Exhaustive valorization of cashew nut shell waste as a potential bioresource material

In this paper, we report extraction of cashew nut shell liquid (CNSL) from cashew nut shell waste (CNSW) and further use of residues for generation of activated carbon for removal of heavy metals and methylene blue (MB). Solvent extraction yielded 24.6 ± 0.4%, 38.2 ± 0.4% and 40.1 ± 0.9% for petroleum ether, hexane and ethanol respectively. Phytochemical screening showed presence of alkaloids, carbohydrates, saponins, phenols, tannins, flavonoids, amino acids, terpenoids, proteins, steroids, glycosides and carboxylic acids. The CNSL had a pH of 3.2, viscosity (104.6 ± 1.8 mPa s), moisture (6.5%), ash (1.6 ± 0.1%), refractive index (1.52 ± 0.001), specific density (0.9561 ± 0.0002 g/cm3), acid value (118.7 ± 9.2 mg KOH/g), free fatty acid value (60.1 ± 4.7%), saponification number (138.1 ± 3.2 mg KOH/g) and iodine value (188.1 ± 2.3 mgI 2/100 g). The average percentage removal of Cu (II), Pb (II), Cd (II) and Zn (II) was 99.4 ± 0.5, 95.4 ± 1.5, 99.5 ± 0.1, 98.4 ± 0.1%, and removal efficiency of MB at 50, 150, 250 and 350 mg/L was 99.63, 97.66, 96.48 and 94.81%, respectively. Equilibrium data were best described by the Freundlich isotherm model. The maximum monolayer adsorption capacity was 12.1 mg/g. The adsorption kinetics conformed to pseudo-second-order model. ∆G° was negative and a ∆H° of + 22.76 kJ/mol indicated that adsorption was endothermic. The ΔS° (+ 0.086 kJ/mol/K) showed that there was spontaneous interaction of the solution and adsorbate. These results show that CNSW is a potential bioresource for CNSL production for use in the paints, varnishes, surface coatings, agrochemicals and ethnomedicine industries. Residual shells can be exploited as fuels or converted to activated carbon for use as low-cost filters in water purification.

Methods. Sample preparation. Pre-processed (roasted) CNS were collected from small-scale cashew nut processors in Mongu District of the Western Province of Zambia. Upon their arrival in the laboratory, the shells were washed several times with tap water and twice with enough distilled water to remove all the dirty, contaminants and debris. After washing, the shells were air-dried under the shade for 7 days. Once dry, the shells were ground to homogeneity using a Thomas-Model-4-Wiley-Mill fitted with a 2 mm sieve, placed in airtight bags and stored in a refrigerator at 4 °C to avoid biological and chemical degradation of the constituents.
Extraction of cashew nut shell liquid. The extraction of CNSL from CNSW was carried out by using a soxhlet extractor system as described by 6,39 . Briefly, 20.00 g of ground CNSW was put into a clean 33 × 100 mm cellulose thimble (Whatman) and extracted with a particular solvent for 8 h. After several cycles of extraction, the soxhlet apparatus was disassembled and the remaining solvent in the extracting chamber was added to the other extract in round bottomed flask, and evaporated under mild conditions with a Buchi Rotavapor until a constant oily mass remained.
Phytochemical screening. Twenty grams of ground CNS were exhaustively extracted under cold conditions in 200 mL acetone, ethanol and hexane respectively for 72 h with interval shaking in the dark. The organic solvents were recovered under mild pressure with a Buchi Rotavapor. The effect of solvent on phytochemical is presented in Table 1. The water extract was warmed at 60 °C for 10 min and left to stand for a total of 24 h with interval shaking. Phytochemical analysis for alkaloids, flavonoids, glycosides, phenols, saponins, steroids, tannins and terpenoids was done according to Refs. [40][41][42][43][44][45] and for amino acids, carbohydrates, carboxylic acids and proteins 46 with minor modifications. All reagents used in this process were prepared fresh before use.
