**A Sugarcane-Bagasse-Based Adsorbent Employed for Mitigating Eutrophication Threats and Producing Biodiesel Simultaneously**

**Wan Nurain Farahah Wan Basri 1,2, Hanita Daud 1, Man Kee Lam 2,3, Chin Kui Cheng 4, Wen Da Oh 5, Wen Nee Tan 6, Maizatul Shima Shaharun 1, Yin Fong Yeong 3, Ujang Paman 7, Katsuki Kusakabe 8, Evizal Abdul Kadir 9, Pau Loke Show <sup>10</sup> and Jun Wei Lim 1,2,\***


Received: 25 June 2019; Accepted: 5 August 2019; Published: 28 August 2019

**Abstract:** Eutrophication is an inevitable phenomenon, and it has recently become an unabated threat. As a positive, the thriving microalgal biomass in eutrophic water is conventionally perceived to be loaded with myriad valuable biochemical compounds. Therefore, a sugarcane-bagasse-based adsorbent was proposed in this study to harvest the microalgal biomass for producing biodiesel. By activating the sugarcane-bagasse-based adsorbent with 1.5 M of H2SO4, a highest adsorption capacity of 108.9 ± 0.3 mg/g was attained. This was fundamentally due to the surface potential of the 1.5 M H2SO4 acid-modified sugarcane-bagasse-based adsorbent possessing the lowest surface positivity value as calculated from its point of zero charge. The adsorption capacity was then improved to 192.9 ± 0.1 mg/g by stepwise optimizing the adsorbent size to 6.7–8.0 mm, adsorption medium pH to 2–4, and adsorbent dosage to 0.4 g per 100 mL of adsorption medium. This resulted in 91.5% microalgae removal efficiency. Excellent-quality biodiesel was also obtained as reflected by the fatty acid methyl ester (FAME) profile, showing the dominant species of C16–C18 encompassing 71% of the overall FAMEs. The sustainability of harvesting microalgal biomass via an adsorption-enhanced flocculation processes was also evidenced by the potentiality to reuse the spent acid-modified adsorbent.

**Keywords:** eutrophication; sugarcane bagasse; adsorption; harvest; biodiesel; reusability

#### **1. Introduction**

"Eutrophication is the enrichment of water with nutrient salts that causes structural changes to the ecosystem, namely, the increase in production of microalgae and aquatic plants, depletion

of fish species, general deterioration of water quality and other effects that reduce and preclude use". This is one of the definitions given to the eutrophic process by the Organization for Economic Cooperation and Development [1,2]. Eutrophication is a serious environmental threat since it gives rise to deterioration of water quality, and it is also one of the major impediments in achieving the quality objectives established by the Water Framework Directive in the EU as well as those in other countries [3]. Intrinsically, all water resources are subjected to a natural and slow eutrophication process. However, in recent years, the eutrophication threat has undergone very rapid progression due to the presence of various human activities, particularly the farming of cash crops in agriculture. When the eutrophication threat becomes intense, undesirable effects and environmental imbalances arise. The two most acute phenomena to stem from eutrophication are hypoxia in the deep parts of lakes (or lack of oxygen) and microalgal blooms that may produce harmful toxins. Both occurrences can plausibly bring severe devastation to the aquatic life living in the afflicted water bodies [4–8].

Malaysia and Indonesia are growing countries with steadily improving economies due to huge contributions from many agricultural sectors such as palm oil, rice and paddies, sugarcane, and other planted cash crops. Although oil palm is a major cash crop that can promote the gross domestic product (GDP), another rising cash crop in Malaysia and Indonesia is sugarcane, with total productions of 28.1 and 28.0 million metric tons, respectively, in years 2016/17 [9–12]. This gives rise to gigantic levels of agricultural waste from the juice crushing process, i.e., sugarcane bagasse which is usually left to decay on the fields. The accumulated sugarcane bagasse waste without proper management will potentially lead to the spreading of diseases that adversely affect humans, animals, and the environment. Of late, a fraction of this bagasse has been used as fuels in sugar factories, and some raw materials have been exploited in pulp and paper making. Sugarcane bagasse is a type of lignocellulosic biomass that generally consists of cellulose (50%), hemicellulose (25%), and lignin (25%) [13]. The hemicelluloses are made up of C5 and C6 sugar, while lignin comprises about one-fourth of the lignocellulose biomass. These biological polymers have hydroxyl and/or phenolic functional groups that can be chemically activated to produce materials with new properties. Accordingly, the carboxylic and hydroxyl groups can improve the capacity of adsorption via ion exchange and complexation processes. Owing to its low ash content (approximately 2.4%), sugarcane bagasse can offer many advantages when compared with other crop residues such as rice straw and wheat straw, which have about 17.5% and 11.0% ash contents, respectively [14]. In addition, sugarcane bagasse is a porous material with a relatively high fixed carbon content. Considering these advantages, sugarcane bagasse has been among the preferred waste materials chosen to be used as an adsorbent by many researchers previously to resolve issues associated with water pollution via adsorption processes [5,8,9,15,16].

Nevertheless, the application of adsorption processes to concentrate planktonic microalgal cells and subsequently flocculate and separate them from a liquid medium has been documented little of late. Accordingly, the potential use of a fabricated sugarcane-bagasse-based adsorbent to mitigate the eutrophication phenomenon served as the prime objective of this comprehensive study. The adsorptive operating conditions were initially optimized to spur the removal of microalgal biomass via adsorption-enhanced flocculation processes from the eutrophic water. In tandem with this study, the fundamental rationale describing the attachment of microalgal cells onto the surface of the fabricated sugarcane-bagasse-based adsorbent was also unveiled. Since microalgal biomass is conventionally lauded for containing myriad valuable biochemical compounds, the harvested sugarcane-bagasse-based adsorbent loaded with microalgal biomass was subsequently exploited for biodiesel production. The plausible reusability of the spent sugarcane-bagasse-based adsorbent was lastly assessed to confirm the sustainability of the fabricated sugarcane-bagasse-based adsorbent in mitigating the eutrophication threat.

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

#### *2.1. Fabrication of Acid-Modified Sugarcane-Bagasse-Based Adsorbents*

The residual sugarcane bagasse generated after juice crushing was used as a precursor in fabricating various acid-modified sugarcane-bagasse-based adsorbents. Residual sugarcane bagasse was initially amassed from the local market and cut into pieces of about 3–4 cm each. The pieces of bagasse were then boiled and washed thoroughly to remove the entrapped sugar. The wet sugar-free bagasse was later air-dried to partially remove the water content before drying to constant weight at 102–105 ◦C in an oven. The dried sugarcane bagasse was pulverized using a 0.5 mm blade grinder and subsequently activated using various H2SO4 acid concentrations: 0.1, 0.5, 1.0, 1.5, 2.0, and 2.5 M. This was achieved by individually introducing 15 g of pulverized sugarcane bagasse into 1 L of each of the H2SO4 acid concentrations. Each mixture was finally stirred for 24 h and washed with distilled water until a neutral pH was attained in discarded washing water prior to once again drying to constant weight at 102–105 ◦C in an oven. All the fabricated acid-modified sugarcane-bagasse-based adsorbents were separately kept in a desiccator ready for use.

#### *2.2. Determination of Adsorbent Point of Zero Charge*

Points of zero charge (pHPZC) were initially measured from each of the fabricated acid-modified sugarcane-bagasse-based adsorbents. The pHPZC is determined as the adsorption medium pH that causes the surface charge density of the adsorbent to become zero. The pHPZC values of all the acid-modified sugarcane-bagasse-based adsorbents were measured following the solid addition method as described by Ofomaja and Naidoo [17] with modifications. Briefly, to a series of 100 mL Erlenmeyer flasks, 45 mL of KNO3 solution with a concentration of 0.01 M was added to every flask. The pH in each flask was then individually adjusted from pH 2 to 9 and filled up to 50 mL using a similar KNO3 solution. The initial pH values of all flasks were accurately recorded, and 0.1 g of specific acid-modified sugarcane-bagasse-based adsorbent was administered into each flask. All flasks were securely capped and agitated manually and sporadically for the next 2 days, allowing the charges to equilibrate before recording the final pH values from each flask. Subsequently, the differences between the initial and final pH values (ΔpH) of each flask were plotted against the respective initial pH values. The pHPZC of the adsorbent was finally determined from the point of intersection of the resulting curve at ΔpH = 0. Later, these procedures were repeated using other acid-modified sugarcane-bagasse-based adsorbents to determine their respective pHPZC values.

#### *2.3. Characteristics of Eutrophic Water*

A water sample was collected from a eutrophic body at a site which receives discharged streams from the aquacultural activities in the vicinity (Figure 1). The in situ measurements indicated a pH of 3.8 ± 0.3, temperature of 31.6 ± 1.7 ◦C, and dissolved oxygen concentration of 0.1 ± 0.1 mg/L. The sample was instantly transported to the laboratory, and ex situ determination of the microalgal biomass concentration was immediately executed via gravimetric analysis. The eutrophic water was found to be loaded with 0.86 ± 0.2 g/L of microalgal biomass. The remaining sample was stored in the cold room upon reaching the laboratory at 2.0 ± 1.5 ◦C to minimize biological and chemical changes after isolating the sample from the eutrophic body. A predetermined volume of eutrophic water was later siphoned from the homogenized stock sample in the cold room and allowed to reach ambient temperature prior to use in the experiments outlined hereafter.

**Figure 1.** Eutrophic body which currently receives discharged streams from the aquacultural activities in the vicinity.

#### *2.4. Setups for Eutrophic Water Treatment*

Six 150 mL Erlenmeyer flasks were each initially filled with 100 mL of the eutrophic water sample. The sample was homogenized via vigorous shaking before the predetermined volume was transferred into each Erlenmeyer flask. A quantity of 0.5 g of acid-modified sugarcane-bagasse-based adsorbent with respective H2SO4 concentrations of 0.1, 0.5, 1.0, 1.5, 2.0, and 2.5 M was individually administered into each of the Erlenmeyer flasks containing a eutrophic water sample. All the mixtures were immediately adjusted to pH 3, and sufficient agitation was provided for each flask to prevent the settlement of adsorbents to the bases of all flasks. Sampling of microalgal biomasses present in the eutrophic water over time from each flask was executed to determine the residual microalgal biomass concentrations as well as for the time course studies. Samplings were also performed to determine the adsorption efficiencies (Equation (1)) and capacities (Equation (2)) of each acid-modified sugarcane-bagasse-based adsorbent loaded into each flask at equilibria as confirmed from the time course studies.

$$\text{Adsorptions efficiency} = \frac{\left[ \text{Initial microbial biomass} \right] - \left[ \text{Microload gain mass at any time} \right]}{\left[ \text{Initial microbial biomass} \right]} \times 100\% \tag{1}$$

$$\text{Adsorptions capacity} = \frac{\text{Weight of adsorbed microlagonal biomass}}{\text{Weight of adsorption}} \tag{2}$$

The best acid-modified sugarcane-bagasse-based adsorbent was then selected and sieved into four different sizes, namely, <4.0 mm, 5.6–6.7 mm, 6.7–8.0 mm, and >8.0 mm. A quantity of 0.5 g of each size of adsorbent was individually administered into the Erlenmeyer flasks containing the eutrophic water sample. Similar procedures to those described above were later executed in selecting the best size of acid-modified sugarcane-bagasse-based adsorbent. Afterward, stepwise optimization of the pH of the adsorption media and the adsorbent dosage were also performed to enhance the adsorptive removal of microalgal biomass from the eutrophic water. The studied ranges of the pH of adsorption media and adsorbent dosage were from pH 2 to 10 and 0.1 to 0.7 g in 100 mL of adsorption medium, respectively.

