Extractant efficiency in the solubilization of alternative sources of potassium

Knowledge of a rock’s composition allows for inferences regarding several properties, ranging from its physical characteristics to its solubility. This study aimed to evaluate the use of different extractants to solubilize the K present in rocks as a potential source of nutrients and the effects of extractant contact time and temperature on rock solubilization. Samples of two rocks and a mineral concentrated from a granitic rock were treated with ammonium dihydrogen phosphate (NH4(H2PO4)), calcium dihydrogen phosphate (Ca(H2PO4)2), sodium hydroxide (NaOH) and water (control). Sample-extractant treatments were performed using a water bath shaker at temperatures of 25 and 50°C for periods of 3, 7, 10, 20, and 30 days. The amounts of K extracted from rocks using the extractants were in the following order: NH4H2PO4>Ca(HPO4)2>NaOH>water. The sequence of K release (ppm) based on the rocks studied was as follows: nepheline syenite>green banded argillite>concentrated biotite. Increasing the extractant contact time and temperature enhanced the solubilized K content.


Introduction
Stonemeal (rocks for crops) (Luz, Lapido-Loureiro, Sampaio, Castilhos, & Bezerra, 2010;Straaten, 2007) is defined as crushed rocks or materials containing naturally occurring soil fertilizers used as a soil amendment.The application of stonemeal to soils may represent an alternative to minimizing dependence on imported fertilizers in Brazil, which, in the specific case of potassium (K), is the world's fourth largest importer (Associação Nacional para a Difusão De Adubos e Corretivos Agrícolas (ANDA, 2012).
The use of unconventional and globally available geological K sources that can be weathered to provide sufficient K for agronomic benefits might be considered in agricultural systems in which K is the limiting nutrient and oxisols predominate (Manning, 2010), a common situation in Brazil.Whereas phosphatic rocks are feedstock for both conventional and unconventional phosphate fertilizers, potassium silicates, such as feldspars, biotite, phlogopite, and muscovite, and rocks with feldspathoids, such as leucite, nepheline and clayrich sediments, especially illite, are the basis for alternative potassium fertilization (Straaten, 2010).
Mineralogical knowledge of rocks as potential sources of plant nutrients is essential to predicting their weathering behavior according to the Goldich dissolution series (Goldich, 1938), which, in turn, aids in the development of methods to promote mineral dissolution or solubilization with a consequent nutrient release.Among the factors that should be considered regarding mineral dissolution are the extractant solution's effect on the rate of dissolution, solution pH, ionic strength, concentrations of individual elements, temperature, and reactive mineral surface (Lasaga, Soler, Ganor, Burch, & Nagy 1994).
The use of rocks as alternative sources of potassium in the form of stonemeal, combined with phosphate reagents or even phosphate rocks, whose properties can promote the solubilization of potassium and the release of phosphorus and other nutrients, such as nitrogen, in satisfactory quantities, represents an alternative to the use of commercial fertilizers for use in organic agriculture and to minimize environmental damage caused by mining wastes.
In the present study, we aimed to characterize the mineralogy and evaluate the effects of using different extractants, contact times, and temperatures on the solubility of K present in rocks considered to be potential alternative sources of this nutrient.
The rocks were first subjected to grinding in a rotating ball mill for 20 minutes, and the material was separated using a sieve of 0.354 mm (Brasil, 2007), repeating the procedure until all samples reached this granulometry.
Major oxides were analyzed by Acme Analytical Laboratories Ltd. (Vancouver) using methods FullSuite 4A (major oxides) and 4B (trace elements) on a inductively coupled plasma emission spectrometer (ICP-ES).Ignition losses were determined by the weight difference of the sample before and after heating to 1000°C (Table 1).
For the X-ray diffraction (XRD) analysis of rocks obtained by the powder method, XRD patterns were obtained on a Shimadzu XRD-6000 diffractometer operating at 40 kV with a current of 20 mA and CuK radiation with a graphite monochromator.The amplitude sweep was 3 to 70 (2θ), and the recording speed was 1.5° 2θ min -1 .
Scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDS) were performed at Geoscience Laboratories (Sudbury) using a Zeiss Evo SEM scanning electron microscope, model EVO 50.
The experiments were conducted in the laboratories of the School of Environmental Sciences at the University of Guelph.The rocks were initially crushed in jaw crushers, followed by a disk mill.Rock size separations were performed using a WS Tyler Ro-Tap Sieve Shaker, Model RX-29, to homogenize the particle size to less than 125 microns (0.125 mm).
For the solubilization experiment, protocols described in the literature (Zhou & Huang, 2006;2007) were followed with adaptations related to rock/extractant relationships, molar concentrations, rock-extractant contact times, and temperature.The procedure consisted of mixing 1 g of each rock with 10 mL of the following solutions: 1 mol L -1 ammonium dihydrogen phosphate (NH 4 (H 2 PO 4 )), 0.07 mol L -1 monocalcium phosphate (Ca(H 2 PO 4 ) 2 ), 0.1 mol L -1 sodium hydroxide (NaOH), and distilled water (control) in 15 mL centrifuge tubes.Pure chemical salts were used to prepare the solutions, and the term "extractant" will be adopted to refer to solubilizing potassium and other elements in the rocks and minerals studied.
Treated samples were subjected to agitation at 120 rpm in an Eppendorf water shaker bath, model Innova 3100, at a water temperature of 25°C for periods of 3, 7, 10, 20, and 30 days of rock-extractant contact.Each treatment was performed three times.For the extractant Ca(H 2 PO 4 ) 2 , a molarity that corresponded to the maximum salt solubility (1.8 g in 100 mL) was used.The second experiment consisted of the same conditions as the first, with the water shaker temperature at 50°C.
At the end of each contact time period, treated samples in centrifuge tubes were removed from the shaker.Separation of the extractant-rock samples was performed by centrifugation at 3,000 rpm for 20 minutes in a Thermo Scientific centrifuge model Heraeus Megafuge 16.Solubilized potassium (K sol) in extractants was determined by spectrometric flame atomic absorption (FAAS) on a Varian spectrometer, model SpectrAA 50, equipped with an air-acetylene flame atomizer and HLA-4S cathode lamp.
The soluble K (K sol) was calculated based on the amount of K released by the sample (mg kg -1 ), dividing this value by the total K amount present in the rock, as determined by the geochemical analysis (Table 1).The quantification of Al, Si, Fe, Ca, Mg, Mn, Zn, Sr, Ba, Cu, As, Cd, Pb, and Cr contents for treated samples under both temperatures at thirty days was performed by optical emission spectrometry with an induced plasma source (ICP-OES, Varian Vista Pro Model).Na was determined by FAAS.Ca was determined only for the extractant NH 4 (H 2 PO 4 ) and water.
The experiment was performed using a randomized block design.Soluble K values for extractants, temperatures, and reaction times for each rock were submitted to analysis of variance, and when relevant, means grouping by the Tukey probability test was performed.Regression models correlating the amount of solubilized potassium (dependent variable) versus contact time for each extractant at both temperatures were generated at SISVAR (Ferreira, 2011).
In the studied rocks, some minerals identified by XRD were visualized by SEM, and the chemical composition was also confirmed by dispersive X-ray spectrometry (EDX) of the mineral phases present (Figure 2).
For biotite, the minerals quartz, sodic and potassic feldspars albite and microcline and mica biotite were observed, with the last two considered as potential sources of K.
Syenite contained the minerals nepheline, sanidine and titanite.DRX also indicated the minerals biotite, analcime, aegirine and balliranoite (Figure 1).As K sources, the feldspathoids nepheline and balliranoite and biotite mica could be considered, with high weathering rates of the latter according to the Goldich dissolution series (Goldich, 1938).
In argillite rock, the minerals glauconite, phengite and illite, observed by DRX, constitute the main potential K sources (Figure 1C).Phengite, a dioctahedral mica similar to muscovite, and glauconite and illite belong to the phyllosilicate group.Glauconite and titanite were also observed by SEM according to EDX (Figure 2).