Physicochemical characterization of cashew nut shell liquid. Methods by Refs. 1,11,47 with minor modifications were followed for characterization of the CNSL extracted from the roasted CNS. Moisture content was determined by heating 2.0 g of sample to a constant weight in a crucible placed in a Memmert oven (Memmert GmbH + Co. KG) maintained at 105 °C for 3.5 h. The crucible was cooled in the desiccator and reweighed, the mass change in the sample was recorded. Ash was determined by incinerating 1.0 g sample in a Carbolite muffle furnace (HTF ELP4, Bamford, Sheffield UK) maintained at 550 °C for 5 h 1 using sintered glass crucibles. Specific gravity was determined using a standard pycnometer bottle with a stopper. The 25 mL bottle was filled with distilled water and the CNSL respectively and weighed independently. The acid and free fatty acid values were determined using the methods of Refs. [48][49][50] . The saponification number and iodine values were determined by the method of Ref. 51 . Refractive index at 20 °C was determined using Bellingham Stanley Abbe refractometer. Viscosity was determined by the Oswald viscometer using distilled water as a reference at 24 °C. The pH was determined with a calibrated pH meter (Crison base 20).
Preparation of activated carbon adsorbent. The defatted cashew nut shells were pre-heated at 110 °C for 2 h using a Carbolite AAF 11/7 Furnace at a heating rate of 10 °C/min. Chemical activation with 50 wt% sulphuric acid was carried out using an impregnation method. The impregnation ratio of sulphuric acid to the raw materials was 2: 1. Thus, 60 g of the pre-heated precursors were soaked in 86 mL of 50 wt% sulphuric acid for 24 h. www.nature.com/scientificreports/ After soaking, the precursor was dried in an oven at 110 °C. The dried precursors were carbonized in the same furnace as before at 400 °C for 3 h at a heating rate of 10 °C/min. The carbonized material was cooled to room temperature and washed severally with hot distilled water until the pH was neutral. The cooled activated carbon was then dried in an oven for 4 h, ground and sieved using a 0.5 mm sieve and stored in airtight bottles until use.

Analysis of heavy metals and methylene blue (MB).
Heavy metal concentrations were measured on a Perki-nElmer Analyst 400 Atomic absorption spectrophotometer and a Shimadzu UV-2600 spectrometer was used to determine the concentration of Methylene blue standard before and after adsorption.
where; C o is the initial concentration of adsorbate (mg/L), C e is the final concentration of adsorbate after adsorption (mg/L), q e is the amount of adsorbate adsorbed at equilibrium (mg/g), m is the mass of activated carbon used (g) and V is the volume of adsorbate solution used (mL).
Adsorption isotherm models. Three commonly models used to fit adsorption experiment results are the Langmuir, Freundlich and Temkin adsorption isotherm models 53,54 .
The Langmuir isotherm model. The isotherm assumes that the monolayer adsorption process happens between the adsorbate and homogenous surface of the adsorbent 55 . The binding sites have the same affinity for adsorption. The linear equation is given below; where; q e is the metal ions adsorbed (mg/g) at equilibrium, C e is the equilibrium concentration (mg/L), q max is the monolayer adsorption capacity (mg/g) and K L is the Langmuir adsorption constant which is related to the energy of adsorption and is a measure of the metal ions affinity to the adsorption sites. If the magnitude of K L is large, the interaction of the adsorbent with the adsorbate molecules will be more while a smaller value indicates a weak interaction. The Langmuir parameters q max and K L were calculated from the slope (1/q max ) and intercept (1/q max K L ) of the plot of C e /q e versus C e. An important characteristic of the Langmuir isotherm can be expressed in terms of the dimensionless equilibrium parameter or the separation factor, R L, which is defined as; where; K L is the Langmuir adsorption constant and C o is the initial metal ion concentration. The value of the separation factor gives an indication of the shape of the isotherm and the nature of the adsorption process. The values of the R L between 0 and 1 indicates favourable adsorption, unfavourable adsorption occurs when R L is greater than 1 and adsorption is linear when R L is equal to 1 56 .
The Freundlich isotherm model. The Freundlich isotherm model is an empirical model that explains that adsorption occurs on an unevenly distributed or heterogeneous surface of the adsorbent. The adsorbent surface has different affinity and energy for adsorption. Stronger binding sites are occupied first and then the binding strength decreases with the rise in the degree of site occupation. It is represented by the equation below; where; q e is the metal ions adsorbed at equilibrium (mg/g), C e is the equilibrium concentration (mg/L), and K F is the Freundlich constant and n is the adsorption intensity. The value of n indicates the degree of non-linearity between metal ions concentration and its adsorption in the following manner; if n is equal to 1 (n = 1) then adsorption is linear, adsorption becomes a favourable physical process when n is greater than 1 (n > 1) and when n is less than 1 (n < 1) then adsorption is a chemical process 52,54 . From the slope (1/n) and intercept (log K F ) of the plot of log q e versus log C e , the constant K F and n can be calculated.
