The Advantages of Soaking with Aqueous Ammonia Pre-Treatment Process of Oil Palm Empty Fruit Bunches

Oil palm empty fruit bunches (OPEFB) as one of lignocellulosic biomass, consists of three main components: cellulose, lignin, and hemicelluloses. Cellulose may occur in a crystalline form in addition to amorphous form. Cellulose and hemicelluloses can be converted into fermentable sugar, i.e. glucose and xylose, by chemical or enzymatic hydrolysis. This sugars is usually used as a feedstock for bioethanol production. Lignin and high-crystallinity of cellulose inhibit the performance of enzymatic hydrolysis process. Therefore pretreatment is necessary to remove lignin, decrease the crystallinity and improve the yield of enzymatic hydrolysis. One of the developed pre-treatment processes is soaking with aqueous ammonia solution or called by SAA process. In this study, the advantage of SAA pre-treatment process was investigated using ammonium solution prior to the enzymatic hydrolysis. The pre-treatments were carried out in various ammonium concentrations (5%, 7.5%, and 10%) at mild conditions (25 C and 1 atm) for 24 hrs, and with or without a following additional diluted-acid pre-treatment (92 – 98 C and 1 atm for 1 hour). The pre-treated materials were then enzymatically hydrolyzed by cellulose and β-glycosidase for 96 hrs and anaerobically digested by inoculums microbial. The changes in cellulose crystallinity were analyzed by FTIR Spectroscopy. The OPEFB pre-treated by 10% ammonium solution without a following dilute-acid pre-treatment shows low crystallinity with a crystallinity index (CI) of 0.80 compared to that of the untreated material (2.11). The method could increase the yield of hydrolyzed glucose to 79% compared with that of untreated material (13 g/g glucan added). The results also show increasing the methane production from 0.18 Nm /g volatile solid to 0.35 Nm/g volatile solid via anaerobic digestion. Keywords—ammonia pre-treatment; dilute-acid pre-treatment; crystallinity index; enzymatic hydrolysis; anaerobic digestion


