Reversibility of the Hardening Process of Plinthite and Petroplinthite in Soils of the Araguaia River Floodplain under Different Treatments

Although ferruginous materials occur frequently in soils of tropical regions, information about the reversal of the hardening process of these materials is scarce. This study assessed the influence of different chemical treatments and periods of immersion on the reversibility of the hardening process of plinthite and petroplinthite in soils of the Araguaia River plain. Soil samples were collected from the plinthic horizons in 0.10 m high and 0.15 m diameter PVC cylinders and divided into subsamples with a rock hammer. Homogeneous petroplinthite samples were also collected and broken into subsamples with a rock hammer. The plinthite and petroplinthite subsamples were subjected to five immersion treatments: distilled water, calcium carbonate solution, sodium hydroxide solution, sodium hydroxide solution + sodium hexametaphosphate, and acidic solution. The subsamples were immersed for 20, 40, 80, and 160 days. The dispersion and stability degrees and compressive strength in these subsamples were assessed. The wet aggregate stability test indicated no impact on the structural stability of plinthite and petroplinthite subsamples subjected to the treatment with different chemical agents, demonstrating the non-reversibility of the hardening process of these materials of the Araguaia River floodplain, under the studied conditions.


INTRODUCTION
Plinthite occurs in soils of flat or slightly sloped areas, rich in iron, and exposed to seasonal variations of the groundwater level.It is the result of subsurface accumulation of iron hydroxides, kaolinite, and quartz, which may undergo irreversible hardening and consequently be transformed into petroplinthite (Daniels et al., 1978;Gardi et al., 2014).Thus, petroplinthite is an extremely consolidated and firm material when moist and extremely hard when dry (Alexander and Cady, 1962;Santos et al., 2013).
The progressive hardening of ferruginous materials is related to their higher iron content that results in increased crystallinity (Santos and Batista, 1996).Soil iron can originate from other horizons, higher sites or result from the elimination of silicon cations and base ions through leaching of compounds dissolved by weathering (Gardi et al., 2014;Soil Survey Staff, 2014).Iron is transported by water under anaerobic conditions as ferrous ion and precipitates in soft clayey concretions of iron oxides that harden after drying (Gardi et al., 2014).
Ferruginous crusts containing only Fe and Al oxides do not provide plants with nutrients and have low water retention capacity (Alexander and Cady, 1962).On the other hand, ferruginous crusts with kaolinite contain water and channels that can be penetrated by roots.Some laterites contain encapsulated weathering minerals, such as micas and feldspars which, when ruptured, can provide nutrients.The presence of vegetation in lateritic areas is an important factor to reverse the hardening process of this material, which is caused by reduction and chelation with organic matter, by the mechanical action of roots, and mitigation of temperature and drought (Alexander and Cady, 1962).Coelho et al. (2001) studied the development of nodular ferricretes (iron-rich crusts) in sandstone of the Bauru Group, and concluded that the cortex of the petroplinthic profiles was yellowish, suggesting dissolution and removal of hematite, favoring the presence of goethite.This fact characterizes the destruction of the ironstone layer and may be related to climate or lithological factors, indicating that these conditions contrast with paleoclimatic conditions that favor ferricrete formation, such as prolonged droughts or high iron contents.
Apart from some studies on plinthite, Plinthosols (WRB, 2014) and Plintossolos [Brazilian Soil Classification System (Santos et al., 2013)] and related issues (Daniels et al., 1978;Zhang and Karathanasis, 1997;Ogunwole et al., 2001;Eze et al., 2014), there is very scarce or no reliable information at all on the reversal of hardening of ferruginous materials, which occurs frequently in soils of tropical regions.This study addressed the impact of different chemical, mechanical, and immersion treatments on the reversibility of plinthite and petroplinthite hardening in soils of the Araguaia River floodplain, to contribute to a more accurate classification of these materials.

