NC State
BioResources
Abdul Latif, A., Shuhaida, H., Sajab, M. S., Markom, M. (2024). “Lab to pilot scale assessment on the pretreatment of empty fruit bunch using anhydrous ammonia,” BioResources 19(3), 5764-5792.

Abstract

Laboratory and pilot scale pretreatment processes were assessed for empty fruit bunch (EFB) at different morphologies using anhydrous ammonia pretreatment (AAP). The AAP was used to deconstruct the complex structure of EFB through physical and chemical reaction to promote efficient conversion of the carbohydrates to monomeric sugars. Different morphologies of EFB samples used were unpressed EFB (UE), pressed EFB (PE), pressed and shredded EFB (PES), and lastly pressed, shredded and ground EFB (PESG). The APP process was optimized using a 1.0 L laboratory scale reactor and further scaled up to a 22 L pressure vessel (AAPB). AAP-PESG contained 76.2%, and AAP-PES contained 75.5% of structural carbohydrates, showing no significant difference. AAP-UE showed the lowest glucan conversion of 28%. The optimal laboratory conditions adopted were 135 °C, 30 min, moist to dry EFB loading of 2:1, and ammonia to dry EFB loading of 1:1. Glucan conversion of AAP-PES were 87%, 80%, and 62% at 1%, 3%, and 6% glucan loadings, respectively. The AAP-PES detected acetamide concentration at 7.3 mg/g, while AAPB-PES was only at 4.4 mg/g. Chemical composition, Fourier transform infrared spectroscopy, Brunauer Emmett Teller surface area, and scanning electron microscopy supported the assessment of AAP and AAPB processes.


Download PDF

Full Article

Lab to Pilot Scale Assessment on the Pretreatment of Empty Fruit Bunch Using Anhydrous Ammonia

Azuan Abdul Latif,a,b Harun Shuhaida,a,* Mohd Shaiful Sajab,a and Masturah Markom a

Laboratory and pilot scale pretreatment processes were assessed for empty fruit bunch (EFB) at different morphologies using anhydrous ammonia pretreatment (AAP). The AAP was used to deconstruct the complex structure of EFB through physical and chemical reaction to promote efficient conversion of the carbohydrates to monomeric sugars. Different morphologies of EFB samples used were unpressed EFB (UE), pressed EFB (PE), pressed and shredded EFB (PES), and lastly pressed, shredded and ground EFB (PESG). The APP process was optimized using a 1.0 L laboratory scale reactor and further scaled up to a 22 L pressure vessel (AAPB). AAP-PESG contained 76.2%, and AAP-PES contained 75.5% of structural carbohydrates, showing no significant difference. AAP-UE showed the lowest glucan conversion of 28%. The optimal laboratory conditions adopted were 135 °C, 30 min, moist to dry EFB loading of 2:1, and ammonia to dry EFB loading of 1:1. Glucan conversion of AAP-PES were 87%, 80%, and 62% at 1%, 3%, and 6% glucan loadings, respectively. The AAP-PES detected acetamide concentration at 7.3 mg/g, while AAPB-PES was only at 4.4 mg/g. Chemical composition, Fourier transform infrared spectroscopy, Brunauer Emmett Teller surface area, and scanning electron microscopy supported the assessment of AAP and AAPB processes.

DOI: 10.15376/biores.19.3.5764-5792

Keywords: Anhydrous ammonia pretreatment; Empty fruit bunch; Pilot-scale; Chemical composition; Enzymatic hydrolysis

Contact information: a: Department of Chemical and Process Engineering, Faculty of Engineering and Built Environment, Universiti Kebangsaan Malaysia, 43600 Bangi, Selangor, Malaysia; b: Renewable Energy and Green Technology Unit, TNB Research Sdn. Bhd, 43000 Kajang, Selangor, Malaysia;

* Corresponding author: harun.shuhaida@ukm.edu.my

GRAPHICAL ABSTRACT

INTRODUCTION

As one of the world’s biggest palm oil exporters, Malaysia produces abundant empty fruit bunch (EFB) from an oil palm tree (Elaeis guineensis) in the palm oil mill. The EFB accounts for 25% to 30% of solid waste generated for every tonne of fresh fruit bunch (FFB) processed. The volume increases annually, with only a small fraction of EFB being employed for value-added products, while most are underutilized (Barlianti et al. 2015). The abundance of EFB produced has created many environmental issues, such as fouling and attraction of pests, which can pose challenges for processing mills to handle. Managing and disposing of large quantities of EFB can be time-consuming and costly (Suhartini et al. 2022). Additionally, if not managed properly, the high amounts of EFB can lead to operational issues or interfering with production processes. To address these environmental issues, it is important for the palm oil industry to implement effective EFB management practices. On top of current efforts, which utilize soil mulching using EFB and composting treatment as enriched biofertilizer, pretreatment of EFB can lead to more efficient and environmentally friendly ways to process and utilize EFB in the palm oil mill plant.