I. INTRODUCTION
The technological development of the production of biomass-based energy as a renewable fuel is influenced by fossil energy depletion; dampen price volatility in transportation fuels; mediate economic and security concerns related to importing oil; and reduced CO 2 , SO x , and NO x emissions along with associated risks of climate change and global oceanic acidification [1]. Agriculture residue, municipal and/or industrial waste, wood chips, and any edible and non-edible biomass were converting to fuels and another chemical to solve energy issue and the waste disposal problems [2]. The main concern of many researchers is in producing bioethanol and biogas. Both of two products can be produced from carbohydrate content of the material, such as lignocellulosic biomass.
Oil palm empty fruit bunches (OPEFB) is the largest waste fraction from oil palm plantation in Indonesia. The possibility of using OPEFB as a raw material in a variety of applications was supported by many researchers, including power generation [3] citric acid [4], composites [5], and paper production [6]. Recently, OPEFB was utilized as raw material for bioethanol and biogas production.
As a lignocellulosic biomass, OPEFB is made up of a complicated matrix of cellulose and lignin bound by hemicellulose chains [7]. This matrix should be broken during a suitable pre-treatment step to increase the digestibility of enzymatic by removing lignin as inhibition, reducing the crystallinity of cellulose, and increasing the fraction of amorphous cellulose. Pre-treatment was the most critical steps in lignocellulosic biomass-based energy as renewable fuel production involving the overall cost of the bio-conversion process.
The pre-treatment process is main processing challenge in the bio-conversion production from lignocellulosic biomass [8]. Numerous biomass pre-treatment methods have been developed and tested in laboratories and pilot plants [9]. Dilute acid pre-treatment, liquid-hot water, ammonia fiber explosion, lime pre-treatment or ionic liquid are the most studied technologies. Physical pre-treatment combined with a chemical such as microwave-NaOH pretreatment was also reported as an effective method [10]. However, most of these pre-treatment methods require extreme conditions such as high pressure and/or temperature. The extreme conditions could result in the formation of biologically inhibitory compounds derived from lignin and carbohydrate decomposition. Furan, formic and levulinic acid are generated from acidic pre-treatment (above 120 o C) and require downstream detoxification process.
Wet-alkali oxidized pre-treatment (>90 o C) also reported generating the phenolic products and various carboxylic acid [11]. Improvement in pre-treatment method will reduce the number of enzymes or microbial used and decreased an overall cost of lignocellulosic conversion [12], [13]. Additionally, removing lignin and hemicelluloses also desired to improve the enzymatic and anaerobic digestibility. Enzymatic hydrolysis is necessary to liberate sugars for the subsequent fermentation to ethanol, and the rates are typically 3-30 times faster for amorphous cellulose than for the high-crystalline cellulose. Anaerobic digestion is one of technique to liberate methane from biomass in the absence of oxygen, whereas methane can be used as biogas fuel 14 . However, pre-treatment has been viewed as one of the most expensive processing steps. Therefore intensive research and development efforts are carried out for improvement of the efficiency and lowering the cost of different pre-treatment technologies, such as steam explosion, liquid hot water, dilute acid, lime, and ammonia pre-treatments.
One of the developed pre-treatment methods, soaking in aqueous ammonia (SAA) at the mild condition, has not yet been widely studied for OPEFB as lignocellulosic biomass material [15]. This method is highly selective for lignin removal and shows the significant swelling effect on lignocelluloses [11]. The removed lignin can be then utilized as a fuel or for the production of other chemicals. The SAA at room temperature has higher selectivity retaining the hemicelluloses onto solids fraction as well as removing lignin content. Retained hemicelluloses can be then hydrolyzed by a following dilute acid treatment and may be used for several applications as for producing chemicals, fuels, and in the food industry. Fractionation of lignocellulosic biomass into the three main constituents, i.e. lignin, hemicelluloses, and cellulose, is a concept being developed as a means to improve the overall biomass utilization [16].
The main purpose of the present study was to investigate the advantages of SAA pre-treatment process to OPEFB as a lignocellulosic material, compare to other feedstock of lignocellulosic materials. Furthermore, the effect of ammonia pre-treatment and the synergistic combination effect of the ammonia steeping at mild condition combined with acid hydrolysis of high crystalline cellulose and for the removal of lignin and hemicelluloses of OPEFB were investigated in this paper. The goals of this treatment were to enhance the digestibility of enzymatic and anaerobic; together with the separation lignin, hemicelluloses, and cellulose as part of materials.

A. Raw Material
The OPEFB was kindly supplied by Gersindo Minang Plantation-POM Company, Pasaman Barat (West Sumatera, Indonesia). The material was washed by water, oven-dried at 40 o C, cut into pieces with size less than 3 cm, and stored in plastic bags at room temperature until further processing.

B. Delignification
In order to remove lignin, 20 grams of OPEFB was soaked in aqueous ammonia at 5, 7.5, and 10 % (w/w) concentration and incubated at room temperature for 24 hrs. The mixture was then filtered with vacuum filter and washed to remove residual solvent until a normal pH (+ 7) was achieved. The solid residue was oven-dried at 40 o C and keeps it in a plastic bag until further use. The ammonia in lignin-rich solution was recovered by vacuum rotary evaporation thus it can be then reused for the next delignification process. This solution also can be used for any other purpose, such as fertilizer and synthetic rubber.

C. Hemicelluloses Hydrolysis
Delignified OPEFB (solid residue) was treated with a 0.1M sulphuric acid solution at 92 -98 o C for 1 hour. The pre-treated cellulosic residue was then washed with deionized water to remove residual acid until a normal pH (+ 7) was achieved, oven-dried at 40 o C, and keep it in a plastic bag for next step.

D. Enzymatic Hydrolysis
50 ml 0.05M citrate buffer (pH 4.8) was mixed with an appropriate amount of treated and untreated samples to achieve a 3% w/v concentration in 100mL screw-capped Erlenmeyer flask. Then an enzyme mixture was added to achieve enzyme loadings equivalent to 30 FPU cellulase enzyme and 60 IU β-glucosidase enzyme for each gram of OPEFB. The enzyme was obtained from Novozyme. The enzymatic hydrolysis was carried out at 48 o C for 96 hours, while samples were taken regularly after 0, 12, 24, 36, 48, 72, and 96 h, and then analyzed for liberated sugar concentrations. Total released glucose after 96 hours of hydrolysis was used to calculate the enzymatic digestibility. The yield of enzymatic digestibility is defined as: The dehydration factor (1.111) is used to convert the cellulose chains to glucose monomers.

