.G., R.J.G., L.B. and T.F.; methodology, D.G.; software, T.J.F.; validation, D.G., T.F., L.B. and R.J.G.; formal analysis, T.J.F.; investigation, T.F., A.W.D. and P.C.Z.; resources, D.G.; data curation, T.F., C.S., M.P.D. and L.F.S.; writing—original draft preparation, T.F., D.G., L.B. and R.J.G.; writing—review and editing, D.G., T.F., C.S., M.P.D. and L.F.S.; visualization, T.F. and D.G.; supervision, D.G.; project administration, D.G. and R.J.G.; funding acquisition, D.G. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico), CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior), FAPEMIG (Fundação de Amparo à Pesquisa do Estado de Minas Gerais),and (INCT – Café) Instituto Nacional de Ciência e Tecnologia do Café.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** The authors thank the CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico), CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior), FAPEMIG (Fundação de Amparo à Pesquisa do Estado de Minas Gerais), and (INCT – Café) Instituto Nacional de Ciência e Tecnologia do Café.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **Nitrogen Fertilizers Technologies for Corn in Two Yield Environments in South Brazil**

**Bruno Maia Abdo Rahmen Cassim 1,\*, Marcos Renan Besen 1, Wagner Deckij Kachinski 1, Celso Rafael Macon 1, João Henrique Vieira de Almeida Junior 1, Rodrigo Sakurada 2, Tadeu Takeyoshi Inoue <sup>1</sup> and Marcelo Augusto Batista 1,\***


**Abstract:** Improvements in nitrogen use efficiency (NUE) in corn production systems are necessary, to decrease the economic and environmental losses caused by loss of ammonia volatilization (NH3-N). The objective was to study different nitrogen (N) fertilizer technologies through characterization of N sources, NH3-N volatilization losses, and their effects on the nutrient concentration and yield of corn grown in clayey and sandy soils in south Brazil. The treatments consisted of a control without N application as a topdressing, three conventional N sources (urea, ammonium sulfate, and ammonium nitrate + calcium sulfate), and three enhanced-efficiency fertilizers [urea treated with NBPT + Duromide, urea formaldehyde, and polymer-coated urea (PCU) + urea treated with NBPT and nitrification inhibitor (NI)]. The losses by NH3-N volatilization were up to 46% of the N applied with urea. However, NI addition to urea increased the N losses by NH3-N volatilization by 8.8 and 23.3%, in relation to urea alone for clayey and sandy soils, respectively. Clayey soil was 38.4% more responsive than sandy soil to N fertilization. Ammonium sulfate and ammonium nitrate + calcium sulfate showed the best results, because it increased the corn yield in clayey soil and contributed to reductions in NH3-N emissions of 84 and 80% in relation to urea, respectively.

**Keywords:** urea; ammonia volatilization; enhanced-efficiency fertilizers; nitrification inhibitor; plant nutrition; X-ray diffraction; scanning electron microscopy

#### **1. Introduction**

Producing food sustainably and sufficiently for humanity has been a challenge over time. At the global level, corn (*Zea mays* L.) is the most produced grain, with 1.2 billion tons produced per year [1], and it will be responsible for a 45% increase in cereal production in the coming years [2], driven by an estimated population expansion to 9.7 billion people by 2050. Although the production potential of corn hybrids has increased through genetic improvements and the development of more technically advanced crops, the world average yield is 5980 kg ha−<sup>1</sup> [1], far below the productive potential of the crop.

Nitrogen fertilization management is one of the factors that contributes the most to increasing corn yield. In plants, nitrogen (N) is the mineral element required in greatest quantities and is responsible for the synthesis of amino acids, proteins, and enzymes, and for photosynthetic processes [3]. Urea [CO(NH2)2] is the most commonly used source to meet the N needs of plants, because it has industrial advantages, such as a high N concentration per unit mass (45 to 46%) and lower production costs than other N sources [4]. It is estimated that in 2023, the global demand for N will be 155 Mt yr<sup>−</sup>1, of which 53% will be supplied by urea [5]. However, once applied to soil, urea is hydrolyzed by the action

**Citation:** Cassim, B.M.A.R.; Besen, M.R.; Kachinski, W.D.; Macon, C.R.; de Almeida Junior, J.H.V.; Sakurada, R.; Inoue, T.T.; Batista, M.A. Nitrogen Fertilizers Technologies for Corn in Two Yield Environments in South Brazil. *Plants* **2022**, *11*, 1890. https://doi.org/10.3390/ plants11141890

Academic Editors: Przemysław Barłóg, Jim Moir, Lukáš Hlisnikovský, Xinhua He and Bertrand Hirel

Received: 16 May 2022 Accepted: 20 June 2022 Published: 21 July 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

of urease enzyme, producing ammonia (NH3-N), which is rapidly lost to the atmosphere in the form of gas [6]. This loss may account for more than 60% of the N applied [7], depending on the soil and air temperature [8], soil moisture [9], soil pH [10], soil buffering capacity [11], presence of straw on the soil surface [12], N source [13], and rate of applied fertilizer N [14].

Although NH3-N is not a greenhouse gas (GHG), it can indirectly contribute to nitrous oxide (N2O-N) emissions [15], which are extremely harmful, due to their high global warming potential and permanence in the atmosphere for long periods [16]. Ammonia losses can reduce the N use efficiency (NUE), because less nutrients are left for plant absorption, causing negative yields and economic consequences for farmers [17,18]. In addition, NH3-N losses in agricultural areas affect air quality and contaminate terrestrial and aquatic ecosystems [19]. In the United States, for example, economic losses of approximately 39 billion dollars and deaths of more than 4300 people annually are linked to air pollution, as a result of NH3-N emissions from corn production systems that have both low NUE and nitrogen fertilizer overdoses [20].

The incorporation of urea into the soil is an effective way to reduce losses by NH3-N volatilization and increase NUE. This incorporation can be achieved using mechanical techniques [21] or irrigation [22]. Such practices are not always possible, because less than 20% of the world's areas are irrigated [23] and because they interrupt the no-tillage system, which is an important soil conservation management practice. Therefore, surface application of N is the predominant practice in agricultural production systems. Nitrogen sources such as ammonium sulfate [(NH4)2SO4] and ammonium nitrate (NH4NO3) are not subject to considerable losses by NH3-N volatilization [13,24], but tend to be more expensive, due to their lower N concentration. In addition, ammonium nitrate is subject to purchase restrictions by the military, due to its use as an explosive material [4].

To circumvent these limitations, N fertilizer industries have relied on the use of urea as a raw material, due to its high concentration of N for the development of N enhancedefficiency fertilizers (EEFs), classified as stabilized, slow-release, and controlled-release fertilizers [25]. However, although meta-analysis studies have revealed potential reduction in NH3-N losses by EEFs in relation to urea of between 39.4 and 52.0%, depending on the soil characteristics and climatic conditions before fertilizer application [26,27], the gains in crop yield are low compared to those obtained with conventional urea, ranging from 5.3 to 6.0% [4,26,27].

The N fertilizer industry has developed new stabilizing molecules to inhibit urease activity and has proposed combinations of enhanced-efficiency technologies, to obtain mixed (two or more granules) and/or complex (single-granule) fertilizers. For example, the new Duromide + N-(n-butyl) thiophosphoric triamide (NBPT) stabilization technology reduced NH3-N losses by 33% compared to only NBPT [9]. On the other hand, the addition of nitrification inhibitors (NIs), which aims to reduce N2O-N losses and nitrate leaching (NO3 −-N) [28,29], increases the volatilization of NH3-N by 35.7% and consequently the indirect emissions of N2O-N by up to 15.2% [30], leading to major debates on the use of NIs to increase NUE in EEFs [31].

Therefore, the objective of this work was to study the different technologies of N fertilizers through the characterization of their N sources, NH3-N volatilization losses, and effects on the nutrient concentration and yield of corn grown in clayey and sandy soil in south Brazil.