In general, for successful operation of an oil palm mill with a composting plant facility, the EFB must go through a pressing process to recover the oil palm and subsequently increase the crude palm oil (CPO) recovery. Additionally, the EFB yields valuable byproducts, including nutrient rich content for biofertilizer and as feed for animals (Dolah et al. 2021). The bunch-size unpressed EFB (UE) contains approximately 2% of the oil compared to pressed EFB (PE), with only 0.2% oil content (Chang 2014; Law et al. 2017). Additionally, PE is further subjected to the shredding process prior to the composting process that yields shredded PE (PES) as loose and strand fibre with a length between 50 and 250 mm. Consequently, different morphology and size of EFB in the form of UE, PE, and PES from the oil mill causes limitation in utilizing them all. Although these different types of UE, PE, and PES are abundantly available, thus far these different EFB samples have not been efficiently utilized, and most are used for low-value products that result in quick gain profit. The different morphology and type of EFB will give different effectiveness of utilization due to the characteristics such as moisture content, bulk density, and chemical composition. Moreover, additional processing of the EFB will require higher energy and cost, while the larger size of EFB is difficult to handle for storage and mobilization. Many studies have focused on utilizing the small sizes of EFB (Abdul et al. 2016; Tan et al. 2016), which requires an energy-intensive grinding process that is not economical and tedious to oil palm millers for process of commercialisation. Thus, the most prevailing practice for EFB pretreatment is hydrothermal treatment due to its technology availability and its potential to enhance oil recovery from EFB (Diyanilla et al. 2020).

In the effort for an efficient and sustainable utilization of the biomass as the new green resource, the anhydrous ammonia based pretreatment (AAP) has been developed, in the laboratory as well as in the pilot scale, to mainly alter and open up the EFB cell wall structure. The deconstruction of the lignin-carbohydrate matrix improved the hemicellulose (arabinoxylan) hydrolysability, increased the cellulose crystallinity, and finally made the cellulose more accessible to enzymes in the hydrolysis process to produce fermentable sugars, such as hexose (C6) and pentose (C5), in the biorefinery scheme process (Harun et al. 2013). Previous studies have shown that the AAP has significant effect on the glucose (C6 sugar) yield from different lignocellulosic biomass. Furthermore, because of the high volatility of anhydrous ammonia, it provides a dry-to-dry pretreatment process, and the used ammonia can easily be recovered and reused for the subsequent batches of pretreatment (Lau et al. 2010). However, the referenced study primarily investigated the performance of ammonia pretreatment on standard laboratory size with milled-size EFB, while this current research evaluates EFB with different morphologies at the lab scale and assessing their performance during scale-up.

There are several common issues when utilizing EFB as a source of energy or other products in palm oil mill. The EFB contains high moisture, and drying of the fibre can be energy-intensive and add to the overall cost of operation. Additionally, EFB has relatively low energy density; thus, it requires a large volume of biomass to produce certain amount of energy, which can be a challenge in terms of handling, transportation, and storage. Thus, to achieve the economics of scale and high bioconversion efficiencies can be difficult, especially for smaller scale operations (Chiew and Shimada 2013).

Most of the utilization studies of the lignocellulosic biomass have been evaluated and reported using alkaline, acid, and organic solvent as the pretreatment agent in breaking the plant cell wall structure (Lau et al. 2010; Harun et al. 2013; Abdul et al. 2016; Tan et al. 2016; Law et al. 2017). However, with respect to different morphology and type of EFB for examples, such as UE, PE, and PES, there are limited reports exploring on the utilization of EFB, particularly with this different type of morphology and size using AAP and comparing the results and opportunity for the large-scale utilization. The focus of this work was to verify that the finding observed at small scale can be successfully replicated and maintained consistently at larger scales. A further goal was to observe the potential of continuous process by reusing ammonia, which holds promise for viable commercial applications. Therefore, a comprehensive pretreatment study up to pilot scale system is required so that it can show as a proof-of-concept technology and be resourceful for the industry (Akhlisah et al. 2021; Morais et al. 2022).

This study aimed to assess different EFB morphology using anhydrous ammonia pretreatment from laboratory to pilot-scale. The pretreatment process was optimized at lab-scale and series of experiments were conducted to determine the optimum process parameter in the AAPB system. The effectiveness of pretreatment was evaluated via structural and non-structural carbohydrates content, morphology changes, sugar conversion, and inhibitor formation from the enzymatic hydrolysis. Although some EFB in the processing mill already has been utilised for value-added products, proper and efficient biomass utilisation is important for sustainable and economic impacts in the biomass industry.

EXPERIMENTAL

Raw Material

The EFB was collected from Flemington oil palm mill of Sime Darby Plantation Berhad (Bhd) in Bagan Datuk, Perak, Malaysia. There were three categories of the samples taken, as shown in Fig. 1(a), which were UE, PE, and PES. All samples were freshly produced in the oil palm mill on the same day of collection. Samples were properly sealed in gunny bags and transported to the laboratory within 4 to 5 hours and were immediately keep in a cool room at 4 °C overnight. Moisture content and mass of EFB upon collection were also recorded.

