A biorefinery model for the production of oil and biomethane using castor plants

Castor (Ricinus communis L.) is an important oilseed crop worldwide whose inedible oil is widely used in the industrial, pharmaceutical, and agricultural sectors. Castor plants show high conversion potential for use as biorefining feedstocks. The present study was conducted to investigate the effects of two nitrogen fertilization levels (0 and 120 kg N ha−1) on seed and oil yield. From a biorefinery perspective, the residual biomass of seed processing was analyzed in terms of fiber composition and biomethane production carrying out a biological pretreatment using two white‐rot fungi (Pleurotus ostreatus and Irpex lacteus). Nitrogen fertilization resulted in an increase in seed and oil yields and a difference in capsule husk composition. Fungal pretreatment of capsule husks showed promising effects on anaerobic digestion, increasing the biomethane yield compared to untreated biomass. The highest lignin degradation and the lowest cellulose loss during pretreatment were obtained with I. lacteus, and this fungal pretreatment resulted in the highest biomethane yield (103.2 NmL g−1 volatile solids) for the fertilized biomass.

on marginal lands and in semiarid climates (Falasca et al., 2012;Patanè et al., 2019).Castor is largely cultivated for the extraction of oil from its seeds.The oil is composed of 91%-95% ricinoleic acid, 4%-5% linoleic acid, and 1%-2% palmitic and stearic acids (Severino & Auld, 2013), making it highly desirable for industrial and pharmaceutical uses, such as the production of paints, coatings, inks, lubricants, and cosmetics (Ogunniyi, 2006).Moreover, in the last few years, this nonedible vegetable oil has been investigated as an alternative to the edible oils, such as rapeseed, sunflower, soybean, and palm oil, traditionally used in biodiesel production (Dias et al., 2013;Keera et al., 2018).Castor cultivation in a Csa climate (Köppen-Geiger classification) could be attractive for its capacity to grow in a variety of soils and climatic conditions, reducing the competition with food crops and avoiding the "food versus fuel" dispute (Abdul Hakim Shaah et al., 2021).Furthermore, castor cultivation allows valorization of residual biomass using a biorefinery approach because after the oil extraction from seeds for biodiesel production using transesterification, residual biomass, including stems and leaves, can be used for bioethanol or biogas production, significantly increasing the profitability of the crop (Bateni et al., 2014;Ogunniyi, 2006;Patel et al., 2016).
Biorefineries represent a promising solution for addressing climate change and reducing dependence on fossil fuels in economically viable pathways (Khoshnevisan et al., 2018).Moreover, considering the crisis for food and biofuel markets further to the Ukraine-Russia war, there has been a growing interest in shifting biofuel production from first generation, using food/feed crops as their feedstock, to higher generations, using energy crops and lignocellulosic residue (Esfandabadi et al., 2022).
The potential of using plant-based materials as precious feedstock for biofuel production, replacing a significant fraction of fossil resources to reduce the dependence on fossil fuels and mitigate climate change, is increasingly recognized (de Freitas et al., 2021).Thus, the effective utilization of lignocellulosic biomass as biofuels, such as bioethanol and biomethane, could satisfy the need for energy sources and solve environmental concerns.Several studies have been conducted to evaluate different biomass crops as feedstocks to produce biofuels through a biorefinery approach (Asghar et al., 2022;Clauser et al., 2021;Khounani et al., 2019;Zetterholm et al., 2020).Many of these have analyzed the economic viability of biodiesel production from castor oil as the main product resulting in byproducts such as pellets, fertilizer, biogas, bioethanol, and heat from plant residues (Rahimi et al., 2020;Rahimi & Shafiei, 2019;Sandoval-Salas et al., 2022).
Table 1 lists some studies that adopted castor plants to produce biofuels, but none have used capsule husk residue to produce biomethane.
Biomethane produced by anaerobic digestion (AD) is an excellent technique for the energetic valorization of different types of biomasses, including lignocellulosic residues, as a renewable and cheap resource.Lignocellulosic biomass consists of a cellulose unit covered by hemicellulose and lignin, resulting in an amorphous, highly lignified, and naturally recalcitrant material to energetic conversion.The interaction of these components determines the resistance to anaerobic degradation, limiting the enzymatic hydrolysis phase during the AD process.Thus, pretreatment is essential to make the holocellulose more accessible for the subsequent hydrolysis phase.Effective pretreatment determines the breakdown of the lignocellulosic complex and the decrease in cellulose crystallinity with no inhibitory chemical production.
Most investigated pretreatments, such as physical and thermochemical processes, are costly and can consume large amounts of energy.In contrast, biological pretreatment is more environmentally friendly than other pretreatment methods and has several advantages, including low or no formation of harmful chemicals, low-energy requirements operating under mild conditions, and no special equipment required for the process.Despite its advantages, the use of biological microorganisms to pretreat lignocellulosic biomass is limited by some drawbacks, mainly related to the optimal microorganism choice, considering the high variability rate of culture conditions (i.e., temperature, humidity, substrate concentration, and sterilization), as well as the difficulty in developing industrial and commercial technologies because of the lengthy pretreatment process (Singh, 2021;Wu et al., 2022).Biological pretreatments make use of wooddegrading microorganisms, including white-rot fungi (WRF), brown-rot fungi, soft-rot fungi, and bacteria, that through the action of extracellular enzymes can degrade lignin by decreasing cellulose crystallinity and increasing accessible surface area.WRF can be widely used in many biotechnological sectors, including biofuel and biorefinery, due to their superior ability to selectively degrade lignocellulose, and appear to be very promising for the biological pretreatment of lignocellulosic biomass feedstock.Nitrogen fertilization is an important management practice for optimizing crop growth and obtaining high yields because it is an essential nutrient required by plants; however, it is the most limiting nutrient in the soil (Al-Thabet, 2006;Jamil et al., 2017).In addition, it also has a significant impact on biomethane production, affecting both biomass yield and chemical composition.
With the aim of maximizing the potential of castor plants using a biorefinery approach to combine seed and oil yields for biodiesel production and the valorization of capsule husks through biomethane production, this study evaluated the effects of two nitrogen fertilization treatments (0 and 120 kg N ha −1 ) on castor yield and biomethane production by AD after the biological pretreatment of capsule husks using two WRF (Pleurotus ostreatus and Irpex lacteus).

