Phytolith‐rich biochar: A potential Si fertilizer in desilicated soils

Silicon (Si) is beneficial to plants since it increases photosynthetic efficiency, and alleviates biotic and abiotic stresses. In the most highly weathered and desilicated soils, plant phytoliths make up the reservoir of bioavailable Si. The regular removal of crop residues, however, substantially decreases this pool. Si supply may therefore be required to sustain continuous cropping. Available Si fertilizers are costly and usually poor in soluble Si. Biochar produced from the pyrolysis of phytolith‐rich biomass is thus a promising alternative Si source for plants. Taking into account the challenges of increasing food demand and environmental concerns, we evaluate the global potential of biochar produced from major crop residues and manures in terms of phytogenic Si (PhSi) supply. Crop residues contribute to 80% of the global production of biomass dry matter (8,201 Tg/year) of which 3,137 Tg/year are potentially available after pyrolysis, giving a potential application rate of 1.7 T ha−1 year−1 for highly weathered soils in the tropics. The potential PhSi supply from crop biochar amounts to 102 Tg Si/year. On its own, rice straws produce 57.7 Tg PhSi/year, accounting for 56.6% of the potential annual PhSi production. The Si release from crop biochar depends on inter altere feedstock type, pyrolysis temperature, soil pH, and buffer capacity. Furthermore, the amplitude of plant Si uptake and mineralomass depends on plant species, soil properties, and processes. These factors interact and can exert a decisive influence on the effectiveness of phytolithic biochar in releasing Si into highly weathered soils. We conclude that the use of phytolithic biochar as a Si fertilizer offers undeniable potential to mitigate desilication and to enhance Si ecological services due to soil weathering and biomass removal. This potential must be explored, as well as the conditions for using biochar in the field.