MATERIALS AND METHODS
Plinthite and petroplinthite samples from diagnostic horizons of two profiles (P2 and P5, respectively) of soils of the Araguaia River floodplain and surroundings, in Luiz Alves, district of São Miguel do Araguaia, Goiás (Table 1) were used in the study.The profiles were selected based on field observations related to the occurrence of these materials in the soil, and were described according to the recommendations of the Manual of Field Soil Description and Sampling (Santos et al., 2015) and the Technical Manual of Pedology (IBGE, 2015).The studied soils were classified according to the Brazilian Soil Classification System (Santos et al., 2013) and World Reference Base for Soil Resources (WRB, 2015), to characterize each sampling point.
Regional climate is type AW, defined as tropical with dry winters, according to the Köppen-Geiger classification system.The average annual rainfall ranges between 1,800-2,000 mm and the average annual temperature between 25-26 °C (Cardoso et al., Rev Bras Cienc Solo 2018;42:e0170191 2014).Profile P2 lies on the floodplain and possibly on paleo-channels, owing to its physical nature (mean texture) (Table 2), while P5 lies outside the floodplain, approximately 6.40 km away from Luiz Alves, towards São Miguel do Araguaia, in a higher adjacent area with good natural drainage.
From profile P2, due to its plinthic horizon, undisturbed samples were collected in PVC cylinders (height 0.10 m, diameter 0.15 m), inserted directly into the plinthic horizons (4Btfc2 and 4Btfc3, to a depth between 0.61 and 1.10 m), with three replicates.Afterwards, in a protected environment, plinthite subsamples were randomly extracted by hand from the collected samples, according to visual (color) and physical or mechanical (consistency) criteria (IBGE, 2015).After division, the consistency of most plinthite samples ranged from soft to hard (IBGE, 2015;Santos et al., 2015), when manually examined.
In profile P5, three homogeneous samples of petroplinthite were collected in the form of continuous blocks (lateritic concretion, "canga"), weighing approximately 400 g per sample.The most distinctive criterion was the great hardness (very hard consistency in dry sample).Petroplinthite subsamples with a mean volume of 7 cm 3 were collected from these profile samples in the laboratory, using a rock hammer.
The plinthite and petroplinthite subsamples of both profiles were individually weighed in semi-analytical scale, placed in plastic containers and subjected to five wetting treatments for 20, 40, 80, and 160 days, which simulate field situations, conditions promoted by use and soil management, and alterations that may occur in the laboratory, to investigate  Coarse sand, fine sand, silt, and clay determined by densimeter method (Donagema et al., 2011).CS = coarse sand; FS = fine sand; TC = textural class; CL = clay loam; SL = sandy loam; L = loam; SCL = sandy clay loam; C = clay.
the effect of the different time periods on these materials.The five treatments consisted of: DW -distilled water, simulating normal wetting of soils with plinthic materials caused by flooding or rainfall; CC -calcium carbonate solution (3.33 g L -1 calcium carbonate), simulating a management with liming; SH -10 % sodium hydroxide (40.00 g L -1 sodium hydroxide) solution, simulating the application of conventional soil dispersants; SH + HM -5 % sodium hydroxide solution + 5 % sodium hexametaphosphate (40.00 g L -1 sodium hydroxide + 35.70 g L -1 sodium hexametaphosphate), also simulating the application of conventional soil dispersants; AS -acidic solution (6.80 g L -1 monopotassium phosphate), theoretically simulating a characteristic pH range of soils with plinthic horizons in Brazil, of around 5.0, for maintaining the buffering of the soil solution.The pH of the soils was determined in water and 1 mol L -1 KCl, at a soil-liquid ratio of 1:2.5.During the study period, the pH of the solutions was measured weekly and the mean pH value calculated for each solution at the end of the assessment period (Table 3).
For each immersion treatment, the following properties were assessed sequentially in the subsamples: material dispersion degree, stability and compressive strength of plinthite and petroplinthite, with 10 replications (10 subsamples).Because of the great variability among the materials of the same horizon (highly irregular populations), a selective mean was used for assessment, after exclusion of 30 % of the most dispersed data from the general mean.
To analyze the degree of dispersion, plinthite and petroplinthite subsamples were removed from the solutions, transferred to aluminum containers and oven-dried at 105 °C.The subsamples were then weighed, and the percentage dispersion was calculated using the equation 1: Eq. 1 in which: W0 is the initial sample weight (g) and Wf the final sample weight (g), corresponding to the material retained in 2-mm-diameter mesh.
Proportional stability degree of plinthite and petroplinthite subsamples was analyzed by wet sieving in a Yoder wet-sieving apparatus (Donagemma et al., 2011).The subsamples were individually weighed, pre-moistened by capillarity, placed on moist filter paper, transferred to two sets with five sieves (diameter 2.00; 1.00; 0.50; 0.25; and 0.106 mm) and subjected to vertical oscillation in water containers for 15 min (40 cycles min -1 ).The material retained in each sieve was transferred to aluminum containers and oven-dried at 105 °C.Subsequently, the weight and percentage of stable materials in each diameter class were calculated.
Rev Bras Cienc Solo 2018;42:e0170191 (N), was reached when the equipment deformed constantly 40 % compared to the initial height of the material, at a speed of 0.50 mm s -1 .Due to the irregular shapes, the plinthite and petroplinthite subsamples were measured with a digital caliper, taking three measurements (height, width, and length) to calculate the approximate volume of each sample.Compression test values were expressed in N cm -3 .
Qualitative and quantitative analyses of the chemical composition of representative samples of plinthite and petroplinthite were determined by X-ray fluorescence spectrometry (XRF) to facilitate the characterization of the material.The individual samples were pulverized in an agate mortar and pelleted, and then compacted for 1 min in a 5 ton hydraulic press.The pressed pellets were analyzed in an EDX-700 HS Energy Dispersive X-ray Spectrometer, operated at the following conditions: collimator -10 mm; atmosphere -vacuum; tube voltage -50 kV; tube flow -automatic μA; irradiation time -100 s; acquisition mode -quantitative/FP; analytical line -Kα.
The mean values of dispersion, stability, and compressive strength of a total of 500 subsamples were determined, as follows: 250 subsamples of plinthite and 250 subsamples of petroplinthite.The comparisons were performed using non-parametric statistics, based on the mean confidence interval (CI = 95 %; p<0.05) and the mean standard error (± SE) was used to test the reliability of the calculated sample mean, using software Xlstat (Addinsoft, 2016).The data were organized in figures and tables with Microsoft Excel 2010.