| Field experiment
The field experiment was conducted in 2020-2021, at the experimental farm of the University of Catania, Italy (37°25′ N, 15° 03′ E; 10 m a.s.l.) in typical xerofluvent soil.
The soil in the experimental area was plowed before sowing and fertilized with 70 kg ha −1 P 2 O 5 (as single superphosphate).The effects of two nitrogen fertilization levels, N0 (0 kg N ha −1 ) and N120 (120 kg N ha −1 ) were studied.The amount of nitrogen at the N120 level was 35% before sowing (as ammonium sulfate), 45% before flowering (as ammonium nitrate), and 20% at the beginning of grain formation (as ammonium nitrate).Irrigation was provided using a drip irrigation system, restoring 100% of the maximum crop evapotranspiration (ET m ) with a total 1800 m 3 ha −1 during the growing cycle.Irrigation was scheduled when the sum of the daily ET m corresponded to the volume, subtracting rainfall events from the calculation.The daily ET m (1) was calculated as follows: where ET m is the maximum daily evapotranspiration (mm), E 0 is the evaporation of a class A pan (mm), and K p is the pan coefficient, equal to 0.80 in a semi-arid environment.The crop coefficients (K c ) were determined based on previous observation: 0.4 from emergence to the four-leaf stage, 0.7 from the four-leaf stage to flowering, 1.2 from the beginning of flowering to complete capsule development in the first raceme, and 0.55 from the first raceme to complete capsule senescence.The irrigation volume (2) was calculated using the following equation: where V is the water amount (mm); 0.66 is the readily available water not limiting for evapotranspiration; FC is the soil water content at field capacity (27% of dry soil weight); WP is the soil water content at wilting point (11% of dry soil weight); ɸ is the bulk density (1.1 g cm −3 ); and D is the rooting depth (0.6 m).
A randomized complete block design with three replications was used.
Each plot measured 68 m 2 (8 × 8.5 m) with 1.7 m rowto-row spacing and 1 m plant-to-plant distance (sowing density 0.58 plants per m 2 ) and consisted of five rows.
The seeds used in this experiment were of a local castor variety selected by the Department of Agriculture, Food and Environment (Di3A) from a Tunisian genotype.Sowing was manually carried out on June 24, 2020, by placing three seeds per hole at a depth of 3-4 cm, and the plots were hand-thinned to one plant per hill when the plants were at the four-leaf stage.
Meteorological conditions and potential evapotranspiration (ET0) were continuously measured using a (1) weather station connected to a data logger (Delta-T, WS-GP1 Compact) and a Class A evaporation pan (mm d −1 ).
During the growing season, the dates of the main phenological stages of the crop (seedling emergence, anthesis, and seed physiological maturity) were recorded.The primary racemes of the unfertilized and fertilized plants were harvested on October 30 and November 8, 2020, respectively.
The harvest dates of secondary racemes were Decem- ber 6, 2020, December 12, 2020, and February 28, 2021, for  unfertilized plants and January 7, 27, and March 17, 2021, for fertilized plants, according to the different flowering period.Three central rows of each plot were harvested and used for the measurements.At harvest, the number of primary and secondary racemes per plant and the insertion height of the primary raceme (measured from the ground to the first raceme insertion point) were recorded.
The primary raceme length (from the apex to the end of the peduncle) was measured in the laboratory.The first and second racemes were separated and weighed, the capsules of each raceme were counted and manually separated from the peduncle, and the seeds were separated from their capsules to determine the number of capsules per raceme and seed weight.