| 1265
LI and dELVaUX process increases the capacity of the pyrolyzed matter to retain nutrients, and the resilience of carbon in soil (Liang et al., 2006). Regarding the recent interest in the use of biochars in agricultural soils, Sanchez (2019) notes that burnt organic carbon (OC) is a major component of soil humus in areas where periodic burning is common, such as in many tropical savannas and in slash-and-burn systems. This author further reports that the most iconic effect of biochars in soils occurs in the fertile black earths, so-called Terra preta (Sombroek, Nachtergaele, & Hebel, 1993) widely found in ancient Amazonian settlements, where large quantities of charcoal, organic residues, ashes, etc., were added over long periods of time by previous civilizations (Glaser & Woods, 2004;Lehmann et al., 2003). The production of pyrolyzed organic substances by humans is thus very old, but the scientific interest and agronomic potential were only recently developed. It is clear that the issue of climate change is part of this high interest because of the potential offered by biochars to store carbon in soil (Lehmann & Joseph, 2015). From recent literature data, both biochar production and supply to agrosystems are recommended as a promising multiple-"win" strategy for a number of benefits, inter altere (i.a.) the disposal of crop residue (Jeffery et al., 2015;Van Zwieten et al., 2010;Woolf, Amonette, Street-Perrott, Lehmann, & Joseph, 2010) and the sequestration of OC to mitigate climate change (Lehmann, 2007;Lehmann et al., 2006;Schlesinger & Amundson, 2019;Weng et al., 2017). Besides, biochar application does not only lead to increase soil pH and nutrient retention (Chan & Xu, 2009;Laird et al., 2010;Sohi, Krull, Lopez-Capel, & Bol, 2010;Van Zwieten et al., 2010) but also decreases the toxicity of contaminants in soil (Houben, Evrard, & Sonnet, 2013;Lu et al., 2017;Nie et al., 2018;O'Connor et al., 2018;, thus promoting crop biomass and yield (Biederman & Harpole, 2013;Crane-Droesch et al., 2013;Jeffery et al., 2011;Liu et al., 2013).
However, the potential of biochars as a source of silicon (Si) to alleviate advanced soil desilication in vast tropical croplands is poorly known despite the beneficial effects of Si in terms of plant protection against biotic and abiotic stress (Belanger, 1995;Cooke & Leishman, 2016;Coskun et al., 2019;Fauteux, Rémus-Borel, Menzies, & Bélanger, 2005;Liang, Sun, Zhu, & Christie, 2007;Ma, 2004). In fact, a large part of crop production is lost due to harmful pests and diseases, while the remainder is threatened by increasingly erratic weather and soil fertility decline (Sanchez, 2019;Vanlauwe & Giller, 2006). In challenging these global change issues, the accumulation of Si in cereals provides protection against pests and pathogens, and mitigates the impacts of climatic stresses such as drought and salinity (Coskun et al., 2019; Meunier F I G U R E 1 Schematic illustration of the organic carbon (OC)-phytogenic Si structural interaction relative to pyrolysis process. (A) The pyrolitic process (300-700°C) increases the degree of aromatic OC, highly disordered in amorphous mass, and exposes phytolith particles to the surrounding environment. (B) Further heating above 500-700°C generates growing sheets of turbostratic aromatic carbon, and promotes silica crystallization: phytoliths transform into cristobalite and/or tridymite and eventually quartz. (C) The structure becomes graphitic with order in the third dimension whereas crystalline silicates are amorphized at high temperatures. The heating process from (A) to (C) is accompanied by a decrease in the apparent volume and an increasing organization. Adapted from Chia et al. (2015) et al., 2017). Si-based plant protection could therefore open new avenues to enhance crop yields by addressing current threats and contribute to improving food security, enhancing bioenergy production, and mitigating climate change. Several cultivated plants are high Si-accumulators. In particular, monocotyledons, notably cereals and sugarcane, actively take up dissolved silica (Deshmukh & Bélanger, 2016) which precipitates in plant tissues as amorphous silica bodies named phytolith (McKeague & Cline, 1963), a phytogenic silica (PhSi; Cooke & Leishman, 2016;McKeague & Cline, 1963;Puppe et al., 2017;Sommer et al., 2013). Once deposited into soil within organic debris, phytoliths can dissolve and provide plant-available Si (McKeague & Cline, 1963). Pyrolyzed phytolith-rich biomass can thus be regarded as a source of bioavailable Si (Wang, Xiao, & Chen, 2018;Xiao, Chen, & Zhu, 2014), which can be taken up by plant roots and increase crop biomass (Li, Song, Singh, & Wang, 2019;Li, Unzué-Belmonte, et al., 2019). Besides, the biochar supply improves soil fertility and C sequestration (Houben, Sonnet, & Cornelis, 2014;. Generalizations about biochars performance must, however, be balanced against the fact that biochars are highly heterogeneous in their form and reactivity in soil, which may affect the dynamics of dissolved Si (DSi) in soil. In addition, as illustrated in Figure 1, the pyrolitic process may impact the intrinsic properties of phytoliths (Jindo, Mizumoto, Sawada, Sanchez-Monedero, & Sonoki, 2014;Wu, Yang, et al., 2012;Wu, Wang, & Hill, 2012;Xiao et al., 2014).
Here, we review the available literature to evaluate the prospect of using phytolith-rich biochar to reduce our reliance on expensive Si fertilizer inputs. In particular, we first provide global estimates of PhSi biochar and then evaluate their potential to challenge agronomic and environmental issues. This knowledge is essential to develop guidelines for the selection and appropriate use of biochars to achieve desired agronomic outcomes. We particularly focus on the potential use of PhSi-rich biochar in highly weathered soils. Except for Podzols, the distribution of which is confined to cold and forested areas, the most weathered soils occur in the tropics where desilicated Ferralsols, Acrisols, Lixisols, and Nitisols make up roughly 2,385 Mha (Driessen, Deckers, Spaargaren, & Nachtergaele, 2000;IUSS, 2014), of which approximately 1,890 Mha are cultivated in input-free cropping systems (Sanchez, 2019).