#### *2.5. Microalgal Lipid Extraction and Transesterification into Fatty Acid Methyl Esters (FAMEs) of Biodiesel*

Lipid extraction from the adsorbed microalgal biomass was accomplished following the procedures described by Bligh and Dyer [18] with modifications. Initially, the microalgal biomass adsorbed onto acid-modified sugarcane-bagasse-based adsorbent was harvested from the adsorption medium using a sieving net and dried to constant weight at 102–105 ◦C in an oven. The dried microalgal biomass, together with the adsorbent, was doused with water mixed with chloroform: methanol (1:2 *v*/*v*) at a ratio of 1.6:6.0 *v*/*v*. The mixture was immediately sonicated for 30 min at 40 kHz and 40 ◦C. For further separating the free lipids into the chloroform layer after sonication, an approximately 20% chloroform: water (1:1 *v*/*v*) solution by volume with respect to the total volume of the sonicated sample was added and mixed well manually. Centrifugation for 20 min at 5600 rpm was then employed to separate the sample solution into three layers, namely, a lipid and chloroform mixture (bottom layer), residual microalgal biomass adsorbed onto the spent sugarcane-bagasse-based adsorbent (middle layer), and a methanol and water mixture (top layer). The bottom layer was then collected by suction, and 8% chloroform by volume with respect to the total volume of the sample was added to the remaining mixture, mixed well manually, and centrifuged again to recover the residual lipids. The amassed volume of the lipid and chloroform mixture from both consecutive separation processes was lastly dried under compressed air blow to constant weight prior to the lipid yield determination (Equation (3)).

$$\text{Lipid yield} = \frac{\text{Weight of extracted liquid from adsorbed mineral biomass}}{\text{Weight of absorbed wind liquid was onto absorbed} - \text{Weight of initial absorbed}} \times 100\% \text{ (aligned found } \frac{\text{Weight of alcohol}}{\text{Weight of material produced in the system}})$$

The extracted lipids from the adsorbed microalgal biomass were subsequently transesterified into FAMEs of biodiesel following the procedures as described by Mohd-Sahib and Lim [19]. Briefly, a vortex mixer at 2000 rpm was initially employed to homogenize about 10–15 mg of extracted lipid, 1 mL of chloroform, and 2 mL of KOH in methanol (1.5 mg/mL). The sample was then capped well and transesterified for 3 h at a temperature of 60 ◦C. The temperature was maintained by placing the sample in a water bath until the transesterification process was completed. The FAMEs mixture of the biodiesel produced was later purified by repeatedly washing with 5 mL of distilled water followed by separation using a separating funnel until a neutral pH was measured from the aqueous phase. Next, the chloroform layer loaded with FAMEs was dried under compressed air blow to constant weight. The FAMEs mixture was finally injected into a gas chromatograph to analyze the individual FAME composition forming the biodiesel. The FAME profile obtained determines the quality of the biodiesel derived from the harvested microalgal biomass from eutrophic water. The details of the operating conditions of the employed Shimadzu brand gas chromatograph (Model GC-2010 plus) equipped with a flame ionization detector (GC-FID) can be acquired from Mohd-Sahib and Lim [19]. Ultimately, the individual FAME composition was calculated as follows:

$$\text{Compression of specific FAME species} = \frac{A\_S}{A\_{I\text{STD}}} \times \frac{C\_{I\text{STD}} \times V\_{I\text{STD}}}{m} \times 100\% \tag{4}$$

where *AS* = Peak area of specific FAME species;

*AISTD* = Peak area of internal standard; *CISTD* = Concentration of internal standard; *VISTD* = Volume of internal standard; *m* = Mass of sample.

#### *2.6. Assessment of the Reusability of Spent Sugarcane-Bagasse-Based Adsorbent*

Upon lipid extraction, the spent sugarcane-bagasse-based adsorbent or lipid-exhausted sugarcane-bagasse-based adsorbent was dried to constant weight at 102–105 ◦C in an oven. The dry adsorbent was then administered into 100 mL of fresh eutrophic water sample, and the optimum operational conditions to remove microalgal biomass from the adsorption medium were employed according to the first cycle, termed Cycle-1. Cycle-2 was allowed to continue until the equilibrium was attained as confirmed from the time course measurement of the residual microalgal biomass concentration. To that end, the adsorption efficiency and capacity were recorded following Equations (1) and (2), respectively. The harvested microalgal biomass adsorbed onto spent

sugarcane-bagasse-based adsorbent was subsequently subjected to lipid extraction again, followed by drying of the lipid-exhausted sugarcane-bagasse-based adsorbent to constant weight at 102–105 ◦C. The dry adsorbent was later reused to remove microalgal biomass from fresh eutrophic water samples in Cycle-3 to Cycle-5.

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

#### *3.1. Enhancement of Microalgal Biomass Adsorption from Eutrophic Water*

The potential of various acid-modified sugarcane-bagasse-based adsorbents for removing microalgal biomass via adsorption-enhanced flocculation is presented in Table 1. In essence, the fresh negatively charged microalgal cells are hypothesized to be electrostatically attracted to the positively charged supporting materials. The acid-modified treatment mode was employed in this study in order to economically intensify the positive surface potential of the sugarcane-bagasse-based adsorbents. Increasing the H2SO4 acid concentration during the treatment of sugarcane-bagasse-based adsorbents progressively increased the separation of the microalgal biomass from the liquid medium, reaching the highest adsorption efficiency of 89.6% while using 1.5 M of H2SO4 to acid-modify the sugarcane-bagasse-based adsorbent. The adsorption efficiency and capacity were observed to gradually decrease thereafter when concentrations of H2SO4 higher than 1.5 M were employed to treat the sugarcane-bagasse-based adsorbents. Many studies have associated the reduction of adsorption with increasing acid concentration employed for treating adsorbents to the destruction of adsorbent structures [20,21]. Instead of agreeing with this typical rationale, the surface potentials of all fabricated acid-modified sugarcane-bagasse-based adsorbents were measured and exploited to elucidate the adsorption phenomena observed in this study. The positivity value of each adsorbent calculated by subtracting the pH of the adsorption medium from their pHPZC is demonstrated in Figure 2. A mirror image trend against either the adsorption efficiency or capacity (Table 1) was obtained for the positivity values of increasingly H2SO4 acid-modified sugarcane-bagasse-based adsorbents. In this regard, the 1.5 M H2SO4 acid-modified sugarcane-bagasse-based adsorbent presented the lowest positivity value, attracting more negatively charged microalgal cells to adsorb onto the adsorbent surfaces. On the flipside, the higher positivity values of the surface potentials of other acid-modified sugarcane-bagasse-based adsorbents would attract more counter ions from the adsorption medium, forming a layer of negatively charged counter ions. This layer would shield the microalgal cells from interacting with the acid-modified sugarcane-bagasse-based adsorbents, indirectly preventing the adsorption-enhanced flocculation processes. In the case of putting the 1.5 M H2SO4 acid-modified sugarcane-bagasse-based adsorbent to use, the formation of this counter ions layer was inconspicuous, thereby permitting the adsorption of greater microalgal biomass than the other fabricated adsorbents.


**Table 1.** The performance of various acid-modified sugarcane-bagasse-based adsorbents in removing microalgal biomass from eutrophic water.

To examine adsorbent sizes, the 1.5 M H2SO4 acid-modified sugarcane-bagasse-based adsorbent was sieved into four different sizes, namely, <4.0 mm, 5.6–6.7 mm, 6.7–8.0 mm, and >8.0 mm. The residual biomass concentrations of microalgae along with the adsorption time and the consequential adsorption capacities from the use of each adsorbent size are shown in Figure 3. Although the adsorption capacities attained by the adsorbent sizes of 6.7–8.0 and >8.0 mm were comparable, the use of 6.7–8.0 mm sized adsorbent could achieve a faster adsorption-enhanced flocculation equilibrium than the >8.0 mm sized adsorbent—specifically, 0.0576 and 0.0522 g/L h, respectively. As the targeted adsorbate in

this study was a suspended microalgal biomass having a size range of 20–50 μm rather than the dissolved adsorbate, a large adsorbent size offered more macropores for accommodating microalgal cells. For comparison with dissolved adsorbate, the powdered form of adsorbents is generally preferred due to the presence of more micropores and mesopores capable of capturing the targeted dissolved adsorbate [22,23]. However, continuous increase of adsorbent size, especially beyond 8.0 mm, would culminate in the reduction of external surface area, impoverishing the frequency of contact between the adsorbent and microalgal cells. Therefore, acid-modified sugarcane-bagasse-based adsorbent with a size range of 6.7–8.0 mm was regarded as the best size for adsorption-enhanced flocculation of microalgal biomass; its adsorption capacity increased to 143.6 ± 1.7 mg/g, as compared with only 108.9 ± 0.3 mg/g using the unsieved adsorbent (Table 1). Besides this, the flocculated microalgal biomass adsorbed onto this size range of adsorbent could also be easily harvested from the liquid medium via a simple sieving net.

**Figure 2.** The point of zero charge (pHPZC) values of various acid-modified sugarcane-bagasse-based adsorbents. The values in parentheses indicate the positivity of pHPZC subtracted by pH 3, i.e., the pH of the adsorption medium.

The adsorption capacities were maintained in a range of 143.0–148.0 mg/g when the adsorption-enhanced flocculation processes were performed in adsorption media with pH values between 2 and 4. Increasing the pH of the adsorption medium to 6 caused a decrease in the adsorption capacity to 127 ± 4.3 mg/g. The adsorption capacities were noticed to plummet to merely 119.9 ± 0.5 and 104.3 ± 3.3 mg/g when the pH was shifted to basic in the adsorption media, namely, at pH values of 8 and 10, respectively. As the 1.5 M H2SO4 acid-modified sugarcane-bagasse-based adsorbent possessed a pHPZC value of 4.50 (Figure 2), increasing pH value of the absorption medium above 4.50 would gradually increase the adsorbent surface potential negativity. Since microalgal cells are negatively charged, repulsion between the adsorbent and adsorbate arising from charge similarity would debilitate the adsorption-enhanced flocculation processes of the microalgal biomass. In acidic adsorption mediums, adsorption efficiencies of more than 80% could be easily attained at pH values between 2 and 4. Indeed, the pH value of the eutrophic water of 3.8 ± 0.3 also fell within this pH range, therefore safely circumventing the necessity to pretreat the eutrophic water through pH adjustment prior to adsorption-enhanced flocculation processes.

**Figure 3.** The time courses of residual biomass concentrations of microalgae adsorbed onto various sizes of acid-modified sugarcane-bagasse-based adsorbent and the consequent adsorption capacities from the use of each adsorbent size.

Increased amounts of adsorbent in terms of grams per 100 mL of adsorption medium increased the adsorption capacity, as revealed in Figure 4. At an adsorbent dosage of 0.1 g/100 mL, all the macropore active sites were swiftly occupied by microalgal cells, as demonstrated by the rapid attainment of adsorption-enhanced flocculation equilibrium during the early time course study. The insufficient active sites were later offset by the increasing adsorbent dosage, reaching a maximum adsorption capacity of 192.9 ± 0.1 mg/g when 0.4 g of adsorbent was introduced into 100 mL of adsorption medium, equivalent to an adsorption efficiency of 91.5%. Further increment of adsorbent dosages beyond 0.4 g/100 mL gave rise to the presence of increasing numbers of free active sites, thereby reducing the adsorption capacities. Increasing the adsorbent dosage from 0.4 to 0.7 g/100 mL slightly prompted more microalgal biomass to adsorb onto the adsorbent, steadily increasing the adsorption efficiency from 91.5% to 94.9%, respectively. This negligible rise in adsorption efficiency was due to the presence of excessive free active sites available to capture more microalgal cells from the diluted adsorption medium. As the percentage point increase was only about 3% (from 91.5% to 94.9%) but required a 75% increase in adsorbent dosage (from 0.4 to 0.7 g/100 mL), the 0.4 g/100 mL adsorbent dosage was considered ideal for executing adsorption-enhanced flocculation processes for the remediation of eutrophic water.