Temkin isotherm model. The Temkin isotherm model considers the effect of indirect adsorbate-adsorbent interaction on the adsorption process. It is based on the assumption that the heat of adsorption of all the molecules in a layer decreases linearly due to increase in surface coverage of the adsorbent. The decrease in heat of adsorption is linear rather than logarithmic, as implied in the Freundlich isotherm. Further, the adsorption is characterized by uniform distribution of binding energies, up to a maximum binding energy. The Temkin isotherm model is represented by the following equation 53 : where; K T is the equilibrium binding constant (L/mol) corresponding to the maximum binding energy, b is related to the adsorption heat, R is the universal gas constant (8.314 J/K/mol) and T is the temperature at 298 K. The constants K T and b can be calculated from the slope (RT/b) and intercept (RTIn K T /b) of the plot of q e versus ln (C e ) 53 .
Ethics approval and consent to participate. Ethical approval and waiver was obtained from the Natu-

Results and discussion
Cashew nut shell liquid was extracted from CNSW with different organic solvents using a soxhlet extractor system. The percent yields are represented in Fig. 1. Extraction methods involving organic solvents under mild conditions have been reported to preserve the natural composition of CNSL 6 . The quality and percent yields of extracted CNSL varied among the organic solvents. The hight yield was recorded from ethanol 40.1 ± 0.9% followed by hexane 38.2 ± 0.4% and petroleum ether 24.6 ± 0.4% respectively. Although, ethanol recorded the highest yield, the quality of CNSL it extracted was poor, as it extracted more of undesirable polar coloured compounds from the shell 57 . Even during the carbonization process, ethanol defatted shells produced toxic fumes, which led to a conclusion that it did not completely remove CNSL from the shell 58 . Phytochemical analysis of the aqueous, ethanol, acetone, and hexane extracts of CNSW (Table 1) revealed the presence of phenols, tannins, flavonoids, saponins, steroids, terpenoids, glycosides, carboxylic acids, carbohydrates, proteins and amino acids. Various extracts from CNSW have been reported to have antimicrobial, antifungal 59 , insecticidal (Acero, 2018) and antioxidant properties 60 . Phyto-compounds such as saponins have been reported to have anticancer and anticholesterol activity 61 . Flavonoids and other polyphenolic acids show antioxidant, anti-inflammatory, antidiabetic and anticarcinogenic activities 62,63 . Alkaloids are natural anticancer and analgesic agents 64,65 . Steroids and terpenoids have anti-tumor, neuroprotection, antihypertensive, antimicrobial and insecticidal properties 66,67 . Glycosides are better known for their physiological effect on the cardiovascular system, with cardiac glycosides being the drug of choice for treating congestive heart failure 68 . Thus, the presence of these essential phyto-compounds in CNSW extracts signifies the importance of this wasted raw material. The physico-chemical properties of CNSL are presented in Table 2, and were all determined using hexane extracted CNSL, because the quality and yield were better than the ethanolic and petroleum ether extracted-CNSLs respectively. The extracted CNSL was a reddish-brown viscous oil with a pH value of 3.2, probably due to a high concentration of anacardic acid and other phenolic compounds 8 . The moisture and ash content of the CNSW biomass expressed in percentages were 6.5% and 1.6 ± 0.1% respectively. These values were in line with 6.7% and 1.3% reported by Ref. 1 . The specific gravity, and refractive index values were 0.9561 ± 0.0002 (g/cm 3 ) and 1.52 ± 0.001 respectively. These values were higher than 0.9118 (g/cm 3 ) and 1.4325 reported by Ref. 39 , but lower than 1.686 and 0.9999 (g/cm 3 ) reported by Refs. 1,69 respectively.