Fig. 1. (a) Flow of different categories of EFB samples from oil palm mill, (b) Schematic diagram of anhydrous ammonia pretreatment process

Preparation of EFB Samples

There were four categories of EFB, as shown in Fig. 1(a), to be analyzed in this study. These were UE, PE, PES, and PESG. Each morphology the EFB was dried for three to four consecutive days until the moisture content reduced to 10 ± 3% (dry weight basis, DWB). The EFB was sun-dried in an open environment, with daytime temperatures ranging from 30 to 35 °C (9 am to 4 pm) and nighttime temperatures between 25 and 28 °C. During the night, the drying process occurred under a roof but in an open area. To facilitate the drying process, all EFB samples were exposed to direct sunlight. Given the substantial quantity collected (approximately 100 kg wet weight) and the large size of the whole fruit, drying them inside a laboratory oven was not feasible. The PESG was ground into approximately 2-mm particle size using a universal cutting mill (Pulverisette 15 Fritsch, Germany) with a 2-mm screen mesh to achieve uniform particle size. Then, PESG was packed in a sealed bag and stored at 4 °C to maintain the moisture content until further use.

Moisture Content of EFB Samples

The moisture content of each EFB category was recorded at a different stage of preparation which are as follows: (1) collection at the mill, (2) followed by receiving the EFB at the laboratory, (3) after the drying process, and (4) after the pretreatment process. The purpose of moisture sampling was to observe the loss and gain of moisture content from the EFB throughout the experiment. Samples were periodically flipped up and down every 2 h to quicken the drying process. Any spoilt samples were removed immediately to ensure no biological contamination in the samples. The activity for the moisture content profiling of UE, PE, and PES was taken every 12 h to observe moisture loss during the drying process. The moisture content was determined via a Moisture Analyser (Denver Instrument, Germany) and recorded based on DWB.

Anhydrous Ammonia Pretreatment

Laboratory-scale experiment

As in ammonia fiber explosion (AFEX) pretreatment (Lau et al. 2010), a similar set up was also applied in the AAP process to pretreat these different morphological EFB samples. The pretreatment process was performed in a 1-L closed stainless steel vessel reactor (Ammar Reactor, India). Each pre-wetted EFB sample was loaded into this reaction vessel and clamped shut for the preheating process. The ammonia was transferred into this vessel using 1-L transfer cylinder prior to preheating process. Both the transfer cylinder and pressure vessel were preheated simultaneously until stable heat and pressure was achieved. The pretreatment conditions were based on the previous AFEX study for EFB pretreatment (Abdul et al. 2016) with (1) pretreatment temperature of 135 °C, (2) ammonia to EFB loading ratio of 1 to 1, (3) water to EFB loading ratio of 1 to 1, and (4) pretreatment time of 45 min. The 45 min residence time started once the pressure vessel reached the pretreatment temperature of 135 °C. Once the time was completed, the pressure in the vessel was slowly released by gradual opening of the vessel valve, causing the ammonia vapour to be evaporated out of the vessel. The vessel cooling process started once the pressure inside reached the atmospheric pressure, and it was cooled to the room temperature or until it was safe to unclamp and open the vessel. Subsequently, the pretreated EFB was unloaded from the pressure vessel and allowed to dry overnight under the fume hood for removal of excess residual ammonia. Pretreated EFB was then stored at 4 °C until further use. Figure 1(b) illustrates the schematic diagram of the AAP process.

Fig. 2. (a) The schematics of pilot scale diagram of batch AAP (AAPB) system, (b) Configuration of laboratory and pilot scale pretreatment vessels

Determination of different pretreatment parameter conditions at laboratory scale.

The optimum pretreatment parameters in this study were determined through a one factor at a time method, where different components within the pretreatment condition were optimised while the other conditions remained constant. Due to high resistance from contamination and storage easiness PES morphology was chosen as the biomass for the determination of optimal conditions study. The parameters involved were pretreatment time (15, 30, 45, 60, and 75 min), the ratio of moisture content to dry EFB (0.5:1, 1:1, 2:1, 3:1, and 4:1), pretreatment temperature (95, 115, 135, 155, 175, 195, and 215 °C), and the ratio of ammonia loading to dry EFB (0.5:1, 1:1, 1.5:2, 2:1, and 2.5:1). Each parameter pretreatment efficacy was evaluated based on the sugar concentration and conversion from 1% glucan loading enzymatic hydrolysis.