In summary, 1.5 M H2SO4 acid-modified sugarcane-bagasse-based adsorbent with a size range of 6.7–8.0 mm was employed to spur the separation of microalgal biomass from eutrophic water via adsorption-enhanced flocculation processes. The adsorption capacities achieved from sequential studies were 108.9 ± 0.3 and 143.6 ± 1.7 mg/g, respectively. Subsequently, the adsorption capacity was maintained in the range of 143.0–148.0 mg/g when adsorption-enhanced flocculation processes were carried out in adsorption media with pH values between 2 and 4. Finally, the adsorption capacity was further improved to 192.9 ± 0.1 mg/g when an adsorbent dosage of 0.4 g was introduced into 100 mL of adsorption medium. A high adsorption efficiency of microalgae of 91.5% and low residual biomass concentration of microalgae of 0.064 g/L in the adsorption medium were attained.

**Figure 4.** The time courses of residual biomass concentrations of microalgae adsorbed onto various dosages of acid-modified sugarcane-bagasse-based adsorbent and the consequent adsorption capacities from the use of each adsorbent dosage.

#### *3.2. Biodiesel Derived from Harvested Microalgal Biomass*

The microalgal biomass harvested from the eutrophic water was then sequentially subjected to extraction to obtain the lipid biocompounds and perform transesterification into a FAMEs mixture of biodiesel. The lipid yield acquired from the extraction was 30.3 ± 0.0 wt %, which was very close to the lipid yield of the control microalgal biomass of 31.0 ± 0.2 wt %. The control microalgal biomass was harvested via the centrifugation of fresh eutrophic water. With a standard deviation value of merely 0.5%, the insignificant difference between the two lipid yields showed that the adsorption-enhanced flocculation processes employed to remove the microalgal biomass from the liquid medium did not have an obvious deleterious impact on the microalgal lipid content. Finally, the FAME profile obtained from the transesterification of extracted lipids is presented in Table 2. The FAMEs are important components of biodiesel; thus, the quality of biodiesel can be concluded from the FAME profile study. There were 21 species of FAMEs identified in the biodiesel derived from microalgal biomass adsorbed onto the sugarcane-bagasse-based adsorbent. Among the FAMEs, C16 to C18 were the dominant species, making up approximately 71% of the overall FAMEs mixture. These species are also naturally found in many oil-bearing crops, e.g., soybean, sunflower seed, cottonseed, and palm oil, which are suitable for use as biodiesel [24]. According to Song and Pei [25], feedstock suited for the production of biodiesel must contain palmitic (16:0), stearic (18:0), oleic (18:1), linoleic (18:2), and linolenic (18:3) acids, which were all found in the FAME profile of the microalgal biomass adsorbed onto the sugarcane-bagasse-based adsorbent. The saturation degrees of the FAMEs mixture were calculated afterward and showed the mixture to contain 44.87% of saturated fatty acid (SFA), 32.55% of monounsaturated fatty acid (MUFA), and 22.58% of polyunsaturated fatty acid (PUFA). Biodiesel with higher levels of SFA will generally have a high cetane number and oxidative stability but poor low-temperature properties. The presence of MUFA species of palmitoleic (C16:1), oleic (C18:1), eicosenoate (C20:1), and erucate (C22:1) acids also essentially give rise to biodiesel with suitable oxidative stability, besides ameliorating cold flow [26]. The presence of high levels of PUFA, on the other hand, will offer excellent cold flow. However, this biodiesel type is easily oxidized [27]. The low degree of PUFA (<30%) and high degree of MUFA and SFA (>65%) as reported by Mohd-Sahib and Lim [19] were also attained in biodiesel derived from microalgal biomass adsorbed onto the sugarcane-bagasse-based adsorbent. According to Mohd-Sahib and Lim [19], this type of FAMEs

mixture has great potential for the production of high-quality biodiesel with acceptable oxidative stability and cold flow properties.


**Table 2.** The fatty acid methyl ester (FAME) profile of biodiesel from microalgal biomass harvested via adsorption-enhanced flocculation processes.

#### *3.3. Potential Reusability of Spent Sugarcane-Bagasse-Based Adsorbent*

From the point of view of sustainability, the potential to reuse spent sugarcane-bagasse-based adsorbent after the first cycle of removing microalgal biomass via adsorption-enhanced flocculation processes was evaluated. The performance results of spent sugarcane-bagasse-based adsorbent for five consecutive cycles of reuse are presented in Table 3. Cycle-1 represents the first use of virgin sugarcane-bagasse-based adsorbent upon fabrication to carry out adsorption-enhanced flocculation processes. After extracting the lipids from the microalgal biomass adsorbed onto sugarcane-bagasse-based adsorbent, the lipid-exhausted sugarcane-bagasse-based adsorbent was then used to carry out adsorption-enhanced flocculation processes again in Cycle-2, and these procedures were reiteratively repeated until Cycle-5. The adsorption efficiency diminished about 10% in Cycle-2 when compared to Cycle-1, though the adsorption capacity was still maintained above 100 mg/g. This could be undoubtedly rationalized by the presence of vacant and unexploited active sites left unoccupied after Cycle-1. When the same adsorbent was employed for the third time, parts of the loose cellulosic materials were noticed to inevitably detach from the sugarcane-bagasse-based adsorbent. Accordingly, the total weight of the lipid-exhausted sugarcane-bagasse-based adsorbent plunged about 20% to merely 0.33 g in Cycle-3. This could be due to continuous mechanical abrasion among the adsorbent solids, stemming from the agitation provided during the adsorption-enhanced flocculation and lipid extraction processes. As a result, the adsorption efficiency and capacity also dropped to lowest values of 38.2% and 59.8 mg/g, respectively, in this cycle. As the material structure of the remaining lipid-exhausted sugarcane-bagasse-based adsorbent was more firm and stable after Cycle-3, the adsorption efficiency was observed to improve to about 48% in Cycle-4, later plateauing in Cycle-5. As the total weight of the lipid-exhausted sugarcane-bagasse-based adsorbent did not differ much in Cycle-4 and Cycle-5, the adsorption capacities were measured in the range of approximately 70–80 mg/g in these two cycles.


**Table 3.** Performance of spent sugarcane-bagasse-based adsorbent for reiterative removal of microalgal biomass via adsorption-enhanced flocculation processes.

#### **4. Conclusions**

Acid-modified sugarcane-bagasse-based adsorbent was successfully employed to remove microalgal biomass from eutrophic water via adsorption-enhanced flocculation processes. By activating the adsorbent with only 1.5 M of H2SO4, a microalgal biomass adsorption capacity of 108.9 ± 0.3 mg/g was achieved at equilibrium. This is due to 1.5 M H2SO4 acid-modified sugarcane-bagasse-based adsorbent having the lowest surface positivity value among the adsorbents tested, minimizing negative counter ion formation whilst maximizing the negatively charged microalgal cell interaction. In enhancing microalgal biomass separation from eutrophic water, the employment of a 6.7–8.0 mm adsorbent size resulted in an increase of the adsorption capacity to 143.6 ± 1.7 mg/g. Further optimizing the adsorbent dosage permitted the adsorption capacity to reach 192.9 ± 0.1 mg/g with a dosage of 0.4 g of acid-modified adsorbent in 100 mL of adsorption medium. This was equivalent to a 91.5% microalgae removal efficiency from eutrophic water. The harvested microalgal biomass also produced excellent-quality biodiesel, as manifested by the high levels of C16–C18 components (71%) in the FAME profile. The biodiesel quality was also proven by the low degree of PUFA (22.58%) and high degree of MUFA (32.55%) and SFA (44.87%). From the sustainability viewpoint, the spent acid-modified adsorbent also could be reused immediately after lipid extraction from the adsorbed microalgal biomass without the necessity to regenerate.

**Author Contributions:** Conceptualization, M.K.L. and J.W.L.; methodology, W.N.F.W.B. and H.D.; validation, C.K.C., W.D.O. and W.N.T.; formal analysis, M.S.S. and W.N.F.W.B.; resources, K.K. and E.A.K.; data curation, U.P.; writing—original draft preparation, W.N.F.W.B.; writing—review and editing, P.L.S. and J.W.L.; visualization, Y.F.Y.; supervision, J.W.L.; project administration, J.W.L. and M.K.L.; funding acquisition, J.W.L.

**Funding:** Funding from Ministry of Education Malaysia through HICoE awarded to the Centre for Biofuel and Biochemical Research, Universiti Teknologi PETRONAS.

**Acknowledgments:** The financial supports from International Grant—Universitas Islam Riau (UIR), Pekanbaru, Indonesia with the cost center 015ME0-039 and Universiti Teknologi PETRONAS via YUTP-FRG with the cost center 0153AA-E48 are gratefully acknowledged. Funding from Ministry of Education Malaysia through HICoE awarded to the Centre for Biofuel and Biochemical Research, Universiti Teknologi PETRONAS is as well duly acknowledged.

**Conflicts of Interest:** All authors declare that they have no conflict of interest.

#### **References**


© 2019 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* **Thermal Analysis of Nigerian Oil Palm Biomass with Sachet-Water Plastic Wastes for Sustainable Production of Biofuel**

**Bello Salman 1, Mei Yin Ong 1, Saifuddin Nomanbhay 1,\*, Arshad Adam Salema 2, Revathy Sankaran <sup>3</sup> and Pau Loke Show 4,\***


Received: 17 June 2019; Accepted: 8 July 2019; Published: 23 July 2019

**Abstract:** Nigeria, being the world's largest importer of diesel-powered gen-sets, is expected to invest in bio-fuels in the future. Hence, it is important to examine the thermal properties and synergy of wastes for potential downstream resource utilization. In this study, thermal conversion as a route to reduce the exploding volume of wastes from sachet-water plastic (SWP) and oil palm empty fruit bunch (OPEFB) biomass was studied. Thermogravimetric (TGA) and subsequent differential scanning calorimeter (DSC) was used for the analysis. The effect of heating rate at 20 ◦C min−<sup>1</sup> causes the increase of activation energy of the decomposition in the first-stage across all the blends (0.96 and 16.29 kJ mol−1). A similar phenomenon was seen when the heating rate was increased from 10 to 20 ◦C min−<sup>1</sup> in the second-stage of decomposition. Overall, based on this study on the synergistic effects during the process, it can be deduced that co-pyrolysis can be an effective waste for energy platform.

**Keywords:** sachet-water plastic waste; oil palm empty fruit bunch; TGA-DSC analysis; activation energy; physio-thermal analysis; co-pyrolysis

#### **1. Introduction**

Concern about the growing demand for energy, with emphasis on the developing economies, has prompted the urgent calls to implement renewable energy planning and advancement in reducing solid waste by utilizing it for energy production. The use of conventional energy sources is largely responsible for increasing CO2 emissions in the atmosphere [1]. Thus, it is regarded as the underlying cause of greenhouse gas emissions and global warming. It was reported that the African energy-related CO2 emissions are projected to increase by about 40% by the year 2030, with Nigeria contributing significantly to this growth [2].

According to the Power Africa Fact Sheet in Nigeria, only 45% of the nation has access to the national power grid [3]. This electricity mix percentage is mostly generated using 80% natural gas and 20% hydropower [4]. Another electrification method is the use of diesel-powered generator sets (gen-sets). Self-generation of electricity became necessary to avoid blackouts, especially in the densely-populated northern region of Nigeria. Furthermore, it is estimated that demand for energy in Nigeria will grow by at least 500% by 2035. However, the current trajectory of the electricity supply will increase by around just 1% in the same period [5]. The current situation of the power generation requires a new approach to impede the growing threats of CO2 emissions from gen-sets. In this context, alternative energy generation from local biomass could be one of the most viable solutions to instantly minimize the intake of fossil fuels to reduce environmental complications [6–8].

