Absorptive Removal of Palm Oil Free Fatty Acids Onto Silica/smectite Composite: A Statistical Study using Box-Behnken Design in Response Surface Methodology

doi:10.21203/rs.3.rs-3224990/v1

Download PDF

Research Article

Absorptive Removal of Palm Oil Free Fatty Acids Onto Silica/smectite Composite: A Statistical Study using Box-Behnken Design in Response Surface Methodology

https://doi.org/10.21203/rs.3.rs-3224990/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 26 Nov, 2023

Read the published version in Chemical Papers →

You are reading this latest preprint version

Response Surface Methodology was used to evaluate the main and interactions effects of contact time (20–60 min), temperature (60–90°C) and adsorbent dosage (1–2%) of the palm oil free fatty acids (FFAs) adsorption onto a silica/smectite composite. The regression model was found to be able to predict more than 99% of the targeted response variation in the studied range (R² = 99.31%). An ANOVA was used to evaluate the accuracy of the regression model and it was found to predict more than 99% of the response variation. The Pareto Charts of Standardized Effects showed that almost all the explanatory factors were influential. The factorial plots evidenced that time/temperature interaction was most effective and the corresponding contour and surface plots were drawn. The second-order kinetic model was better followed with the highest value of the determination coefficient (R² = 0.996) and the lowest value of Root Mean Square Deviation of 0.555. Response Optimizer Function showed an optimum for FFAs removal of more than 90% (27.18 mg of KOH/g) at the adsorbent dosage of 2%, contact time of 60 min, and temperature of 90°C. Moreover, a good correlation between the predicted and experimental values was observed within reasonable limits.

Adsorption

Free Fatty acids

Box-Behnken Design

Response Surface Methodology

Factors’ effects

Palm oil is mainly made up of triglycerides of fatty acids and glycerol. Some of those fatty acids which have escaped to be “imprisoned” in glycerides are called free fatty acids (FFAs). The so-called FFAs can be produced by many means such as hydrolysis of triglycerides during palm oil processing, natural release in crude palm oil, and action of enzymes in the palm fruits by microbial lipases. The reaction of oil with water during storage, lengthy storage of products and contamination of fungi in palm oil also produce FFAs (Che Man et al. 1999; Atinafu and Bedemo 2011). Frying oils have high FFAs contents due to the repeated heating that favours various chemical reactions including oxidation, hydrolysis, cyclization, polymerization and degradation. Such reactions give rise to many degrading products like monoglycerides, diglycerides, volatile compounds, lipid peroxides, hydrocarbons, oxygenated products and FFAs (Kheang et al. 2006; Sukiran et al. 2011; Guo et al. 2016) which deteriorate the quality of the oil. Despite the important role of FFAs in a wide range of industrial fields (anticorrosive agent, antimicrobial protection…) (Foucaud et al. 2021), their content in some products like cooking oil should be 5% at the highest (Azeman et al. 2015) in order to assure a good quality for human health.. In fact, an abnormal increase in FFAs concentration due to one or more of the aforementioned means limits the quality of oil and makes it unhealthy for human consumption. On one hand, it may lead to various health issues such as risk of high blood pressure, cardiovascular disease and atherosclerosis and limit its conservation as well as environmental issues (Ebongue et al. 2008; Azeman et al. 2015; Gadiraju et al. 2015; Sarkar and Mohan 2017). On the other hand, low FFAs contents produce good physicochemical properties for the palm oil products and make it valuable for food applications, industrial applications such as production of cosmetics and biodiesels. Oils containing more than 1% of FFAs will form a difficult emulsion of soap during the separation of biodiesel and provoke the clogging of the engine (Ismaila et al. 2017; Japir et al. 2017; Tan et al. 2018; Susilowati et al. 2019a). Due to the high demand in palm oil industry market nowadays and the fact that palm oil quality and price is indeed dependent on the FFAs content in palm oil, it is thus imperative to match the standards of palm oil quality. That reduction during refinery will provide the consumers with attractive products having a long shelf life and fulfilling the standards of the production of some products. Therefore, many adsorbent types ranging from natural biosources to synthetic materials have been already used. Extensive results so far put emphasis on the effect of some experimental parameters such as dosage of the adsorbent, temperature, contact time and on the kinetic study of the adsorption process (Sukiran et al. 2011; Pohndorf et al. 2016; Ismaila et al. 2017; Baptiste et al. 2020; Kurtulbaş et al. 2021). Response Surface Methodology (RSM) is a collection of tools developed in the 1950s for the purpose of determining optimum operating conditions in applications in the chemical industry (Myers et al. 1989). The response of interest is influenced by several variables and the objective of the study is to optimize this response by minimizing the total number of experimental runs (Arenas et al. 2007; Turan et al. 2011; Ranganath and Vipin 2015; Zolgharnein et al. 2015). The so-called technique uses quantitative data from appropriate experiments to determine and simultaneously solve multi-variable equations correlating a response with independent parameters (input variables) (Annadurai and Sheeja 1998; Resende et al. 2020). The graphical representations of these equations are called response surface and they can be used to describe the individual and cumulative effect of the input variables on the response. To date, the mathematically evaluation of the effects of some fundamental processing parameters such as time, temperature and adsorbent dosage as well as their interactions on the removal of FFAs has not been investigated enough (Ayu Putranti et al. 2018; Susilowati et al. 2019b). On this line, the novelty of the present study focuses on the palm oil FFAs adsorption onto a new silica/smectite clay based composite previously synthesized by an eco-friendly pathway using low cost natural resources, clay and rice husks (Kepdieu et al. 2023). Also, emphasis was put on the optimization of the fundamental parameters affecting the FFAs adsorption using RSM as well as the non-linear kinetic modelling and mechanism of the process.

