*Article* **Application of a Liquid Biphasic Flotation (LBF) System for Protein Extraction from** *Persiscaria Tenulla* **Leaf**

**Hui Shi Saw 1, Revathy Sankaran 2,3, Kuan Shiong Khoo 4, Kit Wayne Chew 5,\*, Win Nee Phong 6, Malcolm S.Y. Tang 2, Siew Shee Lim 4,\*, Hayyiratul Fatimah Mohd Zaid 7, Mu. Naushad <sup>8</sup> and Pau Loke Show 4,\***


Received: 10 January 2020; Accepted: 17 February 2020; Published: 21 February 2020

**Abstract:** *Persiscaria tenulla*, commonly known as *Polygonum*, is a plant belonging to the family Polygonaceae, which originated from and is widely found in Southeast Asia countries, such as Indonesia, Malaysia, Thailand, and Vietnam. The leaf of the plant is believed to have active ingredients that are responsible for therapeutic effects. In order to take full advantage of a natural medicinal plant for the application in the pharmaceutical and food industries, extraction and separation techniques are essential. In this study, an emerging and rapid extraction approach known as liquid biphasic flotation (LBF) is proposed for the extraction of protein from *Persiscaria tenulla* leaves. The scope of this study is to establish an efficient, environmentally friendly, and cost-effective technology for the extraction of protein from therapeutic leaves. Based on the ideal conditions of the small LBF system, a 98.36% protein recovery yield and a 79.12% separation efficiency were achieved. The upscaling study of this system exhibited the reliability of this technology for large-scale applications with a protein recovery yield of 99.44% and a separation efficiency of 93.28%. This technology demonstrated a simple approach with an effective protein recovery yield and separation that can be applied for the extraction of bioactive compounds from various medicinal-value plants.

**Keywords:** extraction; leaf; liquid biphasic flotation; polygonum; protein

#### **1. Introduction**

The usage of traditional herbs for preventive health care is widespread, and plants are the source of numerous natural antioxidants that could be utilized for the development of novel medicines. Natural antioxidants and bioactive compounds that originate from traditional herbal medicines have received escalating attention for their ability in treating specific human diseases. For example, traditional herbal medicine has been used widely in treating cancer patients [1] and to treat neurodegenerative disorders [2]. Plants comprising high medicinal value are currently screened for a variety of pharmacological properties.

*Persiscaria tenulla* (formerly known as *Polygonum* (*P. minus*)) or frequently recognized as "kesum" in Malaysia, have been used as a flavoring ingredient and food additive in Malaysia. Polygonum plant has been reported to contain a wide range of pharmacological properties and many studies have been conducted to evaluate the phytochemical and pharmacological aspects of the plant, which include anti-inflammatory activity [3], antiproliferative effects [4], anti-microbial activity [5], cytotoxic activity [6], gastric cytoprotective activity [7], and antiviral activity [8]. It has been proven that *P. minus* comprises many high-value components that include protein, flavonoids, and antioxidant vitamins, such as carotenes, retinol equivalents, and vitamin C. However, there are limited studies concerning the effective extraction techniques of the biomolecules from the plant extract.

The major challenge in the extraction of the high-value components from herbal plants is the downstream processing. Up to date, there is a lack of effective and efficient techniques for high yield and cost-effective biomolecule extraction. In this study, a novel method known as liquid biphasic flotation (LBF) system is introduced to extract protein from *P. tenulla* leaves. The LBF process comprises the incorporation of two processes, which are an aqueous two-phase system (ATPS) and a solvent sublation (SS). The conventional flotation system that is commonly known as SS was first introduced by Sebba [3,9]. The SS process is a type of non-foaming adsorptive bubble separation technique in which the surface-active or hydrophobic compounds in aqueous phase are adsorbed on the bubble surfaces of an ascending gas stream and then collected in an immiscible apolar organic solution layer placed on top of the aqueous phase. The mass transfer of SS involves the air bubbles that are produced from the sublation column. The air bubbles drag a sheath of water into the top organic solvent, which eventually drains as water droplets, depleted of solute, and descend back into the bottom aqueous phase via the gravitational force [5,10]. As for the LBF process, the mass transfer comprises the integration of ATPS and SS, which are the utilization of aqueous two-phase systems as a liquid medium for facilitating the mass transfer of biomolecules and the involvement of mass transport from SS. The LBF system is known to be a newly developing separation process that has many benefits over conventional processes. LBF has several advantages, such as a high separation efficiency, high yield, simple separation, and is an economical technique [11]. Recently, the LBF process has gained much attention and many studies have been conducted using this technique for various biomolecules extraction, which include lipase from bacteria, penicillin G, puerarin, a-lactalbumin, lincomycin, C-phycocyanin, polyphenols, betacyanins, and protein from microalgae [12].

In this current study, the LBF method is utilized to extract protein from *P. tenulla* and the optimization of the technique is performed to obtain optimal operating conditions for protein recovery. Parameters that were evaluated in this research were the effect of types of alcohol, types of salt, the concentration of alcohol and salt, amount of kesum biomass, pH, flotation time, and scale-up capability. This study aimed to assess the feasibility of LBF in protein extraction from kesum leaf and to demonstrate that the LBF system is an effective method that has a high possibility to be employed in large scale productions.

#### **2. Materials and Methods**

#### *2.1. Materials*

Ammonium sulphate, dipotassium hydrogen phosphate, magnesium sulphate, disodium hydrogen phosphate, sodium hydroxide, hydrochloric acid, and Bradford reagent were purchased from R&M Chemicals, (Selangor, Malaysia). Food grade 99.8% ethanol and 2-propanol were acquired from R&M Chemicals,(Selangor, Malaysia). Fresh Kesum leaves were obtained from Aeon supermarket (Selangor, Malaysia, distributed by Edsam Trading Sdn. Bhd.).

#### *2.2. Equipment*

The LBF equipment was made up of a glass column, where the small LBF column had an internal diameter of 2.4 cm, height of 15 cm, and could accommodate a solution with a maximum volume of 50 mL. For the large-scale LBF system, the internal diameter was 7 cm, 15 cm in height, and the capacity of the system was 500 mL. The base of the column was built using a G4 sintered disk with a pore size of 10 μm manufactured by DONEWELL RESOURCES SDN. BHD, (Selangor Malaysia) and the base was connected to an air pump for air bubble generation. A Dwyer flowmeter (model RMA-26-SSV) with the range of 25–250 cc/min was used to measure the flow rate of air supplied to the column. Figure 1 displays the experimental framework of the LBF method that was used in this study.

**Figure 1.** Schematic diagram illustrating the apparatus set-up of liquid biphasic flotation (LBF) system for protein extraction. 1: Air pump; 2: flowmeter; 3: sintered disk; 4: LBF column; 5: top alcohol phase; 6: bottom salt phase.

#### *2.3. Methodology*

#### 2.3.1. Preparation of Kesum Leaf Powder

Fresh *Persiscaris tenulla* obtained from a supermarket were used in this study. The leaves with petiole attached were removed from the stem and were cleaned. The leaves were cut into smaller pieces and were desiccated in a silica gel box overnight. The dried leaves were then ground with a mortar and pestle into powder form.

#### 2.3.2. LBF Extraction

This study was performed in batches and was repeated thrice. A one variable at a time (OVAT) method was used in this study to assess the effect of different parameters on the protein recovery. The initial condition of 250 g/L of ammonium sulphate was dissolved in 15 mL of distilled water. Fifteen milliliters of salt solution, which served as the bottom phase, was pipetted into the flotation system. Fifteen milliliters of 100% ethanol was added to 300 mg of ground leaves and vortexed. The top phase was then poured into the LBF tube gently from the edge. The system was capped with a lid and then immediately timed using a stopwatch. The amount of bubbles that formed was maintained by adjusting the pressure of the flotation system. Adjustment of the pH was made via the addition of hydrochloric acid (1 M) or sodium hydroxide (1 M). After the system had settled for 10 min, the top phase was pipetted into a tube and the remaining bottom phase was poured into another tube. The volumes of the top and bottom phases were measured.

#### 2.3.3. Protein Assay

The protein concentration was obtained by applying the Bradford method. An extracted protein sample with a volume of 0.25 mL was mixed with 2.5 mL of Bradford reagent in a cuvette. After 10 min of reaction time, the absorbance was measured using a UV-Vis spectrophotometer at the wavelength of 595 nm. The absorbance reading obtained was converted to a protein concentration by using a standard calibration curve that was established using a standard protein, namely BSA. The results were expressed as a mean of triplicate readings.

#### 2.3.4. Calculation of the Separation Efficiency (E) and Recovery Yield (R)

The separation efficiency (E) describes the concentration of protein being extracted in the alcohol phase (top phase). The efficiency is obtained by calculating the concentration of protein present in the bottom phase before and after the flotation process and it was evaluated by employing Equation (1):

$$\mathbf{E} = \left( 1 - \frac{\mathbf{C}\_B}{\mathbf{C}\_{Bi}} \right) \times 100\% \,\mathrm{s} \tag{1}$$

where *c*<sup>B</sup> represents protein concentration in bottom phase after flotation, while *cBi* signifies protein concentration in bottom phase before flotation. The E value determines the concentration of protein being successfully recovered in the alcohol-rich top phase.

The total recovery yield (R) of protein was assessed by applying Equation (2). The CT describes the protein concentration that is recovered in the top phase, while VT defines the volume of the top phase. Based on the protein concentration obtained from the top phase, the amount is compared with the theoretical protein content in mg to obtain the recovery yield. The amount of protein present in the leaf is based on theoretical value obtained from Revathy Sankaran et al. [11].

$$\text{R (\%)} = (\text{C}\_{\text{T}} \times \text{V}\_{\text{T}}) \text{(Amount of protein content based on theory)} \times 100\% \tag{2}$$

#### **3. Results and Discussion**

#### *3.1. E*ff*ect of Alcohol Types on the Protein Recovery and Separation E*ffi*ciency*

In this study, water-miscible pure alcohols (100%), namely ethanol and 2-propanol, demonstrated the ability to form LBF with 250 g/L of ammonium sulphate (NH4)2SO4. Based on the results, LBF formed using (NH4)2SO4/ethanol showed a 74.93% separation efficiency and a 32.35% protein recovery. In contrast, the separation efficiency and recovery yield achieved using LBF containing (NH4)2SO4/2-propanol were 57.95% and 28.2%, respectively. The results clearly show that LBF formed using (NH4)2SO4/ethanol was more efficient in recovering protein from kesum leaves than LBF containing (NH4)2SO4/2-propanol. Ethanol exhibited a better performance possibly due to its property of high solute solubility that can assist in the desorption of the solute from the substrate [12]. Additionally, by comparing both solvents, ethanol is more environmentally friendly compared to propanol [12]. From the industrial point of view, ethanol is a better selection for large-scale production as it can be easily reused [13]. The findings suggest that the efficiency of LBF in protein separation

is dependent on the type of alcohol used in the system. In this case, LBF formed using 250 g/L of (NH4)2SO4 and 100% of ethanol was chosen for the subsequent studies.