| Meteorological conditions
The weather conditions during the field experiment reflected those typical of a Csa climate (Köppen-Geiger classification).
Total rainfall of 231.6 and 208.8 mm were recorded before sowing (March-June 2020) and from sowing until the first harvest (June-October 2020), respectively.Cumulative rainfall during the harvest season (October 2020-March 2021) was 345.8 mm (Figure 1).
During the growing season, annual average temperatures were a maximum of 22.9°C, a minimum mean of 11.9°C, and a mean of 17.2°C.The reference evapotranspiration (ET0) was 1560.1 mm, with an average of 4.39 mm day −1 .

| Characterization of feedstock
Capsule husks were oven-dried at 65°C, milled, and used for chemical analysis and biological pretreatment.The total solids (TS), volatile solids (VS), and chemical compositions of the capsules were determined before pretreatment.TS content was measured after drying samples to constant weight at 105°C, and VS content was estimated after incineration in a muffle furnace at 550°C for 5 h (Sluiter, 2008).TS and VS measurements were performed in triplicate.
The fiber composition was determined as neutral detergent soluble (NDS), neutral detergent fiber (NDF), acid detergent fiber (ADF), and acid detergent lignin (ADL) according to the Van Soest method (Van Soest et al., 1991) using fiber analyzer (Fibertec model FIWE, Velp Scientifica).The hemicellulose and cellulose contents were calculated as the difference between NDF and ADF and ADF and ADL, respectively.The ADL residue was ignited in a muffle furnace at 550°C for 5 h to determine the ashes, and lignin content was calculated as the difference between ADL and ash.

| Oil extraction
Seeds were ground and subjected to oil extraction using a GM200 blade mill (Retsch GmbH).Oil content was determined according to the Randall method using a quantitative solvent extractor SER 148/6 (Velp Scientifica) (Randall, 1974).
Extraction was performed by immersing a cellulose thimble containing 3 g of the sample in boiling n-hexane solvent, followed by rinsing and recovery.Extraction vessels were placed in an oven at 100°C for 30 min to eliminate the solvent residues.All analyses were performed in triplicate.

| Inoculum preparation
The fungal strains used in this study (Pleurotus ostreatus (MUT00002977) and Irpex lacteus (MUT00005918)) were purchased from the Mycotheca Universitatis Taurinensis (MUT) of the Department of Life Sciences and Systems Biology at the University of Turin (Italy).
Fungi were activated on Malt Extract Agar plates and incubated at 26°C for 7 days.
The inocula were prepared as described by Piccitto et al. (2022).Four agar discs of 7-day-old mycelia were added to the reactors containing the sterilized capsule husks with a moisture content of 70% and incubated at 26°C.Full substrate colonization occurred after 4 weeks of incubation, and sterile capsule husks colonized with Pleurotus ostreatus and Irpex lacteus were used as inocula for subsequent fungal pretreatment experiments.

