3.1. Properties of Biochars
The chemical properties of the two fruit waste biochars (mangosteen shell and pineapple peel) were similar, as shown in
Table 3. The pH was slightly alkaline and EC values were high, indicating that these biochars contained much soluble salts. Total K concentrations were high and total Ca, Mg, P, and S concentrations were moderate to high. Eucalyptus wood biochar differed considerably from fruit waste biochars as it was highly alkaline, had low EC, much less K, and more Ca.
The proportions of elements soluble in water, as shown in
Table 4, were quite diverse for these biochars. Little (<21%) of the Ca, Mg, and P was soluble, whereas larger to major proportions (26–63%) of the K and (0.17–63%) of the S were soluble. The K/Cl ratio (0.69, 0.87) in the water extract for fruit waste biochars was consistent with much of the soluble K being in water-soluble KCl, whereas the higher ratio (1.57) for eucalyptus wood biochar indicates that some of the soluble K was not in KCl.
X-ray diffraction patterns, as shown in
Figure 1, of original biochars show that mangosteen shell and pineapple peel biochars contain sylvite (KCl), which is consistent with their high K concentrations (2.9%, 3.3%, respectively). Sylvite was not detected in the eucalyptus wood biochar. Calcite (CaCO
3) was abundant in eucalyptus wood biochar [
11], which is consistent with its high Ca (1.4%) content. No XRD peaks for other K or Ca minerals or for Mg, Mn, P, or S minerals were detected, which is partly a consequence of the low sensitivity of the XRD technique and the high background due to amorphous carbon [
11]. It must be noted that water-soluble elements were determined in finely ground biochar and would not indicate the elements that would dissolve from biochar particles.
XRD patterns of mangosteen shell and pineapple peel biochars particles recovered after various times in the soil show that all sylvite had left the biochars within one month. The calcite in eucalyptus wood biochar remained after eight months. All biochars had acquired quartz from the soil during incubation.
3.2. Changes in the Chemical Composition of Biochar during Incubation
Changes in the amounts of soluble and insoluble elements in biochars during the eight months of incubation are shown in
Figure 2,
Figure 3 and
Figure 4. The points in the graphs represent average data for the nine soils with standard deviation values shown as error bars. The error bars are small, indicating that there were no major differences in amounts of elements lost for the different soils.
For all three biochars, most of the soluble and insoluble K had been lost from the particles within the first month of incubation in soil, with continuing losses up to eight months. Readers should note that the term insoluble K indicates that K was not extracted from ground biochar during an extraction in water possibly because it was protected from dissolution within the microporous fabric of biochar. This insoluble K may be present in water-soluble compounds, but these are located in protective pores within the microporous fabric of biochar. The same nomenclature is used for other elements [
12].
Amounts of insoluble Ca increased moderately or remained constant during incubation, whereas amounts of water soluble Ca increased greatly, although remaining much smaller than the amounts of insoluble Ca. Thus, the biochars had absorbed Ca from the soils. The soils contained considerable exchangeable Ca, as shown in
Table 2, some of which may have been replaced by K released from biochar. Insoluble Mg remained almost constant for all biochars over the eight month incubation period, as shown in
Figure 3. Water soluble Mg increased greatly for mangosteen shell and pineapple peel biochars, but remained almost constant for eucalyptus wood biochar, as shown in
Figure 3. Water-soluble Mg was much less than insoluble Mg for all biochars, so some of this additional water soluble Mg could have been derived from the initially insoluble Mg fraction rather than by sorption from soil. Insoluble Mn decreased greatly within one month for mangosteen and eucalyptus biochars, while remaining almost constant for pineapple peel biochar. Water soluble Mn increased over time for all three biochars but was a much smaller amount of Mn than insoluble Mn. Much of the initially insoluble and soluble P remained in all three biochars after eight months of incubation, as shown in
Figure 4. Much of the insoluble S remained in mangosteen shell and pineapple peel biochars, whereas insoluble S increased greatly for eucalyptus wood biochar. Water soluble S remained almost constant for mangosteen shell and eucalyptus wood biochars and decreased greatly for pineapple peel biochar.