#### *3.2. E*ff*ect of Types of Salt on Protein Recovery and Separation E*ffi*ciency*

Selecting a proper phase-separation salt with a high salting-out ability is a key step in developing an efficient LBF system for maximum protein recovery from kesum leaves. The type of salt selected is considered to be an important variable to take into account when designing an LBF for protein separation owing to their strong effects on the salting-out effect and the partition coefficient of protein [14]. While keeping the other variables constant, such as alcohol type, alcohol concentration, salt concentration, flotation time, and the amount of starting material constant, the relative salting-out effectiveness of salt types was investigated in this study. The salts used were ammonium sulphate, di-potassium hydrogen phosphate, magnesium sulphate, and disodium hydrogen phosphate.

In this study, the protein separation efficiency was found to be varied with the types of salts used. There was no value for the separation efficiency in the LBF formed using magnesium sulphate/ethanol and disodium hydrogen phosphate/ethanol due to the precipitation that occurred in these two systems (Figure 2b). A similar conclusion was also reached by Phong et al. [14], who reported that the protein separation efficiency as a result of the salting-out effect is greatly influenced by the types and complexation of the cations and anions of salt in the LBF system.

**Figure 2.** Effect of different conditions on the protein recovery and separation efficiency: (**a**) Effect of the alcohol type, (**b**) Effect of the types of salt, (**c**) Effect of the ethanol concentration, and (**d**) Effect of the ammonium salt concentration.

Among all the salts, LBF formed using ammonium sulphate achieved the highest separation efficiency and recovery yield, with the values of 74.93% and 32.35%, respectively, as shown in Figure 2b. This trend supports previous findings in the literature. The relative effectiveness of salt types was found to follow the well-known Hofmeister series [14], in which ammonium sulphate forms two ions at the ends of their respective Hofmeister series [15]. Apart from this, ammonium sulphate has been reported as the most commonly used salt for salting out of proteins due to its effectiveness, high solubility,

cheapness, availability of pure material, lack of toxicity, and their ions possess the stabilizing effect on protein structure and bioactivity [15]. All these characteristics have made ammonium sulphate a popular choice for use in protein precipitation [16]. As such, LBF formed using 250 g/L of (NH4)2SO4 and 100% ethanol was selected for the next experiment.

#### *3.3. E*ff*ect of Concentration of Alcohol on Protein Recovery and Separation E*ffi*ciency*

While keeping the other variables constant, the relative effectiveness of salting-out at different concentrations of ethanol was investigated in this study. It is stated that a high extraction yield can be obtained by using ethanol-water compared to pure ethanol [17]. In this research, the influence of ethanol-water with several different concentrations and pure ethanol on the protein extraction was examined. The result obtained is similar to the study done by the Machado group in which they discovered greater extraction yields of blackberry residues attained by applying pressurized liquid extraction when ethanol-water was utilized compared to pure ethanol [18].

Based on Figure 2c, the results show that the protein separation efficiency increased from 80% to 90% of ethanol concentration and reached an optimum level of 82.33% at 90% ethanol concentration. However, the separation efficiency started to show a decreasing trend with ethanol concentration higher than 90%. In the case of the recovery yield, the highest yield obtained was 43.19% in LBF containing 80% ethanol, followed by a 40.2% recovery yield at 90% ethanol. The findings indicate that there was no correlation relationship between the two variables of separation efficiency and recovery yield in the same system. The addition of water to the organic solvent in this case ethanol possibly created a relatively polar medium that facilitated the extraction of protein [18]. In this study, LBF formed using 250 g/L of (NH4)2SO4 and 90% ethanol was identified as being the most proficient at protein separation and was thus chosen for further optimization.

#### *3.4. E*ff*ect of Salt Concentration on Protein Recovery and Separation E*ffi*ciency*

The concentration of salt in the liquid biphasic system is another important factor to consider because different salts interact differently with the protein, water, and other chemicals. The effects of the concentration of salt on protein recovery and separation are well documented. The presence of salt in the solution will impact the surface tension of water, which will then increase the hydrophobic interaction between the protein and water [19]. Following this change, the targeted protein will shift to or from the aqueous phase depending on the nature of the protein [20]. For this work, the ammonium sulphate salt concentration was varied from 150 g/L to 350 g/L with 50 g/L increments. The alcohol content was set at 90%, while the mass of leaves used was 300 mg. The flotation time for this experiment was set at 10 min.

One important observation for this experiment is that the volume of top and bottom phases changed as the concentration of salt increased. The volume of the two phases reached a plateau when the concentration of salt reached 300 g/L and above. Another important observation to note is that at a low salt concentration (150 g/L), the two phases did not form. This suggests that the lower boundary for salt concentration that allows for two-phase formation is higher than 150 g/L. This is because an increase in salt or alcohol concentration in an alcohol/salt system could increase the tie-line-length (TLL), which could facilitate phase separation [21].

A plot of the effect of salt concentration on the recovery yield and separation efficiency is shown in Figure 2d. From Figure 2d, it is seen that the recovery yield and separation efficiency decreased in tandem with the increase in salt concentration. At a concentration of 200 g/L, the highest recovery yield of 40.69% and separation efficiency of 87.81% was recorded. As the concentration of salt increased, however, the recovery yield and separation efficiency gradually decreased. As the concentration of salt increased, the solubility of protein decreased. This is commonly known as the salting-out effect. Since different proteins salt-out at different salt concentrations, the effect can help us to determine the upper boundary of salt concentration for the LBF system. A higher salt concentration results in a

higher salting-out effect. This could then lead to a higher protein partition coefficient Ke [21]. A high Ke would result in a low yield.

At concentrations of 300 g/L and 350 g/L, however, there was a slight increase in separation efficiency compared to a decrease in recovery yield. There was a sharp decrease in both separation efficiency and recovery yield at a concentration of 300 g/L. This could be the concentration at which the salting-out effect occurred. As the concentration of salt increased, it caused more water to enter the bottom phase. This way, the protein was then impelled to the top phase [21], which caused the slight increase in the recovery yield at a concentration of 350 g/L. Therefore, we can conclude that the optimum salt concentration for the extraction of protein from kesum leaf was between 150 g/L and 300 g/L. The concentration of 200 g/L was then used in the experiments with other parameters in this work.

#### *3.5. E*ff*ect of the Kesum Leaf Biomass Amount on the Protein Recovery and Separation E*ffi*ciency*

The influence of kesum leaf biomass, or the mass of protein source, is another important factor to consider. Generally, increasing the concentration of protein sources can have a profound effect on the performance of phase separation due to the specific partition behavior of the target protein [22]. For this work, the concentration of salt was set to 200 g/L, ethanol concentration was set to 90%, and the flotation time was set to 10 min. The mass of kesum leaves was varied between 100 mg and 400 mg with 100 mg increments.

From Figure 3, it is seen that as the mass of leaves increased, the yield decreased. There was a significant drop in yield (over 40%) when the mass of leaves increased from 100 mg to 200 mg, and the drop continued gradually as the amount of leaves continued to increase. For the separation efficiency, the highest efficiency occurred at 300 mg, which then dropped to its lowest point at 400 mg. It was interesting to see that the yield decreased as the mass of leaves increased. In general, increasing the biomass concentration of the LBF also increases the number of contaminants and impurities in the system, thereby reducing the performance of the LBF separation [23]. In addition, a higher biomass content also increases the precipitation at the interface of the two phases, which could adversely affect the yield [23].

**Figure 3.** Effect of the kesum biomass amount on the protein recovery yield and separation efficiency.

Based on the definition highlighted in the materials section, separation efficiency is highly dependent on the protein activity in the bottom phase after the LBF process. It determines the concentration of protein extracted and this represents the efficiency of the system in extracting the

protein. As for the recovery yield, it represents the total amount of protein being recovered in the top phase. At 300 mg, the high separation efficiency was obtained with 87.81%; however, the protein recovery yield was low with only 40.69%. The low recovery yield was possibly due to high impurities present in the top phase. Several possibilities contributed to the low recovery yield: (1) protein retrieved in the upper phase was low (CT), or (2) the low phase volume of the top phase (VT) at 300 mg. These could be caused by the decreasing LBF performance as the level of impurities increased due to the increase in kesum leaf mass. Due to the high recovery yield value (89.58%) and separation efficiency of 80.68%, a 100 mg kesum leaves mass was used for the next step of this experiment.

#### *3.6. Influence of pH on the Recovery Yield and Separation E*ffi*ciency*

The pH value of an LBF system affects the separation outcome by altering the surface properties of the target protein, including the surface net charge, molecular shape, surface hydrophobicity, and the presence of specific binding sites [24]. A simple example is the case of a biomolecule with both polar and non-polar groups that experiences changes in its net charge and surface properties with changing pH values [25]. For this work, the pH of the system was varied from pH 4 to pH 8.

From Figure 4, it is seen that despite the high absorbance at pH 5, the highest recovery yield and separation efficiency occurred at pH 6. The lowest yield and E, on the other hand, occurred at pH 7. This suggests that for the extraction of biomolecules from kesum leaves, the LBF system should be kept in an acidic condition. Both the recovery yield and separation efficiency did not fluctuate much as the pH of the system increased. An interesting observation is that there was no formation of two-phases at pH 8. This experiment shows that as the pH approached basic pH, the hydrophobicity of the system was impacted to the point where it induced a salt-out. In order to prevent a salt-out, the pH of the system should be kept below the neutral level. The optimized pH condition that gave the maximum separation efficiency and recovery yield was at pH 6 with 87.19% and 96.37%, respectively.

**Figure 4.** Effect of pH on the protein recovery yield and separation efficiency.

#### *3.7. The Influence of the Flotation Time on the Recovery Yield and Separation E*ffi*ciency*

The effect of flotation time is one of the most important factors that could affect the LBF recovery yield and separation efficiency. The flotation time affects the outcome of the process by influencing

the area of the air–water interface per unit volume of aqueous solution over time [26]. For this part, the concentration of salt was set to 200 g/L, the pH of the system was maintained at pH 5.0, the ethanol concentration was maintained at 90%, and the mass of kesum leaves consumed was 100 mg. The flotation time was varied from 5 min to 15 min with 2.5 min increments. The results of the experiment are provided in Figure 5.