Energy from waste, mainly lignocellulosic, is advantageous due to its widely recognized social, economic, environmental and renewable properties [9]. At the same time, plastic waste poses an ecological dilemma due to its long life-time, and symbolizes an essential element of waste management [10]. According to Ben-Iwo et al. and Stoler, biomass and sachet water plastics produced from agricultural and industrial activities in Nigeria are estimated to be nearly 144 million ton y−<sup>1</sup> and 70–100 million ton y<sup>−</sup>1, respectively [11,12]. Sachet water, primarily known as 'pure water', is symbolic to the sub-Saharan region of Africa, sold in mechanically-sealed 500 mL plastic sleeves at a unit price of not more than 0.10 US \$. This water is regarded as a multibillion-dollar industry. The primary concern related to this water industry is the generation of plastic waste that continues to be one of the most significant threats to the region because of clogging gutters, causing routine flooding and exposing residents to a variety of health risks. In Nigeria, waste from plastics comprises about 65% of national solid waste streams. Therefore, in this study, the thermal and kinetic behavior of plastic waste will be examined, and the possible application of this waste for energy generation and as a platform for waste management will be explored.

Several methods, including gasification, hydrothermal liquefaction, combustion and pyrolysis, have been explored to treat the waste biomass [13–15]. Among these, pyrolysis involves the thermochemical decomposition of hydrocarbon or organic materials (usually biomass) at elevated temperatures in the absence of oxygen. Through the pyrolysis process, biomass will be converted into energetic products, such as bio-oil, syngas and bio-char. Researchers have explored a possible solution by co-pyrolysis of plastic waste with biomass [16]. The co-pyrolysis with plastics is preferred due to its ability to balance the ratios of carbon, oxygen, and hydrogen in the bio-oil derived from biomass pyrolysis [17]. In general, plastics are known to have high hydrogen compounds which make them suitable, and the potential substrate to improve the quality of the bio-oil product. Previous investigations claim apparent interactions and synergistic effects between biomass and plastics in the co-pyrolysis [18]. The inherent individual characteristics observed were due to the lower thermal stability of biomass being caused by the presence of plastics. These contradictions may have been caused by significant variations in composition, the origins of biomass, and complexity of chemical reactions during pyrolysis [19]. Thus, the need for increased research on the thermal and physio-chemical anatomization of biomass and secondary waste resources to predict the downstream product, as well as the associated environmental impacts, is necessary.

To the author's knowledge, no data is available on the pyrolysis of Nigerian oil palm empty fruit bunch (OPEFB), sachet-water plastic wastes (SWP) and their blends. In view of the large amount of OPEFB and SWP produced in Nigeria, it is necessary that a study on waste-to-value added product should be undertaken. Further, the results are expected to provide useful information for individuals and institutions who are interested in using Nigerian biomass and plastics for thermochemical conversion. Therefore, the objective of the study is to determine the thermal behavioral properties and synergy (if any) between OPEFB and SWP wastes as a potential application in renewable biofuel and biomaterials.

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

#### *2.1. Biomass Samples*

The OPEFB samples were provided by the Nigeria Institute for Oil Palm Research (NIFOR), Nigeria and SWP (Low Density Polyethylene) were acquired from Deezor Pharmaceuticals Limited, Nigeria. The samples were dried in an oven at 105 ◦C for 24 h and then ground in Fristch Pulverisette (model 19) to the size of 250 μm and further to 200 μm size in a Fristch Pulverisette (model 16).

The biomass-plastic (OPEFB and SWP) blends were prepared with different weight percentages (wt.%) of 20%, 40%, 50%, 60%, 80%, and 100%. However, for DSC analysis, the blends of 10, 20 and 30 wt.% of biomasses to plastic blends were used. The ratio of SWP to OPEFB was capped at maximum of 30% (w/w) to retain OPEFB as the major component of the sample based on results from TGA analysis of this work. To make certain of the blending homogeneity, samples were put through vortex shaker (IKA Vortex 3) for two minutes at 2500 rpm.

#### *2.2. Elemental and Proximate Analysis*

The ultimate/elemental analysis of the sample was performed using a CHNS/O analyzer (2400 model by PerkinElmer, Waltham, MA, USA), followed ASTM D5373-93 method. The volatile matter, fixed carbon and ash content of the sample (proximate analysis), however, were determined according to ASTM E 897-82 and ASTM D 1102-84 method. All experiments were conducted in triplicated and averaged values are reported in Table 1.


**Table 1.** Ultimate and proximate composition of samples.

#### *2.3. Thermal Analysis Using Thermogravimetric Analysis (TGA) and Di*ff*erential Scanning Calorimetry (DSC)*

Pyrolysis of original and blend samples (~4 mg) was carried out in a programmable TG analyzer (DSC-TGA Q Series instrument and SDT Q600 thermal analyser, manufactured by TA Instrument, New Castle, DE, USA) from room temperature to 800 ◦C and at two heating rates (10 and 20 ◦C min<sup>−</sup>1). Isothermal DSC measurements can be successfully applied for information about the heat capacity as a function of temperature during phase transitions, autoxidation, thermal decomposition and adsorption of different kinds of fuels [20]. For DSC (DSC823 manufactured by Mettler Toledo) experiments, the samples were heated from 25 ◦C to 600 ◦C at a heating rate of 15 ◦C min−<sup>1</sup> to measure the heat flow during pyrolysis. For all experiments (TGA/DSC), nitrogen was used as an inert carrier gas with a flow rate of 50 mL min−1. The instrument continuously recorded TG and DTA data which were used to analyze its thermal characteristics and to calculate the kinetic parameters. All tests were performed in triplicates to ensure reproducibility.

#### *2.4. Kinetic Reaction*

Several models have been used in kinetic analysis [21–23]. The Coats-Redfern is one of the most widely-used approach to obtain the kinetic parameters of carbonaceous materials [24]. It is a model-fitting method to determine the activation energy, pre-exponential factor and reaction order from a single measurement of thermogravimetric [25]. In general, biomass pyrolysis can be characterized using an infinite number of reaction mechanisms, described by the *n*-order law:

$$\frac{d\alpha}{dt} = k(T)f(\alpha) \tag{1}$$

$$f(a) = (1 - a)^n \tag{2}$$

where *k(T)* is the reaction rate constant, α indicates the amount of conversion or the fractional weight loss (Equation (3)), and *n* is the reaction order.

$$\alpha = \frac{m\_i - m\_o}{m\_i - m\_f} \tag{3}$$

where *mi*, *mo* and *mf* are the initial mass, the current mass at time 't' and the final mass of the sample respectively. Note that α value is always between 0 and 1.

The reaction rate constant, *k(T)* is a function of temperature, *T* with a unit of *K* and can be expressed as (Equation (4)) based on Arrhenius relationship.

$$k(T) = A e^{-\frac{k}{R\hbar}}\tag{4}$$

where *A* symbolizes the pre-exponential factor (min<sup>−</sup>1). On the other hand, *Ea* indicates the activation energy of the decomposition reaction (kJ mol<sup>−</sup>1) and *R* is the universal gas constant (8.314 J mol−<sup>1</sup> K<sup>−</sup>1). By substituting (Equation (1)) and (Equation (3)) into (Equation (4)), the kinetic equation for the sample decomposition is expressed as follow:

$$\frac{d\alpha}{dt} = A e^{-\frac{E\_q}{kT}} (1 - a)^n \tag{5}$$

For the non-isothermal case (at constant heating rate, β), however, the above equation can be further modified to:

$$\frac{d\mathbf{a}}{d\mathbf{T}} \frac{d\mathbf{T}}{d\mathbf{t}} = A e^{-\frac{E\_{\mathbf{q}}}{RT}} (1 - a)^{n} \tag{6}$$

As

$$
\beta = \frac{\text{dT}}{\text{dt}} \tag{7}
$$

Hence, the final kinetic equation in non-isothermal TG experiments is:

$$\frac{d\alpha}{d\mathcal{T}} = \frac{A}{\beta} e^{-\frac{F\_d}{RT}} \left(1 - \alpha\right)^n \tag{8}$$

According to Coats and Redfern method, Equation (8) was then rearranged, integrated and finally expressed as:

$$\ln\left\{\frac{-\ln(1-\alpha)}{T^2}\right\} = \ln\left\{\psi\left(1-\frac{2RT}{E\_d}\right)\right\} - \frac{E\_d}{RT} \tag{9}$$

where ψ = *AR* <sup>β</sup>*Ea* and the reaction is assumed to be first-order.

By assuming <sup>2</sup>*RT Ea* 1, ln ψ <sup>1</sup> <sup>−</sup> <sup>2</sup>*RT Ea* - ≈ ln(ψ) [26]. Thus, Equation (9) can be further modified to:

$$\ln\left\{\frac{-\ln(1-\alpha)}{T^2}\right\} = \ln(\psi) - \frac{E\_a}{RT} \tag{10}$$

By plotting this equation, the activation energy and the pre-exponential factor can then be determined.

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

#### *3.1. Characteristic properties of OPEFB and SWP*

In Table 1, the proximate analysis shows that the volatile matter (OPEFB: 81.4%; SWP: 99.6%) and fixed carbon (OPEFB: 18.6%; SWP: 0%) yielded high-level reactivity and volatility benefits which were appropriate for production of liquid fuels [27]. Nonetheless, the values in this study differ slightly, i.e., by around 2%, from the experimental analysis of OPEFB reported by Oyedun et al. but are comparable to those in other agricultural residues [27]. This could be attributed to the erratic nature of biomass due to species of plant being used, the age of those plants, or the climate of the plantation. Interestingly, crop origin did slightly affect the (O) element property in biomass material that is in the scope of 40 wt.% to 44 wt.%. Additionally, the ash content was at 4.6%, which was more consistent with the results from [28,29]. Therefore, this low value of ash is expected to have a positive effect on heating value since high-level ash quotas in biomass are known to lead to unfavorable yield conditions in terms of liquid by-product in fast pyrolysis [30].

The pure samples (OPEFB: SWP) contained traces of Carbon (C), 54.4: 86.93%, Oxygen (O) 36.44: 1.39% and Hydrogen (H) 7.64: 16.54% respectively. The plastic material had higher C and lower O contents, and therefore lower O/C ratio. On the other hand, the OPEFB biomass had higher sulphur contents compared to the SWP plastics. The content of sulphur (0.48%) in OPEFB is slightly higher than in the SWP (0.12%) of plastic wastes. It has been reported that the typical amount of sulphur in oil palm empty fruit bunch biomass materials is in the range of <0.1 to 0.68% [27,29]. Thus, the 0.48 wt.% value determined in this study is consistent, irrespective of the geographical location. The trace of sulphur in plastic could be because waste plastic contains some contamination, e.g., in the pigment used to impart color in the plastic material [31].

#### *3.2. Thermal Characteristics*

Figures 1 and 2 show the TG and DTG curves of OPEFB, SWP, and their blends at a heating rate of 10 and 20 ◦C min−<sup>1</sup> in a nitrogen environment. The pyrolysis process in the present study was differentiated into three main stages; (a) drying stage (30–120 ◦C), (b) pyrolysis stage (120–500 ◦C), and (c) char stage (500–800 ◦C).

The initial weight loss that occurred during the drying stage from room temperature to about 120 ◦C is due to the removal of free and bound moisture content. SWP showed almost no mass loss in the drying stage as compared to OPEFB (which suffered about 7.1 wt.% mass loss in this region) [32]. The drying stage can also be detected from the presence of the first small peak in the DTG curves (Figures 1B and 2B). A similar phenomenon was also reported by Xu et al. in their work on pine sawdust with and without polyvinylidene [19].

**Figure 1.** (**a**) TG, and (**b**) DTG curves for OPEFB and SWP samples and their blends at heating rate of 10 ◦C min<sup>−</sup>1.

**Figure 2.** (**a**) TG, and (**b**) DTG curves for OPEFB and SWP samples and their blends at heating rates of 20 ◦C min<sup>−</sup>1.