The results shown in
Figure 2,
Figure 3 and
Figure 4 resemble those reported from a similar experiment by Limwikran et al. [
4], for which incubation of the same biochars was limited to two months. In the present work, the much longer (8 months) incubation time has resulted in greater losses of soluble K and insoluble K from all three biochars. The gains in water soluble and insoluble Ca by all three biochars continued from two to eight months. The losses of insoluble P and the small change in soluble P for mangosteen shell biochar persisted for 8 months. For pineapple peel biochar, the relatively minor loss of insoluble P and moderate loss of soluble P continued from two to eight months. For eucalyptus wood biochar, the absence of systematic changes in soluble and insoluble P concentrations persisted over 8 months.
The divergent behavior of different elements and of different biochars is probably associated with the various minerals containing these elements that are present in biochar, as is investigated below. It is evident that the limited release of some plant nutrient elements from biochar to soil is quite different from the rapid loss of most K and P from chemical fertilizers [
13,
14,
15].
The slower release of nutrients from biochar might be beneficial in sandy soils that experience rapid leaching of nutrients from the root zone, but in general, the limited release over eight months of several nutrient elements is likely to greatly reduce the effectiveness of biochar as a fertilizer.
3.3. SEM/EDS Analyses of Biochar Particles Recovered from Soil
The diverse dissolution behaviors of plant nutrient elements within biochars incubated in soil are at least partly due to the various mineral forms of these elements and their locations within the microfabric of biochar particles. These two aspects were investigated by scanning electron microscopy (SEM) and energy dispersive X-ray spectrometry (EDS) analysis of the interior of recovered biochar particles after zero, one, two, four, and eight months of incubation in the soil. Average data for nine soils and three biochars after zero, one, two, four and eight months of incubation are shown in
Figure 5. Spectra were quite variable both for different fracture surfaces of individual biochar particles and biochars from different soils. Consequently, the error bars in
Figure 5 are often quite large, but the major trends are still clear. For all three biochars, the relative abundance of K had decreased greatly during one month of incubation, with further losses up to eight months of incubation. This trend mirrors the changes in insoluble and soluble K discussed earlier, as shown in
Figure 2. The rapid decrease in the abundance of K is matched by the decrease in Cl for pineapple peel and mangosteen shell biochars and is associated with the dissolution of sylvite (KCl), the mineral identified by XRD analysis, as shown in
Figure 1. Eucalyptus wood biochar also lost much K, but there was no associated loss of Cl, which is consistent with XRD not detecting sylvite in this biochar, and K was mostly present in other minerals [
2]. Much of the K remaining in all three biochars after one month was likely to be in one or more of the moderately to poorly soluble K-compounds that are common constituents of burnt plant materials (e.g., fairchildite (K
2Ca(CO
3)
2), archerite ((K,NH
4)H
2PO
4), pyrocoproite ((Mg(K,Na))
2P
2O
7) [
16]). These minerals generate weak XRD patterns so would not be evident in
Figure 1, but highly sensitive synchrotron XRD has been used by other researchers to identify pyrocoproite and archerite in biochars produced from tropical plant wastes [
2].
For mangosteen skin and pineapple peel biochars, the relative amounts of Ca and Mg exposed on fracture surfaces increased after two months of incubation, which is consistent with the values of insoluble and water soluble Ca and Mg shown in
Figure 2 and
Figure 3. Eucalyptus wood biochar contained much Ca, and its relative abundance on fracture surfaces decreased with incubation period due at least partly to increasing amounts of clay (Al, Si, Fe) and quartz (Si) diluting the observed concentration of Ca. The relative amount of Mg remained almost constant as did insoluble and soluble Mg, as shown in
Figure 3.
For pineapple peel biochar, the relative amounts of P and S determined by EDS analysis of fracture surfaces remained approximately constant as did insoluble P and S, as shown in
Figure 4; although, water soluble P and S, which represent about half of the P and S in pineapple peel biochar, decreased in abundance. For mangosteen shell biochar, the relative amounts of P and S determined by EDS on fracture surfaces had increased after eight months; whereas insoluble P and S had decreased to a moderate extent and water soluble P and S (minor proportions) remained approximately constant, as shown in
Figure 4.
For eucalyptus wood biochar, the relative amount of S on fracture surfaces decreased; whereas the insoluble S increased over time, as shown in
Figure 4. It is likely that the high initial S value (5%) calculated from EDS spectra for original eucalyptus wood biochar is too high and the variation in this value for different surfaces is very large, indicating a large range of S concentration existed (±3%). There was no systematic trend over time for the amount of P in fracture surfaces of eucalyptus wood biochar, which is consistent with the amounts of insoluble and soluble P present in this biochar, as shown in
Figure 4.