Pilot-scale set-up

The AAP process was scaled up in a pilot set up at Demo Plant Zero-Waste technology at KKS Tennamaram, Selangor. The pilot scale of batch AAP (AAPB) system consists of 22-L pretreatment vessel, equipped with a booster pump, a steam generator, and a condenser, as shown in Fig. 2(a). Two identical pretreatment vessels were used in the process design, aiming for continuous operation with ammonia reuse. Notably, the operating pressure at the pilot scale was significantly lower (8 to 10 bar) compared to the laboratory scale, which operated at 35 bar. The vessel was equipped with top and bottom flange covers, three cylindrical perforated baskets for biomass loading, and electrical heating jacket. Sample of PES were loaded into each basket in the vessel and densified by hand pressure. The top flange cover was closed once loaded with EFB for the pre-steaming process using saturated steam (pressure of steam). The outlet at the bottom flange cover was opened to allow displaced air to escape from the pressure vessel. Once the inside vessel temperature reached 70 °C, the steam supply was stopped and the outlet at the bottom flange cover was closed. Immediately, the anhydrous vapour ammonia was directly loaded into the pressure vessel through the booster pump. This mixture of EFB and ammonia in the vessel was heated using the electrical jacket once the ammonia to EFB loading ratio was attained. The pretreatment conditions for the critical parameters were according to the optimised conditions obtained using the laboratory scale set up and was further tuned for the best operating conditions. Pretreatment process stops by releasing ammonia into the water container underneath the pretreatment vessel, preventing ammonia from escaping to the environment. The vessel was allowed to cool for a while until safe to open, and EFB samples were unloaded for drying purposes, letting the ammonia residue evaporate overnight. The pretreated samples were labelled according to the position of cylinder basket filter as B1-AAPB-PES, B2-AAP-PES, and B3-AAP-PES, respectively. The schematic diagrams for the configuration of both laboratory and pilot scale pretreatment vessels are shown in Fig. 2(b).

Chemical Composition Analysis of EFB

The chemical composition analysis was performed on all samples of untreated and pretreated EFB according to the Laboratory Analysis Protocol (LAP) developed by the National Renewable Energy Laboratory (Golden, CO, USA). All samples were subjected to a two-stage extraction process in Accelerated Solvent Extraction (ASE, Dionex 350, Sunnyvale, CA, USA) with water and 95% methanol as solvents for the respective extractions in the protocol. The extractives from water and ethanol extractions, including non-structural sugar, inorganic compound, nitrogenous compound, and waxes, must be removed to avoid disturbance in the structural carbohydrates and lignin determinations during acid hydrolysis. The acid hydrolysis was a two-stage process begins with a higher concentration of 75% HCl, followed by a subsequent with 4% HCl. Then, the hydrolysate was drawn and filtered for monosaccharides content analysis using high-performance liquid chromatography (HPLC) (Dionex Ultimate 3000, Thermo Scientific, Waltham, MA, USA). The remaining hydrolysate was collected via vacuum pump as acid-soluble lignin, and the residue was dried, weighed, and recorded as acid-insoluble lignin.

High-Performance Liquid Chromatography Analysis

The HPLC method was used to analyze the concentrations of the monosaccharides, such as glucose, xylose, and arabinose, in the hydrolysate samples obtained from the chemical compositional analysis. The HPLC device was equipped with an autosampler and refractive index (RI) detector (RefractoMax 520, ERC, Germany). Rezex ROA-Organic acid column (300 mm × 7.8 mm, Phenomenex, USA) with guard column (50 mm × 78 mm) was used for the separation of monosaccharides. The column temperature was maintained at 60 °C, and 0.005 M sulphuric acid (filtered and degassed) was used as the mobile phase with a flow rate of 0.6 mL/min and 10 µL injection volume. The standard curve was generated using different concentration of analytical grade sugar mixture (glucose, xylose, and arabinose).

Furfural, hydroxymethyl furfural (HMF), formic acid, acetic acid, and acetamide content were analyzed from the pretreatment hydrolysate. Rezex ROA-Organic acid column (300 mm × 7.8 mm, Phenomenex, USA) and C18 column (Phenomenex, USA) were used to separate organic acids and furfural, respectively. The column temperature was maintained at 40 °C using 0.005 M sulphuric acid (filtered and degassed) as the mobile phase with a flow rate of 0.6 mL/min and 20 µL injection volume.

Structural Carbohydrate Content

The structural carbohydrate content in the EFB samples was determined based on the concentration of monosaccharides in the hydrolysate samples using the following equations:

where 0.9, 0.88, and 0.88 are the conversion factor of glucose, xylose, and arabinose to the equivalent molecular weight of glucan, xylan, and arabinan, respectively.

Fourier Transform Infrared Spectroscopy

Fourier transform infrared (FTIR) analysis on untreated EFB, pretreated EFB, and unhydrolyzed pretreated EFB samples were performed using a Nicolet 6700 FTIR spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) with the Attenuated Total Reflectance (ATR) method. All samples were scanned from 4000 to 400 cm-1 at a wavelength resolution of 2 cm-1 and recorded in transmission mode. The aim was to determine the lateral order index (LOI), total crystalline index (TCI), and hydrogen bond index (HBI). The LOI, TCI, and HBI were obtained using height absorbance ratios at 1437/899, 1378/2900, and 2208/1330, respectively.

Brunauer Emmett Teller Analysis

Brunauer Emmett Teller (BET) surface area, total pore volume, and average pore diameter of PES, OAAP-PES, and AAPB-PES were analyzed using the benchtop ChemBET PULSAR TPR/TPD analyser (Quantachrome Instruments, USA) with adsorption-desorption of nitrogen. The sample analysis was conducted using liquid nitrogen at temperature of 77.35 °K. The specific surface area was calculated from the adsorption isotherms changes in pressure using pressure transducers by the BET method, while the volume of the pore was estimated from the volume of nitrogen held at the highest relative pressure (P/Po).