Extractants and K solubilization
The percentages of soluble K (K sol) of biotite, syenite, and argillite were significantly different under the effects of the extractors 1 mol L -1 NH 4 (H 2 PO 4 ), 0.07 mol L -1 Ca (H 2 PO 4 ) 2 , 0.1 mol L -1 sodium hydroxide (NaOH) and water at temperatures of 25C and 50°C (Figure 3).For the same extractor, there were also differences in the K sol between rocks.
The highest values of K sol for syenite and argillite were observed with the use of the extractant NH 4 (H 2 PO 4 ) at 25 and 50°C.The similarity between the ionic radii of potassium and ammonium and the high concentration of ammonium in solution (ionic strength) might have contributed to the displacement and solubilization of the K present on the outer surfaces and exchange points of the external surfaces of mineral components of the rocks.For biotite, the extractant Ca(H 2 PO 4 ) 2 promoted the release of more soluble K (Figure 3).
A hypothesis to explain the lower soluble K verified for biotite (< 1 mg kg -1 ) may be related to the phenomenon of Fe 3+ oxidation, promoting the release of some cations at octahedral sites, leaving them empty.This activity causes the reorientation of H + and OH -ions away from K, occasioning a high retention of K and increasing mineral stability (Melo, Castilhos, & Pinto, 2009).
For water, the highest K sol was observed for argillite, which is due to the mineralogical composition of this rock, with the phyllosilicates glauconite, phengite and illite (XRD).In phyllosilicates, the protonation of aluminol (Al-OH) and silanol groups (Si-OH) with ruptures of metaloxygen bonds followed by the dissolution of basic cations from interlayers and octahedral sheets predominates in hydrolysis reactions (Kampf, Curi, & Marques, 2009), releasing high amounts of K.
The average pH levels of the extractant solutions were 4.05 (NH 4 (H 2 PO 4 )), 3.5 (Ca(H 2 PO 4 ) 2 ), 12.8 (NaOH), and 6.0 (water).In the initial stage of mineral dissolution, the rates of surface complexation reactions increase with the increased concentration of protons (acidic conditions) and some ligands as well as under alkaline conditions (OH -groups) (Sokolova, 2013), with the minimum dissolution at pH values near the zero point of charge (pH ppzc ) from minerals (e.g., some oxides and aluminosilicates) (Oelkers et al., 2008;Rozalén et al., 2008;White, 2008).
Higher K sol was observed under acidic solution pH, with the extractants NH 4 (H 2 PO 4 ) and Ca(H 2 PO 4 ) 2 .Even with a pH difference of only 0.55, the high K sol observed for the more acidic extractant could be related to its ionic strength.
In mineral dissolution, a higher molar concentration (ionic strength) of reagents (extractants) can promote higher reaction velocities (law of mass action) and, therefore, higher element solubilization.The ionic strength of the extractants NH 4 (H 2 PO 4 ): 1 mol L -1 , Ca(H 2 PO 4 ) 2 : 0.07 mol L -1 , and NaOH: 0.01 mol L -1 implies higher solubilization rates for the NH 4 (H 2 PO 4 ) extractant, as was observed for K sol and other elements quantified (Table 2).
Means followed by the same lowercase letter for each extractant (rocks) and uppercase letter on each rock (extractants) do not differ by the Tukey test at 5% probability.
With respect to phosphorus extractants, H 2 PO 4 ions are the predominant form of phosphorous in soil at pH values near neutrality (5.5-6.5),whereas the HPO 4 2form present in the extractant Ca(H 2 PO 4 ) 2 prevails at higher pH values and is less soluble due to the presence of Ca 2+ (Lindsay, 2001;Mello & Perez, 2009).
In addition to the similar ionic radii of the ions NH 4 + and K + , the binding effect of phosphate ions (Zhou & Huang, 2006;2007) and the predominant ion form H 2 PO 4 -for NH 4 (H 2 PO 4 ) may also explain the higher soluble K + levels for this extractant.
Rock sample preparation (crushing and milling) may affect the reactive site density and distribution (Baere, François, & Mayer, 2015), and the presence of ultrafine particles on the mineral surface may promote an initial intense element release, decreasing mineral dissolution over time (Sokolova, 2013).