**Figure 5.** Effect of the the flotation time on the protein recovery yield and separation efficiency.

The resulting yield and separation efficiency, as shown in Figure 5, shows that the flotation time of 7.5 min gave the highest recovery yield with 98.36%, while the flotation time of 5 min gave the lowest recovery yield of 91.42%. For the separation efficiency, the highest occurred at a 10 min flotation time with 87.19% but at 7.5 min, 79.12% was recorded. One possible explanation for this phenomenon is that longer flotation time allowed for the accumulation and build-up of biomolecules in the LBF phases with the movement of gas bubbles [25]. As the flotation time increased, the concentration of targeted biomolecules in the top and bottom phases changed based on the kinetic processes. This explains the general trend of the separation efficiency, which showed a positive gradient, i.e., increasing with increasing flotation time. However, as the flotation time continued to increase, the level of impurity carried upward by the gas bubbles also increased, which then affected the separation performance of the LBF system.

According to Iqbal et al., the flotation force is highly dependent on the flow properties of the phases [27]. As the flotation time increases, the concentration of targeted biomolecules in the top and bottom phases changes based on the kinetic processes [25]. Based on Figure 5, as the flotation time increased, the yield was reduced. From this study, although 10 min gave the highest separation efficiency, 7.5 min was selected as an optimized condition because the focus of this study was to obtain a high protein recovery yield. Additionally, a long flotation time requires high energy consumption, which is costly, non-environmentally friendly, and it is not suitable for large-scale processes.

#### *3.8. Large-Scale Protein Extraction Using the LBF System*

In this section, a scale-up study of the protein extraction using the optimized conditions were assessed. By using the operating conditions that were optimized previously, a large-scale study was conducted. This assessment was performed to corroborate the consistency and efficiency of the LBF technique on a large scale. In the large-scale study, it was discovered that the amounts of both top organic phase and bottom phases increased ten-fold compared with the small-scale experiment. A total of 300 mL of working volume with 150 mL of bottom phase and 150 mL of top phase were employed. Following the results achieved from Table 1, it can be seen that a comparatively higher recovery yield and separation efficiency of 99.44% and 93.28%, respectively, were obtained. Many reported studies have shown that the separation efficiencies of LBF for the recovery of different kind of biomolecules in the comparative study were between 85%–98.5%, which were much higher than SS, which achieved separation efficiencies of 48%–70%. LBF is preferable compared for this case as the separation efficiency achieved was more than 90%.

**Table 1.** Comparison between small- and large-scale LBF systems for the protein extraction of *P. tenulla* leaf.


Based on the results achieved, the large-scale version of the LBF system was validated for its reliability, which is beneficial for the extraction of other biomolecules on an industrial scale. Other studies that demonstrated that LBF can be an alternative technology that can be utilized in industries for the extraction of various medicinal components include ortho-phenylphenol [28], puerarin [29], antioxidant peptides from trypsin hydrolysates of whey protein [30], baicalin [31], lipase enzyme [32], C-phycocyanin [33], and betacyanins [34]. All these studies have proven that the separation efficiency and recovery of the biocomponents were enhanced with the utilization of LBF system as their extraction method.

#### **4. Conclusions**

The findings from this study revealed that a high protein recovery and separation efficiency can be obtained using this LBF approach. Based on the experiment conducted, the optimized conditions for highest protein recovery and separation efficiency were 90% ethanol, 200g/L of ammonium sulphate, 100 mg of kesum leaf biomass, pH 6, and a flotation time of 7.5 min. The highest protein recovery yield achieved was 98.36% and the separation efficiency was 79.12% for the small-scale system. The study on the large-scale LBF system demonstrated the reliability and consistency of the system in which a recovery yield and separation efficiency of 99.44% and 93.28%, respectively, were attained. The application of the LBF system for protein extraction from herbal leaves involves a simple procedure, short processing time, and cost-effective process with a high recovery yield. This study shows the significance of LBF in downstream processing, especially for the extraction of value-added biomolecules. LBF can be an alternative technology that can be utilized in industries for the extraction of various medicinal leaf extracts. This LBF system is beneficial in the pharmaceutical and nutraceutical industries for the improvement of overall production and biotechnology fields.

**Author Contributions:** Conceptualization, K.W.C.; data curation, H.S.S. and K.S.K.; methodology, software, and resources: H.S.S., K.S.K., and M.N.; original draft writing: R.S., W.N.P., and M.S.Y.T.; review and editing: R.S., K.W.C., and P.L.S.; supervision: S.S.L., P.L.S., and R.S.; project administration: K.W.C., S.S.L., P.L.S., and H.F.M.Z.; funding acquisition: P.L.S. and H.F.M.Z. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Fundamental Research Grant Scheme, Malaysia (FRGS/1/2019/STG05/UNIM/02/2). The APC was funded by Yayasan Universiti Teknologi PETRONAS (YUTP 015LC0-047).

**Acknowledgments:** One of the authors (M. Naushad) is grateful to the Researchers Supporting Project number (RSP-2019/8), King Saud University, Riyadh, Saudi Arabia, for the support.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Review* **Potential of** *Jatropha curcas* **L. as Biodiesel Feedstock in Malaysia: A Concise Review**

**Nurul Husna Che Hamzah 1, Nozieana Khairuddin 1,\*, Bazlul Mobin Siddique <sup>2</sup> and Mohd Ali Hassan <sup>3</sup>**


Received: 24 April 2020; Accepted: 2 June 2020; Published: 6 July 2020

**Abstract:** Fluctuation in fossil fuel prices and the increasing awareness of environmental degradation have prompted the search for alternatives from renewable energy sources. Biodiesel is the most efficient alternative to fossil fuel substitution because it can be properly modified for current diesel engines. It is a vegetable oil-based fuel with similar properties to petroleum diesel. Generally, biodiesel is a non-toxic, biodegradable, and highly efficient alternative for fossil fuel substitution. In Malaysia, oil palm is considered as the most valuable commodity crop and gives a high economic return to the country. However, the ethical challenge of food or fuel makes palm oil not an ideal feedstock for biodiesel production. Therefore, attention is shifted to non-edible feedstock like *Jatropha curcas Linnaeus* (*Jatropha curcas* L.). It is an inedible oil-bearing crop that can be processed into biodiesel. It has a high-seed yield that could be continually produced for up to 50 years. Furthermore, its utilization will have zero impact on food sources since the oil is poisonous for human and animal consumption. However, Jatropha biodiesel is still in its preliminary phase compared to palm oil-based biodiesel in Malaysia due to a lack of research and development. Therefore, this paper emphasizes the potential of *Jatropha curcas* as an eco-friendly biodiesel feedstock to promote socio-economic development and meet significantly growing energy demands even though the challenges for its implementation as a national biodiesel program might be longer.

**Keywords:** non-edible; oil; biodiesel production; fuel

#### **1. Introduction**

The depletion of crude oil reserves coupled with the awareness of environmental issues and escalating petroleum prices have stimulated the search for alternatives to reduce overdependence on conventional fossil fuels [1,2]. Historically, researchers have substituted conventional fuels with renewable energy resources (e.g., biofuels) since the invention of diesel engines [3]. These technologies have since advanced to this day. Typically, conventional diesel and petroleum fuels release harmful gases into the atmosphere, thereby causing global warming and climate change. Furthermore, fossil fuels contribute to pollution through the emission of major greenhouse gases (GHG). In principle, a biofuel is cleaner than any fossil fuel, since it can reduce carbon dioxide (CO2) emissions by 78% and carbon monoxide (CO) emissions by 50% [4].

The global human population is predicted to increase by 34% by 2050 [5]. This increase in human population has become a contributing factor to the high demand for energy consumption. Moreover, the continuous exploitation and rapid depletion of the Earth's natural and mineral resources will significantly increase the energy demand for transportation, industrial, and other purposes. According to the International Energy Agency (IEA), it is estimated that global energy consumption will soar by 53% by the year 2030 [6]. Another study also reported that global petroleum resources will be completely depleted within 40 years [7]. This situation could result in annual oil price escalation in the near future. Therefore, renewable energy resources such as biofuels urgently need to be adopted as substitutes for conventional fuels.

Biodiesel derived from *Jatropha curcas*, which is suitably planted in tropical or subtropical countries such as Malaysia and Indonesia, could potentially reduce the use of conventional fuels. However, there are many challenges faced during the cultivation, harvesting, and processing of the crop yield. Hence, finding the root cause is important for resolving the issues and ensuring a higher quality of harvested Jatropha seeds. Consequently, Jatropha's enormous potential in the financial, agricultural, environmental, sustainable energy production, and industrial fronts could attract the attention of researchers and policymakers.

#### **2. Distribution and Physicochemical Properties of Biodiesel Feedstock**

Bioethanol, biodiesel, and biogas are the main biofuel components of various agricultural biomasses produced from the different biochemical routes [8]. The first generation feedstock was produced from edible oils such as corn, sugarcane, sugar-beet, and others. Unlike the first generation of biofuels, second-generation biofuels targeted non-food biomass and agricultural residues [9]. Biodiesel originated from the first and second-generation of biofuels, as it is processed from agricultural crops and residues. Typically, the sources of biodiesel differ in many regions or countries. European nations use rapeseed due to the surpluses from edible oil production. On the other hand, soybean is commonly utilized for biodiesel production in the United States, which is becoming the main biodiesel-producing country [10]. Singh, V. et al., had summarized all classification of biofuels production starting from the first until the fourth generation as shown in Figure 1.

**Figure 1.** Classification of biofuels (adapted from [11]).

On the contrary, the excess palm oil and coconut oil in Malaysia, Indonesia, and Thailand could be utilized for the synthesis of biodiesel. However, the food versus fuel competition could be overcome by exploring non-edible seed oils such as *Jatropha curcas* and *Karanja* (*Pongamia* second-generation) as raw materials for biodiesel production [12]. Other than vegetable-based biodiesel, waste cooking oil (WCO) is becoming more popular for biodiesel production in Malaysia due to the low cost and the high volume of waste generation in each household. Kabir et al. reported that the average WCO generated in Malaysia per household is 2.34 kg/month [13].

Biodiesel processed from animal fats and vegetable oils is defined as fatty acid alkyl esters or fatty acid methyl esters (FAME) [14]. It is typically synthesized from the reaction of triglycerides with alcohol and a catalyst in a process known as transesterification (Figure 2). The transesterification will produce FAME and simultaneously cause saponficiation or soap formation to occur. Common catalysts usually employed during the reaction are homogeneous alkaline catalysts—for instance, potassium hydroxide (KOH), sodium hydroxide (NaOH), potassium methoxide (CH3OK), and sodium methoxide (CH3ONa). However, it is necessary to control the amount of alkali catalyst, because excess alkali enhances the saponification reaction which reduces the yield of product [15]. Nevertheless, the saponification value is important as an indicator of the oil as normal triglycerides, making the oil useful in the soap and shampoo industries [16]. It is reported that the nature of the catalyst employed is crucial to converting triglycerides into biodiesel. For instance, using a homogeneous catalyst will produce glycerol or soap as a by-product, which could risk consuming or even deactivating the catalyst. As a result, the biodiesel purification process would be hampered by the loss of catalytic process [17]. As for heterogeneous catalysts, a higher quality of FAME can be generated after transesterification.