#### **2. Materials and Methods**

#### *2.1. Description of the Sites and Soils*

The experiments were conducted in two locations belonging to the Technology Diffusion Unit (UDT) of Cocamar Cooperativa Agroindustrial: one located in the municipality of Floresta (23◦35 37 S; 52◦04 06 W), and another in the municipality of Guairaçá (22◦56 48 S; 52◦43 22 W), at 392 and 478 m above sea level, respectively. The climate of the study area is classified as subtropical humid mesothermal (Cfa) according to the

Köppen-Geiger classification system [32]. The rainfall, temperature, relative air humidity, and irrigation depth data during the experiments are shown in Figure 1.

**Figure 1.** Pluviometric precipitation (mm), air relative humidity (%), maximum and minimum temperature (◦C), and irrigation (mm) for corn grown on clayey soil (**a**) and corn grown on sandy soil (**b**). Bef. appli. N: Before application nitrogen day in topdressing. N appli. Day: Day of nitrogen application in topdressing.

The experiments were conducted in the municipalities of Floresta and Guairaçá, located in the state of Paraná, Brazil, in no-till areas, with previous crops of *Brachiaria ruziziensis* and *B. brizantha*, respectively. The soils of the experimental areas in Floresta and Guairaçá were classified as a Latossolo Vermelho distroférrico with clayey texture (clayey soil) and Argissolo Vermelho-Amarelo distrófico (sandy soil) [33], corresponding to an Oxisol and Ultisol, respectively, according to the soil taxonomy of the USDA [34]. Soil samples from the 0–20 m layer were collected for chemical characterization and the determination of particle size (Table 1).

#### *2.2. Experimental Design, Treatments, and Crop Management*

A randomized block experimental design was applied, with five replicates and seven treatments. The treatments consisted of a control without N application as topdressing; three conventional nitrogen sources: urea (46% N), ammonium sulfate (21% N and 24% S), and ammonium nitrate + calcium sulfate (27% N, 3.7% S and 5% Ca); and three fertilizers with increased efficiency: one fertilizer stabilized to inhibit the activity of urease enzyme [urea treated with NBPT + Duromide (46% N)], one slow-release fertilizer [urea formaldehyde (37% N)], and one fertilizer consisting of a mixture of granules of different technologies [polymer-coated urea (PCU) (42% N) + urea treated with NBPT and nitrification inhibitor (Ur-NBPT + NI) (46% N) + 3.0% S and 0.3% B in the form of elemental sulfur

(99% S) and ulexite (10% B), respectively]. The experimental units were 4 m wide and 10 m long, yielding a total area of 40 m2.

**Table 1.** Chemical and granulometric analysis of an Oxisol (clayey texture), Ultisol (sand texture) and interpretation of values for the surface layer (0.00–0.20 m).


pH(CaCl2) (0.01 mol L<sup>−</sup>1) at a soil:solution ratio of 1:2.5; H + Al was determined by the Shoemaker–McLean–Pratt (SMP) method; Ca2+, Mg2+, and Al3+ extracted with KCl 1 mol L−1; OM: soil organic matter content obtained by organic carbon × 1.724 (Walkley-Black); P, K+, Zn, Cu, Fe, and Mn: Mehlich-1 extraction; SO4 <sup>2</sup><sup>−</sup> was extracted by calcium phosphate in acetic acid; B was extracted with hot water; sum of bases (SB): Ca2+ + Mg2+ + K+; CEC: cation exchange capacity at pH 7 (SB+H+ Al); ECEC: effective cation exchange capacity (SB + Al3+); BS: base saturation [(SB/CEC) × 100]; and particle size distribution (sand, silt, and clay) by densimeter method. <sup>1</sup> Soil attribute interpretation of the clayey soil. <sup>2</sup> Soil attribute interpretation of the sand soil. <sup>3</sup> Interpretation of soil attributes, according to SBCS/NEPAR [35].

Corn (*Zea mays* L.) was sown at the Floresta (clayey soil) and Guairaçá (sandy soil) sites on 14 October and 6 November 2020, under a dry mass cover of 3.15 and 4.76 Mg ha−<sup>1</sup> *B. ruziziensis* and *B. brizantha*, respectively. The corn hybrids used in Floresta and Guairaçá were Brevant 2433 PWU and FS512 PWU, respectively, with a distribution of 2.7 seeds m−<sup>1</sup> and a spacing of 0.45 m, totaling 60,000 pl ha−1. Sowing fertilization was performed with the application of 535 and 400 kg ha−<sup>1</sup> of 10-15-15 (N-P2O5-K2O), and when the corn was at phenological stage V4 (four leaves with collars visible), 60 and 40 kg ha−<sup>1</sup> K2O as KCl were applied in Floresta and Guairaçá, respectively. The N fertilizers were applied as a topdressing, according to the expected yield of the Floresta (clayey soil) and Guairaçá (sandy soil) sites at phenological stage V5 (five leaves with collars visible) at doses of 200 and 150 kg ha−1, respectively, as recommended by the Parana State Fertilization and Liming Manual (SBCS/NEPAR) [35].

#### *2.3. Capture and Determination of Ammonia Volatilization*

To determine the ammonia volatilization, the N fertilizers were weighed separately with an analytical balance and manually applied in a semi-open static chamber allocated within each experimental unit. For the treatment consisting of a mixture of granules of different technologies, PCU was physically separated from Ur-NBPT + NI, and three chambers were installed in the experimental unit for each treatment; with the first chamber for PCU, the second chamber for Ur-NBPT + NI, and the third chamber for granules mixed at a ratio of 50% PCU and 44% Ur-NBPT + NI, as the product is marketed.

Immediately after the application of N fertilizers in Floresta (clayey soil) and Guairaçá (sandy soil), N losses via ammonia volatilization were quantified through sample collection, performed at 1, 2, 4, 6, 8, 11, 15, 19, 22, 26, 33, 40, 47, 54, 61, 68, 76, and 83 days and 1, 2, 4, 6, 9, 15, 21, 28, 36, 44, 51, 58, 65, 71, and 78 days after application, totaling 18 and 15 collection times, respectively. The chambers were constructed from plastic bottles (polyethylene terephthalate, PET) with a total area of 0.007854 m2. Each chamber contained a 2.5 cm wide and 25 cm long strip of filter paper with a base immersed in a 50-cm<sup>3</sup> flask with 20 mL 0.05 mol L−<sup>1</sup> H2SO4 and a solution of 2% glycerine (*v*/*v*) [36,37]. The used vials were replaced with new vials until the ammonia loss stabilized.

After each collection, the chambers were rotated between the three bases within each experimental unit, to minimize the effects of environmental factors such as rainfall and temperature. Subsequently, the samples were sent to the Soil Fertility Laboratory of the Maringá State University, Paraná, Brazil, and refrigerated until analysis. Ammonia captured in the form of ammonium sulfate was determined by UV/VIS spectrophotometry, using the salicylate green method [37]. During the experimental period of ammonia volatilization sampling, no irrigation was performed.

#### *2.4. Characterization of Nitrogen Fertilizers*

The N fertilizers were finely ground and characterized by X-ray diffraction (XRD) analysis (XRD 6000, Shimadzu, Kyoto, Japan). X-ray diffractograms were obtained with a scanning interval of 3◦ to 70◦ 2θ, sampling step of 0.02◦, and time of 1.2 s using CoKα radiation with a nickel filter (40 kV, 30 mA). The values obtained were exported to X'Pert Highcore Plus software to determine the intensity, peak position, and crystallographic *hkl* plane. To identify the coating layer and coating thickness of the PCU + Ur-NBPT + NI treatment with controlled-release technology, the granules were physically separated into PCU and Ur-NBPT + NI and cut longitudinally with the aid of a scalpel blade. Subsequently, the granules were fixed to a stub microscope support with the aid of carbon tape and then metallized with gold. The samples were then analyzed using scanning electron microscopy (SEM), using a Quanta FEG 250 microscope.