Degree of dispersion of plinthite and petroplinthite in response to different treatments and immersion periods
Dispersion of ferruginous materials under different immersion treatments is generally affected (Figure 1).Petroplinthite indicated less dispersion, i.e., a smaller amount of material was dispersed than for plinthite, in all treatments.This result is mainly related to the higher content of Fe oxides in petroplinthite (Table 4).Iron oxides are cementing agents and, therefore, do not facilitate the dispersion of ferruginous materials (Donagemma et al., 2003;Rodrigues et al., 2011).Factors such as isomorphic substitution (Fe for Al) and crystallinity degree of Fe oxides possibly contributed to the low degree of dispersion (Santos and Batista, 1996;Ker, 1997).
For both materials, the highest dispersion degree was observed in the subsamples immersed in sodium hydroxide solution + sodium hexametaphosphate, however not differing from petroplinthite in the sodium hydroxide treatment (Figure 2).This result is associated to the action of the reagents, sometimes in increasing the pH of the solution (sodium ions) and sometimes resulting in negative charges (phosphate ions) on the surface of the oxides by the reaction of ions with the hydrogen of the functional groups, with consequent improvement in dispersion (Rodrigues et al., 2011).
Plinthite and petroplinthite are not completely inert and react or are dispersed by hydrolysis, indicating a certain degree of reversibility of the hardening process when subjected to the action of different agents.Medium acidification (use of acid solution) did not induce a high degree of dispersion of plinthite and petroplinthite.In this case, both materials are more stable at these acidity levels, close to those of their formation environment (Table 3).Regarding plinthite, there was also a considerable degree of dispersion after immersion in calcium carbonate and sodium hydroxide solutions.These results indicate that in limed or salinization-prone environments, the dispersion process of plinthite or petroplinthite can be accelerated.
Our studies did not allow the identification of the sections of plinthite and petroplinthite subsamples with highest dispersion degree, since the hardness, composition, and arrangement of these materials in a given soil horizon is variable, due to the great variability Rev Bras Cienc Solo 2018;42:e0170191 in the subsamples.In spite of the precautions taken with the separation (division) of the materials, the compositions and shapes of plinthite and petroplinthite subsamples were variable.Therefore, it is believed that the dispersion process observed in these materials is primarily related to forms of petroplinthite in more recent development stages or with lower Fe contents, which occur in each sampling unit.This fact also explains the larger amount of dispersed material of plinthite than petroplinthite.
The immersion periods of each treatment had no effect on the dispersion of plinthite and petroplinthite, but the dispersion values between the materials differed (Table 5).
The variations observed among the different periods for each treatment are attributed to the differences in the natural composition of each subsample.
In short, the duration of immersion had no effect on the amount of dispersed material in all treatments.Therefore, the time factor has little or no effect on the dispersion of plinthite and petroplinthite in comparison with the reagent factor.