| Fungal pretreatments
Sterile capsule husks and inoculum (fungal-colonized capsules) were mixed and added to 0.5 L reactors.Fungal pretreatment was performed at an inoculum ratio of 30% (dry weight).Deionized water was added to achieve a moisture content of 70%.Reactors were covered with cotton plugs and incubated at 26°C for 30 days.
Subsamples were taken at predetermined periods (10, 20, and 30 days) to determine the cellulose, hemicellulose, and lignin contents.
Dry matter loss and cellulose, hemicellulose, and lignin degradation during fungal pretreatment were calculated as percentages of the initial dry weight and fiber fractions before fungal pretreatment.

| Biochemical methane potential tests
The biochemical methane potential (BMP) test was conducted using an automatic methanogenic potential detection system (AMPTS II; Automatic Methane Potential Test System; Bioprocess Control).
The experiment was conducted with an inoculum substrate ratio of 1:3 in terms of grams of VS at mesophilic conditions (38 ± 1°C) in reactors of 500 mL each with continuous mixing.All tests were performed in triplicate.The inoculum was originally obtained from an anaerobic digester in Sicily.The inoculum was filtered through a 2-mm porosity sieve to remove large and undigested particles, and stabilized in an incubator at 38°C for 5 days.
TS and VS were determined for both the organic substrate and the inoculum, as reported above.
Each reactor was connected to an 80 mL trap bottle of 3 M sodium hydroxide solution used to absorb CO 2 from the raw gas.After scrubbing, the remaining gas was passed through ultra-low gas flow meters, which were connected to the data analysis and acquisition system.The BMP test was conducted for 40 days.
Blank samples containing only the inoculum were also incubated.The methane production of the substrate was determined by subtracting the methane production of the blank (inoculum) from that of the substrate (substrate + inoculum).The final value of cumulative methane production at the end of the test was defined as the experimental BMP of the substrate.

| Statistical analysis
Agronomic data were subjected to statistical analysis using analysis of variance (ANOVA) according to randomized blocks with three replicates.The biomass contents of hemicellulose, cellulose, ADL, ash, and NDS were analyzed using one-way ANOVA with fertilization as a fixed effect.The pretreatment laboratory data were statistically analyzed using ANOVA with a completely randomized factorial design with three factors (pretreatment, nitrogen fertilization, and time of measurement), considering all fixed factors.The daily and cumulative biomethane of untreated and fungalpretreated capsule husks after 40-day incubation were analyzed using two-way ANOVA with fungal pretreatment and fertilization as fixed effects.The biomethane yield (BMY) per hectare was statistically analyzed using a completely randomized factorial design with two factors (pretreatment and nitrogen fertilization), considering all fixed factors.Before conducting ANOVA, Bartlett's test was performed to verify the assumption of homogeneity of variance.Mean separation was performed using the least significant difference LSD test (p < 0.05).Data were statistically analyzed using CoHort Software (CoStat version 6.003).

| Phenological stages
ANOVA showed a significant effect of nitrogen fertilization on the physiological maturity of the first raceme (p ≤ 0.001) that was completed 128 days after sowing for unfertilized treatment N0 and after 137 days for fertilized treatment N120.Seedlings emerged after 7 days, and anthesis of the primary raceme occurred 61 days after sowing in both nitrogen treatments (Figure 2).

| Morphological characters and yield components
Plant height up to the primary raceme, length of the primary raceme, number of secondary racemes, number of capsules per raceme, and seed weight are presented in Table 2.
ANOVA showed a significant effect of fertilization on the number of secondary racemes, the number of capsules per raceme for both primary and secondary racemes, and the seed weight of the primary raceme.Nitrogen fertilization had no significant effect on plant height up to the primary raceme, length of the primary raceme, and seed weight of the secondary racemes.

| Seed yield and capsule husk yield (kg ha −1 )
ANOVA showed a significant effect of fertilization on castor seed and capsule yield (p ≤ 0.01).
The seed yields were 1104 and 1991 kg ha −1 for the unfertilized and fertilized treatments, respectively.Capsule yields were 903.6 and 1629 kg ha −1 for N0 and N120, respectively (Figure 3).