**Figure 2.** Transesterification of triglycerides with alcohol and catalyst [18].

The physicochemical properties of biodiesel depend on numerous factors—for example, the composition of the fatty acid in the raw feedstock, the chain length of the fatty acid, the saturation degree, and branching. Other factors include the production technique and operating conditions for the biodiesel synthesis [19]. The quality of biodiesel may also differ due to the impurities from unreacted feedstock glycerides, the fraction of non-fatty acids, or runaway reactions during the process of transesterification [20]. Typically, a longer fatty acid chain will enhance the synthesis of biodiesel products with a higher cetane number, which results in lower emissions of toxic NOx [21]. The composition of the fatty acid determines the level of saturation with higher compositions, resulting in a higher degree of saturation and viscosity [10].

### **3. Potential of** *Jatropha curcas* **as Biodiesel Feedstock**

Energy crops are specifically grown for fuel and energy production. Currently, these crops only contribute to a comparably small percentage of the total biomass energy produced each year. However, this percentage is expected to increase over the next few decades. Nevertheless, energy crops compete for land earmarked for food production, environmental protection, and forestry or nature conservation. Generally, the characteristics of ideal energy crops are high yield (maximum dry matter per hectare production), low cost, low energy input, low nutrient or fertilizer requirements, pest resistance, and the composition with the least contaminants generated [22].

In Malaysia, palm oil is considered the most valuable energy crop due to its abundance and productivity. The Malaysian Palm Oil Board (MPOB) reports that the total area of oil palm planted in 2019 was 5.9 million ha [23]. Recently, the Ministry of Primary Industries and the MPOB launched the B20 Biodiesel program (20% palm methyl esters blended with 80% petroleum diesel) to offset the palm vegetable oil demand and stabilize the market price of these products [24]. However, this crop is still perceived as a major source of vegetable oil all over the world rather than as a fuel, and the demand for it as a food ingredient is increasing. Therefore, the idea of utilizing inedible food crops such as *Jatropha curcas* could help overcome the major problems faced by the first generation of biodiesel feedstock.—for example, the food vs. fuel dilemma, the issues of scaling, and the inability to grow on peripheral areas of land, etc. Its average productive life span is also longer than that of oil palm

(50 years and 30 years, respectively). Besides, its oil content is reportedly between 63.16% and 66.4%, which is higher than that of soybean (18.35%), linseed (33.33%), and palm kernel (44.6%) [25].

The tree is a drought-resistant perennial and grows well in marginal land that has little or no agricultural or industrial value due to poor soil and other undesirable characteristics. Correspondingly, unlike some other conventional edible feedstock, the planting of this crop is no threat to existing arable land land and the food chain. In 2012, it was reported that Forest Research Institute Malaysia (FRIM) has successfully planted 6000 *Jatropha curcas* plants in Terengganu, since the state has about 71,000 ha of problematic land along the coast that is left without commercialized agricultural activities [26].

Besides FRIM, other government agencies such as the Malaysian Rubber Board (MRB) estimate that approximately 50 hectares (ha) of *Jatropha curcas* were planted in Sungai Buloh, Selangor, and Kota Tinggi as of June 2012. In the early phase, about 1712 ha of land in total was earmarked for the principal cultivation of this crop in the country. Likewise, a small number of local private companies have indicated their willingness to cultivate *Jatropha curcas* on the scale of 400 ha to 1000 ha. Some stakeholders are planning to expand the cultivation to 57,601 ha in total by the year 2015. The Plantation, Industries, and Commodities Ministry (MPIC) has also initiated an experimental project on Jatropha, for which 300 ha has been allocated [4].

Biodiesel fuel has been widely adopted in most countries because it is biodegradable, non-toxic, and environmentally friendly with lower greenhouse gas (GHG) emissions. Furthermore, biodiesel adoption is considered an excellent method for reducing noise and potentially scaling down air pollutants such as carbon monoxide (CO), sulfur, polycyclic aromatic hydrocarbon (PAH), smoke, and particulate matter (PM) [27]. The most significant fuel characteristics considered for biodiesel application in diesel engines are density, viscosity, cetane number, and flash point [28].

The Cetane number is the principal indicator of fuel quality, particularly ignition and combustion in diesel engines. Typically, a high Cetane number indicates a lower ignition delay time—i.e., the time interval from the injection of the fuel to initialization of ignition in the combustion chamber. Typically, the parameter ensures a good quality fuel combustion, cold start, and engine performance, along with low white smoke formation and emissions [29]. The Cetane number of Jatropha is reported to be as high as 55, which is similar to that of diesel (Table 1). Hence, any biodiesel to be effectively substituted for diesel should retain a higher cetane number.


**Table 1.** Comparison of vegetable oil with biodiesel specification [30–33].

In practice, the blend of any vegetable-based biodiesel with petroleum diesel need to comply with the two most referred biodiesel standards, namely, the American Standard Specifications for Biodiesel Fuel (B100) Blend Stock for Distillate Fuels (ASTM 6751) and European Standard for Biodiesel (EN 14214). According to the both standards, biodiesel must meet the minimum flash point—i.e., above 120 ◦C. The flashpoint is the temperature at which a fuel begins to burn after interaction with fire. Typically, any fuel with a high flash point could result in the deposition of carbon in the combustion compartment. Since Jatropha oil has a lower flash point compared to palm oil (162 ◦C and 181 ◦C, respectively), it has a higher potential compared to palm oil as a biodiesel.

Based on Table 1, Jatropha oil has a medium viscosity between palm oil and diesel, which is good for biodiesel utilization. Typically, most vegetable oils have higher viscosities due to their high fatty acid compositions relative to petroleum diesel. The higher viscosity indicates a better lubrication of the fuel, which reduces wear on the moving mechanical parts of the engine. Ultimately, the reduced wear prevents leakage and reduces issues related to power losses and the durability of engines. Viscosity plays an important roles in the atomization efficiency of fuel injection inside the combustion chamber, fuel droplet size distribution, and the mixture uniformity. If viscosity is too high it may lead to pump damage, filter clogging, poor combustion, and increased emissions. A higher viscosity will also lead to greater surface tension and will influence the dissolution of a liquid jet into smaller fuel droplets, which will impose a bad effect on the spray characteristic of the fuel spray injector in a diesel engine. As a result, larger size fuel droplets are injected from an injector nozzle instead of a spray of fine droplets, leading to inadequate air–fuel mixing [34,35].

Furthermore, the price of biodiesel feedstock derived from *Jatropha curcas* is considered to be the most affordable compared to other biodiesel feedstock as shown in Table 2. Its low end price will attract consumers to be using biodiesel on the road. Eventually, it would increase the market demand for biodiesel used in the transportation vehicles and will enhance the economy from people living in rural areas. The data on the price of B100 biodiesel for different feedstock in Table 2 was reported by Lim, S. and Teong, L.K. [36].

**Table 2.** A comparison of biodiesel prices from different feedstock (adapted from [36]).


### **4. Biodiesel Processing from** *Jatropha curcas*

The general processing of biodiesel from *Jatropha curcas* oil involves three major steps, namely seed drying, oil extraction, and transesterification (the processing of pure vegetable oil into biodiesel), as shown in Figure 3. There are also other minor steps that are considered significant, such as the cleaning of the seeds, dehulling, and post-harvest storage. The conventional technique for recovering oils from Jatropha seeds is through the use of a mechanical screw press machine. However, a large proportion of the oil is retained in the kernel, which requires more effective ways to extract the residual oil. The most notable extraction techniques include ultrasound-assisted systems, enzyme extraction, and the utilization of catalytic materials [37]. The catalyst materials are chemicals that enhance the process of transesterification. The extraction method is closely related to the cost of mass biodiesel production in a biorefinery plant.

**Figure 3.** Biodiesel processing of *Jatropha curcas* [38–41].

The extracted oil subsequently undergoes a purification and transesterification process for the production of crude biodiesel. However, the crude biodiesel cannot be directly used as a transportation fuel due to limitations such as the standard requirements for biodiesel in the industry. Therefore, the crude biodiesel is usually blended with pure diesel in certain percentages before utilization in diesel engines. Before blending, the crude biodiesel is purified to remove unwanted moisture and the chemical waste produced during the transesterification process. The most popular method of purification is water washing since it is cheap and easy, although this time-consuming [42] process needs to be run several times until no more glycerol is produced.

The composition of the fatty acid significantly affects the fuel properties of the biodiesel [18]. Typically, inedible oils such as Jatropha comprise high compositions of detrimental free fatty acids (FFA) (>1% *w*/*w*), which reduces the biodiesel yields. Likewise, the high amount of fatty acid hampers the direct conversion of the oil into biodiesel since the high FFAs promote soap formation, which can hamper the separation of products during or after transesterification. Jatropha oil comprises nearly 14% FFA, which exceeds by far the standard limit of 1% FFA. Therefore, the pretreatment stage is required to lower the feedstock FFAs for an enhanced yield of biodiesel [30]. The typical unwanted saponification reaction that forms soap and water when NaOH catalyst is utilized is presented in Equation (1).

$$\text{R1 -COOH (FFA)} + \text{NaOH (sodium hydroxide)} \rightarrow \text{R1} \text{COONa (soap)} + \text{H}\_2\text{O (water)}.\tag{1}$$

Therefore, the two-step transesterification process is an efficient method used extensively to process crude oil from *Jatropha curcas* that contains significant FFAs. Furthermore, the pretreatment or esterification process using the acid- base catalyst is performed to reduce the FFA content of *Jatropha curcas* oil. Hence, transesterification subsequently results in an optimal yield of 90% methyl ester after two hours [43]. In addition, the acid catalyst reduces the FFA content to <1% through conversion into esters by esterification. The second step involves the transesterification of the triglycerides in *J. curcas* oil into biodiesel in the presence of an alkaline catalyst. The unsaturation of fatty acids in oil is an important factor that determines the biodiesel quality. In this aspect, Malaysia, however, sits on the favourable side, as polyunsaturated fatty acids are lower in Malaysia-grown *Jatropha* oil than in the varities found in neighbouring countries [44]. Interestingly though, the Triacylglycerol (TAG) profile among these different varities of *Jatropha* oil from Malaysia and neighbouring countries did not show significant differences [45]. The kinematic viscosity of *Jatropha* oil is higher than that of regular diesel fuel, which indeed imparts a problem for all use in a diesel engine without blending. On the other side, it is much safer to handle and for storage than regular diesel fuel at higher temperatures [39]. Considering all these promising factors, its a highly promising seed oil to be taken seriously when it comes to boosting the socio-economic conditions in Malaysia.