The pyrolysis stage follows immediately after the moisture was released in two stages; (a) pyrolysis of light volatization between 120 and 220 ◦C. At this stage, the chemical structure of the polymer and biomass starts to depolymerize and soften though without any loss in mass [16]. The next significant weight loss occurred during (b), the main pyrolysis in the temperature range between 220 ◦C and 500 ◦C due to the removal of heavy organic compounds. The pyrolysis of original OPEFB and SWP showed only one major peak in the central pyrolysis region. The total weight loss in this region for OPEFB and SWP was about 62 wt.% and 94 wt.%, respectively. Basically, during this stage, the biomass gets converted into volatiles which contain condensable and non-condensable gases. Original SWP plastic (100%) depicted a single peak at both heating rates (10 and 20 ◦C min<sup>−</sup>1) in this region, as evident in Figures 1A and 2A. The major weight loss (~94 wt.%) for SWP started at 456 ◦C (at 10 ◦C min−1) and 398 ◦C (at 20 ◦C min−1) and completed at 505 ◦C (at 10 ◦C min−1) and 497 ◦C (at 20 ◦C min−1). This thermal behavior of plastic was similar to that in previous studies [27,33]. The thermal decomposition of plastic is vastly different among biomass materials due to different chemical bonds. The degradation mechanisms of polymers are due to the rapid primary hemolytic scission and intra-molecular hydrogen transfer in macro-radical forming oligomers at low temperatures, i.e., around 300 ◦C [34]. This random scission and de-polymerization occur at intermediate temperatures of 500 ◦C, further b-scission, and intra-molecular hydrogen transfer to form ethylbenzene, toluene, and a-methylstyrene at ~600 ◦C [34]. In theory, lignocellulosics such as oil palm empty fruit bunch is acknowledged to contain significant amounts of cellulose, hemicellulose and lignin [35], and the devolatilization has been shown to primarily correspond to the degradation of the components occurring in this region of significant weight loss [36]. Within this temperature range, the devolatilization stages of OPEFB can be divided into the lower temperature stage between, 195 ◦C and 300 ◦C, representing hemicellulose decomposition, and the intermediate temperature stage, 270–370 ◦C representing cellulose decomposition. It was reported that the significant devolatilization of biomass is because of the thermal formation of more stable and stronger bonds which replace the weaker bonds in the biomass structure [37]. The findings of the present study are in agreement with those obtained by previous researchers [38,39].

Some interesting facts were revealed when the OPEFB biomass was blended with SWP at different weight percentage ratios. Obvious differences in the pyrolysis of the original OPEFB/SWP from blend were noticed in the central pyrolysis region. The blends showed two peaks, with the DTG peak height and positions depicting the reactivity of the materials [40]. For instance, with the increase in biomass weight percentage in the blend, the reactivity and the mass loss rate of blends decreased significantly, depicting some degree of interaction between the two samples. Thus, it seems that the decomposition of biomass is, to some extent, affected by the presence of plastic and vice versa. According to Fang et al., the pattern of the fractured surface characteristic of OPEFB may provide substantial information about the adhesion and interfacial compatibility between the OPEFB fiber and SWP during co-pyrolysis [41]. The fractured surfaces in OPEFB also implies the potential of the higher hygroscopic behavior of the material. Hence, simultaneous degradation may be observed during the co-pyrolysis process, and thus, might re-orientate the chemical and thermal features of the co-pyrolysis.

Furthermore, it is reported that plastic starts softening at around 365 ◦C, but does not decompose completely [32]. In this study, based on Figures 1A and 2A, it can be observed that there was a slight drop in the weight loss of the SWP, which indicated that within the heating range there was an effect on the heat and mass transfer process. The peaks in this stage might overlap due to the simultaneous devolatilization of both OPEFB and SWP materials, resulting in an amplified synergetic effect. Also, a gap exists between the 2nd and 3rd peaks of the DTG curves (Figures 1B and 2B), which is similar to the results obtained by Oyedun et al. [27]. Worthy of note in Figure 2B is the fact that the OPEFB (50) +SWP (50) blend and SWP (100) depicted tails in the second peak at temperatures beyond 500 ◦C. This feature is different to the pyrolysis under 10 ◦C min−<sup>1</sup> shown in Figure 1B. This might be due to the uneven pyrolysis of the constituents. In many cases, this is difficult to avoid due to the preparation procedures of the samples.

Table 2 displays the initial and final temperatures, peak temperatures and total mass loss in the main pyrolysis stage for biomasses and their blends. A shift in the pyrolysis range and peak temperature was observed due to biomass with SWP blending. A similar observation was reported in a previous work [42], and can be attributed to the thermal resistance between the reacting and evolved species which occur at higher heating rates. The initial degradation temperature in Table 2 shows the minimum temperature where the feedstock starts to decompose, which is also important because it gives information about the minimum ignition temperature required to decompose the material. For instance, the initial degradation temperature of OPEFB was 225 ◦C, as reported by Abdullah and Gerhauser [28]. However, in the present study, it was about 216 ◦C, a slightly lower than the values reported by previous authors. This could be due to the difference in the analysis method such as particle size at 250–355 μm; furthermore, a 100 mL min−<sup>1</sup> nitrogen flowrate was applied in the work of Abdullah and Gerhauser [28]. On the other hand, the initial degradation temperature for SWP was in close agreement with the study of Banat and Fares [43].


**Table 2.** Properties of active pyrolysis zone at different heating rates.

The last stage is called char or the carbonization stage, in which the carbon-rich residual is formed due to the relatively slow degradation of lignin. In this study, slow and steady weight loss could be observed at temperatures of around 500 ◦C. Typically, the degradation of lignin starts from the very

early stage, i.e., from 300 ◦C and continues until 800 ◦C [44]. In this study, the carbonization stage of the lignin component of biomass OPEFB takes place within a temperature range of approximately 375–800 ◦C. The mass loss in this stage was about 22 wt.% for OPEFB. Meanwhile, the final residue from the carbonization stage of SWP was about 5 wt.% at 500 ◦C and completely melted at 650 ◦C. The leftover char residues at the end of pyrolysis (~700 ◦C) were about OPEFB20+SWP80 (17.2 &17.7 wt.%), OPEFB40 + SWP60 (14.9 & 13.8 wt.%), OPEFB50+SWP50 (11.3 & 2.8 wt.%), OPEFB60+SWP40 (11.3 & 10.5 wt.%), OPEFB80+SWP20 (7.8 & 7.6 wt.%) at 10 and 20 ◦C min−<sup>1</sup> respectively. Among the blends, the char residues decreased in mass loss with increasing portions of OPEFB with OPEFB80+SWP20 showing the highest degradation rate. These variations further suggest a synergetic effect among the biomass and plastic mixtures in that the chemical components of the plastic are acting as catalyst [38,45]. This gives new insights into the behavior of co-adding plastics in the co-pyrolysis of biomass which have not been widely discussed in previous studies. However, the effect of the heating rate on the char residues is less pronounced, except in the case of OPEFB50+SWP50 at 20 ◦C min<sup>−</sup>1, which could be attributed to mixing errors.

#### *3.3. TGA Kinetic Analysis*

The kinetics parameters of OPEFB, SWP and blends co-pyrolysis were determined using Coat Redfern's (CR) Method. The activation energy, *Ea* was estimated from the slope of -(*Ea*/RT) by a linear fit of the experimental points. By substituting this value back into the CR Equation (Equation (10)) gives the pre-exponential factor. The peak temperature (*Tp*) of each reaction, as revealed in Figures 1 and 2, is to be expected within a small range around the local maximum of mass loss rate. Listed in Table 3 are the calculated values for the kinetic parameters, including activation energy, *Ea*, pre-exponential factor, A, and reaction rate constant at peak temperature, of all the samples at the applied heating rates of 10 and 20 ◦C. In all cases, the value of R2, (correlation coefficient) of the fitting straight line was above 0.90; this indicates that the corresponding non-isothermal model-fitting is in good agreement with the pyrolysis analysis and kinetics [46]. Furthermore, a comparison of the selected heating rates on pyrolysis of blends on solid-state kinetics data is presented. There is limited evidence in the literature of the effects of different heating rates on the kinetics of thermal decomposition to describe the devolatilization process in co-pyrolysis of biomass and plastics.

From Table 3, the thermal decomposition of OPEFB and SWP can be described by a single step reaction, while their blends showed two consecutives first and second step reactions. In view of the results of the kinetic parameters of the isolated samples tabulated in Table 3, it may be seen that a pure sample of SWP at varying heating rates (10 and 20 ◦C min<sup>−</sup>1) has the least conversion and slower reactive characteristics due to its higher (*Ea*) value of 346.93 and 234.36 kJ mol−1. Nonetheless the pre-exponential factor which expresses the probability of colliding molecules resulting in a reaction revealed 2.95 <sup>×</sup> 1023 min−<sup>1</sup> and 9.13 <sup>×</sup> <sup>10</sup><sup>15</sup> min−<sup>1</sup> respectively. It was also expected that higher heating rates reduce the complex energy required to decompose the polymer atoms. This confirms that plastics decompose at higher temperatures (480–486 ◦C). As to the isolated OPEFB sample the (*Ea*) values of 46.83 and 44.21 kJ mol−<sup>1</sup> was revealed at 10 and 20 ◦C min−<sup>1</sup> and decomposes at a much lower temperature range (227–334 ◦C), compared to isolated SWP. Thus, a similar trend could be deduced regarding the effect of heating rate on their activation energies. Nonetheless, the effect of the heating rate is more apparent in the degradation mechanism of solid-state reactions of the SWP. The kinetics parameters in Table 3 also show that the temperature and energy needed to decompose the blends is higher with a higher weight percentage of SWP or plastic. Meanwhile, Nyakuma examined decomposition of pelletized oil palm empty fruit bunch through thermogravimetric analyzer and calculated kinetic parameters in the range from 36.60 kJ mol−<sup>1</sup> to 233.92 kJ mol−<sup>1</sup> through *Popescu* method [47]. Also, the apparent (*Ea*) and (A) obtained with the CR method employed in this study were in accordance with different biomass as demonstrated by the *Ea* values for oil palm empty fruit bunch, 50.37 kJ mol−<sup>1</sup> [27] and those of other biomass species including almond and hazelnut shells (11.2–254.4 kJ mol−1) and (40.3–144.9 kJ mol−1) [48]. Similarly, in comparison with the polystyrene

parameters reported in the literature, the activation energy obtained in this study at 10 and 20 °C min−<sup>1</sup> falls within the reported ~213.78 kJ mol−<sup>1</sup> [49] to 253.69 kJ mol−<sup>1</sup> [50].


**Table 3.** Kinetics parameters of samples at different heating rates.

In the case of kinetic results of the blended mixtures generally as expected, the (*Ea*) manifested a significant increase alongside the corresponding (A) with an increasing percentage of SWP in blends. The experimental data shown in Table 3 revealed that the values obtained for (*Ea*) and (A) of the SWP/OPEFB blends are relatively different from those of the individual materials. This means that a synergetic effect is observed in the SWP/OPEFB blends that might have an overlapping degradation temperature, which creates the opportunity for free radicals from biomass pyrolysis to participate in reactions of plastic decomposition. For example, the activation energies of 80 and 60% composition of SWP were much higher than the value obtained from the pyrolysis of pure samples, and presented obvious changes with the increase of plastics weight percentage. However, the value of the activation energy and pre-exponential factor decreases when the weight percentage of SWP in the blend decreases. Similar results have been reported in previous studies [17,33]; however, interestingly, the addition of SWP in the OPEFB biomass revealed a negligible effect on the activation energy at the first reaction order of decomposition.

However, at the second stage, an appreciable increase in action energy was revealed. At higher compositions of SWP 80% *Ea* was 40.08 kJ mol−<sup>1</sup> and 240.26 kJ mol−<sup>1</sup> at heating rate of 10 ◦C min−<sup>1</sup> in the first and second stages, respectively. A similar trend was observed for the heating rate of 20 ◦C min<sup>−</sup>1, where the activation energy of 60% SWP composition was 61.75 kJ mol−<sup>1</sup> and 195.99 kJ mol−<sup>1</sup> in the first and second stages, correspondingly. On the other hand, the kinetics parameters in Table 3, remarkably show that less energy is required to decompose the blends (below 50%), depicting a favorable synergistic effect between lower mass percentages blends of SWP lowering the activation energy. This technically supports the discrepancy that in co-pyrolysis, the ratio of feed is the most significant variable in product yield and economics [16,51].