#### *2.5. Evaluation of Nutrient Concentration, Morphological, and Yield Status*

During the flowering period of the corn, corresponding to phenological stage R1 (silking), indirect readings of the chlorophyll leaf content (SPAD index) were performed using SPAD 502 Plus® instrument (Konica Minolta, Tokyo, Japan). The nutrient concentration of the corn was evaluated at phenological stage R1, by randomly collecting 15 plants from the middle third of the first leaf opposite and below the upper ear [35]. After collecting the leaves, the samples were sent to the Soil Fertility Laboratory of the Maringá State University, Paraná, Brazil; washed with distilled water; dried in an oven with forced air circulation at 65 ◦C for 72 h; and ground in a Wiley mill. Subsequently, the samples were weighed and subjected to sulfuric acid and nitric-perchloric acid digestion for the extraction of N, phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), sulfur (S), zinc (Zn), copper (Cu), iron (Fe), and manganese (Mn). For analysis of leaf boron (B) content, the samples were subjected to dry digestion via calcination in a muffle furnace [38].

The content of Ca, Mg, Zn, Cu, Fe, and Mn were determined by atomic absorption spectrophotometry (AAS) with a mixture of air:acetylene. Phosphorus was determined by vanadate yellow spectrophotometry, S by spectrophotometry using the barium sulfate turbidimetry method, K by flame emission photometry, N by the micro-Kjeldahl method, and B by azomethine-H spectrophotometry [38].

The plant height was determined at phenological stage R2 (milky grains) by measuring from the soil surface to the insertion of the tassel using a tape measure. After the physiological maturation of corn, corresponding to phenological stage R6, manual harvesting of the ears was performed in a useful area of 10.8 m2, the kernel moisture was corrected to

13%, and the kernel mass of each experimental unit was extrapolated to obtain the yield in kg ha<sup>−</sup>1.

#### *2.6. Statistical Analysis*

The data obtained for corn yield, height, and nutrient concentration were subjected to homogeneity of variance (Bartlett) and error normality (Lilliefors) tests, thus meeting the assumptions of analysis of variance [39]. Subsequently, the data were subjected to joint analysis of variance, provided that the quotient between the largest and smallest residual mean squares of the analysis of individual variances was less than 7 [40]. Treatments and places were considered fixed factors, and their interaction was subdivided into treatments within places and places within treatments (*p* < 0.05), as shown by the follow statistical model. Subsequently, the means were compared using Tukey's test at 5% probability, using the statistical software GENES [41].

$$\mathbf{Y}\ddot{i}\ddot{j}\mathbf{k} = \mu + \mathbf{G}\dot{\mathbf{i}} + \mathbf{B}/\mathbf{A}\dot{j}\mathbf{k} + \mathbf{A}\dot{\mathbf{j}} + \mathbf{G}\mathbf{A}\dddot{i}\mathbf{j} + \varepsilon\dot{i}\mathbf{j}\mathbf{k}$$

where Y*ijk* is the observed value for treatment *i* (nitrogen fertilizer) in place *j* (clayey or sandy soil) in block *k*; *μ* is the effect of the mean; G*i* is the fixed effect of treatment i; B/A*jk* is the block nested in place *j*; A*j* is the fixed effect of place *j*; GA*ij* is the interaction between treatment *i* and place *j*; and ε*ijk* is the experimental random error in treatment *i*, place *j*, and block *k*.

For the NH3-N volatilization data, model selection was performed according to the Akaike information criterion (AIC) [42], and the model with the lowest AIC was chosen. After selecting the model, the data were subjected to nonlinear regression using SigmaPlot software, using the logistic model of three parameters (α, β and γ) represented by Equation (1), as described by Seber and Wild [43]. This model has traditionally been used to estimate plant growth and nutrient uptake rates [44] and has been more recently used to estimate losses by NH3-N volatilization [9,13].

$$\hat{Y} = \frac{\alpha}{1 + \exp^{\left[-\left(\text{time} - \beta\right)/\gamma\right]}}\tag{1}$$

where Yˆ is the amount of N volatilized as NH3-N (kg ha−1) at time t; α is the maximum accumulated volatilization; β is the time at which a 50% loss occurs, corresponding to the inflection point of the curve (the day when the maximum daily loss of NH3-N occurs); t is the time (days); and γ is a parameter of the model used to calculate the maximum daily loss (MDL) of NH3-N, as shown in Equation (2).

$$\text{MDL} = \frac{\alpha}{4\gamma} \tag{2}$$

#### **3. Results**

#### *3.1. Scanning Electron Microscopy and X-ray Diffraction*

The electron micrographs revealed the transverse morphology of the granules that compose the PCU + Ur-NBPT + NI granule mixture, indicating the presence or absence of coating (Figure 2). Thus, the Ur-NBPT + NI granules do not have a coating layer (Figure 2a), whereas the PCU granules have a polymer coating layer with an average thickness of 34.90 μm (Figure 2b).

The X-ray diffractograms showed typical reflections of the chemical species of each nitrogen fertilizer (Figure 3). The urea-based fertilizers (urea and Ur-NBPT + Duromide) showed characteristic reflections of urea (110) and biuret (Figure 3a,d). The only phase found in ammonium sulfate was the characteristic reflection of this fertilizer (Figure 3b). Conversely, the ammonium nitrate + calcium sulfate-based fertilizer contained dolomite, ammonium nitrate, and calcium sulfate, with more intense *hkl* planes at 104, 111, and 020, respectively (Figure 3c). In Ur-formaldehyde, at least two phases were identified,

urea and methylenediurea (MDU), indicated by more intense reflections at 110 and −311, respectively (Figure 3e).

**Figure 2.** Electron micrographs of the separate Ur-NBPT + NI (**a**) and PCU granules (**b**). P1 is the thickness of the coating layer of the polymer coated granule.

**Figure 3.** X-ray diffraction of the urea (Ur) (**a**), ammonium sulfate (**b**), ammonium nitrate + calcium sulfate (**c**), Ur-NBPT + Duromide (**d**), Ur-formaldehyde (**e**), PCU + Ur-NBPT + NI (**f**), elemental sulfur (**g**) and ulexite (**h**) for conventional and enhanced efficiency nitrogen fertilizers characterization.

For the mixed fertilizer, the granules were separated into PCU, Ur-NBPT + NI, S granules (elemental sulfur), and B granules (ulexite). The XRD patterns of PCU and Ur-NBPT + NI indicated reflections characteristic of urea and biuret (Figure 3f). The S granule was identified as elemental sulfur with a more intense plane at 222 (Figure 3g). The B granule showed several phases, such as ulexite, gypsum, glauberite halite, and bassanite, with more intense planes at -2-12, 020, 311, 042, and 301, respectively (Figure 3h).

#### *3.2. N Losses through Ammonia Volatilization*

The climatic conditions before the application of the N topdressing fertilizers are shown in Figure 1. The fertilizers were applied 24 h after 4.6 and 12 mm rainfall to the clayey and sandy soil, respectively. The maximum and minimum temperatures during the first 76 h after fertilization were 31.1 and 19.9 ◦C, respectively, for Floresta (clayey soil) and 34.0 and 20.2 ◦C for Guairaçá (sandy soil). The volatilization of NH3-N followed a sigmoidal pattern, increasing at the beginning, reaching the maximum daily loss, and then stabilizing (Figure 4). The maximum accumulated loss (α) of NH3-N according to the adjusted model decreased in the following order: urea (41.0 and 69.2 kg ha−<sup>1</sup> NH3-N; 20.5 and 46.1% of the N applied), PCU + Ur-NBPT + NI (40.1 and 62.5 kg ha−<sup>1</sup> NH3-N; 20.0 and 41.7% of the N applied); Ur-NBPT + Duromide (33.4 and 44.5 kg ha−<sup>1</sup> NH3-N; 16.7 and 29.7% of the N applied), Ur-formaldehyde (18.3 and 30.5 kg ha−<sup>1</sup> NH3-N; 9.1 and 20.3% of the N applied), ammonium nitrate + calcium sulfate (10.7 and 13.6 kg ha−<sup>1</sup> NH3-N; 5.3 and 9.1% of the N applied), and ammonium sulfate (8.5 and 11.0 kg ha−<sup>1</sup> NH3-N; 4.2 and 7.3% of the N applied) for clayey and sandy soils, respectively (Table 2).

**Figure 4.** Cumulative volatilization of NH3-N after topdressing applications of urea, ammonium

sulfate, ammonium nitrate + calcium sulfate, Ur-NBPT + Duromide, Ur-formaldehyde, and PCU + Ur-NBPT + NI for clayey soil at a rate of 200 kg ha−<sup>1</sup> of N (**a**) and for sand soil at a rate of 150 kg ha−<sup>1</sup> of N (**b**). Data with overlapping vertical bars with 95% confidence interval in the curve.