### **5.** *Jatropha curcas* **Planting Challenges in Malaysia**

*Jatropha curcas* is a highly promising crop for biodiesel production, although supportive innovative technologies are required for planting, harvesting, and oil extraction. Furthermore, mechanized crop operations are limited, and hence Malaysia needs to import the knowledge and machines from other countries such as India. Other notable challenges of Jatropha production are the poor seed yield, low impute crop, and pest and disease vulnerability. Although it is a promising crop for biodiesel production, the unavailability of a high-yielding cultivar is a major failure factor [46]. Besides this, high-yield fluctuation among trees and the lack of disease resistance could also hamper the commercialization of Jatropha biodiesel. In addition, it also needs appropriate nutrients and irrigation for growth and maturity, although it could flourish under limited conditions. Recent studies have reported that *Jatropha curcas* is susceptible to virus-related contaminations such as the Cucumber mosaic virus, powdery mildew, leaf spots, and soil fungous diseases. Other notable challenges are insect and rodent attacks, which result in the extensive defoliation of the plant [47,48].

In some environs, Jatropha creates complications such as weeds, which could require higher labour costs during cultivation [49]. Lastly, technologies for harvesting and post-harvesting, such as oil extraction, are still lacking. Furthermore, biodiesel is vulnerable to oxidation when exposed to air, selected storage conditions, and high levels of unsaturated fatty acids [50]. As a result, the oil content deteriorates due to inappropriate handling and storage. In addition, the main ester components of biodiesel could rapidly undergo hydrolysis to form carboxylic acids in the presence of water. Hence, both materials along with the chemical structure of biodiesel affect the swelling characteristics of the elastomer, which in turn depends on its composition and the preparation of the compound [51].

#### **6. Approaches to Enhance the Jatropha Seed Oil**

One of the most critical solutions is the cultivation of high oil yield *Jatropha curcas*, although such commercial varieties are lacking. The existing Jatropha breeding scheme is restricted to the traditional approach, which involves the collection of wild plant germplasm capital of Jatropha [52]. Furthermore, the review of modern applications of biotechnology for improving Jatropha is minimal [53]. In particular, research is largely absent on the expression, cloning, and annotation of biotic roles for Jatropha genes, which are responsible for its economic characteristics [46]. The main purpose of cultivation should be to advance the unit seed yield of Jatropha for commercial uses. Therefore, Jatropha cultivation techniques must require the application of numerous field practices such as planting, site planning, tree density, irrigation management, and cropping treatments.

Other notable practices involve fertilization and canopy protection, along with the control of pests and diseases. However, there is limited research that precisely and systematically validates the effect of field activity on the seed yield of *Jatropha curcas*. Selected methodological studies on planting base and management restrict the commercial cultivation of *J. curcas* [54]. Furthermore, there are limited comprehensive field or empirical reports on seed yield under different agronomic or treatment methods. For instance, data on the cultivated Jatropha tree density, the strength and interval of pruning its canopies, the insecticide impact, fertilization, and irrigation efficiency are mostly lacking in the literature [46].

#### **7. Economic and Business Perspectives of Biodiesel from Jatropha Oil**

Malaysia is amongst the world's premier biodiesel manufacturers. The immense profit of biodiesel in terms of rising fossil fuel values and the intention of decreasing the emission of greenhouse gasses (GHG) are factors that contribute to the development of the biodiesel market in Malaysia. Additionally, the community also is personally involved in production in order to improve income and eliminate poverty. In Malaysia, the growth of the biodiesel industry has been maintained at the top of its agenda by granting a significant amount of subsidies and farmer support programs. In fact, the government is encouraging private companies to launch more treating plants and improve biodiesel for vehicles and electricity generation. This is parallel with the post estimation of diesel vehicles, which accounts for approximately 5% of the motor vehicle population in Malaysia. The number of vehicles used is as indicated by the registered vehicles from 1996 to 2009. Based on the post estimation, diesel vehicles may possibly provide a larger share of the total in the forthcoming prior to the commencement of B5 and the campaign of the government incentives.

To date, the majority of countries have declared the standards and policies of their biodiesel. All countries have regulated their mandate or aim for biodiesel consumption success and proclaimed the exploitation of biodiesel energy fusion in their policies. As recapped in Figure 4, the national biodiesel policy of Malaysia stated on 21 March 2006 [24] objectives are as follows:


c) Reducing the country's dependence on depleting reserves of fossil fuels, promoting the demand for palm oil, and stabilizing its prices.

**Figure 4.** Strategic five thrusts of Malaysia's national biofuel policy and implementation (Adapted from [24]).

#### **8. Conclusions**

In conclusion, *Jatropha curcas* has a bright future as the next important biodiesel feedstock, considering the problems currently faced by the oil palm industry in Malaysia. It is necessary to create a higher value of its by-products in order to make Jatropha a viable biofuel in the market. Therefore, other parts of this crop, such as wood, fruit shells, seed husks and kernels, could be used to produce renewable energy [55]. The waste generated after the oil extraction process, such as the pressed cake, could also be utilized as organic fertilizer. Thus, this crop has the same characteristics as the oil palm crop, which can be used as a whole package. However, more research and development needs to be undertaken by researchers to find solutions to the existing challenges outlined in this review.

Biodiesel from *Jatropha curcas* has great potential to be implemented because it has lower carbon and emissions of GHG. It also has a lower cost compared to palm biodiesel. However, our dependency on the foreign workers in the plantation is unavoidable. The mechanization and automation specifically for maintaining the good health of *Jatropha curcas* must be improved and tested in the field beforehand. The workers must be careful while harvesting because the oil yield is dependent on the right timing of harvesting. As we know, this fruit's ripening is uneven, making harvesting a strenuous and time-consuming process. Until 2015, it has been stated that Malaysia has a total of 259,906 hectares of Jatropha crops plantation, and the current planted crops are capable of producing 4.27 tons of dry seeds every year [56,57].

As the world has been affected by global warming and the alarming threat of food security for the growing human population, much attention should be focused on non-edible oil bearing crops as biodiesel feedstock. This renewable green energy will protect the environment from the emission of harmful gases due to the combustion of fossil fuels and become an effective substitution for the depletion of the mineral resources of Earth. While many challenges await as this crop is introduced as a new biodiesel feedstock, it will never be impossible to cope with them when there is ample research and development undertaken and more expertise involved in joint ventures for the research project on *Jatropha curcas*.

**Author Contributions:** Investigation, N.H.C.H.; writing—original draft preparation, N.H.C.H.; Funding acquisition, N.K.; Supervision, N.K.; M.A.H., and B.M.S.; writing—review and editing, N.K., M.A.H. and B.M.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Geran Putra—Inisiatif Putra Berkumpulan grant number (9671301) to support the research and development activities in Universiti Putra Malaysia Bintulu Sarawak Campus, Malaysia.

**Acknowledgments:** The authors acknowledge the funding received from the Geran Putra—Inisiatif Putra Berkumpulan by Universiti Putra Malaysia, Malaysia.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Enzymatic Saccharification with Sequential-Substrate Feeding and Sequential-Enzymes Loading to Enhance Fermentable Sugar Production from Sago Hampas**

**Nurul Haziqah Alias, Suraini Abd-Aziz, Lai Yee Phang and Mohamad Faizal Ibrahim \***

Department of Bioprocess Technology, Faculty of Biotechnology and Biomolecular Sciences, Universiti Putra Malaysia, Serdang 43400, Selangor, Malaysia; nhaziqahalias95@gmail.com (N.H.A.); suraini@upm.edu.my (S.A.-A.); phanglaiyee@upm.edu.my (Y.L.P.) **\*** Correspondence: faizal\_ibrahim@upm.edu.my; Tel.: +603-9769-1936

**Abstract:** Sago hampas composed of a high percentage of polysaccharides (starch, cellulose and hemicellulose) that make it a suitable substrate for fermentation. However, the saccharification of sago hampas through the batch process is always hampered by its low sugar concentration due to the limitation of the substrate that can be loaded into the system. Increased substrate concentration in the system reduces the ability of enzyme action toward the substrate due to substrate saturation, which increases viscosity and causes inefficient mixing. Therefore, sequential-substrate feeding has been attempted in this study to increase the amount of substrate in the system by feeding the substrate at the selected intervals. At the same time, sequential-enzymes loading has been also evaluated to maximize the amount of enzymes loaded into the system. Results showed that this saccharification with sequential-substrate feeding and sequential-enzymes loading has elevated the solid loading up to 20% (*w*/*v*) and reduced the amount of enzymes used per substrate input by 20% for amylase and 50% for cellulase. The strategies implemented have enhanced the fermentable sugar production from 80.33 g/L in the batch system to 119.90 g/L in this current process. It can be concluded that sequential-substrate feeding and sequential-enzymes loading are capable of increasing the total amount of substrate, the amount of fermentable sugar produced, and at the same time maximize the amount of enzymes used in the system. Hence, it would be a promising solution for both the economic and waste management of the sago hampas industry to produce value-added products via biotechnological means.

**Keywords:** sago hampas; amylase; cellulase; substrate feeding; saccharification; biomass

#### **1. Introduction**

Sago palm, scientifically known as *Metroxylon sagu*, can be found in tropical Southeast Asia. This plant grows healthily in the environment with an average temperature of 25 ◦C and an approximate humidity of 70% [1]. Its ability to thrive in a swampy area and grow naturally without the need for pesticide or herbicide has made sago palm cultivation increase in recent decades [2]. Approximately 90% of commercially grown sago palm in Malaysia is in Sarawak, a state located in the east of Malaysia. Sago palm became an important economic species and resource for this region as the production of sago starch was reported to be approximately 15–25 tons/ha. The starch composition in sago palm is the highest (25 tons/ha) as compared with other types of the plant such as rice (6 tons/ha), corn (5.5 tons/ha), wheat (5 tons/ha) and potato (2.5 tons/ha) [3]. The commercial production of sago starch was established in Malaysia in the 1970s and became one of the most important industries in terms of its contribution to the export revenue [4]. Due to the upward trend in sago starch production, the amount of waste generated from this industry has significantly increased due to the numerous of sago processing mills. The industry of sago palm has generated an extensive amount of waste including sago bark,

Phang, L.Y.; Ibrahim, M.F. Enzymatic Saccharification with Sequential-Substrate Feeding and Sequential-Enzymes Loading to Enhance Fermentable Sugar Production from Sago Hampas. *Processes* **2021**, *9*, 535. https:// doi.org/10.3390/pr9030535

**Citation:** Alias, N.H.; Abd-Aziz, S.;


Academic Editor: Pietro Bartocci

Received: 31 December 2020 Accepted: 27 January 2021 Published: 18 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

sago hampas and sago wastewater [5]. The polluting effects caused by these agro-wastes have become the main concern and started to generate attention among the researchers attempting to find a solution with a sustainable approach.