Increasing temperature results in an increase in soluble K for most rock-extractor combinations (Table 2).Elevated temperatures suggest a high energy (agitation) of ions and molecules dissolved in the solutions, enhancing the number of collisions at the mineral surface and occasioning the release of elements (including K).For each rock and extractant, means followed by the same lowercase letter (column) and uppercase letter (line) do not differ by the Tukey test at 5% probability.
The lack of differences in soluble K for biotite for the extractor NaOH with both temperatures and reaction times may be related to the pH of NaOH, as acidic conditions promote higher potassium release (Bray et al., 2015).
In a dissolution kinetics study of illite under the influence of phosphates (0.5 M Ca(H 2 PO 4 ) 2 at pH 2.5, 1 M NH 4 (H 2 PO 4 ) at pH 4.0, and 1 M (NH 4 ) 2 HPO 4 at pH 8.0) at temperatures of 25 and 45°C, the effect of H + on K + release was weaker for pH values greater than 4.0, with an increase in K + release for the NH 4 (H 2 PO 4 ) solution due to the combined effect of phosphate ions and H + .With increasing temperature, the differences in the K + release rates between the NH 4 (H 2 PO 4 ) and (NH 4 ) 2 HPO 4 solutions decreased, with no differences observed at a temperature of 45°C, with phosphate ions being the most important factor for K + release at high temperatures (Zhou & Huang, 2007).
The effect of phosphate ions with increasing temperature resulted in a greater solubility of syenite with the extractant NH 4 (H 2 PO 4 ) in the present study, whereas for other rocks and extractants, no effects of ions and temperature on K sol were observed, and there was even a reduction in K sol with increasing temperature in some treatments.
In the present study, argillite rock, similar to Verdete, presented a maximum K sol value of 0.0234 mg kg -1 for the extractant NH 4 (H 2 PO 4 ).The reduced evaluation time (30 days) and lack of sequential extractions on samples after each contact time may explain the reduced values compared to the work cited above.
The relationship between contact time and K sol (mg kg -1 ) was linear for biotite and the extractant NH 4 (H 2 PO 4 ), with coefficients of determination (R 2 ) of 0.9541 and 0.9874 for temperatures of 25 and 50°C, respectively, with more soluble K at 25°C (Figure 3A).
For the extractants Ca(H 2 PO 4 ) 2 and water, respectively, at 50°C, linear and quadratic correlations best described the amount of K sol with reaction time (Figure 3B).In addition to the negative effect of increased pH on dissolution rates, the grain size, superficial mineral area and chemical mineral composition may also influence the release of K and other nutrients, such as Fe, Mg and Al, from biotite (Bray et al., 2015).
Quadratic equations best described the relationship between K sol and contact time for the extractant NH 4 (H 2 PO 4 ) at both temperatures (25 and 50°C) for syenite.More K sol was observed at 25°C (Figure 4A).The extractants NaOH (25°C) and Ca(H 2 PO 4 ) 2 (50°C) showed linear and quadratic relationships between the variables, with R 2 values of 0.8162 and 0.7682, respectively (Figure 4B).A linear K release over the period of 1 year was observed for the feldspathoid nepheline at pH 5, with reduced K release at pH 7. The K release from nepheline (38 meq g -1 ) was an order of magnitude higher than for K feldspars (lower than 5 meq g -1 ).Rocks containing nepheline are more effective K sources than rocks with K feldspar alone (Manning, 2010).The presence of nepheline in syenite rock (Figure 1) and the effects of extractant pH effects on this mineral might also have influenced the amounts of K released from this rock (Figure 5).The relationship between K sol and contact time for the extractants NH 4 (H 2 PO 4 ) and NaOH, at both temperatures studied, for argillite rock were best described by quadratic equations, with R 2 values of 0.9975 and 0.9986 for NH 4 (H 2 PO 4 ) and 0.9854 and 0.9697 for NaOH at 25 and 50°C, respectively (Figures 6A and B).For the extractants Ca(H 2 PO 4 ) 2 and water, linear and quadratic equations best described the relationships between K sol and contact time (Figure 6C).