Ammonium sulfate, ammonium nitrate + calcium sulfate, Ur-formaldehyde, Ur-NBPT + Duromide, and PCU + Ur-NBPT + NI reduced NH3-N losses by 79.3 and 84.1, 73.9 and 80.3, 55.4 and 55.9, 18.5 and 35.7, and 2.2 and 9.7% compared to urea for clayey and sandy soil, respectively (Table 2). Ammonium sulfate and ammonium nitrate + calcium sulfate were the sources that most reduced the NH3-N volatilization losses in both locations compared to urea, with average reductions of 81.7 and 77.1%, respectively. The peak NH3-N volatilization (β) of urea, ammonium sulfate, ammonium nitrate + calcium sulfate, Ur-NBPT + Duromide, Ur-formaldehyde, and PCU + Ur-NBPT + NI occurred at 8.4 and 1.2, 8.3 and 7.1, 7.3 and 3.5, 8.5 and 3.6, 6.0 and 1.6, and 11.1 and 2.2 days after application of topdressing fertilizers to clayey and sandy soil, respectively (Table 2).


**Table 2.** Nonlinear regression parameters adjusted (logistic model) for NH3-N volatilization cumulative losses for conventional andefficiencynitrogenfertilizersandreductionofNH3-Nemissioninrelationforclayeyandsandysoils.

 enhanced

to calculate the MDL (maximum daily loss of NH3-N).

The environment in the sandy soil area provided lower β compared to the environment in the clayey soil area, and the NH3-N volatilization peaks advanced by 8.9, 7.2, 4.9, 4.4, 3.8 and 1.2 days for PCU + Ur-NBPT + NI, urea, Ur-NBPT + Duromide, Ur-formaldehyde, ammonium nitrate + calcium sulfate, and ammonium sulfate, respectively. The maximum daily loss (MDL) of NH3-N decreased in the following order: urea (3.10 and 6.92 kg ha−<sup>1</sup> NH3-N), Ur -NBPT + Duromide (2.69 and 6.18 kg ha−<sup>1</sup> NH3-N), PCU + Ur-NBPT + NI (1.82 and 3.32 kg ha−<sup>1</sup> NH3-N), Ur-formaldehyde (1.43 and 2.82 kg ha−<sup>1</sup> NH3-N), ammonium nitrate + calcium sulfate (0.54 and 0.64 kg ha−<sup>1</sup> NH3-N), and ammonium sulfate (0.56 and 0.47 kg ha−<sup>1</sup> NH3-N) for clayey and sandy soil, respectively (Table 2). The greatest reductions in MDL were obtained with the use of ammonium sulfate and ammonium nitrate + calcium sulfate in both locations.

#### *3.3. Ammonia Volatilization in Granules with and without Nitrification Inhibitor*

According to the fitted model (Figure 5), the maximum accumulated losses of NH3-N for the granules decreased in the order Ur -NBPT + NI (44.6 and 85.3 kg ha−<sup>1</sup> NH3-N; 22.3 and 56.9% of the N applied) followed by PCU (40.3 and 44.9 kg ha−<sup>1</sup> NH3-N; 20.1 and 29.9% of the N applied) for clayey and sandy soil, respectively (Table 3). Urea-NBPT + NI granules increased the NH3-N volatilization losses by 8.8 and 23.3% compared to urea for clayey and sandy soils, respectively.

**Figure 5.** Cumulative volatilization of NH3-N after topdressing applications of the separate granules PCU and Ur-NBPT + NI for clayey soil at a rate of 200 kg ha−<sup>1</sup> of N (**a**) and for sand soil at a rate of 150 kg ha−<sup>1</sup> of N (**b**). Data with overlapping vertical bars with 95% confidence interval in the curve.

The time of peak NH3-N volatilization of the granules occurred in the order PCU (30.6 and 12.6 days) followed by Ur-NBPT + NI (9.0 and 0.7 days), corresponding to delays in the peak of NH3-N volatilization of 21.6 and 11.9 days with PCU compared to Ur-NBPT + NI for clayey and sandy soil, respectively (Table 3). The MDL of Ur-NBPT + NI was 2.93 and 8.53 kg ha−<sup>1</sup> NH3-N, followed by PCU with 0.89 and 0.98 kg ha−<sup>1</sup> NH3-N for clayey and sandy soil, respectively (Table 3). The fertilizer Ur-NBPT + NI increased MDL by 69.6 and 88.5% relative to PCU for clayey and sandy soil, respectively.


**Table 3.** Nonlinear regression parameters adjusted (logistic model) for NH3-N volatilization cumulative losses of the separate granules PCU and Ur-NBPT + NI and the increase of NH3-N emission in relation to urea for clayey and sandy soils.

α: maximum cumulative volatilization; β: time at which 50% of the losses occur, corresponding to the curve inflection point; γ: parameter of the equation used to calculate the MDL (maximum daily loss of NH3-N).

#### *3.4. Leaf Macronutrient Content in Corn*

Regarding the leaf contents of macronutrients (Figure 6), the application of nitrogen fertilizers in topdressing increased leaf N contents by 6.31 (23.6%) and 3.05 (11.0%) g kg−<sup>1</sup> in comparison to the control for the clayey and sandy soils, respectively. However, there was no difference between the nitrogen sources in either soils. The results for nitrogen fertilizers applied as a topdressing in clayey soil were 2.12 (6.9%) g kg−<sup>1</sup> higher than those for the same applications performed in sandy soil, with the exception of the control, which showed no differences in leaf N content between the clayey and sandy soils (Figure 6a). The nitrogen sources increased the levels of leaf P and K only in a clayey soil environment, with a mean increase in relation to the control of 1.02 (37.9%) and 1.68 (11.7%) g kg<sup>−</sup>1, respectively, and no differences between the sources. All treatments, with or without application of N in topdressing, conducted in sandy soil had higher levels of leaf P and K than those conducted in clayey soil, with a mean difference of 1.35 (37.9%) and 4.78 (29.2%) g kg<sup>−</sup>1, respectively (Figure 6b,c).

In the clayey soil, corn without N fertilization as a topdressing had a higher leaf Ca content than the PCU + Ur-NBPT + NI, ammonium sulfate, and ammonium nitrate + calcium sulfate treatments, with increases of 1.22 (33.5%), 1.31 (36.9%), and 1.45 (42.5%) g kg<sup>−</sup>1, respectively. There was no significant difference between treatments for leaf Ca levels in corn grown in sandy soil and no significant differences between locations (Figure 6d). Regarding the levels of leaf Mg, in clayey soil, ammonium nitrate + calcium sulfate and the control had higher levels than ammonium sulfate, with differences of 0.76 (34.8%) and 0.71 (37.3%) g kg−1, respectively. There was no significant difference between treatments regarding leaf Mg content in the corn grown in sandy soil. Regarding the locations, the values for ammonium nitrate + calcium sulfate and the control in clayey soil were 0.44 (19.0%) and 0.48 (20.7%) g kg−<sup>1</sup> higher than the corresponding values in sandy soil, respectively (Figure 6e).

The ammonium sulfate source provided the highest leaf S levels in corn in both clayey and sandy soil, with an average increase of 0.55 (21.9%) and 0.85 (38.5%) g kg−<sup>1</sup> for clayey soil and 0.27 (12.0%) and 0.50 (24.7%) g kg−<sup>1</sup> for sandy soil compared to the other nitrogen sources and the control, respectively. Although ammonium sulfate provided the largest increases in leaf S, the sources urea, ammonium nitrate, Ur-NBPT + Duromide, Ur-formaldehyde, and PCU + Ur-NBPT + NI also increased the levels of leaf S, but only in relation to the control, with average values of 0.30 (13.6%) and 0.23 (24.7%) g kg−<sup>1</sup> for clayey and sandy soils, respectively. Corn, with or without application of N as a topdressing, grown in clayey soil had, on average, 0.29 (12.8%) g kg−<sup>1</sup> more leaf S than corn grown in sandy soil (Figure 6f).