We conclude, from these integrated EDS analyses of the many compounds on fracture surfaces of fruit waste biochars recovered from soils, that rapid dissolution of sylvite (KCl) had greatly reduced the amounts of K and Cl in mangosteen shell and pineapple peel biochar particles. Much K was also lost from eucalyptus wood biochar, but there was no associated loss of Cl, so K compounds other than sylvite had dissolved. At the same time, the amounts of Ca and Mg increased in pineapple peel and mangosteen shell biochars due to migration of these elements from soil into particles. The amount of Ca in eucalyptus wood biochar was initially high and decreased slightly with incubation. Phosphorus and S were largely retained within fruit waste biochar particles, but eucalyptus wood biochar lost much S. Silicon, Al, and Fe concentrations in particle fracture surfaces increased greatly during incubation, as fine soil particles migrated through pores and cracks into the interior of biochar particles. Further explanations for these diverse behaviors were sought through SEM analysis of individual mineral particles and mineral mixtures within biochar particles.
3.4. Point and Particle Analyses of Minerals in Biochar
SEM images of fracture surfaces are dominated by the complex fabric of biochar carbon, which faithfully preserves the diverse tissues in the parent plant materials, as shown in
Figure 6 and
Figure 7. Tissues may have contained solutions that precipitated into various compounds when the plant material was air dried, and these compounds may have persisted or altered during heating to create biochar. For example, some of the K in solution in stem exudate and xylem sap is present as dissolved KCl [
17], although for tree tobacco (
Nicotiana glauca) concentrations of 3673 and 204 μg/mL K and 486 and 64 μg/mL Cl, respectively, indicate that only a minor proportion of the K would crystallize as KCl (sylvite). Micron-sized crystals of sylvite would have precipitated on the surfaces of cells during drying of cell solutions.
Figure 6a shows sylvite in nonincubated mangosteen biochar. These crystals seem to be unaffected by biochar manufacture, which converted cell walls to carbon. Calcium is incorporated in several plant constituents [
18], including cell walls, and also occurs as crystalline calcium oxalate within cells [
19]. Heating to produce biochar at temperatures above about 400 °C dehydrates calcium oxalate to produce porous calcium carbonate particles which often preserve the shape of the parent calcium oxalate particle, as is seen in
Figure 6b [
20]. During heating of plant material, the Ca in cell walls reacts with associated elements to form diverse compounds and especially calcium phosphate minerals, as shown in
Figure 6d and
Figure 7a.
Fracture surfaces of biochar particles recovered from soil commonly contain aggregates of soil clay with diverse compositions representing various mixtures of kaolin, illite, and iron oxides mixed with Mg/K salts derived from biochar, as shown in
Figure 6c. Silt-size particles of quartz with attached clay also occur on fracture surfaces, which represent the surfaces of cracks that provided a pathway for silt size particles to migrate into biochar from the soil, as shown in
Figure 7c. Phosphorus is present in plant materials within several organic compounds, including pectates and proteins, which are oxidized by heating and lose their volatile constituents (C, N), so that calcium phosphate compounds remain as discrete particles within pores in biochar, as is shown for eucalyptus wood biochar in
Figure 6d. Calcium phosphates are also finely disseminated throughout former cell walls, as shown in
Figure 7a.