Scanning Electron Microscopy

Morphological changes analysis of PES, OAAP-PES, and AAPB-PES were performed using scanning electron microscopy (SEM) (EVO MA10, Carl Zeiss, UK). All samples were freeze-dried prior to scanning analysis. The samples were positioned on aluminium stubs and sputter-coated with gold using sputter coater system (Model Q150 RS, Quorum Technologies, UK). The result was observed at a magnification range of 250 to 5000X. Elemental analysis was performed using an energy-dispersive X-ray (EDX). The X-ray map was recorded based on the penetration of electron beam into sample.

Enzymatic Hydrolysis

Sugar conversion for both untreated and pretreated EFB were assessed through enzymatic hydrolysis process using the NREL standard protocol (LAP-009). The conditions of enzymatic hydrolysis used were at 50 °C in an incubated shaker (Infros HT Ecotron, Switzerland) at 150 rpm for 48 h. The enzymatic hydrolysis condition was optimized by the method of Farahin Abdul Rahman et al. (2018). All samples were loaded with a low glucan mass equivalent to 1% (w/v) glucan loading suspended in 20 mL scintillation vials with 50 mM citrate buffer (pH 4.8), loaded with enzymes (Novozymes, Denmark) of Cellic CTec2 cellulase in 15 FPU/g of glucan and combined with Cellic HTec2 hemicellulase at a ratio of 1:1 (v/v). The enzyme activity for Cellic CTec2 was 142 ± 2.97 FPU/mL (279.32 ± 1.10 mg/mL protein concentration). Enzymatic hydrolysis process was ended by immersing the hydrolysis vial in cold ice for 60 min. The process causes enzyme activity to cease quickly, making it possible to continue separating the hydrolysate and unhydrolyzed EFB by centrifuging at 7000 rpm for 10 min. The hydrolysate samples were filtered through 0.22 μm Whatman membrane syringe filter, and the concentration of the sugar were determined by HPLC using the same conditions as describe above. Sugar conversion of glucan and xylan were represented in percentages using the following equations:

RESULTS AND DISCUSSION

Moisture Content of EFB Samples

The EFB produced in the mill was wet and warm (40 to 45 °C), with bunches and shredded strands. The warm touch of EFB is attributable to the sample being taken after sterilization and stripping of fresh fruit bunches in the oil palm mill. The initial moisture content of the UE, PE, and PES showed a range of moisture content from 59% to 65%. The UE had the highest moisture content, which was due to the influence of oil content in the fibre compared to the PE and PES. Once recorded, the EFB was securely sealed and transferred to the Universiti Kebangsaan Malaysia (UKM) laboratory. No significant amount of water was lost from EFB during the transfer process. Although the moisture attempted to evaporate, the EFB was kept inside the sealed bag, thereby trapping the moisture, such that the water would be re-absorbed by the EFB. The initial moisture content recorded were 64.6%, 62.6%, and 59.4% for UE, PE, and PES. Approximately 12 h after samples collection, the UE, PE, and PES batches showed 62.8%, 61.6%, and 58.3% moisture content, respectively.

The moisture content decreased with water removal from the EFB via evaporation as the time exposure increased. After 24 h, the moisture content for both UE and PE was approximately 35% to 40%. However, after 48 h, it decreased to 26.4% and 20.3% respectively (Fig. 3).

Meanwhile, PES showed a significant drop in moisture content to 12.1% at 48 h. The significant difference was merely due to the loose structure of shredded EFB, making it have more surface area to dry off. For the next 60 and 72 h, the PES showed stagnant moisture content at 9.4% and very minimal reduction until the 96th h. In contrast, the UE and PE took longer than 84 h to reach below 10% and maintain its moisture content. The moisture content inside EFB had reached equilibrium, where the EFB neither gained nor lost any more moisture due to the surrounding environment factors: humidity and temperature. The remaining moisture inside the EFB can be held either as a vapor incorporated in the cell wall cavities or as a compound bound within the cell wall structure. Figure 3 shows the decreasing trend of moisture content for EFB over 96 h.

Moisture content is the most cardinal parameter before particle size reduction. The EFB with high moisture content makes the structure soft and resistant to grinding. In addition, this study observed that EFB with less exposure during drying was contaminated easily with rot fungi within 24 h of the process was worse for EFB in whole bunch size because the stranded fiber held closely together and trapped the moisture favoring fungal growth (Kamcharoen et al. 2014; Hau et al. 2020). Therefore, it is impossible to handle EFB in large amounts and have it prepared in the laboratory unless using industrial equipment for drying.