| Oil content and oil yield
ANOVA showed a significant effect of N fertilization on oil content and oil yield (p ≤ 0.01 and p ≤ 0.05, respectively).
The seed oil contents were 43.06% and 40.05% in the N0 and N120 treatments, respectively (Figure 4).
The oil yields were 476.1 and 796.8 kg ha −1 in the N0 and N120 treatments, respectively (Figure 3).

| Capsule husk composition
ANOVA of nitrogen fertilization as the main effect on capsule husk composition is shown in Table 3. Fertilization significantly affected cellulose and lignin content, whereas hemicellulose, NDS, and ash content did not differ.
F I G U R E 2 Phenological stage duration (days) from sowing to emergence (S-E), from emergence to anthesis (E-A), from anthesis to seed maturation (A-M) of primary raceme under fertilization levels (N0-0 kg N ha −1 and N120-120 kg N ha −1 ).Different letters indicate significant differences according to LSD test at p ≤ 0.05.
The NDS content was higher in the unfertilized (38.08% w/w) treatment than in the fertilized treatment (32.66% w/w), whereas the hemicellulose and cellulose contents were higher in the fertilized residue (21.63% and 36.49%w/w, respectively) than in the unfertilized treatment (20.74 and 29.31% w/w, respectively).The unfertilized residue had higher lignin (11.03% w/w) and ash (0.8365% w/w) contents than the fertilized treatment (8.526% and 0.6904% w/w, respectively; Figure 5).
The ratio of structural carbohydrates (hemicellulose and cellulose) to lignin was determined.This measurement was used to estimate the digestibility of the tested substrate.The highest ratio (6.9) was recorded for the fertilized residue.

| Pretreatment effects on lignocellulosic biomass
ANOVA revealed a significant effect of pretreatment and fertilization on dry matter, cellulose, and lignin degradation.The degradation of dry matter, hemicellulose, cellulose, and lignin was significantly influenced by time.Significant interactions "Pretreatment × Time" and "Pretreatment × Fertilization × Time" were also observed on dry matter degradation and lignin degradation, respectively (Table 4).
The degradation of dry matter, cellulose, hemicellulose, and lignin in the capsule residues increased with time for both fungi (Figure 6).A high percentage of dry matter degradation was observed for I. lacteus for both unfertilized and fertilized substrates after 30 days of incubation (31.15% and 31.33%,respectively).The percentage of degradation of dry matter for P. ostreatus was 27.66% and 26.95% in the N0 and N120 treatments, respectively (Figure 6a).For hemicellulose degradation, the highest loss was observed in P. ostreatus fertilized treatment (33.81%) (Figure 6b).For cellulose, the highest degradation was observed in P. ostreatus, with a loss of 34.45% in the unfertilized treatment (Figure 6c).The highest lignin loss was observed for I. lacteus in the fertilized sample (25.08%), followed by the unfertilized biomass pretreated with I. lacteus (20.64%) (Figure 6d).

| Methane production
ANOVA showed that daily biomethane production was significantly influenced by pretreatment, fertilization level, and incubation time (Table 5).ANOVA of pretreatment and nitrogen fertilization as main effect on cumulative biomethane production (∑CH 4 ) is shown in Table 6.Daily production (NmL g −1 VS d −1 ) and cumulative methane production (NmL g −1 VS) during the AD of untreated and fungal-pretreated capsule husks are shown in Figure 7.
The daily biomethane production of capsules pretreated with P. ostreatus showed the highest peaks for the fertilized treatment (9.109 NmL g −1 VS d −1 ) after 17 days of digestion (Figure 7a).Capsules pretreated with I. lacteus showed a maximum peak (6.409 NmL g −1 VS d −1 ) on Day 21 for fertilized biomass, whereas untreated capsules showed peaks, reached after 20 days, of daily methane production lower than the other treatments (4.605 and 4.734 NmL g −1 VS d −1 for N0 and N120, respectively).
Cumulative biomethane production was observed for 40 days until the BMY reached a plateau at the end of the exponential phase (Figure 7b; Table 7).The initial lag phase lasted approximately 3 days until the complete adaptation of the bacterial flora to the lignocellulosic substrate.
Methane production obtained for the untreated N0 and N120 was 66.83 and 81.23 NmL g −1 VS, respectively.P. ostreatus pretreated capsules reached values of 69.24 and 93.88 NmL g −1 VS for N0 and N120 fertilization, respectively.The methane yield reached by I. lacteus-pretreated capsule husks was 98.78 and 103.1 NmL g −1 VS for the N0 and N120 fertilization levels, respectively.
T A B L E 3 One-way ANOVA for main effect (fertilization) on hemicellulose content (H), cellulose content (C), lignin content (ADL), neutral detergent soluble content (NDS), ash content (ASH).Lignin negatively impacts BMY by significantly contributing to lignocellulose biomass recalcitrance.The selectivity index, defined as lignin degradation over cellulose loss, is an important tool for assessing the ability of WRF to degrade lignin selectively.
Figure 8 shows the correlation between the selectivity index for fungal pretreatment and the cumulative methane yield for P. ostreatus and I. lacteus, highlighting an increase in the BMY with a selectivity index higher than 0.5.