**Figure 6.** Concentration of nitrogen (**a**), phosphorus (**b**), potassium (**c**), calcium (**d**), magnesium (**e**), and sulfur (**f**) in the corn leaf after fertilization in topdressing with nitrogen sources urea, ammonium sulfate, ammonium nitrate + calcium sulfate, Ur-NBPT + Duromide, Ur-formaldehyde, and PCU + Ur-NBPT + NI for clayey and sand soil. Treatments followed by the same lowercase letter do not differ using a Tukey test (*p* < 0.05). Environments followed by the same capital letter do not differ using a Tukey test (*p* < 0.05). LSDTrat: least significant difference for treatments. LSDEnvi: least significant difference for environment.

#### *3.5. Leaf Micronutrient Content and SPAD Index*

Regarding the micronutrient contents (Figure 7), leaf Fe and Mn were not influenced by N fertilization as a topdressing for either the clayey or sandy soil. However, the levels of leaf Fe and Mn in corn cultivated in a clayey soil environment were higher, at 55.32 (47.6%) and 263.99 (313.9%) mg kg−1, than the values for corn cultivated in sandy soil, respectively (Figure 7a,b). For leaf Zn, there was no difference between treatments in clayey soil. However, in the ammonium sulfate treatment in sandy soil, the leaf Zn content increased in relation to the control by 7.34 (38.1%) mg kg<sup>−</sup>1. All treatments, with or without application of N as a topdressing, conducted in clayey soil were higher to those conducted in sandy soil, with a mean difference of 19.46 (88.5%) mg kg−<sup>1</sup> in leaf Zn (Figure 7c).

Leaf Cu levels increased with topdressing N fertilization only in relation to the control, with an average increase of 4.13 (40.1%) mg kg−<sup>1</sup> for clayey soil; but in sandy soil, there was no difference between treatments. Similarly to Fe, Mn, and Zn, the levels of leaf Cu in corn grown in a clayey soil environment were also higher than those in corn grown in a sandy soil environment, with a mean difference of 5.16 (59.5%) mg kg−<sup>1</sup> in leaf Cu (Figure 7d). The application of PCU + Ur -NBPT + NI increased the leaf B content in both the clayey and sandy environments; but in clayey soil, an increase of 3.18 (30.9%) mg kg−<sup>1</sup> of B occurred only in relation to the control, while in the sandy soil, an increase of 3.41 (32.9%) and 4.46 (45.0%) mg kg−<sup>1</sup> in leaf B occurred in relation to the other nitrogen sources and the control, respectively. There were no significant differences in leaf B content between clayey and sandy soil (Figure 7e).

Regarding the indirect chlorophyll content (SPAD), the application of N fertilizers as topdressing increased the SPAD index by 11.58 (20.2%) and 5.27 (11.2%) compared to the control for clayey and sandy soil, respectively. However, there was no difference between the nitrogen sources in either production environment. The SPAD indexes of all treatments, with or without application of N as topdressing, conducted in clayey soil were higher than those in sandy soil, with a mean difference of 15.67 (30.4%) (Figure 7f).

#### *3.6. Yield and Height of Corn Plants*

Nitrogen fertilizers increased the corn yield only in a clayey soil environment. Increases in yield were obtained with the use of ammonium sulfate, PCU + Ur-NBPT + NI and ammonium nitrate + calcium sulfate, which resulted in increases of 1722 (20.6%), 1838 (21.9%), and 2088 (24.9%) kg ha−<sup>1</sup> corn, respectively, compared to the control that did not receive N fertilization as a topdressing. All treatments, with or without application of N as topdressing, conducted in clayey soil were higher than those conducted in sandy soil, with a mean difference in yield of 3654 kg ha−1, equivalent to 37.4%. However, when considering only the treatments that received N as a topdressing, the difference in yield became 370 kg ha−1, equivalent to 38.4% (Figure 8a). Topdressing N fertilization did not influence plant height in either the clayey or sandy soil environment. However, the plant height in clayey soil was 23 cm higher on average than that in sandy soil (Figure 8b).

**Figure 7.** Concentration of iron (**a**), copper (**b**), zinc (**c**), manganese (**d**), boron (**e**), and SPAD index (**f**) in the corn leaf after fertilization in topdressing with nitrogen sources urea, ammonium sulfate, ammonium nitrate + calcium sulfate, Ur-NBPT + Duromide, Ur-formaldehyde, and PCU + Ur-NBPT + NI for clayey and sand soil. Treatments followed by the same lowercase letter do not differ using a Tukey test (*p* < 0.05). Environments followed by the same capital letter do not differ using a Tukey test (*p* < 0.05). LSDTrat: least significant difference for treatments. LSDEnvi: least significant difference for environment.

**Figure 8.** Corn yield (**a**) and plant height (**b**) after fertilization in topdressing with nitrogen sources urea, ammonium sulfate, ammonium nitrate + calcium sulfate, Ur-NBPT + Duromide, Ur-formaldehyde, and PCU + Ur-NBPT + NI for clayey and sand soil. Treatments followed by the same lowercase letter do not differ using a Tukey test (*p* < 0.05). Environments followed by the same capital letter do not differ using a Tukey test (*p* < 0.05). LSDTrat: least significant difference for treatments. LSDEnvi: least significant difference for environment.

#### **4. Discussion**

#### *4.1. X-ray Diffraction and SEM of N Fertilizers*

The industrial production of N fertilizers in amidic form (Figure 3a,d,f) produces biuret as a by-product, resulting from the increase in temperature above the melting point of urea, which is 132 ◦C [45]. Although biuret is a common impurity, Brazilian legislation allows only up to 2% in solid N fertilizer [46], because it is a toxic chemical compound that interferes with the protein synthesis of plants [47]. Currently, biuret toxicity is insignificant in crops, due to advances in the technology used to manufacture urea fertilizers [48]. Although Ur-formaldehyde comes from an amidic source, biuret was not found in the XRD analysis, probably due to strict controls in the production process.

Dolomite, which was identified in the X-ray diffractogram of ammonium nitrate + calcium sulfate (Figure 3c), inhibits the exothermic and undesirable decomposition of ammonium nitrate, thus improving the safety of the fertilizer [49]. According to standard NFPA 490 of the National Fire Protection Association [50], ammonium nitrate is not considered flammable or combustible. However, factors such as a high temperatures under confinement (260 to 300 ◦C) and contamination by organic or inorganic materials, such as chlorides or powdered metals, can lead to explosive detonation with the production of nitrous oxide, which is decomposed into nitrogen and oxygen [51,52].

Urea-formaldehyde was the first synthetic nitrogen fertilizer with low solubility to be marketed for slow release of N. The production process consists of condensation within a reactor with controlled pH, temperature, molar ratio, and reaction time between urea and formaldehyde [25]. The final product of the reaction consists of a mixture of methylene urea polymers (methylene urea, MDU and polymethylene) with differences in the degree of polymerization (insolubility) and molecular weight (chain length) [53,54]. Thus, the presence of MDU in Ur-formaldehyde (Figure 3e) will provide an intermediate molecular weight and degree of polymerization, contributing to the slow release of N. This compound, combined with certain amounts of unreacted urea, results in an intelligent fertilizer for

agricultural use within the cultivation time of the plants. For example, Cassim et al. [55] observed no increase in yield in cultures with the application of Ur-formaldehyde with 70% slow-release compounds; but, with proportions of 55 and 60% of slow-release compounds, the yield increases were significant.

The lack of sulfur in the mixture of PCU + Ur-NBPT + NI granules (Figure 3f) indicates that the coating layer of the PCU granules (Figure 2b) is covered only by polymer and not by elemental sulfur (S0). This configuration provides better nutrient release kinetics than an S0 coating, because it is independent of the activity of microorganisms responsible for the oxidation of S0 [54]. However, the production cost of PCU is higher. The source of B present in the PCU + Ur-NBPT + NI + B + S mixture was identified as ulexite, an evaporite formed under arid conditions in saline lakes supported by hydrothermal vents and linked to volcanic activity [56]. For this reason, ulexite may be associated with other evaporites, such as halite, gypsum, glauberite, and bassanite, as described in Figure 3h.