Fig. 3. The profile of moisture contents of different EFB at UKM processing laboratory before pretreatment

Morphology Appearance of Untreated and Pretreated EFB

Pretreated EFB showed a very distinct physical appearance compared to untreated EFB. The shape remained the same, but the color changed slightly to dark brown, and a powerful ammonia residue odor was detected. Figure 3 shows the physical structure of untreated and pretreated EFB under different conditions. The dark brown color was caused by the lignin condensation and re-localization process that increased the residual lignin to the surface of the solid particle (Abdul et al. 2016; Latif et al. 2018). The significant presence of residual lignin contributes to the browning effect. The study reported that lignin modification (demethylation) occurs at high temperature and pressure by melting, coagulation, and reprecipitation of cellulose fibers (Syed Abdullah et al. 2009). Upon completion of pretreatment, the pretreated EFB was put to rest and dried under a fume hood to release all the ammonia residue. The moisture content of pretreated EFB was found in the range of 45% to 65%, indicating the loosening of the cell wall structure and showing that the pretreatment works on EFB. The AAP-PES and AAP-PESG (also the control group) showed an even distribution of darker tone, while AAP-UE and AAP-PE showed uneven darker tone distribution. The ammonia may not penetrate and react with the entire fiber strand due to the bunch structure. Figures 4(b) of AAP-PE shows some lighter color of the internal structure, which indicates the pretreatment may not completely cover the entire bunch. If the EFB was found dry in the pressure vessel, the ammonia failed to penetrate the cellulose structure, thus giving off the ammonia effect. The moisture inside the EFB acts as a competing catalyst for breaking the acetyl bond and solubilizing the ammonia (Bals et al. 2010; Zhao et al. 2020). As a result, the AAP had made the physical characteristics of EFB to be more pliable, which increased the digestibility during enzyme hydrolysis.

Fig. 4. Physical structures of: (a) untreated EFB, and (b) pretreated EFB

UE: Unpressed EFB; PE: Pressed EFB; PES: Pressed and shredded EFB; PESG: Pressed shredded and ground EFB; AAP: Anhydrous Ammonia Pretreatment

Laboratory Scale Experiment

Structural constituent analyzes of morphologically different EFB samples

The analysis of carbohydrate contents in EFB is the most important value for the sugar conversion. In general, untreated EFB and all pretreated EFB samples did not show significant changes in the structural carbohydrate contents of glucan, xylan, and arabinan (Elgharbawy et al. 2018). This finding of unchanged sugar composition was supported by previous studies, which validated the promising use of AAP (Teymouri et al. 2005; Lau et al. 2010; Harun et al. 2015; Abdul et al. 2016). The main objective of performing this pretreatment was to preserve all the structural components and utilize the EFB as much as possible. Figure 5 shows that the native EFB is composed of 36.8% glucan, 18.7% of xylan, and 2.4% of arabinan; an almost similar composition was found in AAP-UE, AAP-PE, AAP-PES, and AAP-PESG for their respective components. This finding shows that the mechanism of ammonia pretreatment would result in swelling and delignification effect towards cellulose, thus preventing the lignin from absorbing cellulases and beneficial during enzymatic hydrolysis. Typically, the lignin fragments are linked to the carbohydrate fragment in the lignin-carbohydrate complex via the ether and ester linkages. Through this AAP, the glucan composition remained constant due to the dry nature condition of the pretreatment that prevented the loss of holocellulose during pretreatment (Chundawat et al. 2010; Yoo et al. 2011).

Fig. 5. Main structural compositions of native and morphologically different pretreated EFBs

U UE: Unpressed EFB; PE: Pressed EFB; PES: Pressed and shredded EFB; PESG: Pressed shredded and ground EFB; AAP: Anhydrous Ammonia Pretreatment

Ammonia had cleaved the lignocellulosic chemical linkages, particularly the lignin-carbohydrate complex (LCC). These can be seen through the disappearance of carboxylic linkage and the carbonyl linkage in Fourier transform infrared spectroscopy (FTIR) after AAP (Sun and Cheng 2002). This study found that the content of lignin was significantly reduced after pretreatment. Native EFB recorded 27.9% of lignin. Lignin contents of 17.8%, 17.3%, 17.5%, and 16.3% were obtained for AAP-UE, AAP-PE, AAP-PES, and AAP-PESG, respectively, which is approximately 40% of lignin loss. Pretreatment had caused the detachment of lignin and re-deposited on the surface, which dissolved into the extractive component of pretreated EFB (Harun et al. 2015; Latif et al. 2018). The re-deposited lignin can be found in acid-soluble lignin that is significantly shown in the increment of total extractives. AAP-PESG showed the highest lignin loss compared with other pretreated EFBs, mainly due to more surface area being exposed. However, the lignin loss from long fibres, such as AAP-PES, only showed 1% less than AAP-PESG. Thus, this insignificant value does not reflect the need for grinding the EFB into a smaller size.

Liquid extractive analysis of morphologically different EFBs

The liquid extractives contain extraneous components from the two-stage extraction process apart from the insoluble cell wall based on the soluble components namely, non-structural sugars, debris, and protein. Meanwhile, the ethanol extraction would remove waxes, fat acids, phenolic substances, and chlorophyll (Tan et al. 2016). Extractives must be removed to avoid cross reaction, with acid hydrolysis that could be a part of the components in acid insoluble. In Fig. 6, all conditions of pretreated EFB showed an increase in total extractive component compared with native EFB, which only showed 13.4% of total extractives.