In Malaysia, the mass production of sago starch from 600 logs of sago palm per day was 15.6 tons of woody bark, 237.6 tons of wastewater and 7.1 tons of starch fibrous sago pith residue [6]. Starch fibrous sago pith residue or commonly known as sago hampas composed of starchy and lignocellulosic components, which are 54.6% of starch, 31.7% of cellulose and hemicellulose, and 3.3% of lignin [7]. The high polysaccharide content and low lignin composition in sago hampas make this agricultural residue a promising feedstock for fermentation operation. More importantly, there is no pretreatment required before the saccharification process due to the low lignin content in sago hampas [8]. The pretreatment process is one of the crucial and costly processes in the bioconversion of agricultural residue into fermentable sugar before fermentation [9]. This process is important to reduce and/or alter the lignin component, expose the internal structure of cellulose to be accessible by the cellulase [10]. Eliminating this step from the whole process could save huge operational cost. In addition, a high percentage of remaining starch in sago hampas can be easily hydrolysed by amylase to produce fermentable sugar. Therefore, the utilization of sago hampas as a raw material for fermentation operation could be cost-effective for the downstream processing of sugar production and eventually for the production of fermentation-based products, and at the same time, prevent the environmental pollution that is caused by the underutilization of sago waste.

In the production of fermentable sugars from sago hampas, this material must be gelatinized before saccharification. Gelatinization needs to be carried out to break down the hydrogen bond in the sago starch, thus, allowing the amylase to attack the α-glycosidic bond of the polysaccharides into glucose monomer [11]. Gelatinization is a simple process that applies the heat to the starch in the presence of water. As a result, the water is gradually absorbed and caused the starch granules to swell [12]. The addition of glucoamylase (EC 3.2.1.3) with a debranching enzyme such as pullulanase (EC 3.2.1.41) is practically useful as they can hydrolyse the α-1,6-glycosidic bond that links the polysaccharides chain into branches [13]. The hydrolysis process of starch by amylase takes less than 24 h [14]. Meanwhile, to fully degrade the sago hampas into fermentable sugar, the cellulase is also being used to degrade the cellulosic component. Cellulase is a mixture of enzymes composed of endoglucanase (EC 3.2.1.4), exoglucanase (EC 3.2.1.91) and β-glucosidase (EC 3.2.1.21) that act synergistically on the degradation of the β-glycosidic bond of cellulosic component into glucose monomers [15].

The low sugar concentration always obtained from enzymatic saccharification is usually not enough to initiate the fermentation process. This problem can be overcome by increasing the insoluble solid load that will enhance the fermentable sugar production, and thus, improve the efficiency of the downstream processing. However, increasing the substrate concentration can reduce the hydrolysis yield due to the high viscosity, which subsequently causes poor mixing and mass transfer [16,17]. In addition, the current process also suffers from a high cost of enzymes used in the saccharification process, especially cellulase. Therefore, the improvement of the enzymatic saccharification step is required from an economic perspective and for process feasibility. The mixture of amylase and cellulase used in the saccharification of sago hampas has been previously reported by Husin et al. [7] for the production of biobutanol. It was found that the mixture of amylase and cellulase produced higher fermentable sugar as compared to a single enzyme, either amylase or cellulase alone. However, the process has been done in batch for simultaneous saccharification and fermentation (SSF) to produce biobutanol. Although a high biobutanol production yield was obtained, the low sugar concentration produced by this operation can be improved.

Therefore, in this present study, saccharification with sequential-substrate feeding and sequential-enzymes loading has been introduced with the aim of enhancing the fermentable sugar production, and at the same time, maximize the usage of enzymes in the saccharification process. Sequential-substrate feeding is expected to provide sufficient time for enzymes to digest solid material into soluble sugar components, thus improving the capacity of the system to be loaded with a higher amount of substrate. Meanwhile, sequential-enzymes loading is expected to maximize the amount of enzymes used in the system, and technically reduce the cost of enzymes and make the process more feasible.

#### **2. Materials and Methods**

The experimental design of this study is shown in Figure 1. The sequential-substrate feeding and sequential-enzymes loading were conducted in comparison with the batch process. The process began when the gelatinized, dried and ground sago hampas mixed with acetate buffer were added with the enzymes (amylase and cellulase). Sago hampas in a total of 20 g/L was fed sequentially based on the feeding interval followed with the study on the sequential-enzymes loading by loading amylase and cellulase at a different amount. Saccharification was conducted at 60 ◦C, 150 rpm for 6 days or until a stationary production of fermentable sugars was obtained. Then, to optimize the mixing process, the effect of agitation speed was also conducted.

**Figure 1.** Schematic diagram of experimental work for enzymatic saccharification of sago hampas.

#### *2.1. Substrate Preparation*

Sago hampas supplied from River Link Sago Resources Sdn. Bhd. in Mukah, Sarawak was sun-dried for 1–2 days to drain off the excess water naturally. Then, sago hampas was gelatinised by boiling in 0.1 M of acetate buffer for 15 min. Then, the gelatinised sago hampas was oven-dried at 60 ◦C for 24 h and subsequently ground to pass a 1 mm screen. The moisture content of the dried samples was analysed to quantify the buffer to be added prior to the enzymatic saccharification process.

#### *2.2. Batch Saccharification*

Batch saccharification was conducted following the methods by Husin et al. [7]. An amount of 7% (*w*/*v*) of sago hampas was gelatinized in 100 mL of 0.1 M acetate buffer solution at pH 5.5 in comparison with 7% (*w*/*v*) of gelatinised, dried and ground sago hampas. The saccharification was conducted by adding Dextrozyme 71.4 U/gsubstrate of amylase (Novozymes, Bagsvaerd, Denmark) and 20 FPU/gsubstrate of Acremonium cellulase (Meiji Seika Co, Japan) into the mixture. The saccharification process was performed at 60 ◦C and 150 rpm up to 144 h of incubation time. All the saccharification process was performed in

shaker incubator (Labwit, ZHWY-1102C) and samples were drawn out for every 24 h for the analyses.

#### *2.3. Sacchatification with Sequential-Substrate Feeding and Sequential-Enzyme Loading*

The strategies applied in this study were developed for optimising the feeding interval of the gelatinised, dried and ground sago hampas as well as the amount of enzymes loading. For this study, two sets of experiment were conducted, which are the enzymes only initially loaded and the enzymes sequentially loaded according to the amount of substrate feeding. Batch saccharification with the total substrate concentration of 20% (*w*/*v*) with 71.4 U/gsubstrate and 20 FPU/gsubstrate of amylase and cellulase, respectively, was performed as a control.

#### 2.3.1. Feeding Interval

Table 1 illustrates the strategies behind how the feeding interval was applied. The saccharification process consists of five variables for the feeding interval, which are 0 (control), 6, 12, 24 and 36 h of interval time. The substrate was fed sequentially based on the feeding interval that makes it the total substrate loading at 20% (*w*/*v*) with the 2% (*w*/*v*) of the initial substrate for each variable except for the control. The substrate was fed only up to 72 h and prolonged the incubation for another 3 days or until the stationary production of fermentable sugar was obtained. The saccharification was performed with two sets of experiments in order to make the comparison study where the first set was the enzymes that only initially loaded while the second set was the enzymes that were loaded sequentially to per g of substrate feeding.

**Table 1.** Feeding strategies for the feeding interval of sago hampas on the saccharification process.


#### 2.3.2. Enzymes Loading

The effect of enzymes loading were tested on the saccharification with the sequentialsubstrate feeding for both amylase and cellulase, with five variables as illustrated in Table 2. The optimal sequential-substrate feeding was 6% (*w*/*v*) fed sequentially at every 24 h of interval time. The study was also performed with two sets of experiments: the first set was the initially added enzymes only while the second set was the enzymes added sequentially according to the substrate feeding.

**Table 2.** Variables for the enzyme loading of amylase and cellulase on the sequential-substrate feeding saccharification.


#### 2.3.3. Agitation Speed

The effect of agitation speed was performed after the optimal conditions for sequentialsubstrate feeding and sequential-enzymes loading were obtained. The agitation speeds were set at 60, 90, 120, 150 and 180 rpm, and saccharification with no agitation was also conducted as a control.

#### *2.4. Analytical Procedures*

The starch content was determined using iodine starch colorimetric methods by Nakamura [18]. The lignocellulosic biomass of sago hampas were determined by its three major components which are cellulose, hemicellulose and lignin using the standard procedure of acid hydrolysis method and high performance liquid chromatography (HPLC) from the National Renewable Energy Laboratory method, NREL/TP-510-42623 [19]. Total extractives content were determined following the method by the National Renewable Energy Laboratory method, NREL/TP-510-42619 [20]. The sugar monomers obtained by saccharification were analysed by high performance liquid chromatography (HPLC) (Jasco, Tokyo, Japan) equipped with a refractive index (RI) detector and a column (Shodex KS-801, Tokyo, Japan) for ligand exchange chromatography. A 100% ultrapure water was used as a mobile phase with a flow rate of 0.6 mL/min and the temperature of the column was fixed at 80 ◦C using oven column [7]. A statistical analysis was conducted in order to analyse the significant effect from each variable on saccharification process using an analysis of variance (ANOVA) by Statistical Analysis Software (SAS) version 9.4 and verified considering *p* < 0.05.

#### **3. Results and Discussion**

#### *3.1. Characteristics of Sago Hampas*

The composition of raw sago hampas was determined as shown in Table 3. The characterization of sago hampas in this study has been evaluated in order to ensure the quality of the substrate. All the values shown in the table are comparable to those reported previously. Starch content in sago hampas was 56.0%, while the cellulose, hemicellulose and lignin contents were of 20.7%, 11.2% and 3.1%, respectively. The value of starch content in sago hampas depends on the quality of the extraction process conducted by the sago mills [21]. Besides, both water and solvent extractives in sago hampas have a low value of 2.33% and 0.67%, respectively.