For the extractant NaOH, an initial high release of K, decreasing with contact time, suggests K solubilization from external surfaces or exchange points from the clay minerals composing argillite or even the presence of ultrafine particles on the mineral surface.
The effects of acidic pH and phosphate extractants on K solubilization were verified for illite (Zhou & Huang, 2007), a mineral composing argillite rock (Figure 1C), which may explain the higher solubilization for the extractants NH 4 (H 2 PO 4 ) and Ca(H 2 PO 4 ) 2 .The amounts of the main mineral elements included in the studied rocks, quantified at 3 (initial) and 30 d (final) of contact time for the extractants NH 4 (H 2 PO 4 ), Ca(H 2 PO 4 ) 2 , and water at 25 and 50°C, are presented in Table 3.
The extractor NH 4 (H 2 PO 4 ) produced higher solubilization of the elements present in the mineral components of the rocks under study.For syenite, high amounts of calcium (Ca), magnesium (Mg) and sodium (Na) were released for the extractant NH 4 (H 2 PO 4 ) along with a high release of Mg and Na for extractant Ca(H 2 PO 4 ) 2 .
The feldspathoids nepheline and balliranoite, zeolite analcime and feldspar sanidine, which include Na in their chemical composition, may be the minerals that release Na, making it available in solution.Minerals proposed to promote the release of Ca and Mg include balliranoite (feldspathoid), aegirine (pyroxene) and biotite (mica).Based on the mineralogical composition of syenite, a sequence of mineral weathering was proposed based on the Goldich dissolution series: feldspathoids> pyroxenes>micas>zeolites>feldspars.
In the analysis of biotite, for the extractants NH 4 (H 2 PO 4 ) and Ca(H 2 PO 4 ) 2 , considerable amounts of Fe and Mg were present in solution, indicating the weathering of this iron-magnesium phyllosilicate.For the NH 4 (H 2 PO 4 ) extractant, the high amounts of Ca might be related to alteration of the mineral albite, which, compared to other minerals present in the modified biotite rock, such as quartz and microcline, presents lower resistance to weathering according to the Goldich dissolution series.
Other elements that might also be solubilized by the extractants used in this study, such as As, Cd, Pb, Cr, Hg, Ba, and Cu, showed readings below the device detection limit and were not considered in this work.The levels of these elements were below the maximum limits allowed for K mineral fertilizer, according to Normative Instruction No. 27 (Brasil, 2006).
In practical terms, the agronomic application of this rocks to soybean culture as fertilizers, with recommended doses of 38 kg of K 2 O per ton of grain produced (Junior, Castro, Oliveira, & Jordão, 2013) based on the maximum values of soluble K (mg kg -1 ) of rocks studied, would require 4,920,000 tons of biotite, 529,321.6 tons of syenite or 1,714,801.4tons of argillite to supply soybean demand, whereas for KCl fertilizer, less than 0.5 ton is used (0.063 ton).
Considering the high masses of rocks to be applied and the economic costs of both material transport and the processes used to obtain the rocks (e.g., concentrated biotite), the proposal is to adopt these rocks as complementary fertilizers rather than substitute them for conventional fertilizers.Knowledge of the mineralogy of rocks and the main mineral contributors to the release of elements (nutrients) can suggest potential alternative sources for use in soil fertilization and plant nutrition.Future experiments under field conditions with combinations of different rocks, extractors and plants for long-term evaluation periods are suggested.

Conclusion
The contact time, temperature, and extractants enhanced K solubilization from the rocks studied.
The extractants that promoted greater solubility according to the content of soluble potassium (K sol) in biotite, syenite, and argillite were, in the following order, NH 4 (H 2 PO 4 )>Ca(HPO 4 ) 2 > NaOH>water.
The sequence of K release (ppm) based on the rocks studied was as follows: nepheline syenite>green banded argillite>concentrated biotite.

Table 2 .
Effects of contact time (per extractant) at each temperature on the percentage of soluble potassium (mg kg -1 ) for biotite, syenite, and argillite.