Overall, a higher heating rate (20 ◦C min<sup>−</sup>1) will lead to (average 10%) an increase in the activation energy of the decomposition in the first stage across all the blends in the range of 0.96 kJ mol−<sup>1</sup> and 16.29 kJ mol−1. Remarkably, varying the heating rate from 10 to 20 ◦C min−<sup>1</sup> in the second stage of decomposition reaction shows, as in the previous case, an increase in activation energy, with one exception at OPEFB50:SWP50, while with a 50% SWP in blend, a higher heating rate of 20 ◦C min−<sup>1</sup> favors a decline in the activation energy from 183.00 to 112.16 kJ mol−1. This can be explained by the aforementioned phenomenon whereby the increase in the heating rate on isolated OPEFB and SWP result in decreasing activation energy. In this case, the homogeneous blend of equal samples followed the same reaction pathway correspondingly result in much lower activation energy. It can also be concluded that a significant difference can be seen in the distribution of the thermal and elemental composition.

It can also be concluded that significant differences can be seen in the distribution of the thermal and elemental composition. Furthermore, the SWP can be identified as low density polyethylene (LDPE) from the elemental composition and the pattern of thermal degradation [52]. The SWP presented a heat release rate curve typical for intermediately thin non-charring materials of LDPE family [53]. The burning time was very short, with a steady increase in heat release rate after ignition up to the peak heat release rate at the end of combustion. Furthermore, the heat release rate pattern showed hardly any shoulder, with the pyrolysis involving nearly all the LDPE material at once due to the complete melting of the polymer [54], this interpretation was confirmed by visual observation of the thermogravimetric curves in Figures 1 and 2.

#### *3.4. DSC Analysis*

Based on the findings from the TGA analysis, lower portions of SWP in blends do not lead to a significant increase in the energy required to decompose the co-pyrolysis system. Therefore, this section examines further the enthalpy of reaction at 15 ◦C of lower blends (10, 20 and 30 wt.%) of SWP. Differential scanning calorimetry (DSC) is widely employed to assess the heat evolution and spontaneous combustion of carbonaceous materials, due to its sensitive analyses at relative temperatures [55]. The DSC heat evolution curve stipulates quantitative data of the heat flow, enabling the kinetics analysis during thermal heating. There are several DSC based methods for the kinetic analysis including ASTM methods (Kissinger– Akahira–Sunose and Ozawa–Flynn–Wall) [25] and the Roger & Morris method [56]. The ASTM method relies on measuring the exothermic peak temperature at various heating rate, thereby resulting in the kinetics reflecting the main region of combustion. The profiles of the heat evolution rate of samples (OPEFB-100%, OPEFB-90%: SWP-10%, OPEFB-80%: SWP-20%, OPEFB-70%: SWP-30% and SWP-100%), thermally decomposed from 25 ◦C to 600 ◦C at a fixed heating rate of 15 ◦C/min, are illustrated in Figure 3. The single heating rate DSC test at 15 ◦C min−<sup>1</sup> is the best for the application of ASTM approach, which has the advantage of reducing the required time, compared other approaches [55]. DSC analysis of pure OPEFB and SWP samples registered four and two exotherms, respectively, while each of the blended samples registered three stages. The exothermic effect differs from one material to another depending on the chemical composition and decomposition behavior of the material. The maximum temperature of the exotherms in the first, second, third and fourth regions were seen at 87–122 ◦C, 240–251 ◦C, 339–340 ◦C, and 465–491 ◦C, respectively. The first region corresponds to moisture release region, and it was found between 42 ◦C and 131 ◦C for pure OPEFB sample and between 95 ◦C and 128 ◦C for pure SWP sample. The second thermal zone was determined to be between 147 ◦C and 292 ◦C for OPEFB and between 152 and 332 ◦C for sample comprising 70% OPEFB and 30% SWP blends. The third thermal zone was determined to be between 315 and 353 ◦C for pure OPEFB, 287 and 355 ◦C, 295 and 354 ◦C for OPEFB: SWP blended ratios 90:10 and 80:20, respectively. This reflects the heat evolution of the representative samples, in which similar pattern was indicated by the TG and DTG curves in Figures 1 and 2. The melting of the various crystals led to several relatively broad, endothermic and exothermic peaks. The shape of the peak mirrors the size and weight distribution of the crystals and are among the characteristics of a

material. The peak region marked on the experimental thermogram corresponds to the region of some physical or physicochemical processes and is dependent on internal factors, such as the structural nature of the material, quantity and thermal conductivity of the sample and external factors such as crucible shape and material, heating rate, the position of the thermocouples [57].

**Figure 3.** DSC evolutions of sample species evaluated (15 ◦C min<sup>−</sup>1).

The enthalpy of a material has a direct correlation with the material's heating value, i.e., the ratio of the enthalpy of complete combustion to its mass [58]. It is an important thermal parameter in the reactor design and predicting efficiency of bioenergy applications. The temperatures and internal energies (enthalpy) corresponding to the beginning, maximum, and the end of the exothermic reactions are summarized in Table 4. The energy of a phase transition was calculated directly from the thermogram of the investigated material from the experimental function according to the kinetics.

Table 4 displays the heat released (J g−1) through the reaction progress; it is not surprising that with increasing the SWP ratios, a shift towards lower H values was observed during the first stage decomposition. Also, isolated SWP sample exhibited lower values of heat released (117.39 J g−1) as compared to the OPEFB (147.69 J g−1) at the first stage of decomposition. It is no surprise that a high amount of heat was recorded during this stage. For the combustion reaction to proceed, energy is needed to overcome the tightly bonded cellulose and hemicellulose to lignin. The results were in agreement to some extent with previous studies [27,59]. Then, the (H) value for isolated OPEFB declined through the four decomposition stages to 18 J g−1. Typically, it is expected for the (*Ea*) and (H) energies to progressively decrease due to the biomass composition mechanism as each of the components can decompose by parallel exothermic processes. Meanwhile, there was a significant further increase in the (H) of isolated SWP from 117.39 to 1030.19 J g−<sup>1</sup> in its first and forth stages of decomposition, which is generally attributed to β-scissors reaction and in agreement with the TG studies. The net enthalpy of pyrolysis for the blended samples studied in this work were 398.18 J g−<sup>1</sup> (OPEFB, 90: SWP 10); 365.32 J g−<sup>1</sup> (OPEFB, 80: SWP, 20); 317.91 J g−<sup>1</sup> (OPEFB, 70: SWP, 30). This can clearly be seen to be higher than that of pure biomass (276.82 J g<sup>−</sup>1) and much lower than for pure SWP plastic (1147.58 J g−1). The reason for this was that the heat flow of the blended samples at the final stage of decomposition were zero, which would indicate that no degradation reactions had taken place. However, through a comparison of the devolatization stage with the findings of TGA results, it could be observed that enthalpy (40.72) for OPEFB100% was similar to those of TGA. But a significant gap was observed for the case of SWP100%. Nonetheless, blends showed proof of a synergistic effect, and

support the TGA findings in which a lower percentage weight of SWP in the co-pyrolysis mixture does not significantly boost energy required for the decomposition in both of the first stages of delovatization (84.42 J g−<sup>1</sup> for OPEFB90:SWP10, 59.83 J g−<sup>1</sup> for OPEFB80:SWP20, and 88.97 J g−<sup>1</sup> for OPEFB70:SWP30) and second stage decomposition (158.21 J g−<sup>1</sup> for OPEFB90:SWP10, 161.04 J g−<sup>1</sup> for OPEFB80:SWP20, and 145.4 J g−<sup>1</sup> for OPEFB70:SWP30).


**Table 4.** DSC results of the samples.

#### **4. Conclusions**

This study examined the thermal decomposition behavior and kinetics of Nigerian native oil palm empty fruit bunch, sachet-water plastic wastes and their blends under non-isothermal inert conditions at different heating rates. The findings show that the co-pyrolysis of OPEFB biomass and SWP plastic mixture exhibits a diverse pyrolysis reactivities at different temperatures. The co-pyrolysis of OPEFB and SWP clearly showed synergic effects due to the difference in thermal behavior and kinetics parameters. In short, the results show that co-pyrolysis can be an effective method to dispose of wastes (biomass and sachet water, SWP) and convert them into useful energy. The experimental data obtained with a model-free method would be helpful in the design and development of energy systems for sustainable waste utilization for energy.

**Author Contributions:** Conceptualization, S.N. and B.S.; methodology, B.S. and A.A.S.; validation and investigation, B.S. and M.Y.O.; resources, B.S.; writing—original draft preparation, B.S. and M.Y.O.; writing—review and editing, S.N., R.S., and P.L.S.; supervision, S.N. and A.A.S.; funding acquisition, S.N.

**Funding:** This work was funded by TNB Seed Fund (code: U-TR-RD-18-11). A note of appreciation to iRMC UNITEN for the financial support through publication fund BOLD 2025 (RJO10436494).

**Acknowledgments:** The authors would also like to acknowledge Monash University Malaysia and Universiti Tenaga Nasional (UNITEN) for the facilities support.

**Conflicts of Interest:** The authors declare no conflict of interest. Besides, the funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


© 2019 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* **Fabrication of Green Superhydrophobic**/**Superoleophilic Wood Flour for E**ffi**cient Oil Separation from Water**

#### **Xuefei Tan 1,2, Deli Zang 3, Haiqun Qi 1, Feng Liu 3, Guoliang Cao <sup>4</sup> and Shih-Hsin Ho 4,\***


Received: 15 May 2019; Accepted: 19 June 2019; Published: 2 July 2019

**Abstract:** The removal of oil from waste water is gaining increasing attention. In this study, a novel synthesis method of green superhydrophobic/superoleophilic wood flour is proposed using the deposition of nano–zinc oxide (nZnO) aggregated on the fiber surface and the subsequent hydrophobic modification of octadecanoic acid. The as-prepared wood flour displayed great superhydrophobicity and synchronous superoleophilicity properties with the water contact angle (WCA) of 156◦ and oil contact angle (OCA) of 0◦ for diesel oil. Furthermore, the as-prepared wood flour possessed an excellent stability, probably due to the strong adhesion of nZnO, which aggregates to the fiber surface of wood flour with the action of glutinous polystyrene. The maximum adsorption capacity of as-prepared wood flour was 20.81 g/g for engine oil, which showed that the as-prepared wood flour is a potential candidate as an efficient oil adsorbent in the field of water-oil separation. Moreover, it has good chemical steadiness and environmental durability. Taken together, all the information acquired from this research could be valuable in evaluating the potential of as-prepared wood flour as a competitive and sustainable oil-water separation material.

**Keywords:** wood flour; oil adsorption; superhydrophobic; superoleophilic; oil-water separation; sustainable material

#### **1. Introduction**

With the rapid development of modern industry, the growing crisis of global water pollution is severely affecting the environment [1–5]. A representative case is the explosion at British Petroleum's Deepwater Horizon oil rig in 2010, which resulted in the loss of life and property, as well as the spillage of huge amounts of oil into the ocean [6]. To date, various materials and approaches have been adopted to remove spilled oil from water bodies, including physical diffusion [7], activated carbon [8], oil containment booms [9], exfoliated graphite [10], waste barley straw [11,12] and membranes [13,14]. Nevertheless, these traditional techniques have certain deficiencies, such as being time-consuming, economically infeasible, environmentally damaging and non-renewable. Therefore, the exploration and use of novel green materials, which can effectively separate oil contaminants from water, is highly desirable. This is not only important for environmental protection, but also for sustainable urban development.