Starch, cellulose, hemicellulose and lignin are the major components of the sago hampas while extractives are the minor components. Extractives in biomass are usually the non-structural components, which can be extracted by water or other solvents. The solvents can be ethanol, acetone, benzene, hexane, dichloromethane and toluene. The compounds that are commonly extracted out from biomass are fats, waxes, phenolics, resin acids and inorganic compounds. These non-structural components of biomass could potentially interfere with the downstream analysis of the biomass sample. This may result in an error on the structural sugar values where the hydrophobic extractives could inhibit the penetration of the sample that directly caused incomplete hydrolysis [20]. Extractives could also falsely result in high values of lignin when the unhydrolyzed carbohydrates condense with the acid-insoluble lignin. Some studies reported that by removing these extractives, it showed an improvement on the enzymatic digestibility and glucose yield, respectively. Sago hampas has lower total extractives content (3.0%) as compared to other types of biomass such as corn stover (13.5%) and *Artemisia ordosica* (7.78%) [22,23]. Therefore, no pretreatment is required to remove the extractive, as this amount is not significantly affecting the saccharification process. Based on this condition of sago hampas (a high carbohydrate composition with low lignin and extractives content), this substrate has a high beneficial advantage to be used as material for the fermentation feedstock.


**Table 3.** Comparison of the composition of raw sago hampas with a different collection of sago hampas.

n.d indicates not determined.

#### *3.2. Saccharification of Sago Hampas*

To enhance the fermentable sugar production from sago hampas, several saccharification strategies were carried out by identifying the effects of feeding interval, enzymes loading and agitation speed. The saccharification process was performed by determining the effect of preparing a substrate under wet and dried conditions followed by the sequential feeding of the dried substrate. Then, the effects of initially loaded enzymes and sequentially loaded enzymes throughout the saccharification process were also evaluated. The whole strategies were performed to determine the optimum conditions of the saccharification that can produce the highest fermentable sugar production with a low amount of enzymes loading.

#### 3.2.1. Effect of Wet and Dried Sago Hampas

Initially, this particular experiment was carried out to determine the substrate condition used throughout the saccharification process, either in wet or dried condition. This is because in the early study of the saccharification of sago hampas, the sago hampas was gelatinised before the saccharification process, and the gelatinised sago hampas was in the wet condition. However, saccharification of sago hampas with sequential-substrate feeding must be in the dried form to ensure the consistency of the substrate feeding throughout the experiment. The gelatinization process was conducted before the saccharification process due to a high starch content in sago hampas and due to the crystalline structure of the starch. The crystallized structure of starch must be destroyed and change into the amorphous structure in order to make it susceptible to the enzyme action [21]. It works when the substrate suspension is heated in the presence of water and swelling starch granules break down the hydrogen and hydrophobic bonds [24].

In this study, the saccharification profiles of the wet and dried substrates (Figure 2) show that there is no significant difference in the sugar produced from wet and dried substrates, which produced 43.29 g/L (±2.54) and 46.23 g/L (±0.76) of sugar, respectively. However, it can be seen from the graph that the saccharification rate of the wet substrate is slightly faster as compared to the dried substrate with a slightly lower concentration of sugar being produced. It was suggested that drying the temperature also plays an important role in the characterization of sago starch in terms of drying kinetics and the equilibrium of moisture content [26]. The wet substrate might be easily degraded by amylase since the starch structure has been exposed with water, while dried substrates sometimes need the structure to be accessible by the amylase.

**Figure 2.** Effects of the wet and dried substrates used in the saccharification process of sago hampas by the mixture of amylase and cellulase.

#### 3.2.2. Effect of Feeding Interval

In this study, 20% (*w*/*v*) of the total substrate was saccharified with the added mixture of Dextrozyme amylase (71.4 U/g) and Acremonium cellulase (20 FPU/g). However, the high substrate concentration applied in the saccharification process might lead to the high viscosity and subsequently reduced the reaction rate [27]. To saccharify a high amount of substrate, sago hampas must be loaded sequentially throughout the process to maintain the low level of viscosity and increase the accessibility of the enzymes towards the substrate [28]. To examine the effect of feeding interval and the enzymes used, several feeding intervals were conducted in two separate experiments, namely that of the only initially loaded enzymes and enzymes sequentially loaded according to the substrate feeding. All presented data are the means of triplicates ± S.D and stated using the Tukey's test with *p* < 0.05.

Figure 3a illustrates the effect of the feeding interval of the substrate with the initially loaded enzymes. This experiment was also compared with the control (without substrate feeding), where 20% (*w*/*v*) of the total substrate was added at the beginning of the saccharification. From this study, it can be observed that the control produced more sugars (80.33 g/L ± 0.02) as compared with the sequentially added substrate. The sequentially loaded substrate did not show an impressive increment in sugar production, as the enzyme activity might be alleviated throughout the process due to the lower substrate concentration at the beginning [17]. This is because sequential substrate feeding was added with only 2% (*w*/*v*) of substrate loading, whereby, a high amount of enzyme was initially added. Thus, throughout the time, most of the enzyme activity reduced and could not provide sufficient degradation capacity when the substrate was added over time. The extent of the inhibition depends on the ratio of total enzyme to the total substrate. This could be explained by the enzymes' active sites not binding with sufficient substrate at the beginning. Then, the produced sugars might occupy the empty enzyme active site and become an inhibitor to the newly added substrate. In addition, it can be seen that the pattern for the control showed that the saccharification can only be achieved until 48 h of incubation time. In comparison with the sequential-substrate feeding saccharification, the degree of hydrolysis was observed until 96 h of incubation time. Even though the sugar production from the control saccharification was higher than that of the sequential-substrate feeding saccharification, the high substrates used became significant waste, as these cannot be further hydrolysed by the enzymes. It seems that the enzyme–substrate complex has reached its maximum saturation point which is most likely due to the jamming effect phenomenon caused by the overcrowding of the substrate, with the enzyme and substrate obstructing one another [29].

**Figure 3.** Effect of the feeding interval of the substrate in the saccharification process: (**a**) initially loaded enzymes only; and (**b**) sequentially loaded enzymes according to substrate feeding.

Figure 3b shows the effect of the feeding interval with sequentially loaded enzymes according to the substrate feeding. From the graph, it can be observed that the effect of feeding interval on saccharification with the sequentially loaded enzymes produced more fermentable sugars as compared with the control. From the graph, the degree of hydrolysis for sequential-substrate feeding saccharification showed that it increases gradually up to 120 h of incubation time with the addition of substrate compared with the control that reached its maximum saturation point at 48 h of incubation time. The periodical addition of substrate prolonged the process to produce more fermentable sugars [30]. In the comparison with Figure 3a, there is about 34.51% of increment for sugar production. It can also be observed that the viscosity is reduced and more runny solution can be observed. Thus, a greater fermentable sugars yield was produced. This might be due to the rate of reaction which is affected by the total number of enzymes as well as the concentration of substrate loaded accordingly [31]. The result showed that the rate of saccharification did not decrease with the increase in substrate concentration when the enzyme-to-substrate ratio was kept constant. It can be concluded that the optimal interval feeding time for sequential-substrate feeding with the sequential-enzyme loading was every 24 h, which produced the highest sugar concentration of (95.37 g/L ± 0.93) with *p* < 0.05.

There are several studies reported about the crucial parameters that affect the enzymatic saccharification, and one of the parameters is substrate-related. In this present study, the substrate concentration and feeding style have been discussed in terms of how they affect the saccharification process. The substrate features such as the substrate size, lignin structure and substrate pore surface area also play an important role in the accessibility of the substrate to enzyme [32]. This is because the lignocellulosic biomass has a complex

structural arrangement, thus, it is more difficult to hydrolyse as compared to starch-based biomass. Most of the lignocellulosic biomass such as corn stover, switch grasses and forest residue need to undergo a pretreatment process prior to saccharification. This is to ensure that the lignin component was removed or reduced and/or altered to allow the interaction of substrate to enzymes.

#### 3.2.3. Effect of Enzymes Loading

To establish an economically feasible saccharification process, an appropriate amount of enzyme used must be determined as an enzyme used in sugar production generally contributes a significantly high cost in terms of the total operational cost of converting biomass into value-added products [33]. In the previous experiment, the loaded enzyme was 71.4 U/g of amylase and 20 FPU/g of cellulase with the interval feeding time at every 24 h. In this experiment on the effect of enzymes loading, there were five variables for each enzyme range from 7.1 U/g to 285.7 U/g for amylase and 5 FPU/g to 25 FPU/g of cellulase were examined. All presented data are the means of triplicates ± S.D and stated using the Tukey's test with *p* < 0.05.

Figure 4a shows the effect of initially loading the enzyme while Figure 4b shows the effect of the sequential-enzyme loading on the saccharification with the sequential-substrate feeding. Based on Figure 4a, it can be observed that sugar production declined after 96 h of incubation time. In addition, when the enzyme was initially loaded, the inhibitors might have formed from the formation of the product, which subsequently caused the competitive inhibition [34] whereby the substrate and inhibitors compete for the same enzyme's active site [35]. Competitive inhibition occurs in one of the enzymes, in this case cellulase since it is considered the principle bottleneck for practical production from lignocellulosic materials. This situation usually occurs with high substrate concentration as the inhibitors limit the enzyme velocity in their biochemical reaction [36]. Thus, high substrate concentration might escalate the possibility of enzyme inhibition caused by product inhibitors. In addition, the availability of the enzyme's active site is limited at high substrate concentration due to the accumulation of excess substrate. Other factors that may contribute to the low degree of polysaccharide conversion at high substrate concentration, mainly because of the decrease within the reactivity of cellulosic material in the course of hydrolysis, different kinds of enzyme inactivation, and the non-specific adsorption of cellulolytic enzymes onto lignin [37]. From this study, the sugar production significantly showed the difference between the various amounts of loaded enzymes.

Meanwhile, based on Figure 4b, it showed that the performance of hydrolysis was increased with the sequentially added enzyme, as the ratio of enzyme to the substrate used is one factor that affects the saccharification [38]. When comparing these two studies, the effect of sequential-enzymes loading might reduce the inhibition of the product. In addition, the saccharification process has been prolonged up to 120 h of incubation time. However, after 120 h of incubation time, the efficiency of enzyme catalytic reaction has deprived due to the prolonged incubation time, which probably because of the enzyme has achieved its maximum enzyme thermal deactivation process after being exposed at high temperature for a long time [39]. The enzymes' reactivity is mostly associated with the enzyme-related parameters. The maximum utilisation of enzymes during saccharification is important because the enzymes represent the major contribution to the total cost of the bioconversion of biomass to value-added products. The amount of enzyme loading depends on the composition and structural arrangement of the substrate [40]. The effect of sequential-enzyme loading has resulted in approximately a 43% increment in sugar production as compared with the initially loaded enzymes. This study also showed that sequential-enzyme loading with 10 to 20 FPU/g cellulase and 14.3 to 142.9 U/g amylase did not significantly affect sugar production (*p* < 0.05). Hence, the presence of excess enzymes was a waste, since it was usually underutilized and consequently leads to the unnecessarily high cost in the saccharification process. Therefore, a low amount of enzyme

loading (14.3 U/g of amylase with 10 FPU/g of cellulase) can be optimally used to produce a high concentration of (112.48 g/L ± 1.26) with *p* < 0.05.