#### *4.2. Ammonia Volatilization of N Sources in Clayey and Sandy Soil*

The losses by NH3-N volatilization were higher in sandy soil for all N sources tested (Figures 4 and 5). This behavior occurred due to the higher moisture content of the sandy soil, resulting from the higher rainfall volume 24 h before the application of the nitrogen fertilizers as a topdressing (Figure 1). Under dry soil conditions, the urease hydrolysis rate is low; however, the rate increases as the soil water content increases [57]. Above 20% moisture, hydrolysis is practically no longer affected by changes in soil moisture [4].

The previous considerations explain the sigmoidal behavior of NH3-N volatilization losses, which depend on the increase in urease enzyme activity [58], which consumes the H+ resulting from urea hydrolysis, as demonstrated by the reaction CO(NH2)2 + 2H<sup>+</sup> + 2H2O → 2NH4 <sup>+</sup> + H2CO3 [59]. This reaction promotes the increase in soil pH around N fertilizer granules to approximately 8.7, changing the balance between NH4 <sup>+</sup> and NH3 [60]. After reaching the maximum loss (α), NH3-N emissions decrease over time, due to the gradual reduction in pH and stabilization of N in the form of NH4 +-N [24].

In addition to differences in soil moisture, the clay content and, consequently, the cation exchange capacity (CEC) (Table 1) are the main differences between the two soil classes studied that will also influence the intensity of NH3-N volatilization. Clayey soils with a higher CEC have more exchange sites to retain the NH4 <sup>+</sup> produced in the hydrolysis of urea due to adsorption. In addition, the higher buffering capacity of soils with a higher CEC provides greater resistance to changes in soil pH around N granules caused by the urease enzyme, thus decreasing the intensity of NH3 -N volatilization [58,59]. Therefore, the lower loss of NH3-N by volatilization in clayey soils resulted in higher inflection points in the NH3-N curve (β) and a lower MDL, as described in Tables 2 and 3.

Ammonium sulfate and ammonium nitrate + calcium sulfate had the lowest accumulated losses of NH3-N in relation to the other N sources in both production environments (Figure 4), due to the absence of N in the amidic form (NH2-N). Corrêa et al., Minato et al., and Otto et al. [13,14,24] also obtained low losses due to NH3-N volatilization with the use of N sources in the ammoniacal and nitric forms, with losses ranging from 0.7 to 5.2% and 1.0 to 7.7% of the N applied for ammonium sulfate and ammonium nitrate, respectively, depending on the dose of applied topdressing. Following the increasing order of NH3-N emissions, Ur-formaldehyde was the enhanced-efficiency source that most reduced NH3-N volatilization, but it was less efficient than ammoniacal and nitric sources. Although Ur-formaldehyde can reduce the solubility of N fractions through the synthesis of methylene urea groups, it contains some urea that does not react with formaldehyde (Figure 3e), favoring losses by volatilization of NH3-N, even in small proportions.

The next formulation considered is Ur-NBPT + Duromide, a stabilizer that combines two NBPT + Duromide molecules, both of which inhibit the activity of the urease enzyme, but having the advantage of a more stable chemical structure under low pH and high soil temperature conditions [9]. However, the use of Ur-NBPT + Duromide showed higher emissions of NH3-N (mean of 23.5% of the applied N) when compared to meta-analysis

studies that obtained losses by volatilization of NH3-N of 14.8% using NBPT [26]. Under conditions with large amounts of straw on the soil surface, such as those described in the present study (mean of 3.95 Mg ha−<sup>1</sup> Brachiaria straw), the amount of urease enzyme in the soil will be higher, increasing losses by NH3-N volatilization by up to 25.5% [7]. In other words, the stabilizers can reduce, but not eliminate, the activity of the urease enzyme, possibly due to the high amounts of this enzyme in systems with high amounts of straw. Thus, such materials are not the most appropriate technology for production environments with large amounts of straw residues on the soil surface.

The mixture of PCU + Ur-NBPT + NI was inefficient in reducing losses by NH3-N volatilization, with losses very close to those of conventional urea (Figure 4 and Table 2). As PCU + Ur-NBPT + NI is a mixed fertilizer, composed of different granules, the granules were separated to understand the efficiency of each component in reducing or contributing to NH3-N emissions (Figure 5). Although PCU granules are designed to release N at a controlled rate to synchronize with the crop demand and reduce environmental pollution by NO3 −-N, NH3-N and N2O-N [61], factors such as high temperatures, excessive rainfall, the number and thickness of the coating layers, and quality of the coating material may have interfered with the efficiency of PCU, contributing to the release of N in the amidic form and losses in the form of NH3-N.

On the other hand, the addition of NI to Ur-NBPT to mitigate direct emissions of N2O-N [28] and losses by leaching of NO3 −-N [62] was the factor that most contributed to the inefficiency of the mixture PCU + Ur-NBPT + NI, since the addition of NI to NBPT significantly decreased the ability of NBPT to inhibit urea hydrolysis by up to 21% [63], in addition to contributing to the increase in NH3-N volatilization losses [30,31].

#### *4.3. Nitrification Inhibitor Increases Losses Due to Ammonia Volatilization*

The results showed that the use of NI significantly increased the volatilization of NH3-N relative to the PCU granules, especially in sandy soil (Figure 5b). According to Wu et al. [30], there are two main mechanisms associated with increased volatilization: (i) NIs are a group of chemical compounds that inhibit the activity of *Nitrosomonas* spp. bacteria responsible for the oxidation of NH4 <sup>+</sup> to nitrite (NO2 −) and, therefore, increase the soil concentration of NH4 <sup>+</sup> that is converted to NH3; and (ii) NIs induce a liming effect. Qiao et al. [64] found that the application of NI increased soil pH by 0.23 units, due to the decelerated rate of nitrification and increased efficiency of N use by plants resulting from the lower leaching of NO3 −. Thus, unleached NO3 − is absorbed by plant roots, which excrete OH− to maintain the electrochemical balance in the soil, thus increasing the pH of the medium [64,65]. Once the soil pH changes, the balance between NH3 and NH4 <sup>+</sup> is affected; as the soil pH increases, the equilibrium shifts, leading to the transformation of NH3-N and its subsequent loss to the atmosphere in the form of gas [60].

The volatilization of NH3-N can be influenced by several factors, such as dose, N source, climatic conditions, management system, and soil attributes, with the latter being the main factor responsible for altering the efficiency of NIs. For example, a meta-analysis performed by Kim et al. [66] found that treatments with NI increased emissions of NH3-N in soils with higher pH (5.4 to 7.9) and smaller ranges of CEC (5.7 to 16.8 cmolc dm−3) in comparison with lower-pH soils (4.7 to 6.2) and larger CEC ranges (10.0 to 24.0 cmolc dm<sup>−</sup>3). This effect occurs due to the favored formation of NH3-N at basic pH, combined with soils of low CEC, which provide fewer exchange sites for NH4 <sup>+</sup> adsorption, facilitating the loss of N by volatilization [67].

Another important soil attribute is the organic matter (OM) content. Soils with high levels of OM have higher amounts of *Nitrosomonas* spp., which hinders the performance of NIs [29] and leads to a need for higher concentrations of NI in soils with high OM content. In addition, high clay contents will favor lower emissions of NH3-N, due to their contribution to increasing soil CEC [68]. This explains the higher volatilization of urea treated with NI in sandy soil, since the pH, OM, CEC, and clay content were 4.50 and 5.70, 2.89 and 1.42%, 11.06 and 4.19 cmolc dm<sup>−</sup>3, and 78 and 10% for the clayey and sandy soils, respectively (Table 1). Therefore, the use of NIs, especially in sandy textured soils in rainfed agriculture, is not recommended as a strategy to increase the NUE. The INs technology in nitrogen fertilizers is more efficient in flooded agriculture systems, due to the denitrification losses of N2O-N and N2, representing up to 34% of the applied N [69].