In recent years, a number of self-cleaning plant surfaces have drawn significant research attention. The superhydrophobic nature of lotus leaves is a good example of this [15–18]. In general, the water contact angle is a critical index used to characterize the wettability of a surface. Due to their remarkable

waterproofing performance, superhydrophobic materials have been prepared through various approaches, involving a sol-gel process [19], vapor phase deposition [20], chemical etching [21,22], surface fluorination [23] and electrospinning [24]. To date, a number of advanced materials with superhydrophobic and superoleophilic properties have been synthesized and applied to the disposal of spilled oil [25–29]. For example, Wang et al. [30] developed a facile electrochemical deposition method to prepare a novel functional micro-nano hierarchical structured copper mesh film with special superhydrophobic and superoleophilic characteristics for the effective removal of oil from water. Zhang and Seeger [31] successfully synthesized superhydrophobic and superoleophilic polyester textiles using silicone nanofilaments using one-step growth, which could be used for oil/water separation. Cortese et al. [32] described a convenient approach to fabricate cotton textiles with superhydrophobic and superoleophilic properties using plasma-enhanced chemical vapor deposition. Yue et al. [33] developed a kind of superhydrophobic cellulose/LDH (layered double hydroxide) membrane and applied it in an open oil/water two-phase system. Li et al. [34] successfully synthesized superhydrophobic/superoleophilic cotton fabrics combined with polyvinylsilsesquioxanes polymer and nano–Al2O3 particles.

Wood flour is one of the most common forms of agricultural waste, and when burnt, can cause severe environmental problems, such as dust or air pollution. There is an urgent need to reuse waste wood flour in a sustainable way, which could alleviate these problems. Wood flour has the advantages of having a low density, being bio-degradable, eco-friendly and low-cost [35]. However, the synthesis of superhydrophobic and superoleophilic materials using waste wood flour has received little research attention. In this work, wood flour was evenly coated with a layer of octadecanoic acid-modified nZnO aggregates with the assistance of glutinous polystyrene, which enabled it to acquire an excellent performance with regards to both superhydrophobicity and superoleophilicity. The superhydrophobic/superoleophilic wood flour with a WCA of 156◦ and OCA of 0◦ was prepared for an efficient oil removal from contaminated wastewater. Regardless of the oil or organic solvent, the maximum oil adsorption capacity of the as-prepared wood flour was almost three-fold higher than the pristine wood flour. The adsorption efficiency ranged from 98% to 100%. After the adsorbed oils were removed using acetone, the superhydrophobic/superoleophilic wood flour could be reused multiple times with a good chemical stability and environmental durability.

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

#### *2.1. Materials and Chemicals*

The wood flour used in this study was obtained from the Larch species (*Larix gmelini*), provided by the Carpentry Laboratory of Northeast Forestry University, China. Hydrogen peroxide, sodium hydroxide, hydrochloric acid, octadecanoic acid and anhydrous ethanol were provided by Tianjin Kaitong Chemical Reagent Co., Ltd., Tianjin, China. Zinc nitrate was purchased from Tianjin Hengxing Chemical Reagent Manufacturing Co., Ltd., Tianjin, China. Polystyrene (Mw = 101,900) was purchased from Shanghai xibao Biological Technology Co., Ltd., Shanghai, China. Methylbenzene was obtained from Tianli Chemical Co., Ltd., Guangdong, China. The gasoline, crude oil, diesel, engine oil, chloroform, n–hexane and toluene were purchased from Tianjin Hengxing Chemical Reagent Manufacturing Co., Ltd., Tianjin, China. All chemical reagents were of analytical grade, and used without further purification.

#### *2.2. Pretreatment of Wood Flour*

Raw wood flour was first sifted through a filter screen with a uniform grading smaller than 74 μm. Then, 1.5 g wood flour was dipped into 7 mL hydrogen peroxide (30%) and 200 mL sodium hydroxide aqueous solution (0.5 wt.%) at room temperature under constant stirring for 14 h. After that, hydrochloric acid (6 mol/L) was added to adjust the pH within the range of 6.5–7.0. Finally, the wood

flour was washed several times using ultrapure water and dried at 50 ◦C for 3 h until the weight became constant.

#### *2.3. Preparation of nZnO Particles*

2.4 g sodium hydroxide and 200 mL ultrapure water were added to a 250 mL round-bottom flask in a water bath with constant magnetic stirring at 70 ◦C for 10 min. Then, 2.0 g of zinc nitrate was added to the sodium hydroxide solution and continuously stirred for 8 h. The solution was placed at room temperature for some time. Next, the non-reacted reagents or by-products were removed using ultrapure water and anhydrous ethanol. The final solution was dried for 5 h in a vacuum oven at 60 ◦C. Finally, the particles of nZnO were obtained as white powder.

#### *2.4. Fabrication of the Green Superhydrophobic*/*Superoleophilic Wood Flour*

0.1 g wood flour and 0.1 g nZnO particles were mixed and immersed in a beaker containing 10 mL ethanolic octadecanoic acid solution (5%, m/v). The mixture was maintained at 70 ◦C for 5 h, washed with anhydrous ethanol and dried in an oven at 60 ◦C for 4 h to obtain octadecanoic acid modified-nZnO nanoparticles coated on the wood flour surface. Afterwards, the modified-nZnO wood flour composites were mixed with 10 mL (2%, m/v) of polystyrene methylbenzene solution. Finally, the superhydrophobic/superoleophilic wood flour was dried at 50 ◦C for 2 h until the weight became constant.

The mechanism of the interaction of nanoparticles, coated by octadecanoic acid with wood flour fibers, was as follows. Hydroxyl groups on the surfaces of ZnO particles interacted with the hydroxyl groups on the surfaces of wood flour fibers through the formation of hydrogen bonds, thus making the ZnO particles homogeneously cover the wood flour surface. Octadecanoic acid chemically reacted with ZnO to generate modified-ZnO particles, as shown in Figure 1b,c. Polystyrene reagent served as a binder to cover the surfaces of the modified-ZnO particles and wood flour fibers, which resulted in a firm adhesion of nanoparticles onto the fiber surface. Figure 1a shows the generation of ZnO aggregates on the sample surface. In fact, the ZnO nanoparticles on the surface of the final product existed as ZnO aggregates.

**Figure 1.** (**a**) Synthesis route for the preparation of superhydrophobic/superoleophilic wood flour. (**b**) Graphical representation for the modification of the ZnO particle with octadecanoic acid. (**c**) Chemical structure of octadecanoic acid.

#### *2.5. Separation of Oil*/*Water Mixtures*

0.5 g of as-prepared wood flour was added to a beaker containing a 150 mL mixture of water and diesel oil, which was dyed red using Sudan III in order to easily and clearly observe the phenomenon. After the adsorption, the red wood flour was recovered on the surface of the water.

The oil adsorption capability was defined using Equation (1):

$$Q = \left(m2 - m1\right) / m1\tag{1}$$

where *Q* is the oil adsorption capability (g/g); *m*2 is the weight of the wood flour after adsorption (g), and *m*1 is the initial weight of the wood flour before adsorption (g).

#### *2.6. Characterization of the Green Superhydrophobic*/*Superoleophilic Wood Flour*

#### 2.6.1. The Surface Morphologies of Wood Flour Sample

Scanning electron microscopy (SEM, FEI QUANTA200, Hillsborough, Oregon, USA) was used to examine the surface morphologies of the pristine and superhydrophobic/superoleophilic wood flours under the condition that all specimens were pre-coated with a layer of gold.

#### 2.6.2. The Chemical Compositions of the Wood Flour Samples

The chemical compositions of the wood flour samples were investigated using Fourier transform infrared spectroscopy (FT-IR, Magna-IR 560, Nicolet, Madison, Wisconsin, USA), X-ray photoelectron spectrometry (XPS, PHI Thermo Fisher Scientific, Waltham, MA, USA) and energy-dispersive X-ray analysis (EDX, Quantax 70, Billerica, MA, USA). The FT-IR spectra of the wood flour samples were obtained by direct transmittance using the potassium bromide (KBr) pellet technique. For each sample, the wavenumber was measured within the range of 500–3400 cm<sup>−</sup>1, and the spectrum was accumulated from a total of 32 co-added scans at a spectral resolution of 4 cm−1. The preparation of the samples for the FT-IR measurement was performed by fully grinding the mixture of 2 mg wood flour and 200 mg spectroscopic grade KBr powder, which were pressed into a pellet with a diameter of 15 mm. Before analysis, the background spectrum of pure KBr was recorded. For the infrared spectroscopic analysis, the pellets were analyzed directly.

#### 2.6.3. The Water/Oil Contact Angle

A contact angle instrument (CA-A, Hitachi, Tokyo, Japan) was used to measure the water contact angle (WCA) and oil contact angle (OCA). For these measurements, 5 μL of ultrapure water or an oil droplet was dropped at five different locations under ambient conditions. The values of WCA and OCA were obtained as averages of five repeated measurements.

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

#### *3.1. Morphology of the Green Superhydrophobic*/*Superoleophilic Wood Flour*

It has been widely reported that the superhydrophobic property of certain materials is primarily due to their dualistic micro/nano surface structures [36]. As a consequence, it is vital to characterize the surface morphologies of samples before and after the treatments. The morphologies of pristine and superhydrophobic/superoleophilic wood flours were characterized using SEM at different magnifications, as shown in Figure 2. According to Figure 2a–d, after the treatment with octadecanoic acid modified-nZnO particles and polystyrene, the integrity of the fiber-like structure of the wood flour survived, which means that the structures of fibers were the same for both the pristine and as-prepared wood flours. In addition, compared with the pristine wood flour, the as-prepared superhydrophobic/superoleophilic wood flour possessed a rougher surface because of the deposition of nZnO aggregates, which had an average diameter of approximately 80 nm on

the surface of the fibers (Figure 2b,d). A hierarchical structure is thus presented on the surface of the superhydrophobic/superoleophilic wood flour, which included the micron fibers and abundant nanoparticles on each fiber. Due to this structure, the as-prepared wood flour achieved enough roughness to bring about the co-existing features of superhydrophobicity and superoleophilicity. The results showed that the presence of nZnO aggregates played a prominent role in the preparation of the superhydrophobic/superoleophilic wood flour. On this basis, and in combination with the modification of an octadecanoic acid (low-surface energy material), a large amount of air could be trapped in the cavities and interspaces in the as-obtained wood flour surface. Then, as soon as a water droplet was dripped onto the surface of a sample, it would come into contact with the trapped air and bounce off without any residue. The superhydrophobicity of wood flour was due to the combination of trapped air and low-surface energy material, which prevented the adsorption of water at the interface.

**Figure 2.** SEM images of pristine wood flour. (**a**,**b**) without nZnO aggregates and superhydrophobic/superoleophilic wood flour, and (**c**,**d**) with nZnO aggregates at low and high magnifications.

#### *3.2. Surface Wettability of the Green Superhydrophobic*/*Superoleophilic Wood Flour*

In order to verify the superhydrophobicity and superoleophilicity of the as-obtained wood flour sample, its surface wettability was investigated by measuring the values of the water/oil contact angles using a contact angle device at room temperature. As shown in Figure 3a, the water contact angle of pristine wood flour was almost 0◦, indicating that abundant hydroxyl groups were present on the fiber surfaces. With regards to the sample treated with octadecanoic acid modified-ZnO, the water contact angle reached 120◦ (Figure 3c) and therefore had a hydrophobic surface. Furthermore, after coating with octadecanoic acid modified-nZnO and polystyrene (Figure 3b), the water droplet on the as-prepared wood flour became spherical, and the water contact angle became as high as 156◦. Additionally, the scattering of the contact angle along the sample was 5◦, showing that the as-prepared wood flour had excellent superhydrophobic properties. In addition, the oil wettability of the as-prepared wood flour surface was also examined using oil contact angle measurements. When the oil droplets fell onto the surface of the resulting product, all the oil droplets (diesel oil, gasoline and kerosene) were immediately adsorbed onto the treated wood flour at the interface (see Figure 3d), indicating that the oil contact angles of the as-prepared wood flour surface were 0◦. A hierarchical structure is presented on the surface of the superhydrophobic/superoleophilic wood flour, including the micron fibers and abundant nanoparticles on each fiber. Due to this structure, the as-prepared wood flour achieved enough roughness to bring about the co-existing features of superoleophilicity. For the same oil or organic solvent, the maximum oil adsorption capacity of the as-prepared wood flour was almost three times that of the pristine wood flour, indicating the enhancement of superoleophilicity. In short, the dual basic properties of the as-prepared wood flour with superhydrophobicity and superoleophilicity were clearly confirmed.