Fig. 6. Extractive compositions of native and morphologically different pretreated EFBs

UE: Unpressed EFB; PE: Pressed EFB; PES: Pressed and shredded EFB; PESG: Pressed shredded and ground EFB; AAP: Anhydrous Ammonia Pretreatment

Generally, more than 90% of extractive components were extracted after being pretreated. The increase of extractive components demonstrates that the AAP process had successfully broken down the external structure of the cell wall by dignifying and exposing the cellulose and hemicellulose structures. This finding was similar in any AAP using various types of biomasses because of the dry-to-dry process nature (Harun et al. 2013; Shao et al. 2013: Abdul et al. 2016; Mokomele et al. 2018). Furthermore, AAP-UE showed a distinct increase in ethanol extractive because of the oil content present in the EFB extracted. AAP-PESG/control was the only EFB that was pretreated at a standard laboratory size of 2 mm, yielding the highest total extractives, similar to the findings reported in other studies (Lau et al. 2010; Abdul et al. 2016). Moreover, AAP-PE and AAP-PES were almost identical in their yield of the respective extractive component due to a similar group of pressed EFB, except that AAP-PES had the loose free structure. It is expected to have a different composition, but due to chemical composition protocols, the results may represent a small fraction of the pretreated EFB. Nevertheless, the real difference of AAP-PESG was only the other extractive, which was due to higher lignin removal that can be extracted on smaller particle size fibers.

Enzymatic hydrolysis of different morphological EFB

The enzymatic hydrolysis process in this study was run at 1% glucan loading because mixing and mass transfer at lower glucan loading had not significantly affected the hydrolysis performance. Mass transfer efficiency is due to the water content that acts as a medium for enzyme reactions. In a 1% hydrolysis run, the water content is much higher than the substrate content; thus, it gave an added advantage in the process of diluted hydrolysis. Figure 7 shows a sugar conversion profile from zero to 72 h of glucan and xylan for all different EFB morphology where the pretreated EFB was significantly different from the untreated EFB. On average, the conversion has increased more than double after being pretreated with anhydrous ammonia except for AAP-UE, and this finding was found similar with other work (Teymouri et al. 2005; Abdul et al. 2016), both using AFEX pretreatment on EFB. The samples PE, PES, and PESG showed 33.0%, 33.0%, and 33.8% of glucan conversion, respectively, while only 23.5% from UE after 72 h. Findings also showed 17.8%, 20.8%, 22.0%, and 22.5% xylan conversion for UE, PE, PES, and PESG. Both glucan and xylan conversions from UE showed much lower conversion than others due to the oil content inside the fiber that prevents water absorption. The increase in sugar conversion shows that the pretreatment method did have an effect in facilitating the digestibility of carbohydrate content. As predicted, AAP-PESG converted the highest amount of glucan and xylan at 66.5% and 44.7%, followed by AAP-PES (64.4% of glucan, 41.9% of xylan), AAP-PE (63.5% of glucan, 39.2% of xylan), and lowest conversion by AAP-UE (29.5% of glucan, 18.2% of xylan). However, the enzymatic hydrolysis finding in this study was not comparable to the previous studies (Abdul et al. 2016; Farahin Abdul Rahman et al. 2018), which reported a higher glucan conversion at 90% after 72 h of hydrolysis time. The pretreatment in this study may require some process improvement and adjustment (optimization) according to the current resources (EFB) to increase the extents of glucan and xylan conversion. All ammonia-based pretreatment showed a promising result on retaining the sugar inside the pretreated biomass, which eventually increases the total sugar yield during enzymatic hydrolysis (Bals et al. 2010; Farahin Abdul Rahman et al. 2018).

Theoretically, the size of the EFB will affect the performance of enzymatic hydrolysis. The smaller the size of the EFB, the larger the surface area of the EFB will be exposed to the enzyme digestibility, which will increase the extent of sugar conversion (Wang et al. 2022).

Fig. 7. Enzymatic hydrolysis profile on morphologically different EFBs of native and pretreated

UE: Unpressed EFB; PE: Pressed EFB; PES: Pressed and shredded EFB; PESG: Pressed shredded and ground EFB; AAP: Anhydrous Ammonia Pretreatment

This study showed a similar trend as the theory; however, AAP-PE and AAP-PES showed insignificant differences compared to AAP-PESG, which relatively was ground to a significantly smaller size (2 mm). Thus, the size reduction showed no increased pretreatment effectiveness when measuring sugar conversion. The sugar conversion from AAP-PE and AAP-PES showed a comparable value to AAP-PESG. These findings could be a strong justification of larger biomass size producing high sugar conversion but lower energy consumption during biomass size reduction process, which may increase the better biorefinery process and economic value of EFB. Some researchers reported that EFB should be soaked in detergent overnight to prevent fungal contamination and then washed the next day to remove oil and dirt before pretreatment (Shamsudin et al. 2012). The concept is ideal for a small scale, but the industry would not favor practicing additional steps. This process led to lower sugar recovery yield at the end of enzymatic hydrolysis due to some sugar solubilization during the soaking and washing process.