Degree of stability of plinthite and petroplinthite in response to different treatments and immersion periods
Regarding the degree of stability of the plinthite and petroplinthite subsamples previously subjected to the dispersion test, both remained practically unchanged when the subsamples were passed through a 2.00 mm mesh sieve, with little distribution when sieved through smaller mesh.This indicates high stability of the ferruginous materials examined here (Table 6).According to Momoli and Cooper (2016), the cementing action of Fe and Al oxides accounts for the high cohesion between particles, favoring stabilization, which confirms our results.Practically no reversibility of the hardening process of the ferruginous materials was observed in response to the action of different chemical agents, suggesting that this process is irreversible for the region and for the tested periods.However, a small fraction of material detached from the plinthite and petroplinthite subsamples was retained when a smaller mesh sieve (<0.106 mm) was used, and this fact cannot be dissociated from the action of chemical dispersants, with a higher or lower impact on the structural stability of the less cemented minerals present in these materials.Given that the plinthite and petroplinthite units were not entirely purified or uniform, i.e., not free from portions of the soil matrix or mottles adhered internally, it is believed that this impact on the structural stability is related to the action of chemical agents on portions with less or less efficient "welding", an inherent property of this type of material, and thus the most "mature" portions, i.e., the proper plinthite and petroplinthite, remained unchanged.
Regarding the effect of the immersion periods, no behavioral trend could be detected based on the results (Table 7).The small differences observed in the assessed materials can be attributed to the natural variability in composition.