E. Anaerobic Digestion
Smaller flasks (118mL working volume) were used to measure the methane production from OPEFB. A 0.25g dryweight untreated and treated sample was mixed with 20 mL inoculums, where manure is used as raw material, together with 5mL of deionized water. Inoculums were kindly supported by Boras Energy och Miljo AB, Boras, Sweden. In order to remove the gas in a sample, the flasks were flushed with nitrogen, sealed with a rubber septum and placed in a shaker at 55 o C for the thermophilic condition. To measure the methane production, gas from the headspace of each flask was taken regularly and analysed by gas chromatography. The process will be stopped after 50 days digestion.

F. Analysis
The constituent of raw materials and pre-treated materials was determined according to NREL Laboratory Analytical Procedure. The sugar content was determined based on monomer content measured after two-step acid hydrolysis procedure. In the first step, samples were treated with 72% (w/w) H 2 SO 4 at 30 o C for 2 h. The reaction mixture was then diluted to 4% (w/w) H 2 SO 4 and autoclaved at 121 o C for 1hour.The hydrolysis solution was filtered and analysed for sugar content by HPLC on an ion exchange column (Aminex, HPX-87P, Bio-Rad, Hercules, CA). Glucose from enzymatic hydrolysis was analysed by HPLC using another ion exchange column (Aminex, HPX-87H, Bio-Rad, Hercules, CA) at 60 o C with 5mM sulphuric acid as the eluent at a flow rate of 0.6 mL/min.
Gas chromatography (Auto System Perkin Elmer, Waltham, MA) was used to determine the methane produced in anaerobic digestion. This equipment was equipped with a packed column and a thermal conductivity detector (Perkin Elmer). The cellulose crystallinity was examined using Fourier Transform Infra Red (FTIR) spectrometer (Nicolet iS10, Thermo Fisher Scientific), with an average of 32 scans and resolution of 4 cm -1 in the range of 500 cm -1 and 4000 cm -1 . Pre-treatment was performed in duplicate, while enzymatic and anaerobic digestions were performed in triplicate. The analysis results were delivered as an average of the data.

A. Raw Material Contents
The composition of untreated and pre-treated OPEFB is summarized in Table 1. The lignin content of the untreated material was 29.60% of dry-weight. The xylan content that represents the composition of hemicelluloses in the material was 21.97% of dry-weight, and glucose as a cellulose content was 38.74% of dry-weight material. Compare to other lignocellulosic biomass, OPEFB is one of the ligninrich plants. This indicates that OPEFB can be classified as grass type (herbaceous), regarding similarity constituent with grass type of biomass ( Table 2).
The lignin value was decreased to around 23% after the pre-treatment (Table 1). However, there is no significant difference in delignification carried out in different treatment condition. The varying concentration of ammonia reagent and additional treatment, following with acid treatment were not resulting in significant different of lignin removal. In addition, the xylan contents of pre-treated OPEFB samples were lower than that of the untreated sample. 5% ammonia as a reagent was decreased xylan higher than 7.5% and 10% ammonia concentration. It proves that xylan, as a main part of hemicelluloses, was easy to hydrolyse by low base concentration. The degradation resulted as a consequence of cleavage of the linkages between lignin and hemicelluloses (Lignin-Carbohydrate Complexes, LCC). A two-step treatment with 5% ammonia and 0.1M H 2 SO 4 resulted in a removal of xylan by 81.75%, compared to 43.74% removal of xylan during one-step pre-treatment using 5% ammonia. This indicates that greater amount of xylan was released during acid pre-treatment. It proves that base and acid solution, even in a low concentration, could be easier dissolving hemicellulose in biomass content. On the other hand, the acid pre-treatment was reduced the lignin content of lignocellulosic biomass in small number also.
The pre-treatment with a lower concentration of ammonia (5%) also resulted in a reduction of existing glucan content in the lignocellulosic biomass, as a result of the dissolution of sugars contained in the material. But this phenomenon did not appear in a higher concentration of ammonia. Table 1 also shown that synergetic effect of two-step pre-treatment (ammonia and acid) not so significantly to remove lignin and reduce of glucan. The result of two steps pre-treatment showed a similar value with single ammonia pre-treatment. Furthermore, the acid pre-treatment was working well to reduce the amount of xylan in lignocellulosic biomass.