#### *4.4. Nitrogen Sources and Nutrient Concentration of Corn—Macronutrients*

Although there were differences in losses due to NH3-N volatilization between N sources, the N applied as topdressing fertilization that was not lost by volatilization may have been sufficient to meet the N demand of the corn crop. As a result, changes in the concentration N status were not observed between the N sources, but only in relation to the control that did not receive N as topdressing (Figure 6a). According to Cantarella et al. [4], in many cases, most of the N absorbed by crops comes from soil OM, and fertilizer N, although important to increase yield, provides complementary N. Similarly, Oliveira et al. [70], working with 15N isotopes, observed that only 33% of the N absorbed by corn plants was derived from topdressing nitrogen fertilization.

Topdressing N fertilization was important for the maintenance of leaf chlorophyll content, which was indirectly quantified by the increase in SPAD index in relation to the control in both production environments (Figure 7f). According to Taiz et al. [3], chlorophylls are green photosynthetic pigments that have a porphyrin-like ring structure with a Mg atom coordinated in the center, linked to four other N atoms, with a long tail of hydrocarbons. Thus, in the absence of N, the plant degrades chlorophyll molecules to obtain the four N atoms that are part of its structure, developing generalized chlorosis in the leaf and compromising the absorption of light.

The highest leaf N concentration and chlorophyll content (SPAD) being in the corn grown in clayey soil was due to the higher expected yield provided by the better fertility of clayey soils than sandy soils, requiring a higher photosynthetic rate [71], and the higher N uptake by the plants, since each ton of corn grains requires 21.5 kg of N [35]. This effect was not observed in the accumulation of leaf N among the controls, most likely due to the lack of fertilization as topdressing, which inhibited the realization of the productive potential of the clayey soil environment.

After N, K is the nutrient most absorbed by corn plants, followed by P. Thus, the increase in cultivation intensity and, thus, higher yields obtained through N fertilization provided greater absorption of K and P and consequently a higher accumulation of these nutrients in the leaf (Figure 6b,c). Since P is a key element for the synthesis of molecules such as DNA, RNA, ATP, and NADPH [72], and as K is important for the activation of enzymatic systems and protein synthesis [73], synergistic interactions exist between N × P and N × K [74]. For example, Rietra et al. [75] performed a meta-analysis on the interaction between nutrients and found a synergism between N × P and N × K and no case of antagonism, in a total of 77 studies. The highest leaf concentration of P being in corn grown in sandy soil resulted from a lower adsorption to iron and aluminum oxides, which has a positive correlation with clay content [76]. Higher concentrations of leaf K were also observed in corn grown in sandy soil due to the low CEC of the soil, providing lower adsorption of K+ and consequently greater availability in the soil solution.

Elements are absorbed at different rates, due to their affinity with membrane carriers, obeying the decreasing cationic order NH4 <sup>+</sup> > K+ > Na+ > Mg2+ > Ca2+ [38]. Thus, because Ca is absorbed by the roots in the form of Ca2+, its absorption may be compromised by the high concentrations of NH4 <sup>+</sup> in the soil solution, due to competition [77]. Therefore, the highest accumulation of leaf Ca being in the control cultivated in clayey soil, relative to the values obtained with ammonium sulfate, ammonium nitrate + calcium sulfate, and PCU + Ur-NBPT + NI (Figure 6d), occurred due to two factors: (i) high doses of fertilizers containing NH4 +; and (ii) the use of fertilizers with NIs. The use of NIs will inhibit the nitrification process, increasing the level of NH4 <sup>+</sup> in the soil, thus suppressing the absorption of Ca2+ and resulting in a lower accumulation in leaves. This effect was also observed in the reduction in leaf Mg levels in corn grown in clayey soil using ammonium

sulfate compared to the control (Figure 6e). Thus, high concentrations of NH4 <sup>+</sup> can also reduce the absorption of Mg2+ and K+ by plants [78–80].

However, unlike Ca, the use of ammonium nitrate + calcium sulfate promoted an increase in leaf Mg, due to the presence of dolomite in its composition (Figure 3c), which inhibits the undesirable exothermic decomposition process of ammonium nitrate [49], in addition to being an important source of Mg2+ to plants [81]. On the other hand, reductions in leaf K accumulation were not observed with the use of ammoniacal sources, because the suppression is greater with divalent cations (Ca2+ and Mg2+) than with monovalent cations (K+) [82]. Regarding the differences between the production environments, the control conducted in sandy soil obtained lower concentrations of leaf Mg because the soil content was below the critical level (Table 1), which was 1 cmolc dm−<sup>3</sup> [35]. For ammonium nitrate + calcium sulfate, the highest concentration of leaf Mg in corn grown in clayey soil resulted from the highest dose of nitrogen application in topsoil, which thus provided more Mg in the form of dolomite.

Many plant compounds, such as amino acids and proteins, have both N and S, which helps explain the existence of a positive N/S ratio and the increases in leaf S concentration in both corn production environments with N application as topdressing compared to the control (Figure 6f). However, ammonium sulfate was the N source that provided the highest leaf S concentrations, due to the high concentration of S per unit mass (24%) in the form of sulfate (SO4 <sup>2</sup>−), which is the main form absorbed by plants and does not require oxidation by *Thiobacillus*, which, in turn, is dependent on soil temperature and moisture conditions [83]. On the other hand, the sandy soil probably favored more intense leaching of SO4 <sup>2</sup><sup>−</sup> due to the few anionic adsorption sites, decreasing the contact with the root system and, consequently, absorption by the plant [84]. This mechanism explains the higher concentration of S in the leaves of corn grown in clayey soil, as described in Figure 6f.

#### *4.5. Nitrogen Sources and Nutrient Concentration of Corn—Micronutrients*

The nitrogen fertilizers did not alter the leaf concentrations of Fe and Mn. However, the leaf concentrations varied between the production environments. The natural levels of micronutrients in soil depend on the chemical composition of the parent material, pedogenetic processes, and the degree of soil weathering [85]. Thus, the parent material of the clayey soil located in northern Paraná state is basalt, an igneous rock rich in micronutrients such as Fe and Mn, because these elements have the same geochemical formation environments [86]. In contrast, the sandy soil located in the northwestern part of the Paraná state originates from sandstones of the Caiuá Formation, a sedimentary rock whose main constituent mineral is quartz [87]. Consequently, higher levels of Fe and Mn are naturally available in clayey soil, favoring greater absorption and concentration of these nutrients in corn leaves (Figure 7a,b).

The use of nitrogen fertilizers in agriculture promotes the production of H+ through the oxidation of ammonium to nitrate, as demonstrated by the nitrification reaction NH4 <sup>+</sup> + 2O2 → NO3 <sup>−</sup> + H2O + 2H+ [88]. With a reduction in soil pH, the availability of metal cationic micronutrients increases [38], providing higher leaf concentrations of Zn in corn grown in sandy soil (Figure 7c) and Cu in corn grown in clayey soil (Figure 7d). However, the highest concentrations of leaf Zn occurred with the application of ammonium sulfate, which is an exclusively ammoniacal source, and the soil acidification process was intensified because more NH4 <sup>+</sup> was provided as a substrate for nitrifying bacteria [89]. An increase in leaf Zn with the use of ammonium sulfate was not observed in clayey soil, due to the higher CEC and OM content, which favored the acidity buffering reaction through the exchange of H+ ions for basic cations (Ca2+, Mg2+, K+ and Na+) in clay minerals and OM [90], which is the predominant buffering mechanism in the pH range between 4.2 and 5.0 [91].

Regarding the differences in Zn levels between the production environments, the higher grain yield obtained in clayey soil (Figure 8a) required a higher Zn absorption by the plants, because Zn is the most exported micronutrient in corn, with 24.8 g for each ton of grain produced [35]. Cooper, in turn, is the micronutrient that interacts the most with soil organic compounds and forms stable complexes, especially with carboxylic and phenolic groups of OM, in addition to having a strong affinity for clay [92]. Therefore, sandy soils, with low OM levels, are mostly deficient in Cu, due to leaching losses. This explains the higher concentration of leaf Cu in clayey soils than in sandy soils (Figure 7d).