| PhSi content in cultivated plants
The mechanisms of Si uptake by plants are active, passive, or rejective. Si accumulation in plants is caused by active DSi plant uptake, phytolith formation, return to soil, and dissolution define the biological Si feedback loop. PhSi can be exported out of the soil-plant system through crop harvesting and removal of crop residues. It is evident that, in highly weathered soils exhausted in LSi and depleted in PSi, the continuous loss of PhSi results in the increase in soil desilication. Adapted from Cornelis and Delvaux (2016) uptake, which is commonly attributed to the density of Si transporters in roots (Lsi1 and Lsi2) and shoots (Lsi6; Deshmukh & Bélanger, 2016;Exley, 2015;Ma et al., 2006Ma et al., , 2007Ma & Yamaji, 2008;Mitani & Ma, 2005). Thus, Si content in plants is dependent on plant species, and varies from 0.1% to 10.0% of dry matter (DM; Epstein, 1999;Hodson, White, Mead, & Broadley, 2005). Based on Si contents, plant species can be classified as Si high-, low-, and non-accumulator, respectively (Takahashi, Ma, & Miyake, 1990). In this respect, Si high-accumulators generally contain 1.0%-10.0% Si in the DM, and are monocotyledons. Most dicotyledons take up Si passively, and their Si contents range from 0.5% to 1.0%. Some dicots, however, are unable to accumulate Si over 0.5% (Cooke, Degabriel, & Hartley, 2016;Hodson et al., 2005;Liang et al., 2007;Ma, Miyake, & Takahashi, 2001;Ma & Takahashi, 2002). However, biosilicification in plants may also follow passive uptake of Si (Exley, 2015). Crop residues from Si-accumulating plants may contain large amounts of Si so that their removal from croplands may significantly affect the terrestrial Si cycle (Carey & Fulweiler, 2012Conley, 2002;Keller et al., 2012;Song, Parr, & Guo, 2013;Struyf et al., 2010;Vandevenne et al., 2012). Here, we focus on the PhSi content in rice hulls and straws, wheat straws, maize stalks and cobs, barley straws, sorghum straws, millet, rye and oats, sugarcane crop residues, oil crops including sunflower and cotton, legume crop residues including soybean straw and beans. As shown in Table S1, the highest PhSi contents are in rice hulls (10.3 wt%), rice straw (8.4 wt%), sugarcane (8.2 wt%), and millet (7.5 wt%), followed by straws of sorghum, wheat, maize, barley, and rye (1.9-6.0 wt%). In other crop residues, PhSi content is below 1.0 wt%. The differences in PhSi content among soybean straw and beans, sunflower, and cotton could be explained by various transpiration rates given the role of aquaporins in Si flux Deshmukh & Bélanger, 2016;Epstein, 2001;Hodson et al., 2005;Liang et al., 2007). In contrast, for cereals, Si is actively taken up, and the difference in PhSi content has been attributed to the density of Si transporters in plant roots (coded by the low-silicon genes Lsi1 and Lsi2) and in shoots (Lsi6; Liang, Si, & Römheld, 2005;Ma & Yamaji, 2008;Ma et al., 2007;Mitani & Ma, 2005). For example, compared to other crops, rice cultivars are well known to exhibit a stronger ability to accumulate PhSi due to a particularly high density of Si transporters in their roots (Epstein, 1999;Ma et al., 2006Ma et al., , 2007.