| Biomethane yield per hectare
ANOVA revealed a significant effect of fertilization on BMY (Table 8).
The BMYs for the untreated N0 and N120 biomass were 51.97 and 112.2 m 3 CH 4 ha −1 , respectively.Among the fungal pretreated thesis, I. lacteus showed the highest BMY per hectare (75.80 and 144.5 m 3 CH 4 ha −1 for N0 and N120, respectively) while P. ostreatus reached values of for N0 and 129.8 m 3 CH 4 ha −1 for N120 (Figure 9).Many studies have reported positive and significant correlations among seed yield, height at flowering, number of racemes per plant, raceme length, number of capsules per raceme, and 100 seed weight (Mullualem Atinafu et al., 2019;Ramesh & Venkate, 2001).The seed weight observed in the present study was in accordance with other studies, such as Goodarzi et al. (2015), who reported that the 10-seed weight on the primary raceme varied from 2.1 to 3.3 g, whereas Bhardwaj et al. (1996) reported a 100-seed weight variable ranged from 10 to 44 g.The increase in nitrogen dose significantly affected the main yield contributing traits, such as the number of secondary racemes, number of capsules per raceme for both primary and secondary racemes, and seed weight of primary raceme.These results are in line with Chatzakis et al. (2011) andJamil et al. (2017).
Seed yield is considered the most important characteristic for improving crop production (Yousaf et al., 2018).The results showed that it was possible to obtain a satisfactory seed yield using castor, suggesting its adaptability to the Mediterranean climate, as reported in a preliminary field experiment carried out by Anastasi et al. (2015) in southern Italy under rainfed regimes to explore the feasibility of growing castor as a semi-perennial plant.
Regarding the effect of nitrogen fertilization, there was an increase in castor seed yield in the treatment supplied with 120 kg N ha −1 , as also reported by Pari et al. (2022), who obtained similar results for castor hybrids under a Mediterranean climate with highinput field management (100 kg N ha −1 ) and Calcagno et al. ( 2023), who investigated the perennial habit of castor under different irrigation and nitrogen fertilization levels in the Mediterranean area while maintaining the crop over a 2-year period.Weiss (2000) reported that nitrogen fertilization had little effect on seed oil content, which is consistent with our results.The nitrogen fertilization treatment showed a lower seed oil content T A B L E 5 ANOVA on daily biomethane (DCH 4 ) of untreated and fungal pretreated capsule husks after 30-day incubation under two levels of nitrogen fertilization.T A B L E 6 Two-way ANOVA on cumulated biomethane (∑CH 4 ) of untreated and fungal pretreated capsule husks after 30-day incubation under two levels of nitrogen fertilization.than the unfertilized treatment, and similar results were reported in previous studies on other oil crops.Farahbakhsh et al. ( 2006) observed a significant decrease in oil percentage with an increase in nitrogen application in Brassica napus L. Sawana et al. (2007) reported that seed oil content decreased slightly with an increase in the nitrogen rate from 95.2 to 142.8 kg ha −1 in cottonseed.Froment et al. (2000) found that the seed oil content of linseeds decreased from 38.3% to 34.6% when additional nitrogen was applied.Although seed oil content was reduced, seed oil yield was greater for the N120 treatment than for N0 since the main contributor to the total oil yield is the seed yield, as reported by Koutroubas et al. (1999).