The viscosity of CNSL in this work was 104.6 ± 1.8 mPa s. This value was lower than 160 mPa s and 410 mPa s reported by (Mohammed 69 ) and (Rodrigues et al. 11 ). The standard viscosity range for CNSL at 25 °C is 150-600 mPa s 72 . The reason for low viscosity in this work could be that the shells were roasted under uncontrolled conditions by the local cashew nut farmers during processing. However, CNSL with low viscosity presents an advantage, as it can be blended with diesel to form biodiesel for heavy engines 73 . Biodiesels with high viscosity   70 . A high acid value suggests that CNSL cannot be consumed by humans or directly applied on acid sensitive surfaces to avoid corrosion. Ingestion of oils with high acid value leads to human gastrointestinal discomfort, diarrhea and liver damage 76,77 . Cashew nut shell liquid with high acid value is first neutralized with alkaline bases before it is applied in paints, vanishes and others surface coating agents 70 . The saponification number and iodine values for this work were 138.1 ± 3.2 (mg KOH/g) and 188.1 ± 2.3 (gI 2 /100 g) respectively. The obtained saponification number was lower than 161 mg KOH/g reported by 7 . Saponification number depend on the amount of fatty acids present in a fat or oil sample. The higher the number, the higher the amount of fatty acids and vice versa. It is also used to determine the average molecular weight of fatty acid chains in fats/or oils 74 . Fats/ or oils with long fatty acids have low saponification values because they have fewer carboxylic group per unit mass of fat/oil, as compared to short chain fatty acids. Fatty acids with longer chains make good surfactants. Their surfactants have excellent detergent properties and they do not irritate the skin 78 . Iodine value indicates the unsaturation of fats/or oils 74 . The value 188.1 ± 2.3 (mgI 2 /100 g) obtained in this work was in line with and 177.7 (mg I 2 /100 g) reported by Ref. 7 . The higher the iodine value, the more unsaturated the fat/or oil sample is. Highly unsaturated oils or fats are good for paints and surface coating materials, as they dry faster and their conjugated double bonds help to slowdown the oxidation process of painted objects 1 . The difference in the composition and physicochemical properties of CNSL in this work and other literature sources may be due to variation in the species, climate and geography where cashew was grown as well as the operating conditions employed during analysis 79 .

Batch adsorption of heavy metals (copper, lead, cadmium and zinc) onto CNS-AC. The cashew
nut shell activated carbon (CNS-AC) was used to remove heavy metals (lead, cadmium, copper and zinc) from synthetic aqueous solutions. The batch adsorption was carried out at conditions of 1 g adsorbent dosage, 0.002 to 3 mg/L initial metal concentration, 30 mL of adsorbate solution, pH of 6.98, 30 min contact time and agitation speed of 250 rpm. The average percentage removal of Cu (II), Pb (II), Cd (II) and Zn (II) is shown in Table 3.

Batch adsorption of MB onto CNS-AC.
Some of the factors that affects adsorption such as pH, temperature, contact time and concentration were considered in this study.
Effect of solution pH on MB adsorption onto CNS-AC. The influence of initial pH value of the solution on the adsorption process of MB onto CNS-AC was carried out at 50 mg/L initial MB concentration, 298 K temperature   (Fig. 2). The near sigmoidal adsorption pattern under different pH units agrees with other studies done on adsorbates for methylene blue removal from aqueous solutions 80,81 . The adsorption of MB was highly favoured under basic compared to acidic conditions with the highest removal percentage of 99.1 at pH 10. Uptake of MB by CNS-AC was constant at pH 10 and 11 as shown in Fig. 2. The low adsorption efficiency of MB in acidic media could be attributed to high competition for adsorption sites between the excess hydrogen ions (H + ions) in the solution and the cation groups on MB 80 .

Effect of initial dye concentration and contact time on adsorption of MB onto CNS-AC.
The relationship between adsorption of MB and contact time was investigated to establish the rate of MB removal. Figure 3 shows Adsorption isotherms. Analysis of isotherm models is significant in modelling and designing of the adsorption process as they show the distribution of the adsorbate molecules between the liquid phase and the solid phase when an equilibrium state is reached. In this study, Langmuir, Freundlich and Temkin isotherm models were considered. The isotherm constants and regression coefficients (R 2 ) calculated from adsorption experiments are   Table 4. The Freundlich isotherm model was suitable since regression coefficient (R 2 ) was higher than that of the Langmuir and Temkin isotherm models as shown in Fig. 4. Thus, adsorption of MB onto CNS-AC fitted best the Freundlich isotherm model.
Adsorption kinetic models. The adsorption kinetic models are important in evaluating the rate and kinetic behaviour of the adsorption process. The kinetic parameters provide substantial information in designing and modelling of the adsorption process. The kinetic of methylene blue (MB) adsorption onto CNS-AC was analysed using pseudo-first-order and pseudo-second-order kinetic models.