Calcium phosphate compounds remain as disseminated particles within former cell walls, as shown in
Figure 7a, and as discrete particles in pores, as is shown for eucalyptus wood biochar in
Figure 6d. The composition of the particle shown in
Figure 6d for incubated eucalyptus wood biochar is Ca
3.96Mn
0.66Mg
0.48(PO
4)
3OH, which approximates the formula of hydroxyapatite (Ca
5(PO
4)
3OH) [
21]. Many discrete P-rich particles in biochar have similar compositions to this example, but other Ca, Mn, and Mg/P ratios also occur. The disseminated calcium phosphate accumulation within the porous carbon replacing cell wall, shown in
Figure 7a, is from pineapple peel biochar that was recovered from soil after 8 months. It has the approximate composition Ca
4.9Mn
0.1Mg
0.1(PO
4)
3OH and so quite closely resembles ideal apatite. The apatite structure is able to accommodate a wide range of cations [
22] and this capacity is evident in the various cation contents (Ca, Mg, Mn) of apatite particles present in biochar. The persistence of apatite in biochar in soil is to be expected, as apatite fertilizers often persist in soils for long periods as they require an acid soil environment to promote dissolution [
23]. The interior of biochar particles in soil is commonly alkaline, so that apatite will not dissolve until soil acidity has consumed the alkalinity within biochar particles, which is a slow process. Calcium sulphate-rich particles occur in biochar, as for example the particle in pineapple peel biochar recovered from soil after four months shown in
Figure 7b, with the approximate formula of Ca
3SO
4. This formula is not balanced, so the particle is probably a 1:2 mixture of CaSO
4 (anhydrite) and CaCO
3 (calcite). Indeed, many mineral particles in biochar are fine particle mixtures of several compounds, which consequently have complex X-ray spectra.
Figure 7d shows the spectrum of material composed mostly of K (45 percent) and Ca (18.5 percent) with minor amounts of S, Mg, and P. The dominant compound may be fairchildite (K
2Ca(CO
3)
2), a common constituent of plant ash [
16,
24].
The wide range of compositions of particles in biochar is evident when the point analyses of particles or mineralized regions of cell walls exposed in fracture surfaces are plotted in factor diagrams, as shown in
Figure 8. These diagrams contain data for many particles and regions (henceforth called particles) exposed on fracture surfaces of biochars that have been incubated in soils for zero, one, and eight months. Similar results exist for all the soils investigated and for all incubation periods. For mangosteen shell biochar, as shown in
Figure 8a, two factors explain 47% of the variability of the data. This low degree of explanation is indicative of the diverse nature of the mineral particles in biochar. There is a clear segregation of variables (elements) into three groups: (i) Ca, corresponding to calcite; (ii) Mn, Na, Mg, P, Cl, S, and K mostly corresponding to the more labile constituents; and (iii) Si, Al, and Ti, which correspond to soil clay that has entered particles. The factor diagram for mangosteen shell biochar, indicating the distribution of analyses (particles), shows a discrete group of analyses for one and eight month incubated biochar, which correspond to clay. The remaining particles, including those for zero incubation time, extend along an axis that represents various mixtures of the Ca group and the Mn, Na, Mg, P, Cl, S, and K group. The analyses of the original biochar are relatively rich in the labile elements, whereas the one and eight month data are for materials with complex compositions dominated by Ca (calcite).
The factor diagrams for pineapple peel biochar, as shown in
Figure 8b, provide a simple summary of trends in chemical composition explaining 58% of the variation in data. There are three distinct affinity groups of elements which are similar but not identical to those for mangosteen shell biochar: (i) Ca, Mg, P, S (containing calcite, apatite); (ii) K, Cl (sylvite); and (iii) Si, Al, Fe (clay). Many particles in original pineapple peel biochar are rich in KCl with lesser amounts of Ca, Mg, P, and S. After incubation in the soil for one and eight months, little K remained, and the residual particles were mostly mixtures of elements dominated by the Ca, Mg, P, and S groups or a separate group dominated by introduced clay.
In the factor diagram for elements in eucalyptus biochar, as shown in
Figure 8c, the clay elements (Al, Fe, Si) are a tightly associated group, as are the labile elements (K, Na, S but not Cl), whereas Ca, Mg, and P form a diffuse group. Only 47% of the variation in data is explained by the two factors. The analyses of particles from original eucalyptus biochar represent various mixtures of labile elements and the Ca, Mg, and P group. Recovered biochar particles contain many particles with diverse compositions, which correspond to mixtures of clay and the Ca, Mg, and P group. The factor analyses for all three biochars clearly indicate that major changes in the chemical composition of particles had been induced by incubation in soil. The mobile elements (K, Na, Cl, some S) rapidly left the biochar as sylvite and other soluble salts dissolved, and the dissolved ions diffused out of biochar particles into soil. Relatively immobile elements (Ca, P, Mn, some S) were mostly retained in particles in composite particles which have complex compositions. Much soil clay and some silt-size quartz had entered the biochar particles within one month, so that the Si, Al, and Fe group becomes an important constituent of recovered biochar particles.