Adsorbent

The adsorbent used in this study was synthesized from a 20 µm mesh enriched smectite clay fraction and silica from rice husk ash as starting materials. It led to a product which structure was largely described in previous work (Kepdieu et al. 2023) as a delaminated/exfoliated porous material with a specific surface area of 300 m²/g.

Palm oil

The palm oil used in this study is locally purchased from Master Foods Co., Cameroon. It was kept in the dark chamber at a low temperature (˜20°C) to avoid any degradation by oxidizing and photocatalytic reaction (Woumfo et al. 2007; Syamila et al. 2019).

Batch adsorption

For RSM studies, preliminary experiments help to choose the factor’s domains as time (from 20 to 60 min), temperature (from 60 to 90°C) and adsorbent dosage (from 1 to 2%) for design of experiment. The batch tests were performed in a thermostatically controlled reflux reactor following a method described in some previous works (Ayu Putranti et al. 2018; Baptiste et al. 2020). 5 g of oil was introduced in the reactor and when the indicated temperature value (60, 75 or 90°C) was reached. A known amount of adsorbent was added and the mixture was stirred at 250 rpm for a fixed time following the runs provided by the matrix of experiences. The mixture was filtered under vacuum using Whatman paper N°1. The oil was allowed to settle for 20 min and 1000 mg of the supernatant was then used to determine the FFA left in the treated oil.

For kinetics, the study was conducted in a larger range of time (from 0 to 90 min) and the data were compared to three models. On the same line, the influence of temperature and adsorbent dosage on the removal of FFAs was also evaluated. For 5 g of oil, the adsorbent dosage varied as 1, 1.5, 2, 2.5 and 3% at t = 90 min and the temperature values were 60, 75 and 90°C.

Assessment of FFAs content of treated oil supernatant

The oil was tested for acidity using the American Oil Chemists’ Society (AOCS) method Ca 5a–40 (1989) adapted by Japir et al. (Japir et al. 2017). First of all, 1000 mg of the treated oil sample were placed in a dried conical flask. Approximately 10 mL of pre-neutralized ethanol were then added to the sample. Afterward, 4 drops of 1% phenolphthalein indicator was then added to the mixture. The flask was subsequently placed on a hot plate and heated until a temperature of around 40°C was attained. The mixture was then titrated with alcoholic potassium hydroxide solution (0.05 N) until a pink color emerged for at least 30 s. Indeed, if molecules of FFAs are labeled with the formula RCOOH, the acid-base reaction which occurs between them and alcoholic KOH is total and the balance equation of the reaction is Eq. (1).

$$RCOOH + {(K}^{+}+ {OH}^{-})\to RCO{O}^{-}+{K}^{+}+{H}_{2}O$$

For treated sample, the FFA content expressed as Current Acid Value was determined using Eq. (2) (Akinola et al. 2010; Medeiros Vicentini-Polette et al. 2021).