| PhSi amounts in crop residues and manures
Here, we do not consider crop residues belonging to the following three categories. First, the residues that are available in very small quantities (e.g., nut and peanut shells, sesame, peas and orchard prunings) are neglected since they account for only 1.5% of global crop residues, their annual total dry biomass being below 6 Tg/year. Second, some residues exhibit relatively high water contents (e.g., potato) to be unsuitable for pyrolysis. Third, other residues, such as legume crops, exhibit no global significance because they have a very low PhSi content that is below 0.1% (Carey & Fulweiler, 2012Hodson et al., 2005;Song et al., 2013). In Figure 3a, we estimate that the global PhSi annual production, as deduced from selected crops (Table  S1), amounts to 213.8 Tg Si/year (Lal, 2005;Woolf et al., 2010). Our estimated annual PhSi amount is close to the global estimate (204.8 Tg BSi/year) of annual global crop straw provided by Carey and Fulweiler (2016). The total PhSi amount in crop residues depends on DM biomass, and on plant species or plant part ( Since phytoliths are resistant to fungal or animal digestive juices after grass or crop feed ingestion, significant amounts of PhSi accumulate in animal manures Jones & Handreck, 1967;Krishnan, Samson, Ravichandran, Narasimhan, & Dayanandan, 2000). We estimate that the global PhSi amount in selected manures (cattle, pig, and poultry) amounts to 42.0 Tg Si/ year among which 94.5% is provided by cattle manures (Figure 3a). Manure return may thus play a major role in returning PhSi to soil, hence replenishing the pool of DSi in agricultural lands. Besides, the transit of phytoliths through animals can increase phytolith solubility: for example, applications of pig manure from rice straw-fed pigs at the rate of 35 Mg ha −1 year −1 significantly increased the content of plant-available Si in rice-cultivated soils .

| Removal of crop residues and subsequent exportation from croplands
The regular removal of crop straws of Si-accumulator plants is a common practice in agrosystems. The resulting Si exportation can lead to a substantial loss of PhSi in cultivated soils (Carey & Fulweiler, 2016;Guntzer, Keller, Poulton, et al., 2012;Keller et al., 2012;Vandevenne et al., 2012), creating a new loop in the soil-plant Si cycle. The loss of PhSi is likely most pronounced under intensive agriculture, especially for Si high-accumulator crops (Table S1), since they pump large amounts of Si from soils. Carey and Fulweiler (2016) have indeed noted that, on a global scale, agricultural crops contribute to ~35% of PhSi accumulated by terrestrial plants because of their large biomass and PhSi contents (Table S1). They have also pointed out that this contribution will increase during the next decades because of increased agricultural production, predicting that the global removal of PhSi would reach 40 Tmol Si/year by 2050 (Carey & Fulweiler, 2016). Crop harvest and residue removal may thus substantially reduce soil PhSi (Meunier et al., 2008). This statement is supported by several studies. Clymans et al. (2011) report a threefold lower content of amorphous Si (ASi) in cultivated soils than in forest soils (22.8 vs. 66.9 t Si/ha) in southern Sweden, this difference resulting from the annual removal of PhSi from plant products and residues. Vandevenne et al. (2015) have reported that waters drained from arable cropland display heavier Si-isotopic composition (δ 30 Si) than the one from mature forest areas, resulting from the preferential removal of 28 Si in harvested plant parts. Annual exports of wheat straws led to a 10% decrease in the soil PhSi stock over 100 years (Guntzer, Keller, Poulton, et al., 2012), whereas the continuous removal of wheat straws at the rate of 50-100 kg Si ha −1 year −1 over the past 12 years would completely exhaust this pool within next 10-20 years . Exporting PhSi out of rice croplands at 270 ± 80 kg Si ha −1 year −1 would consume the soil PhSi pool in next 5 years (Desplanques et al., 2006). These experimental data suggest that the depletion of the PhSi pool is caused by intensive cropping involving cereal straw removal. In contrast, returning rice straws to paddy soils maintained a constant stock of soil phytoliths and enriched it relatively to neighboring soils under bamboo and coniferous forest (Yang & Zhang, 2018). As recently reviewed (Vander Linden & Delvaux, 2019), the impact of soil weathering stage on the soil-plant Si cycle markedly differs depending on land use in the tropics. In less weathered soils, cultivated plants take up Si in larger amounts than forest trees do whereas in highly weathered soils, Si plant uptake increases in forest ecosystems while it decreases in croplands.