Compressive strength of plinthite and petroplinthite under different treatments and immersion periods
The compressive strength applied to the plinthite subsamples was much lower than that applied to the petroplinthite subsamples in all treatments of this study (Figure 3).This finding was already expected, due to the lower total Fe contents of plinthite, related to a more recent genesis.According to Garcia et al. (2013), plinthite also has secondary minerals in the hardening process, such as not completely crystalline kaolinite or Fe oxides, which are still susceptible to dispersive or disaggregating action.In ferruginous materials, Santos and Batista (1996) found values that allow concluding that the compressive strength applied to the plinthites of this study reflect a negative association between the most crystalline Fe forms and the hardening degree of these structures.Similarity, Martins et al. (2018) verified a strong correlation between the Fe contents and the compressive strength applied in plinthite samples of soil profiles of the Araguaia river floodplain.  1) The stability of plinthite and petroplinthite subsamples was analyzed through wet sieving on a Yoder shaker apparatus (Donagema et al., 2011).
Rev Bras Cienc Solo 2018;42:e0170191 In the case of plinthite, the lowest compressive strength was applied to the subsamples immersed in distilled water and in solutions of calcium carbonate, sodium hydroxide, and sodium hydroxide + sodium hexametaphosphate, without differences between them (Figure 4a).The highest compressive strength (greatest hardness) was observed in the subsamples immersed in acid solution.Thus, it can be seen that plinthite stabilizes in acid solution, or rather, remains stabilized, because the acidity of the medium is similar to the natural conditions of the site of formation, and consequently will not favor its softening (Table 3).On the other hand, the other wetting treatments, all with a pH ranging from neutral to alkaline, indicated a higher softening response, although not directly proportional to the alkalinity of the medium.
For petroplinthite, the lowest compressive strength was observed in the subsamples immersed in distilled water, however not differing from the calcium carbonate solution (Figure 4b).This result may at first indicate that unlike plinthite, petroplinthite is slightly more vulnerable to the action of distilled water (reduced environment) or to low-alkalinity solutions, close to neutrality, and unsusceptible to acid or highly alkaline solutions.Similar results were observed by Alexander and Cady (1962) in Sierra Leone, Africa, where it was found that the softest laterite was found near creeks, due to the higher availability of moisture and vegetation cover.
Considering data of the strength applied to petroplinthite subsamples in the different treatments (Table 8), however, the differences can be explained by differences in the composition of the subsamples rather than by chemical reactions to the treatments, due to the heterogeneity of strengths within a treatment and among treatments as well as the proximity of the values.Moreover, the studied petroplinthite material is completely stabilized and less vulnerable to the action of dispersing agents because its formation occurred in the past.Petroplinthite also has a longer and more uniform maturation stage, which results in higher cotent and crystallinity of Fe and Al (Miguel et al., 2013).
In general, there was no effect of the duration of immersion on the decrease in the compressive strength applied to plinthite and petroplinthite in the treatments.The great variability in compressive strength between immersion periods is related to the great natural variability observed in the composition of subsamples from a same soil horizon and to the impossibility of selecting subsamples with the same degree of purity because of their different stages of development, consistent with observations by Santos and Batista (1996) and Martins et al. (2018).However, Alexander and Cady (1962) and    1) 95 % confidence interval (p<0.05).DW: distilled water; CC: calcium carbonate; SH: sodium hydroxide; SH + HM: sodium hydroxide + sodium hexametaphosphate; AS: acid solution.
Rev Bras Cienc Solo 2018;42:e0170191 Daugherty and Arnould (1982) claimed that the conditions required for the hardening of plinthic materials are extremely variable, particularly the amount of time required, in disagreement with the results of this study.

CONCLUSIONS
Hardness of plinthite and petroplinthite, particularly the former, is susceptible to the action of different chemical agents, especially with alkaline character.
Immersion of plinthite samples in calcium carbonate and sodium hydroxide solutions induced considerable dispersion, suggesting that in limed or salinization-prone environments, the dispersion process of plinthite or petroplinthite can be accelerated.
Structural stability of the plinthite and petroplinthite subsamples, exposed to the action of different chemical agents, was not degraded in the wet aggregate stability test, indicating no reversibility in the hardening process of the materials of the Araguaia River floodplain, under the conditions of this study.
Immersion periods in all treatments did not affect the compressive strength applied to the plinthite and petroplinthite samples.

Figure 3 .
Figure 3. Mean value and standard error of the mean compressive strength of plinthite and petroplinthite of the studied profiles, immersed in distilled water (a), calcium carbonate (b), sodium hydroxide (c), sodium hydroxide + sodium hexametaphosphate (d), and acid solution (e) (95 % confidence interval; p<0.05).

Table 1 .
Location of the profiles and characterization of the study

Table 2 .
Horizon layers, grain size composition, and textural class of the studied profiles

Table 4 .
Main chemical components of plinthite (PL) and petroplinthite (PP) determined by X-ray fluorescence spectrometry (XRF) for each studied profile

Table 5 .
Mean and standard error of the mean dispersion of plinthite and petroplinthite subsamples in the soil profiles, immersed in different solutions for different periods Rev Bras Cienc Solo 2018;42:e0170191

Table 6 .
Mean value and standard error of the mean stability of the plinthite and petroplinthite subsamples in the soil profiles, immersed in different solutions

Table 7 .
Mean value and standard error of the mean stability of plinthite and petroplinthite of the profiles, immersed in different solutions for different periods