$$\text{C}\text{u}\text{r}\text{r}\text{e}\text{n}\text{t} \text{a}\text{c}\text{i}\text{d} \text{V}\text{a}\text{l}\text{u}\text{e}=\frac{{\text{C}}_{\text{K}\text{O}\text{H}}{\text{V}}_{\text{é}\text{q}}{\text{M}}_{\text{K}\text{O}\text{H}}}{{\text{m}}_{0}}$$

Where ${C}_{KOH}$ is the molar concentration of the alcoholic KOH solution (0.05 mol/L), ${V}_{eq}$the equivalent Volume of the pre-neutralized ethanol (mL), ${M}_{KOH}$molar mass of KOH (56.1 g/mol) and ${m}_{o}$the weight of the treated oil sample (1000 mg). The FFA content was determined in oil by titration of the supernatant after treatment and the quantity adsorbed (response) was calculated and expressed as Acid value drop (AV drop) in Eq. (3)

AV drop = Initial Acid value - Current Acid value (3)

Box-Behnken Design/Response Surface Methodology

In this study, RSM was applied to evaluate the effects of three independent variables contact time (X₁), temperature (X₂), and adsorbent dosage (X₃), on the FFAS removal efficiency expressed as AV drop. Experimental adsorption design was established using Box-Behnken Design (BBD) implemented in Minitab 21. It is a three-factors and levels (–1, 0, 1) system consisting of 15 experimental runs. They were designed to optimize Acid Value (AV) drop in palm oil. The ranges of the considered variables as well as the corresponding levels in coded and uncoded values are summarized in Table 1.

Table 1

Designation of factors and levels in coded and uncoded values
Temps		Temperature		Adsorbent dosage
X₁	t/min	X₂	T/°C	X₃	d (%)
-1	20	-1	60	-1	1
0	40	0	75	0	1.5
1	60	1	90	1	2

If the natural variables (uncoded units) are labeled as U₁, U₂ and U₃ corresponding to dimensionless coded variables X₁, X_{2 ,} X₂ respectively, the conversion of natural variables (U_ij) into coded ones (X_ij) is done following Eq. (4) (Merabet et al. 2009; Zolgharnein et al. 2015).

$${X}_{ij}=({U}_{ij}-{U}_{j}^{0})/\varDelta {U}_{j}$$

Where ${X}_{ij}$, ${U}_{ij}$ and ${U}_{j}^{0}$are the values of the coded variable, the natural variable and the natural value at the center of the domain of factor j at the i^th experiment respectively.$\varDelta {U}_{j}$ is the half variation step of the factor domain of the natural variable.

The effect of the processing parameters as well as their interactions were investigated on the adsorption of FFAs by calculating the coefficients of the linear terms b_i and the coefficient of interactions b_ij (i ≠ j) between factors i and j using Equations (5) and (6).

${\text{b}}_{\text{i}}=\frac{{\sum }_{1}^{\text{N}}{\text{X}}_{\text{i}\text{n}}{\text{Y}}_{\text{n}}}{\text{N}}$ (1 ≤ n≤N) (5)

${b}_{ij}=\frac{{\sum }_{1}^{N}{X}_{in}{{X}_{jn}Y}_{n}}{N}$ (1 ≤ n≤N) (6)

where${X}_{in}$ and ${X}_{jn}$ are the values of the coded variables of factor i and j respectively.

The polynomial regression model used to describe the effects of the processing parameters is based on Taylor series and given in Eq. (6) (Khuri and Mukhopadhyay 2010; Turan et al. 2011; Zulfiqar et al. 2016; Kumar and Das 2017; Jawad et al. 2020; Dbik et al. 2022).

$$Y={b}_{0}+\sum _{i}{b}_{i}{X}_{i}+\sum _{i}\sum _{j}{b}_{ij}{X}_{i}{X}_{j}+\sum _{i}\sum _{j}\sum _{l}{b}_{ijl}{X}_{i}{X}_{j}{X}_{l}+\dots$$

Where Y and ${b}_{0}$represent the theorical response of a test and that at the center of the experimental domain respectively. with 1 ≤ n ≤ k, 1 ≤ i ≤ j ≤ k and 1 ≤ i ≤ j ≤ l ≤ k.