| Global amounts of PhSi in biochars derived from crop residues and manures
Using the conversion factor determined by Woolf et al. (2010), we estimate from Tables S1 and S2 the global These amounts were estimated using the data from Tables 1 and 2 amounts of PhSi at 101.9 and 11.9 Tg Si/year in crop and manure biochars, respectively ( Figure 3b; Table 1). They contribute to 82.3% and 17.7% of the global amounts of PhSi in crop and manure biochars, respectively. The amounts of PhSi vary depending on plant species and plant part, in particular on DM and initial PhSi content (Figure 3b). Rice straw is the largest driver of the global PhSi flow into biochar products since it contributes to 48.9% of the PhSi contained in crop biochars (Figure 3b), whereas sugarcane residues contribute to 22.7%. The contributions of wheat straws (10.5%), rice hulls (7.8%), and maize straws (6.3%) to the global PhSi reservoir in crop biochars are relatively low (below 11%).
Maintaining an adequate global PhSi reservoir through biochar supply in agricultural soils thus requires to produce biochar from rice straws and husks, sugarcane residues, wheat and maize straws. Though the global DM production (Tg/ year) is much larger for cattle manure (1,570) than for pig (90) and poultry (134) manures (Table S2), the conversion factor into biochar is 1/3.9 for cattle, 1/1.1 for pig and poultry manures, revealing a much lower potential of cattle manure for biochar production and agricultural use as compared to pig and poultry manures (Hoogwijk et al., 2003;Woolf et al., 2010). However, cattle manure is enriched in PhSi (Table 2), which originates from fodder obtained from Si-accumulating plants (Blecker, Mcculley, Chadwick, & Kelly, 2006;. In contrast, pig and poultry manures are comparatively poor in PhSi (Table 2). Cattle manure is thus the largest driver of the global PhSi flow into manure biochar since it contributes to 83.2% of the PhSi produced by manure biochars (Figure 3b). The PhSi pool in cattle manure biochar may thus affect the soil PhSi pool in grasslands. Unfortunately, this is poorly studied.

| AGRONOMIC POTENTIAL OF PHYTOLITHIC BIOCHAR
Despite its non-essentiality for plants, Si offers a great agronomic potential because it increases the photosynthetic efficiency, and alleviates various biotic and abiotic stresses (Belanger, 1995;Fauteux et al., 2005;Liang et al., 2007). The need of supplying Si is particularly acute in highly weathered soils, which are depleted in weatherable LSi and PSi silicates. In these soils, phytoliths make the pool of plantavailable Si (Alexandre et al., 1997;Li, Unzué-Belmonte, et al., 2019;Lucas et al., 1993;Meunier et al., 1999). Since phytoliths can be exported through the removal of crop residues, PhSi depletion may aggravate soil desilication in these soils, making them candidates for external Si supply. However, in the tropics, most farmers are unable to use Si fertilizers at recommended rates of 1 or 2 t ha −1 year −1 since silicate slag and minerals are expensive and limited worldwide (Savant et al., 1996). Supplying phytolithic biochar produced from crop residues is a potential and environmentally friendly Si fertilizer, which has additional well-known effects on soil fertility through the increase in CEC, pH, and bioavailable nutrients.

| Plant-available Si
The amount of DSi effectively available for plants depends on biochar type, soil properties, and processes since DSi may be involved in PSi formation, adsorbed on secondary oxides or leached out and transferred to watersheds, as well as taken up by various organisms. Houben et al. (2014) have shown that bioavailable Si content significantly differed between distinct biochars (Miscanthus x giganteus straws, coffee husks, and woody material), Miscanthus being particularly rich in PhSi.  have reported significant differences in DSi release from biochars in the following order: rice straw > Miscanthus straw > sugarcane harvest residue > switchgrass, due to their differences in initial amounts of PhSi. Regardless of pyrolysis temperature, Wang, Xiao, et al. (2018) reported that rice husk biochar had a larger PhSi content than rice straw, wood sawdust, and orange peel, and that bioavailable Si in soil was much larger when released from rice husk biochar. Figure 6 shows that the application of wheat straw biochar increases the content of bioavailable Si, and the Si content of rice shoots (Liu et al., 2014).