Source
Composition analysis of the capsule husks showed significant differences in cellulose and lignin content resulting from the different fertilization levels, with fertilized biomass containing more cellulose and less lignin than unfertilized biomass.The presence of a significant amount of carbohydrates in the capsule residue confirmed the potential of this substrate for use in AD for advanced biomethane production.The ratio of structural carbohydrates (hemicellulose and cellulose) to lignin was used as an indicator of the digestibility of the substrate during the AD process (Oleszek & Matyka, 2017).
The highest ratio was reported for the fertilized residue, suggesting that nitrogen fertilization positively affects lignocellulose's susceptibility to AD.However, the recalcitrant nature of capsule husks caused by their high lignin content hinders their direct conversion.Among the pretreatment methods, biological pretreatment plays a key role in reducing the use of chemicals and energy.Several studies have reported that fungal treatment has positive effects on the degradation of lignocellulosic biomass (Noonari et al., 2020;Piccitto et al., 2022;Van Kuijk, Hendriks, et al., 2016;Van Kuijk, Sonnenberg, et al., 2016;Wan & Li, 2010;Zheng et al., 2014).P. ostreatus is one of the most studied WRF, capable of producing a hydrolytic enzyme complex in different lignocellulosic biomasses or under different fermentation conditions (Ding et al., 2019;Elisashvili et al., 2008;Kainthola et al., 2019;Mustafa et al., 2016;Sánchez, 2009).
P. ostreatus is considered a moderately selective lignin degrader because of its significant consumption of cellulose in lignocellulosic substrates, especially with prolonged pretreatment times (Taniguchi et al., 2005).In our pretreatment experiments, P. ostreatus showed the highest hemicellulose (33.8% for N120) and cellulose degradation (34.5% for N0), whereas I. lacteus showed the highest lignin loss (25.1% and 20.6% in the fertilized and unfertilized treatments, respectively).The selective lignindegrading capability of I. lacteus was also investigated by T A B L E 7 Cumulative methane production in relation to different pretreatment and nitrogen levels (N0-0 kg N ha −1 and N120-120 kg N ha −1 ).Note: Average values of all nitrogen levels (last column) and average values of all pretreatment (last row) followed by the same letter do not significantly differ at p ≤ 0.05.Significant interaction (LSD P×F p ≤ 0.05 (10.19)).

F I G U R E 8
Nonlinear regression between the cumulative methane yield and selectivity index of Pleurotus ostreatus and Irpex lacteus.Yu et al. (2010) who reported lignin loss from 75.67% to 80.00% after fungal treatment with I. lacteus during the mild alkaline pretreatment of corn stalks.However, the efficiency of fungal pretreatment depends on the biomass, and fungal selectivity depends on the process conditions (fungal strain, time, and temperature) (Hernández-Beltrán et al., 2019;Wan & Li, 2012).The high holocellulose content lost during pretreatment decreased biomethane production.
Limited studies have been conducted on the potential biomethane production from castor plants and castor seed cakes; however, there has been no research on biomethane production from capsule husks.Quezada-Morales et al. ( 2023) investigated the potential to produce biogas from the residues (leaves, stems, and seed bagasse) of Ricinus communis applying enzymatic and chemical pretreatments at different temperature and humidity levels reporting that residues of the seed bagasse produced the highest biogas yields both at room temperature and at 37°C under cellulase pretreatment (263.41 and 460.63 mL gVS −1 respectively).Bateni et al. (2014) evaluated the effect of alkaline pretreatment at different temperatures and times on BMY from castor stems, leaves, and castor seed cake.Pretreatment increased the biomethane production of castor stem from 80.8 to 145.5 mL/g VS.In contrast, alkaline pretreatment had a negative effect on biomethane production from both leaves and castor seed cakes.
In our experiments, P. ostreatus and I. lacteus pretreatments showed promising results for the AD of capsule husks, increasing BMY compared to untreated biomass.
I. lacteus pretreatment reached the highest BMY (103.2NmL g −1 VS) for the fertilized biomass because of the higher consumption of lignin (25.1%) and the lower cellulose loss (20.4%) that occurred during the pretreatment compared to the other pretreatments.This was confirmed by the highest selectivity (1.25) reached by I. lacteus.
Despite the consumption of cellulose during fungal growth, pretreatment contributes to the degradation of lignocellulosic biomass, making cellulose and hemicellulose more accessible to microbial attack during AD and improving the BMY compared to untreated biomass.
The results highlighted the positive effect of nitrogen fertilization on biomethane production, owing to the differences in composition between unfertilized and fertilized biomass.Fertilized biomass presented a greater hemicellulose and cellulose content and a lower lignin content than unfertilized biomass.Previous studies have reported similar results for fertilizer composition and biomass of different lignocellulosic crops (Blümmel et al., 2003;Mahmood et al., 2022;Oleszek & Matyka, 2017;Scordia et al., 2020).
A positive effect of nitrogen fertilization and fungal pretreatment on the BMY per hectare was also found.Considering the positive effects of fungal pretreatment on biomethane production, there are several variables to consider in a biorefinery system based on lignocellulosic feedstock, such as biomass availability, biomass composition,  and pretreatment which make it difficult to scale up biorefineries.Therefore, extensive studies are required to assess the economic, social, and environmental aspects of process sustainability from crop cultivation to bioproduct conversion in a biorefinery approach.Recently, several studies (Aghbashlo et al., 2022;Gheewala, 2023;Khoshnevisan et al., 2018) adopted different assessment methodologies based on LCA, energy, exergy, emergy, or a combination of these to identify the sustainability and environmental benefits of producing biofuels and bioproducts over their fossil-derived counterparts, showing that these environmental advantages could be improved by using whole feedstock biomass for value-added products and chemicals through biorefineries.