B. Enzymatic Hydrolysis
The effects of various treatment conditions on lignin and hemicellulose removal, crystallinity index, and enzymatic digestibility are shown in Table 3. One step pre-treatment offered higher insoluble solid recovery yield (around 88%) comparing with that of the two-step treatment. The solid lost under the two-step treatment was resulted by the removal of all the constituent of lignocelluloses, mainly hemicelluloses, as a result of the additional dilute-acid treatment step, and by the dissolution of sugars in the raw material. However, the removal of lignin was almost similar in both one-step and two-step pre-treatment methods due to the characteristic of lignin that can only be dissolved in alkaline (Table 1 and  Table 3). Moreover, a small amount of xylan can also be dissolved during the alkaline pre-treatment. This is typical of alkaline pre-treatment, which generally has a stronger effect on lignin than hemicellulose and cellulose.The retaining of hemicelluloses and cellulose into solid fraction was highest in 10% ammonia solution. It indicates that lower concentration of ammonia means a higher concentration of water in the solution, and it has big opportunity to break the chain of hemicelluloses and dilute the monomeric sugar from hemicellulose into solution.
The highest lignin removal was obtained when 10% (w/w) alkaline pre-treatment was combined with 0.1M dilute sulphuric acid pre-treatment. However, this sample did not show the best enzymatic digestibility after following enzymatic hydrolysis. Fig.1 showed the variation of glucose yield from enzymatic hydrolysis of untreated and treated OPEFB samples. The yield of glucose was increased from 13 (g/g glucan added) to 23.23 (g/g glucan added) after the pre-treatments. The highest increase of 78,7%, in glucose yield, was obtained after treatment by 10% (w/w) ammonia without a following additional dilute-sulphuric acid treatment step. This indicates that in addition to the presence of lignin there are other factors that influence the action of the enzymes on the biomass. According to Table 1 and Table  3, the presence of glucan after treatment on treated OPEFB gave a significantly impact to the enzymatic digestibility.
As it is shown in Table 3, pre-treatment with 10% ammonia conditions resulted in the lowest xylan removal, highest glucan recovery, as well as lowest crystallinity index, and as a consequence, the enzymatic digestibility was the highest for that sample. The presence of xylan does not affect the action of the enzymes used in this study because xylanase for the degradation of xylan to xylose was not included. As a result of the high concentration of solvent, the loss of sugars during the pre-treatment was slight. Therefore, a high enzymatic digestibility, as well as a high glucan recovery, could be achieved after this pre-treatment.
In addition, a two-step process with additional dilute-acid treatment step is not necessary, since comparing the result for to 10% ammonia treatment with or without an additional dilute-sulphuric acid treatment. The glucan recoveries in both cases were almost the same, 95,33 and 95,92%, respectively, in Table 3. On the other hand, the following process after ammonia treatment, the acid pre-treatment, was increasing the xylan removal. The percentage of xylan removal in the lignocellulosic residue after ammonia treatment was increased two-fold from the one step process. Therefore, the selectivity of ammonia treatment was proven by the result. The acid pre-treatment will be very effective to separate hemicelluloses from lignocellulosic biomass. Twostep pre-treatment could be the best technology to isolate the fraction of lignocelluloses, in the form of cellulose, hemicelluloses, and lignin.