F I G U R E 6
Rice plant shoot Si content (mg/g) after wheat-straw biochar application as plotted against soil bioavailable Si as extracted by 0.025 M citric acid (field data from Liu et al., 2014) Soil properties and processes affect the release of DSi from biochar phytoliths. Through its regulation of soil pH, soil buffering capacity controls the dissolution of biochar phytoliths in soils (Figure 7; Li, Unzué-Belmonte, et al., 2019). Figure 7 further shows that at a given supply of phytolithic biochar, the liming effect of biochar depends on soil buffering capacity and this effect is crucial to predict the mobility of Si in the soilplant system (Li, Unzué-Belmonte, et al., 2019). Adsorption process may retrieve DSi from soil solution and decrease Si plant availability. H 4 SiO 4 0 adsorption onto Al, Fe oxides is enhanced in highly weathered soils (Meunier, Sandhya, Prakash, Borschneck, & Dussouillez, 2018) where natural soil desilication results in the relative accumulation of Al, Fe oxides, which selectively adsorb H 4 SiO 4 0 . Thus, at given PhSi biochar supply, soil properties and processes chiefly govern the plant availability of Si. The efficiency of PhSi biochar supply is thus expected to vary largely according to soil type.

F I G U R E 7
Plot of CaCl 2 -Si content against pH-CaCl 2 in soil, soil:biochar, and soil:wollastonite-solution systems. Note: Wo: wollastonite (CaSiO 3 ) is a common Si fertilizer; Si − biochar is free of PhSi, whereas Si + biochar is enriched in PhSi. The Cambisol and the Nitisol have a high and a very low buffer capacity, respectively. Wo and Si + biochar supplies are identical in terms of Si addition. Adapted from Li, Unzué-Belmonte, et al. (2019)

| PhSi biochar boosts the biological Si feedback loop
The present review demonstrates the potential of phytolithic biochar to boost the biological Si feedback loop of the soilplant cycle in agroecosystems (Figure 2). The basic principles to achieve this boost are to maximize the restitution of PhSi-bearing biochars to cultivated soils, and to optimize the conditions promoting the release of plant-available Si in the soil-plant system. These conditions are linked to soil properties such as pH as well as soil processes such as i.a. acid buffering, and DSi adsorption onto oxide surfaces, and biosilicification. Boosting the biological Si feedback loop of the soil-plant cycle ( Figure 2) in agroecosystems through phytolithic biochar supply is expected to be particulary efficient in highly weathered soils. Given the global estimates of available PhSi biochar and the surface covered by highly weathered soils in the tropics (Ferralsols, Acrisols, Lixisols, Nitisols), we tentatively estimate a global rate of biochar application of 1.7 ha −1 year −1 . This global estimation is, of course, theoretical, since it supposes free access to PhSi biochars and free circulation of this material.

| CONCLUSION
The pyrolysis of crop straws or residues from Si-accumulating plants provides biochars contributing to feed the reservoirs of stable carbon and plant-available Si in agroecosystems. Feedstock type, pyrolysis temperature, soil pH, soil properties and processes are the main factors that exert a decisive control on the supply of plant-available Si and other benefits (Figure 8). The expected improvement is especially meaningful for highly weathered soils intensively used for cropping Si high-accumulator plants. On a global scale, ample amounts of feedstocks for phytolithic biochar are available. The largest benefits can be expected under the following conditions: (a) the use of cereals (rice, wheat, maize) and sugarcane crop residues as feedstock materials; (b) pyrolytic process performed slowly up to 350-500°C for tree or wood biochar, 500-600°C for crop biochar, but not above 700°C to prevent silica crystallization; (c) an optimal liming effect increasing soil pH to release plant-available Si.
World agriculture faces multiple challenges due to current high demands in food and environmental concerns. Quantifications about the cost and benefits should be compared between phytolithic biochar and traditional Si fertilizers. The analysis should take into account both the agronomic (e.g., crop yield) and environmental aspects (e.g., sustainable ecosystem development) induced by Si ecological services. The supply of phytolithic biochar indeed positively affects the reservoir of plant-available Si and other nutrients, stable OC, as well as soil fertility, given the ample Si ecosystem services offered by biochars.