Fourier transform infrared spectroscopy of morphologically different EFBs

The purpose of FTIR analysis in this context was to show LOI and TCI indices that represent the crystallinity ratio, and the HBI was carried out to show the hydrogen bonding between certain hydroxyl groups in cellulose. The LOI was obtained from the peak-height 1437/899 absorbance ratio, where 1437 cm-1 is associated with crystalline cellulose I and 899 cm-1 is associated with amorphous cellulose. Likewise, the TCI ratio given by 1378 cm-1 is attributed to CH deformation and 2900 cm-1 associated with CH and CH2 stretching. Thus, TCI ratio was proportional to the crystallinity degree of the fiber cellulose. Finally, the HBI ratio was determined using wavelengths 3308 and 1330 cm-1 (Rafidison et al. 2020). Table 1 shows LOI, TCI, and HBI of untreated, pretreated, and unhydrolyzed (UH) pretreated EFB.

Table 1. LOI, TCI, and HBI Indices Ratio from FTIR Spectra

UE: Unpressed EFB; PE: Pressed EFB; PES: Pressed and shredded EFB; PESG: Pressed shredded and ground EFB; AAP: Anhydrous Ammonia Pretreatment

The values for all samples increased after AAP, except that the UE samples showed no changes. The increased value of LOI, TCI, and HBI indicated that pretreatment increased the structure of the crystal in the pretreated EFB (Lee et al. 2014); however, AAP-PESG showed a slightly higher ratio value compared to UE, PE, and PES. UH-AAP-PE, UH-AAP-PES, and UH-AAP-PESG also showed an increase in LOI and TCI due to more crystalline in the structure that failed to be hydrolyzed by the enzyme. The UH samples contained the highest crystallinity degree and highly ordered cellulose structure because the AAP action removed the amorphous portion of the lignin and hemicellulose (Oh et al. 2005; Rafidison et al. 2018). The conditions of PESG may have influenced the pretreatment effect, but it still did not significantly affect the pretreatment process.

Determination of optimal AAP condition – laboratory scale

The biomass used in this experiment was PES, as it is the best and the most economical morphological condition to practice. The aim was to determine the best optimal condition of AAP on PES. Figure 8 shows the effects of residence time, the ratio of moisture content, temperature, and ammonia loading on the sugar concentration and sugar conversion during 1% enzymatic hydrolysis. Experimental data in Fig. 8a show different residence times from 15 to 75 min of pretreatment time. The 30-min residence time achieved 70% glucan conversion and showed no significant increase at 45, 60, and 75 min, while 15-min residence time only gave 53.8% glucan conversion. Thus, 30 min was selected for the subsequent experiment as the shortest pretreatment time given at the best conversion.

Moisture plays an essential role in AAP, as it binds the biomass and ammonium ion and forms hydrogen bonds with cellulose to cause swelling in the crystalline cellulose structure. Figure 8b shows different moisture ratios to the dry weight of PES. The previous ratio by Lau et al. (2010) was a 1:1 ratio, and this study found that the 2:2 ratio resulted in a better glucan conversion of 88.3% compared to 71.4% at a 1:1 ratio. The higher conversion may be due to more moisture inside PES to hold and react with ammonia. However, as the moisture ratio increased, it showed no significant effect on conversion as the biomass may have reached maximum ability to absorb more moisture.

The different temperature effects of AAP are shown in Fig. 8c. Glucan and xylan conversions increased with increasing temperature from 95 to 155 °C, at which they maximized at about 89% and 67%. The increase in temperature, causing an increase in chemical reactions, cleaving the internal bonds within the PES. The sugar conversion showed a decreasing pattern when heating more than 175 °C, and the appearance of AAP-PES at 215 °C showed a slight burn on some of the fiber. A higher temperature than 175 °C can cause depolymerization and hemicellulose solubilization, which could account for the decline in conversion (Zhao et al. 2014, 2020). At extreme temperatures during alkaline pretreatment, higher temperature may convert the xylose to furfural or other degradation products. Pretreatment at 135 °C in this study has shown optimal conversion; further temperature increases did not benefit the process significantly and will cause unnecessarily higher energy consumption.

Figure 8d shows the effect of different ammonia loadings on sugar conversion of AAP-PES. The glucan and xylan conversion increased significantly with ammonia loading from 0.5 to 1.0 and remained about 89% of glucan conversion when ammonia loading was at 1.5 to 2.5 ratio to dry biomass. The AAP-PES showed a great improvement in sugar concentration and conversion compared to PES (24% glucan and 18% xylan conversion). Even though ammonia can cleave the alkaline chemical bonds between lignin and hemicellulose, the higher ammonia loading can disrupt the lignin, causing large pores and deposition hemicellulose and lignin to cell corners and surface of the cell wall (Chundawat et al. 2010). In addition to this, anhydrous ammonia alone can also interact directly with cellulose for phase changes in the crystal structure from cellulose I to cellulose III (Mosier et al. 2005). Moreover, the higher the ammonia loading, the higher the operating pressure involved with the hazardous chemical usage. Therefore, 1.0 ammonia loading was chosen for the most optimal condition on the operation AAP system.