In the, present study, a simple second order form of Eq. (6) for three factors is considered and given in Eq. (7).

$$Y={b}_{0}+{b}_{1}{X}_{1}+{b}_{2}{X}_{2}+{b}_{3}{X}_{3}+{b}_{11}{X}_{1}^{2}+{b}_{22}{X}_{2}^{2}+{b}_{33}{X}_{3}^{2}+{b}_{12}{X}_{1}{X}_{2}+{b}_{13}{X}_{1}{X}_{3}+{b}_{23}{X}_{2}{X}_{3}$$

Regression and graphical analysis were done using Minitab21 Software. Analysis of variance (ANOVA) was done on the regression model and on the model coefficients to test their significance, and adequacy. Thus, parameters such as the correlation coefficient R², R² (Adjusted), F-value, and P-value (probability) were generated to determine the relevance and suitability of the predicated model.

Kinetic studies

For kinetics studies, a non-linear approach was done considering Equations. (8) and (9) where Q_e and Q_t represented the quantities of FFAs adsorbed at equilibrium and at a time t respectively. AVe, the acid value of the oil at equilibrium, AV_t; the acid value of the oil after treatment for time t (min) and m, the mass of oil.

$${Q}_{e}= \frac{{AV}_{0} -{{AV}_{ }}_{e}}{m}$$

$${Q}_{t}= \frac{{AV}_{0} -{{AV}_{ }}_{t}}{m}$$

Using the classical equations of the pseudo-first order, pseudo-second order and intra-particle diffusion models as reported by many authors (Ahmad et al. 2009; Pohndorf et al. 2016; Baptiste et al. 2020; Kurtulbaş et al. 2021) and introducing in the acid value parameter, the kinetics equations in their non-linear form have been re-written and presented in Equations (10)-(12).

$({\text{A}\text{V}}_{0} -{{\text{A}\text{V}}_{ }}_{\text{t}})=({\text{A}\text{V}}_{0} -{{\text{A}\text{V}}_{ }}_{\text{e}})[1-\text{exp}\left(-{\text{K}}_{1}\text{t}\right)]$ (Pseudo-1st order) (10)

$({\text{A}\text{V}}_{0} -{{\text{A}\text{V}}_{ }}_{\text{t}})=\frac{{\text{K}}_{2}{({\text{A}\text{V}}_{0} -{{\text{A}\text{V}}_{ }}_{\text{e}})}^{2}\text{t}}{1+{\text{K}}_{2}({\text{A}\text{V}}_{0} -{{\text{A}\text{V}}_{ }}_{\text{e}})\text{t}}$ (Pseudo-2nd order) (11)

$$({\text{A}\text{V}}_{0} -{{\text{A}\text{V}}_{ }}_{\text{t}})= {K}_{id}*{t}^{\frac{1}{2}}+\gamma \left(Intraparticle diffusion\right)$$

Where K₁ (min^-1), K₂ (mg^-1.g.min) and ${K}_{id}$ (in mg/min^0.5) are the kinetic constants of the pseudo-1st order, pseudo-2nd order and intraparticle diffusion respectively; $\varvec{\gamma }$ (mg/g) is associated to the boundary layer thickness.

In order to validate the best fitting to the model with more accuracy, the determination coefficient and the root mean square deviation designated by R² and RMSD have been calculated using Equations (13) and (14).

$${R}^{2}=1-\frac{{\sum }_{k=1}^{n}{({v}_{exp,k}-{v}_{mod,k)}}^{2}}{{\sum }_{k=1}^{n}{({v}_{mod,k}-{\widehat{v}}_{exp,k)}}^{2}}$$

$$RMSD=\sqrt[2]{\frac{\sum _{k=1}^{n}{({v}_{exp,k}-{v}_{mod,k)}}^{2}}{n}}$$

Where: ${\text{v}}_{\text{e}\text{x}\text{p},\text{k}}$ and ${\text{v}}_{\text{m}\text{o}\text{d},\text{k})}$ are the exprimental and the expected values of the “k^th” test.${\widehat{\text{v}}}_{\text{e}\text{x}\text{p}}$ is the average exprimental value. n is the total number of tests.

Response Surface Methodology

Experimental Design matrix and response

15 experimental runs proposed by the system occurred in repetitious way to optimize the level of chosen variables. The three variables (coded and uncoded), the experimental conditions determined with Eq. (4) as well as the corresponding experimental response values are summarized in Table 2.