| CONCLUSIONS AND PROSPECTS
This study highlighted the suitability of castor, which can provide satisfactory seed and oil yields, for cultivation in a Mediterranean climate.The analysis of the composition enabled the consideration of the castor capsule residue as a potential substrate for AD to produce advanced biomethane.The nitrogen-fertilized biomass showed a higher cellulose content and lower lignin content than the unfertilized biomass, and this composition resulted in the highest biomethane production.However, the presence of lignin is an obstacle in bioenergy conversion processes.Thus, pretreatment to degrade lignin and improve the conversion efficiency is necessary.
The feasibility of using P. ostreatus and I. lacteus, two of the most common strains used for biological pretreatment, to selectively degrade lignin with minimum cellulose loss and to improve BMY were confirmed.This research enables to obtain biomethane from a byproduct generated by castor cultivation, thereby minimizing the quantity of waste released into the environment with effective waste management from the perspective of a circular economy that reduces fossil fuel dependence and greenhouse gas emissions.
Further studies are necessary to ascertain the optimal agronomic practices for improving seed and oil yields.Moreover, from a biorefinery perspective, investigation of other pretreatment methods could help enhance the energy conversion of this residual substrate and other parts of the plant, such as stems and leaves.Techno-economic studies using sustainability assessment tools, such as life cycle assessment and exergy-based approaches, should be adopted to understand the sustainability of the process better.
Yield-related components in castor at different nitrogen fertilization.Different letters in the same columns indicate significant differences between fertilization levels (p ≤ 0.05); values are expressed as mean ± standard deviation (n = 3).F I G U R E 3 Castor seed, capsules and oil yield (kg ha −1 ± SE) under nitrogen fertilization levels (N0-0 kg N ha −1 and N120-120 kg N ha −1 ).Different letters indicate significant differences according to LSD test at p ≤ 0.05.F I G U R E 4 Seed oil content (% ± SE) under nitrogen fertilization levels (N0-0 kg N ha −1 and N120-120 kg N ha −1 ).Different letters indicate significant differences according to LSD test at p ≤ 0.05.
Two-way ANOVA on biomethane yield of untreated and fungal pretreated capsule husks under two levels of nitrogen fertilization.Degree of freedom (df) and adjusted mean square significance: p ≤ 0.001 (***), not significant (ns).

F I G U R E 9
On the left, biomethane yield (m 3 CH 4 ha −1 ± SE) of unfertilized and fertilized pretreated with Pleurotus ostreatus (N0P and N120P, respectively), unfertilized and fertilized pretreated with Irpex lacteus (N0I and N120I, respectively) and untreated unfertilized and fertilized capsule husks (N0 Untreated and N120 Untreated, respectively).On the right, mean separation, with different letters representing statistically significant means according to the LSD test (p ≤ 0.05).
Summary of castor studies for biofuels production.
T A B L E 1