A pseudo-first-order kinetic equation is given as; where; q e and q t (mg/g) are the amounts of methylene blue (MB) adsorbed at equilibrium and at time t (min), K 1 (min −1 ) is the adsorption rate constant. The parameters q e and K 1 were determined from the intercept and slope of a plots of log q e − q t versus t as shown in Fig. 5. The parameters of pseudo-first-order kinetic are tabulated in Table 4. Pseudo-second-order kinetic model is expressed as; where; K 2 (g/mg/min) is second order adsorption rate constant, h (mg/g/min) is the initial adsorption rate. The parameters q e and K 2 were calculated from the slope and intercept of the plots of t q t versus t as shown in Fig. 6. Pseudo-first-order and pseudo-second-order kinetic parameters for different initial concentrations of methylene blue are tabulated in Table 5. The value of the correlation coefficient (R 2 ) for pseudo-second order model is higher than the value of pseudo-first-order adsorption model. Furthermore, pseudo-second-order model has values of q e, cal which are close to q e, exp. It can be concluded that the adsorption of methylene blue onto CNS-AC follows pseudo-second-order kinetic model. This implies that, the rate-limiting step is the surface adsorption that involves chemisorption. Thus, chemical adsorption likely occurs through the formation of a covalent bond.
Effect of temperature and thermodynamic parameters. The influence of temperature on adsorption of MB using CNS-AC was investigated at different temperatures (298, 308, 318 and 328 K) and MB concentrations of 50, 150, 250 and 350 mg/L. Increasing the temperature from 298 to 328 K as shown in Fig. 7  Thermodynamic study of adsorption process of MB onto CNS-AC to estimate the feasibility of the adsorption process was investigated. The Gibbs free energy change (ΔG°) values are useful in determining whether the process is spontaneous or not. A positive value of ΔG° means that the adsorption process is non-spontaneous and a negative value shows that the process is spontaneous. The enthalpy change (ΔH°) differentiates a physical adsorption process from a chemical adsorption process and provides information about the exothermic nature or endothermic nature of the adsorption process 83 . The entropy change (ΔS o ) indicates the disorder of the solid/ solution interface during the adsorption process.
The change in Gibbs free energy (ΔG°) was calculated using the following equations 54,84 ; The enthalpy change (ΔH°) and entropy change (ΔS o ) change was determined from the equation below 82 ; where; T is the absolute temperature (K), R is the universal gas constant (8.314 J/mol/K), ∆G° (kJ/mol) is the Gibbs free energy change, ∆H° (kJ/mol) is the enthalpy change and ∆S° (kJ/mol/K) is the entropy change. The  (Table 6) are obtained from the slope and intercept of the plot of lnK versus 1/T (K −1 ) and are shown in Fig. 8.
The negative values obtained for Gibbs free energy (∆G°) indicates the feasibility and spontaneity of the adsorption process. The positive value (22.76 kJ/mol) of enthalpy change (∆H°) indicates that the adsorption of MB by CNS-AC at different temperatures was endothermic. The value of the entropy change (ΔS°) was 0.086 kJ/mol/K implying that the randomness of solid/solution interface during adsorption process increased. These results were consistent with studies done by Ref. 54 .

Conclusion
In this paper, we analyzed the potential use of cashew nut shell regarded as a waste in Zambia and indeed in many countries growing cashew. Solvent extraction and synthesis of activated carbon experiments were done to evaluate the potential value of cashew nut as a source of both chemical feedstock and activated carbon for use as a low cost filtration system matrix. Best yields of CNSL were achieved by hexane (38.2 ± 0.4%). Physicochemical results showed that CNSL has high potential as an intermediate in the synthesis of paints, varnishes, dyeingstuff, binders, lubricants, nanotechnology 73 and the presence of bioactive compounds such as alkaloids, steroids,   The study showed that increasing the initial pH, temperature and contact time increased the adsorption of MB onto CNS-AC and a decrease in initial MB concentration increased percentage removal of MB. Equilibrium data were fitted to Langmuir, Freundlich and Temkin isotherms models and the equilibrium data were best described by the Freundlich isotherm model. The maximum monolayer adsorption capacity was 12.1 mg/g. The kinetics of the adsorption process conformed to pseudo-second-order model and the negative value of the Gibbs free energy (∆G°) and positive value of enthalpy change (∆H°) indicates that the adsorption process was endothermic and spontaneous. This paper therefore