Table 2

Coded and uncoded values from Box-Behnken’s matrix and experimental data of AVdrop
	Box-Behnken matrix			Experimental plan			Experimental Response (AV drop)
Run order	T/min	t /°C	d (%)	X₁	X₂	X₃	Y
1	1	-1	1	40	60	2.0	20.67
2	1	1	0	60	75	1.0	21.81
3	1	1	0	40	75	1.5	22.44
4	0	-1	1	20	75	2,0	15.13
5	0	1	-1	60	90	1.5	26.08
6	-1	-1	0	60	75	2.0	24.39
7	-1	0	1	40	60	1.0	18.22
8	0	0	0	40	90	2.0	19.23
9	0	0	0	20	60	1.5	12.69
10	-1	0	-1	40	90	1.0	19.87
11	0	-1	-1	20	75	1.0	12.56
12	0	1	1	40	75	1.5	19.23
13	0	0	0	40	75	1.5	19.23
14	-1	1	0	60	60	1.5	21.58
15	1	0	-1	20	90	1.5	14.62

Mathematical modeling and Pareto charts of standardized effects

Using Minitab 21software, an updated resource for RSM; the main coefficients b_i and the interaction coefficients b_ij were calculated from the experimental responses. The regression model of the study in coded values is then given by Eq. (15).

AV drop (mg of KOH/g) = 19.230 + 4,858 X ₁ + 1.23 X₂ + 1.271X₃ – 1.158 X₁X₁ + 0.678 X₂X₂ + 0.400 X₃X₃ + 0.642 X₁X₂ + 0.003 X₁X₃ + 0.030 X₂X₃ (15)

The (+) positive sign of the main and interaction coefficients shows the synergetic behaviour of the processing parameters towards the adsorption process (Kumar and Das 2017). Figure 1 shows the Pareto chart of the Standardized Effects, which gives the relative importance of the individual and interaction effects. The vertical line in the chart indicates the minimum statistically significant effect magnitude at the 95% confidence level (Turan et al. 2011).

Concerning the main effects, temperature, dosage, and time of respective coefficients of 1.231, 1.271 and 4.858 given by the model’s equation are in the order of their increasing influence on the process of adsorption of FFAs on the composite. The overall positive values of the coefficients confirm that when the level of a factor increases independently; it is favorable for the adsorption process. It also emerges from the Pareto diagram that the significance level is 2.57 and therefore, the factors time, temperature and adsorbent dosage are all significant and more for the first. Moreover, in the studied range of parameters, time/temperature interaction is more influential for the process compared to the time-dosage interaction and temperature-dosage interaction which are very weak with 0.03 and 0.003 as coefficients respectively.

ANOVA and Henry Normal plot

Further analysis has been carried out to evaluate the accuracy of the mathematical model of the current study (Merabet et al. 2009). In order to minimize error, all the experimental results were duplicated and the average was taken. The significance of the model is mainly calculated by regression coefficient (R²) value (Burton et al. 1988; Gato et al. 2007; Piñeiro et al. 2008; Kumar and Das 2017). The significance was examined based on the magnitude of the F-values, while the P-value less than 0.05 is considered to be statistically significant with a 95% confidence level (Abdulsalam et al. 2020). Therefore ANOVA for the AV drop of palm oil was done and the details are presented in Table 3.

Table 3

Analysis Of Variance (ANOVA) for the Surface response for the full quadratic model.
Source	Degree of freedom	Adjusted Square Sum	Adjusted Mean Sum	F-Value	P-Value
Model	9	223.267	24.807	79.62	0.000
Linear	3	213.819	71.273	228.76	0.000
t/min	1	188.762	188.762	605.85	0.000
T/°C	1	12.128	12.128	38.93	0.002
d (%)	1	12.929	12.929	41.50	0.001
Square	3	7.793	2.598	8.34	0.022
t/min*t/min	1	4.947	4.947	15.88	0.010
t/°C*T/°C	1	1.657	1.657	5.32	0.069
d (%)*d (%)	1	0.591	0.591	1.90	0.227
2-Way Interaction	3	1.655	0.552	1.77	0.269
t/min*T/°C	1	1.651	1.651	5.30	0.070
t/min*d (%)	1	0.000	0.000	0.00	0.993
t/°C*d (%)	1	0.004	0.004	0.01	0.919
Error	5	1.558	0.312
Lack-of-Fit	3	1.558	0.519	*	*
Pure Error	2	0.000	0 .000
Total	14	224.824
R²	99.31%,

ANOVA shows that R² = 99.31%, meaning that more than 99% of the variation in adsorption can be explained by its relationship with the factors considered and displaying an agreement between experimental and predicted values (Merabet et al. 2009). It is also noticed that the mathematical models for all responses are validated from the statistical test results since the F-value is 79.62 >>> 1 and p-value < 0.001% for the model and such results have been proved by other authors (Zhang and Zheng 2009; Bilici Baskan and Pala 2010; Ojewumi et al. 2021; Dbik et al. 2022). Furthermore, in the Normal Probability plot (Henry normal plot) shown in Fig. 2 all the points are around the straight line, meaning that the residuals follow a normal distribution. This confirms that the model accurately describes the variations in responses and also proves that the values obtained are regular, symmetrical, and have no abnormal or aberrant values (Dbik et al. 2022).