The presence of ulexite in the PCU + Ur-NBPT + NI mixture identified by XRD (Figure 3h) resulted in an increase in leaf B concentration in both corn production environments (Figure 7e). However, soluble B sources such as borax (Na2B4O7·10H2O) and boric acid (H3BO3) are more commonly used to maintain plant growth when compared to lowersolubility sources such as ulexite (NaCaB5O9·8H2O) and colemanite (Ca2B6O11·5H2O) [93]. However, the ulexite present in the PCU + Ur-NBPT + NI mixture is of acid origin and is obtained through the granulation process with the use of sulfuric acid, which provides an increase in water solubility of approximately 90%. This higher solubility favors the faster release of B, making the nutrient available for plant absorption and increasing leaf B concentrations.

#### *4.6. Clay Soils Are More Responsive to Nitrogen Fertilization*

The sandy soil, because it originated from the Caiuá sandstone formation, is characterized by low a CEC, due to its high sand content, especially coarse sand [87]. A low CEC directly affects cation losses, due to leaching, and consequently the expected crop yield. The low fertility of sandy soils was confirmed by the lower growth of corn plants in the sandy soil than in the clayey soil (Figure 8b). Thus, in a production environment with a low response to fertilization, topdressing nitrogen fertilization contributes less to increases in corn yield (Figure 8a). In a meta-analysis performed by Tremblay et al. [94], the authors concluded that corn yield increased by a factor of 1.6 in sandy soils and 2.7 in clayey soils after nitrogen fertilization, showing that corn is more responsive to N fertilization in clay soils.

In addition to the influence of CEC on the response to nitrogen fertilization, soil texture can play a role. For example, clay affects the stabilization of organic N through the protection of OM by aggregates, favoring the preservation of microbial biomass [95]. Ros et al. [96] studied the variation in mineralizable N and its relationship with physical properties in 98 agricultural soils in the Netherlands and observed a lower N mineralization rate in clayey soils than in sandy soils. Ping, Ferguson, and Dobermann [97] found that corn needed less N fertilizer in sandy soils than in clayey soils. This may suggest that soil texture influences the degree of OM stabilization and, consequently, the response to nitrogen fertilization, increasing the chances of an increased yield via N fertilization in clayey soils.

Soil texture also provides different degrees of water storage in the soil. Therefore, sandy soils, due to their higher porosity, store less water, resulting in higher metabolic costs to the plant to absorb water and promote transpiration, which can consequently affect the yield. Although the response to nitrogen fertilization is lower in sandy soils than soils with other textures, studies of fertilization with varying sources and doses of N in these locations should be performed, mainly because such soils are highly susceptible to losses by NH3-N volatilization and because they are the main soils of new agricultural frontiers in Brazil [98,99].

The lower losses by NH3-N volatilization obtained with the use of ammonium sulfate and ammonium nitrate + calcium sulfate increased the corn yield in clayey soil. With a reduction in NH3-N losses, N will be used more efficiently by the plant, favoring the synthesis of biomolecules essential for corn growth and development. However, other characteristics of these N sources may also have influenced the yield and should be mentioned. For example, the higher concentration of leaf S provided by the application of ammonium sulfate may have contributed to the increase in corn yield, because S is closely linked to N metabolism, converting nonprotein N into protein [83]. Thus, all plant metabolism depends

on S compounds, due to the structural functions they perform, such as maintaining the active conformation of proteins through the disulfide bonds between methionine and cysteine (S-S) and metabolic functions, since they constitute amino acids, coenzymes, and proteins with Fe and S [3,100].

Not only did ammonium nitrate + calcium sulfate generate a low loss by NH3-N volatilization, but synergistic benefits for plant growth have been observed if NO3 - and NH4 <sup>+</sup> are provided together [80,101]. The beneficial effect of the simultaneous supply of these two inorganic forms of N occurs due to the lower suppression of the absorption of cationic nutrients, mainly Ca2+ and Mg2+, by the exclusive supply of NH4 <sup>+</sup> [80], lower acidification or alkalinization of the rhizosphere as a result of the absorption in excess of NH4 <sup>+</sup> or NO3 − [65], and lower energy requirements for NH4 <sup>+</sup> assimilation than NO3 − assimilation, given that NO3 − cannot be used directly by plants until it is reduced to NH4 +; a reduction catalyzed sequentially by nitrate reductase enzymes and nitrite reductase [102,103].

The N source PCU + Ur -NBPT + NI also increased the corn yield in clayey soil. However, this effect was not caused by a reduction in NH3-N volatilization, but by the supply of B via ulexite and synchronization of N release with PCU. Boron is responsible for plant functions such as sugar translocation and the regulation of carbohydrate and phytohormone metabolism. However, B plays a very important role in the metabolism of N. This is due to the requirement of B for the synthesis of the uracil nitrogen base, an essential component of RNA, which is also indispensable for ribosome formation and protein synthesis [38]. Therefore, an increased availability of B in soil favors higher yields, mainly because it is a nutrient found in low concentrations in tropical and subtropical soils, due to losses by leaching in the form of H3BO3 0.

Even given the inefficiency of PCU granules in reducing losses by NH3-N volatilization (Figure 5a), in most cases, controlled-release technology can decrease the availability of N at the beginning of corn development, when the absorption is still low [27], and increases the availability of N in phenological stages VT (tasseling) to R1, when the demand for N by corn is high [104]. This behavior in the later release stages of N was observed through the maximum daily loss of NH3-N in PCU granules alone (β), which occurred 30.6 days after topdressing fertilization in the clayey soil (Table 3); with enough time for the corn to be in VT, which usually coincides with the eighth week after emergence [105].

#### **5. Conclusions**

The losses by NH3-N volatilization were up to 46% of the N applied with urea. However, NI addition to urea increased N losses by NH3-N volatilization by 8.8 and 23.3% in relation to only urea for clayey and sandy soils, respectively. This leads to important implications for the use of NI as a mitigation tool for climate change in rainfed agriculture.

The nitrogen fertilizer technologies applied in a topdressing on clayey and sandy soil presented the following, in decreasing order, losses by volatilization of NH3-N: urea > URP + Ur-NBPT + IN > Ur-NBPT + Duromide > Ur-formaldehyde > ammonium nitrate + calcium sulfate > ammonium sulfate.

Soil with a clayey texture was 38.4% more responsive to nitrogen fertilization than soil with a sandy texture. The increase in corn grain yield in the clayey soil did not occur only due to the reduction in losses by NH3-N volatilization, other factors, such as S and B supplementation and N release at a controlled rate, to synchronize with the crop demand, also influenced the increase in corn yield. However, it is always advisable to choose N sources that increase crop yield, while generating the lowest possible losses due to NH3-N volatilization. In this study, these sources were ammonium sulfate and ammonium nitrate + calcium sulfate, which contributed to reductions in NH3-N emissions of 84 and 80% in relation to urea, respectively, thus favoring more profitable and sustainable agriculture.

**Author Contributions:** Conceptualization, B.M.A.R.C., M.R.B. and M.A.B.; data curation, B.M.A.R.C., W.D.K., C.R.M. and J.H.V.d.A.J.; formal analysis, B.M.A.R.C.; funding acquisition, R.S., T.T.I. and M.A.B.; investigation, B.M.A.R.C., M.R.B., W.D.K., C.R.M., J.H.V.d.A.J. and M.A.B.; methodology, B.M.A.R.C. and M.A.B.; project administration, B.M.A.R.C., R.S., T.T.I. and M.A.B.; resources, T.T.I. and M.A.B.; software, B.M.A.R.C.; supervision, B.M.A.R.C. and M.A.B.; validation, B.M.A.R.C. and M.A.B.; visualization, B.M.A.R.C. and M.A.B.; writing—original draft, B.M.A.R.C.; writing—review and editing, B.M.A.R.C. and M.A.B. All authors have read and agreed to the published version of the manuscript.

**Funding:** Coordination for the Improvement of Higher Education Personnel (CAPES—88887.482759/ 2020-00).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in the manuscript are the sole data, and no other data are linked with this data.

**Acknowledgments:** The authors would like to thank COCAMAR Cooperativa Agroindustrial, Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), Complexo de Centrais de Apoio à Pesquisa (COMCAP/UEM) and Grupo de Estudos em Solos da Universidade Estadual de Maringá-GESSO/UEM (in Portuguese).

**Conflicts of Interest:** There is no conflict of interest in terms of performing the research and publishing this manuscript.

#### **References**