**Figure 4.** Effect of loading enzymes on the saccharification process: (**a**) initially loaded enzymes only; and (**b**) sequentially loaded enzymes according to substrate feeding.

3.2.4. Effect of Agitation Speed

The effect of agitation speed has been evaluated in the range of 60–180 rpm together with no agitation at the constant temperature of 60 ◦C for the saccharification of sequentialsubstrate feeding and sequential-enzyme loading. The effect of agitation speed is important to determine the relationship between the saccharification efficiency and liquid viscosity of saccharification. This is because the liquid viscosity from the saccharification process increases with the increase in the saccharification time since the substrate was added sequentially throughout the process.

The trend of fermentable sugar production with different agitation speed for 6 days of incubation is presented in Figure 5 As expected, saccharification with no agitation produced the lowest sugar concentration of (88.38 g/L ± 1.52). Increasing the agitation speed from 60 to 150 rpm had significantly increased the sugar production to (119.9 g/L ± 0.32) with *p*-value < 0.5. Agitation enhances the mass transfer rate during saccharification, thus improving the hydrolysis process and increasing the conversion rate of the substrate into fermentable sugars by the enzymes. It can also be observed that the agitation speed has a moderate effect on saccharification. At a 180 rpm agitation speed, the fermentable sugar production showed a slight reduction probably due to the shear stress or shear forces. It is also reported that vigorous agitation speed could have aggravated cell damage, which in turn led to the mechanical inactivation of the enzymes and contributed to the reduction in enzyme stability. Thus, sugar production was suppressed. In typical saccharification, the agitation speed of 120–150 rpm was found to be optimum. Therefore, as for this study, the agitation of 150 rpm can be considered most suitable for the saccharification of sago hampas to produce fermentable sugars. Even though the agitation speed has a moderate effect on saccharification, it still plays an important role in the cost-effective downstream processing as it can reduce energy consumption.

**Figure 5.** Effect of agitation speed on the saccharification process. All data are the means of 3 replicates ± S.D. The different alphabet indicates significant difference at *p* < 0.05. The data were stated using LSD test.

#### *3.3. Comparison Study*

This study demonstrated the enhancement of sugar production by conducting saccharification with sequential-substrate feeding and sequential-enzyme loading. The comparison of different saccharification strategies in Table 4 shows that this saccharification strategy significantly improved sugar production compared to other studies. It was 13.53% of sugar increment which could be observed when comparing the normal saccharification with sequential-substrate feeding. However, throughout the process, the normal saccharification produced more suspended solution because of the high solid–liquid ratio at the beginning of the saccharification. This highly viscous solution was caused by an ineffective heat and mass transfer due to the solution not being mixed properly, and hence, the reduction in the diffusion of the enzyme and end product [27].

In order to further enhance the fermentable sugar production, the rate of saccharification process needs to be increased. However, due to the high prices of commercial cellulase, the addition of more enzymes is not the best option. Alternatively, the rate of saccharification and fermentable sugar production can be accelerated by implementing a new strategy of feeding for both substrate and enzymes. Therefore, the sequential-substrate feeding and sequential-enzymes loading was implemented in this study. Surprisingly, the small difference in the feeding strategies had significantly improved the sugar production by 20.87% as compared to the process without sequential-enzyme loading. This approach has also reduced the amount of enzyme used for both amylase and cellulase by 20% and 50%, respectively.


The effectiveness of the enzymatic hydrolysis for starch and lignocellulosic components not only depends on the substrate concentration as a whole, but it also requires the optimum synergistic action of the amylase and cellulase components towards the substrate. Sago hampas mainly compose of starch residue (56.0%) and the starch itself is composed of linear amylose and branched amylopectin. Therefore, glucoamylase (EC 3.2.1.3) with a debranching enzyme, pullulanase (EC 3.2.1.41), was used for this study as it is an exo-acting enzyme that mainly hydrolyses an α-1,4 and α-1,6 glycosidic bond from the non-reducing ends of starch chains, which leads to the production of glucose. There are some studies which reported that glucoamylase was able to enhance the efficiency of hydrolysis and increase the substrate concentration at the active site of the enzyme catalytic centre by binding to the raw starch granules and disrupt the surface structure of starch [41,42]. Meanwhile, the lignocellulosic components in this study were hydrolysed by cellulase, which is composed of endoglucanase (EC 3.2.1.4), exoglucanase (EC 3.2.1.91) and β-glucosidase (EC 3.2.1.21) that act synergistically on the degradation of a cellulosic component into glucose monomers. During the saccharification, it is noted that the substrate level should be high enough to provide a sufficient reaction between the enzymes and substrate. Thus, in order to enhance the higher sugar production, it is must proportionally increase the rate of reaction by adding more substrate. However, by relatively adding more substrate, it contributes to the jamming effect due to enzymes needing to act on more portions of the starch and lignocellulosic components [43]. Therefore, the feeding style of substrate and enzymes must be studied in order to achieve the optimal reaction of saccharification. Hence, in this study, the efficiency of saccharification has been improved when the sequential-substrate and enzymes feeding was applied. The enzymes are able to work at their optimal level when they are supplemented according to the amount of fed substrate, instead of being added once at the beginning only.

The correlation between the analysed variables (substrate feeding, enzyme loading and agitation speed) with the sugar production are conducted. From the correlation screening, it can be seen that only enzymes loading was significantly affecting the sugar concentration with *p* < 0.05. The strength of enzymes loading and sugar concentration were associated with the R<sup>2</sup> value of 0.92. This is explained that by the increasing enzymes loading, which produces a higher sugar concentration. However, it should be noted that a further increase in enzymes loading is not economically practical. Both the feeding interval and agitation speed resulted in *p* > 0.05, where it does not give a significant effect on the sugar concentration. The feeding interval showed an insignificant effect in enhancing the sugar concentration, which might be due to the enzymes that engage with the substrate and were not sufficient when the interval time of the substrate feeding was too long. This will cause the alleviation of enzyme activity and result in the slowing down of the reactions [44]. Agitation speed plays an important role for an effective mixing of the substrate and enzymes. However, by increasing the agitation speed too much, there is no significant effect on the sugar concentration, where the shear forces might occur and lead to cell damage [45]. On top of that, the interaction of enzymes loading towards the feeding interval was conducted and it showed that the coefficient of determination (R2) was 0.96. The R2 value indicates that 96% of the variation in enzyme loading is explained by the feeding interval. Thus, the digestibility of the enzymes toward the substrate was improved as the enzymes were loaded based on the amount of substrate feeding.

In comparison with other studies, as shown in Table 5, a high substrate feeding of 20 g/L can be fed in sequential-substrate feeding and fed-batch saccharification as compared to batch saccharification, which is capable of the maximum load substrate feeding at 5–9 g/L. Previous studies have reported the decreased efficiency of hydrolysis when more than 9 g/L of solid substrate were used and this was due to product inhibition, enzyme inactivation and the decrease in substrate reactivity [46]. In addition to that, in the fed-batch saccharification [17], a similar situation was observed whereby a high increment in sugar production was obtained as compared to batch saccharification. However, no feeding strategy for enzyme loading was conducted. In comparison with the sequentialsubstrate-feeding and sequential-enzymes loading, fermentable sugar production from sago hampas was significantly improved even at the low amount of loaded enzymes. It shows that sugar production was affected by how the substrate and enzymes were fed. In order to increase sugar production, the rate of reaction needs to be increased by increasing the amount of substrate and enzymes in the system. However, a high solid–liquid ratio could hinder the effective heat and mass transfer and thus limit the diffusion of enzymes and the formation of end products. Therefore, the efficient saccharification process was dependent on the synergistic feeding of the substrate and enzymes.

**Table 5.** Comparison of sugar production by various substrates, the amount of enzyme used and different saccharification operations.


The properties of the biomass usually affected by their structure that make the hydrolysis difficult to be carried out with a higher substrate feeding. Some of the biomass with high lignin content usually needs to undergo pretreatment and this increases the total cost of bioconversion. The downstream processing usually occurs with a high cost of production. However, this cost can be reduced by minimizing the enzyme usage and maximizing the amount of substrate used. Enzymatic saccharification with a sequentialsubstrate feeding and sequential-enzymes loading was proven to be a promising strategy for efficient and economical saccharification. This present study is possible to implement on a large-scale processing production. It is suggested to study the proper technology that can be integrated with the present study in order to ensure the feasibility of the process.

#### **4. Conclusions**

Sago hampas has been notably known as a promising substrate for the production of fermentation-based products due to its high content of polysaccharides and the low lignin composition. In this study, the saccharification of sago hampas into fermentable sugar has been enhanced by implementing the feeding strategies of the substrate and the enzymes together with the effect of agitation speed. It can be concluded that the sequential-substrate feeding at 6 g/L for every 24 h increased the sugar production by 16% as compared to the batch process. Meanwhile, saccharification with sequential-substrate feeding and sequential-enzymes loading produced a high sugar concentration of 119.90 g/L, and at the same time reduced the amount of amylase from 71.4 U/gsubstrate to 14.4 U/gsubstrate and cellulase from 20 FPU/gsubstrate to 10 FPU/gsubstrate used in the process. Findings from this research suggest that the potential of sequential-substrate feeding and sequential-enzyme loading can be used as an alternative in improving the saccharification process of other types of substrates in order to obtain a significantly higher amount of fermentable sugar derived from biomass.

**Author Contributions:** Conceptualization, M.F.I. and N.H.A.; methodology, M.F.I. and N.H.A.; software, N.H.A.; validation, M.F.I., S.A.-A. and P.L.Y.; formal analysis, N.H.A.; investigation, N.H.A.; resources, N.H.A.; data curation, N.H.A. and M.F.I.; writing—original draft preparation, N.H.A.; writing—review and editing, N.H.A. and M.F.I.; visualization, N.H.A.; supervision, M.F.I., S.A.-A. and P.L.Y.; project administration, M.F.I.; funding acquisition, M.F.I. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was financially supported by the Geran Putra, Universiti Putra Malaysia, project number GP/2017/9559300.

**Data Availability Statement:** All data used to support the funding of this study are included within the article.

**Acknowledgments:** Highly appreciation to all members of the Environmental Biotechnology Research Group, Universiti Putra Malaysia for their kind support and help.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Processes* Editorial Office E-mail: processes@mdpi.com www.mdpi.com/journal/processes

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18