Untreated
Amm 5% Amm 7.5% Amm 10% Amm 5% + H2SO4 Amm 7.5% + H2SO4 Amm 10% + H2SO4 Comparing to the result of other researchers in SAA process at any condition (Table 4), these result has a similarity with other. In other studies, using SAA method with different lignocellulosic feedstock, the enzymatic digestibility of rice straw pre-treated at 69 o C, 10 hours, 21% ammonia was 82% after 48 hours of saccharification time 18 , corn stover treated at 60 o C, 12 hours, 15% ammonia was 85% after 72 hours of saccharification time 20 , and barley hull treated at 75 o C, 48 hours, 15% ammonia was 83% after 96 hours enzymatic hydrolysis 15 . These results thus show that the enzymatic digestibility and hydrolyzability depend on feedstock, pre-treatment temperature, and ammonia concentration. It proves that SAA process is a promised technology in bioconversion process to produce monosaccharide (glucose) from lignocellulosic biomass.
The crystallinity index was obtained from the data of FTIR spectra. FTIR spectroscopy is a successful technique for the characterization the structure of cellulose 21 . Fig. 2 shows the IR spectra of the untreated vs. treated material. The 1,425 and 898 cm -1 absorption bands, which are assigned to the crystalline cellulose I and cellulose II, respectively, were used to study the type of crystalline cellulose and the changes in the crystallinity. Table 3 shows the total crystallinity indexes (TCI), which can be calculated as the absorbance ratio A1,425/A898 from the spectra.
Comparing the FTIR spectra in Fig. 2 shows that pretreatments with ammonia in various concentrations reduced the absorbance band corresponding to cellulose I (crystalline cellulose) and on the other hand increased the absorbance band corresponding to cellulose II (amorphous part of cellulose), resulting in a decrease in total crystallinity indexes after the treatments. As Table 3 shows, one-step pretreatment with 10% ammonia have the lowest index of crystallinity and also gives the highest enzymatic digestibility. The correlation between two effects of SAA process indicates that highest digestibility of enzymatic will obtained by decreasing of cellulose crystallinity.
The acid pre-treatment was not given the real effect to TCI. Table 3 shows that after soaking in acid solution, the TCI of the various sample has a similar value (around 1.28). The result shown that additional step with the acid solution was not given any advantages in reduced lignin content, decreased the crystallinity of cellulose and enhanced the enzymatic digestibility.

C. Anaerobic Digestion
Biogas (methane) production by anaerobic digestion was the effective method to calculate the successful of pretreatment beside enzymatic digestion. Fig. 3 had shown the methane production of untreated and treated OPEFB per day (for 50 days treatment). At normal condition, based on theoretical, 1 gram cellulose will produce 415mL methane. The result shows that methane production was increased after ammonia pre-treatment for almost two-fold. The 10% ammonia concentration gave the highest result for this process after 50 days fermentation. Regarding Table 1 and Table 3, it proves that lignin removal, the crystallinity of cellulose and solid retained has significant effect to produce methane.
Methane production in two-step pre-treatment was not studied in this paper. Based on the result of enzymatic digestion, acid-solution pre-treatment did not have significant effect to enhance the digestibility of anaerobic fermentation, due to the characteristic of acid pre-treatment does not improve the glucan recovery. However, the composition of glucan in treated biomass is the main factor in increasing the digestibility. The result from two analytical methods, enzymatic and anaerobic digestions, has shown a similar trend for untreated and treated material. The digestibility of treated OPEFB was enhancing due to increasing of ammonia concentration as a per-treatment reagent. Furthermore, the correlation of two digestibilities was illustrated in Fig. 4. Fig. 4  Raw EFB 5% Amm onia 7.5% Am m onia 10% A mm onia correlation of digestibility of anaerobic versus enzymatic digestion. There is a linear correlation between anaerobic and enzymatic digestion for untreated and treated OPEFB. The anaerobic digestion will increase while the enzymatic digestion enhanced. It means that pre-treatment had an impact not only to enzymatic digestibility in producing bioethanol but also improve the digestion of anaerobic to produce methane as biogas. In spite of the expectations that the presence of lignin and hemicellulose in biomass can inhibit the enzymatic hydrolysis, our results showed that the main factor in increasing the enzymatic digestibility after aqueousammonia treatment of OPEFB is the reduced the crystallinity in the structure of cellulose and enhanced the glucan recovery. Even the higher removal of lignin or hemicelluloses had an impact to the digestibility, but the effectiveness is not significant. However, the selectivity of ammonia solution as a delignification agent in removing lignin, retaining hemicelluloses and cellulose into a solid fraction, and reducing the crystallinity of cellulose are the advantages of SAA process for bio-conversion by enhancing the enzymatic digestibility and anaerobic digestion. The twostep of pre-treatment process, soaking with ammonia solution and following with acid solution, was not significantly effected to both of digestibility methods. This process was necessary to isolate the component of lignocellulosic biomass into three main fractions, i.e. cellulose, hemicelluloses, and lignin.