**Figure 3.** Images of a liquid droplet on different surfaces: typical photograph of a 5 μL water droplet on the surfaces of (**a**) pristine wood flour and (**b**) superhydrophobic/superoleophilic wood flour; (**c**) a water droplet on a wood flour surface treated with octadecanoic acid; (**d**) an oil droplet on a superhydrophobic/superoleophilic wood flour surface.

On the basis of the theoretical principle of surface wettability, the wettability of a material is a synergistic effect of the chemical composition and surface morphology [37]. The combination of the deposition of numerous nZnO aggregates and the surface decoration through low surface energy octadecanoic acid was required to produce the features of superhydrophobicity and superoleophilicity of wood flour. As a result, when a water droplet was dropped onto the as-prepared wood flour, it bounced off, leaving almost no trace of water. Overall, a novel wood flour with a very good superhydrophobic and superoleophilic performance was successfully obtained in this study.

#### *3.3. Chemical Composition Analysis of the Green Superhydrophobic*/*Superoleophilic Wood Flour*

In order to prove the generation of octadecanoic acid modified-nZnO and polystyrene molecules on the as-prepared wood flour's surface, FT-IR, XPS and EDX were employed to investigate the chemical components of as-prepared wood flour.

The typical FT-IR spectra of the wood flour, coated with octadecanoic acid modified-nZnO and the as-prepared wood flour, are presented in Figure 4. In both the spectra, the absorption peaks at 2954 cm−<sup>1</sup> and 1398 cm−<sup>1</sup> stemmed from asymmetrical stretching vibrations and symmetrical bending vibrations of –CH3, respectively. In addition, the peaks at 2916 cm<sup>−</sup><sup>1</sup> and 2846 cm−<sup>1</sup> were attributed to asymmetrical stretching vibrations and symmetrical stretching vibrations of –CH2, respectively. All four characteristic peaks confirmed that long alkyl-chains existed on the surface of the wood flour coated with modified-nZnO. The absorption peaks at 1537 cm−<sup>1</sup> were due to –COOH– stretching vibrations, whereas those at 1464 cm−<sup>1</sup> were due to –COOH– bending vibrations and were induced by the CH3(CH2)16COO− groups. The presence of octadecanoic acid on both samples' surfaces was evidenced by the characteristic peaks present in both the spectra. In the high frequency region of Figure 4b, the bands at 3024 cm−<sup>1</sup> and 3060 cm−<sup>1</sup> were assigned to the C–H stretching vibrations of benzene ring groups, which were introduced by polystyrene. In the low frequency region of Figure 4b, two typical bands at 696 cm−<sup>1</sup> and 746 cm−<sup>1</sup> were due to the C–H bending vibrations of benzene ring groups of polystyrene. Moreover, compared to Figure 4a, in accordance with the characteristic absorption peaks of the benzene rings from polystyrene, the bands at 1599 cm−<sup>1</sup> and 1452 cm−<sup>1</sup> were visible in Figure 4b. Based on these observations, it was confirmed that nZnO was successfully functionalized by octadecanoic acid, and that polystyrene was successfully used as an adhesive agent to stick octadecanoic acid-modified nZnO to the surface of the as-prepared wood flour.

**Figure 4.** FT-IR spectra of (**a**) wood flour coated with octadecanoic acid modified-nZnO and (**b**) as-prepared superhydrophobic/superoleophilic wood flour.

The XPS spectra of the pristine and superhydrophobic/superoleophilic wood flours are shown in Figure 5. For the pristine wood flour (Figure 5a), peaks corresponding to C 1 s and O 1 s were observed. In comparison, the XPS spectra of the as-prepared wood flour contained a new Zn 2p peak, which accounted for the generation of ZnO. The oxygen content would change when nZnO was introduced into the surface of the wood flour. Based on the XPS spectra, the oxygen intensity in Figure 5a was higher than that in Figure 5b. Therefore, the changes in the C/O proportion were calculated based on the changes in the C and O intensities. In addition, the accurate C/O proportion changed from 65/35% to 83/13%, which resulted in an increase in the C content. In summary, it is inferred that the successful grafting of ZnO with octadecanoic acid was carried out in this study.

**Figure 5.** XPS spectra of the (**a**) pristine wood flour and (**b**) as-prepared wood flour.

In addition to the characterizations of FT-IR and XPS, the elemental composition of the as-prepared wood flour was analyzed using EDX, and the results are shown in Figure 6. The carbon (C) peak and oxygen (O) peak observed in both the spectra are attributed to the sustainable wooden material. Notably, in comparison with the pristine wood flour, there was a new peak of zinc (Zn) in the as-prepared material, which provided clear evidence of the existence of nZnO on the surface of the superhydrophobic/superoleophilic wood flour.

**Figure 6.** EDX spectra of the (**a**) pristine wood flour and (**b**) as-prepared superhydrophobic/superoleophilic wood flour.

#### *3.4. Steadiness and Durability Analysis of the Proposed Superhydrophobic*/*Superoleophilic Wood Flour*

In order to enhance the economic feasibility of the proposed superhydrophobic/superoleophilic wood flour, its environmental durability and chemical steadiness were analyzed. The chemical steadiness was assessed by recording the changes in the contact angles of corrosion solutions on the as-prepared wood flour (see Figure 7a). The as-prepared wood flour reacted with the aqueous solutions at various pH values ranging from 0 to 14 at room temperature for 24 h, and the changes in its contact angles were detected. From Figure 7a, the water contact angles on the sample surface changed within a very narrow range and remained higher than 150◦ (152–156◦). Meanwhile, the oil contact angle remained at 0◦. As such, the great superhydrophobicity and superoleophilicity of the as-prepared wood flour was retained in the corrosive solutions. Moreover, the environmental durability was examined under ambient conditions over 150 days. During the storage period, there were no apparent changes in the values of the contact angles of the as-prepared wood flour, showing that the superhydrophobic/superoleophilic wood flour possessed a remarkable environmental durability (Figure 7b). It is therefore concluded that the superhydrophobic/superoleophilic wood flour prepared in this study has a favorable chemical steadiness and environmental durability, which are due to the viscosity of the polystyrene present on the surface of the as-prepared wood flour fibers.

**Figure 7.** (**a**) The contact angles of the superhydrophobic/superoleophilic wood flour with different pHs of aqueous solution; (**b**) the contact angles of the superhydrophobic/superoleophilic wood flour for the storage evaluation in an ambient environment.

#### *3.5. Application of the Green Superhydrophobic*/*Superoleophilic Wood Flour in Water-Oil Separation*

In view of its dramatic chemical steadiness and environmental durability, the as-prepared wood flour has a great potential for application to oil-containing wastewaters. Figure 8 illustrates the procedure for water-oil separation using the superhydrophobic/superoleophilic wood flour as an oil sorbent. A certain amount of as-prepared wood flour was placed in the mixture of water and diesel oil, which was dyed red using Sudan III for ease of observations. As a result, the superhydrophobic/superoleophilic wood flour could effectively adsorb the diesel oil, while completely repelling water. The red diesel oil was fully adsorbed by the as-prepared wood flour within a few seconds. The transparent region on the water surface indicated that the diesel oil was effectively separated from water. After the adsorption, the red wood flour was recovered from the surface of the water.

**Figure 8.** Photographs of the procedure for water-oil separation using the superhydrophobic/superoleophilic wood flour as an oil sorbent: (**a**) water and diesel oil mixture (diesel oil was dyed with Sudan III); (**b**) right after the addition of the as-prepared wood flour; (**c**) after a few seconds, the wood flour filled with red diesel oil floated on the surface of the water; and (**d**) the red oil-adsorbing wood flour recycled from the surface of the water.

The adsorption capacities of the pristine and as-prepared wood flours for oils and organic solvents are shown in Figure 9a.

**Figure 9.** (**a**) The maximum adsorption capacities of the pristine and superhydrophobic/superoleophilic wood flours for various oils and organic solvents at room temperature. (**b**) The variation in adsorption efficiency (%) for diesel oil and crude oil with various mass ratios of water-to-oil. (**c**) The reusability of the superhydrophobic/superoleophilic wood flour for diesel oil and engine oil applications.

For the same oil or organic solvent, the maximum oil adsorption capacity of the as-prepared wood flour was almost three times that of the pristine wood flour, which may be due to the pretreatment of the wood flour and the composites' coating on the surface of the as-prepared wood flour. As such, the oil sorption capacity of the superhydrophobic/superoleophilic wood flour was greatly enhanced. In addition, the oil adsorption efficiency of the as-prepared superhydrophobic/superoleophilic wood flour was examined. The variation in the adsorption efficiency (%) for diesel oil and crude oil for various mass ratios of water-to-oil is shown in Figure 9b. The adsorption efficiency ranged from 98% to 100%, which was mainly potentially due to the wastage of oil with a high viscosity decreasing with the increase in the proportion of water.

The reusability of the proposed material is an important issue with regards to its practical applications. After the adsorbed oils were removed using acetone, the superhydrophobic/superoleophilic wood flour could be reused for many cycles. As seen in Figure 9c, the sorption capacity of the reused wood flour for both diesel and crude oil decreased to 83% and 77% for the second and third cycles, respectively, which was probably due to the trace residual oils left in the fibers of the as-prepared wood flour. The oil adsorption capacity of the superhydrophobic/superoleophilic wood flour remained almost unchanged after three cycles, exhibiting a good reusability. Moreover, after several cycles, the as-prepared wood flour can still be used as an environment-friendly and low-cost material with satisfactory removal efficiencies for diesel and crude oils. Finally, a comparison

of the oil adsorption capacities between the proposed superhydrophobic/superoleophilic wood flour and the bio-materials reported in the literature is presented in Table 1. It can be seen that the green superhydrophobic/superoleophilic wood flour had a relatively high oil adsorption capacity, demonstrating that the as-prepared wood flour has a great potential for being used as an oil adsorbent due to its excellent superhydrophobic/superoleophilic characteristics.



#### **4. Conclusions**

In this work, green superhydrophobic/superoleophilic wood flour with a WCA of 156◦ and OCA of 0◦ was prepared for an efficient oil removal from contaminated wastewater. The water-resistance and oil-adsorption properties of the as-prepared superhydrophobic/superoleophilic wood flour were comprehensively ascribed to the synergistic effect of octadecanoic acid modified-nZnO on the micron-fiber surface and glutinous polystyrene employed to attach the modified-nZnO aggregates to the surface of fibers. In addition, the as-prepared wood flour could effectively adsorb oil, while completely repelling water, demonstrating that it has a great potential for use as an effective oil adsorbent from wastewater. For the same oil or organic solvent, the maximum oil adsorption efficiency of the proposed wood flour was almost three times that of the pristine wood flour. The adsorption efficiency ranged from 98% to 100% for the proposed wood flour. After the adsorbed oils were removed using acetone, the green superhydrophobic/superoleophilic wood flour could be reused many times. Notably, the superior chemical steadiness and environmental durability of the proposed wood flour with environment-friendly characteristics add to its commercial feasibility.

**Author Contributions:** S.-H.H. conceived designed the experiments. D.Z. performed the experiments. X.T. performed the experiments and wrote the paper. H.Q. and G.C. analyzed the data. F.L. contributed in materials preparation and their surface analysis. All the experiments were performed under the supervision of S.-H.H. All authors read and approved the final manuscript.

**Funding:** This research was funded by Heilongjiang Institute of Technology Doctoral Research Fund (2017BJ31).

**Conflicts of Interest:** Declare conflicts of interest or state "The authors declare no conflict of interest."

#### **References**


© 2019 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*