Factorial plots

The previous data showed that the mathematical model developed is appropriate for the study of the FFAs on the silica/smectite composite. The factorial plots given by main effect, interactions’ effects, contours, and surface plots of the process are given in Figs. 3 and 4.

The main effects and the interaction plots for the adsorption of FFAs onto the composite are in accordance with the predictions of the regression model and the Pareto charts of standardized effects. Concerning Figure (3), it comes out that all the factors seem to affect the FFAs uptake because the lines are not horizontal. Substantially, the relationship between time, temperature, adsorbent dosage, and the response (AV drop) is synergically since the mean value of the response increases as those factors’ levels increase.

Figure 4(i) shows non-parallel curves indicating that time/temperature interaction is observable and almost significant as shown by Pareto charts (Fig. 1). Moreover, Fig. 4(ii) shows that when the temperature is fixed at its lowest value (60°C), the AV drop increases from 12.69 to 14.62 mg of KOH/g when the contact time varies from 10 to 60 min respectively. This corresponds to a half-variation of 0.97. For the highest value of the temperature (90°C), the AV drop increases from 21.68 to 26.08 mg of KOH/g for the same gap of time corresponding to a half-variation of 2.2. The two half-variations are different and graphically, it is also materialized by two parallel lines on top of the charts of Fig. 5(ii). These remarks confirm the existence of interactions between time and temperature.

Figure 4(iii) shows almost parallel curves indicating minor or no interactions between time and dosage. According to Fig. 4(iv), when the dosage is fixed at its lowest value (1%), the AV drop increases from 12.56 to 15.13 mg of KOH/g when the time varies from 10 to 20 min respectively. This corresponding half-variation is 1.29. For the highest value of the dosage (2%), the calculated half-variation corresponding to an increase in mean AV drop from 21.81 to 24.39 is the same. Furthermore, these results are graphically proved by the two parallel lines on top of the charts of Fig. 4(iv).

The temperature-dosage interaction effects presented in Figs. 4(v) and 4(vi) show similar results with Figs. 4(iii) and 4(iv). Indeed, when the temperature is fixed at its lowest value (60°C), the AV drop increases from 18.22 to 19.87 mg of KOH/g when the dosage varies from 1 to 2% respectively. This corresponds to a half-variation of 0.8. For the highest value of the temperature (90°C), the corresponding calculated half-variation is almost the same (0.89) but with a slight difference proving that the interaction is a little bit more significant than the time//dosage interaction. This result is also evidenced in Parato charts where the time/dosage’s coefficient is 10 times greater than the temperature/dosage’s coefficient. Furthermore, these results are graphically proved by the two parallel lines on top of the charts in Fig. 4(iv).

Response surfaces plots in 2D or 3D provide a simple method for optimizing the process and identifying interactions between variables (Ojewumi et al. 2021; Dbik et al. 2022). Since the three independent factors studied are all significant, two of them, time and temperature have been chosen to produce surface plots of AV drop and presented on Fig. 5. On the 3D plot (Fig. 5(a)), the horizontal plane of the figure materializes the domain of variation of time and temperature, and the vertical axis the variation of the mean values of AV drop. The 3D plot is in accordance with the previous results since it can be evidenced that the mean AV drop values rises when time or temperature increase.

For better understanding, a projection along the z-axis (Acid drop‘s axis) at different values of AV drop has been done and the results obtained are presented in Fig. 5(b) as contours plots.

The so-called contour plots represent an infinity of combinations between time and temperature levels with adsorbent dosage held at a constant level of 1.5% corresponding to 0 in coded value (Merabet et al. 2009). Considering that the recommended acid value should be equal to or less than 10 mg of KOH/g in palm oil and that one of the untreated oils is around 30 mg of KOH/g, it can be noticed in Fig. 5(b) that the hatched area is satisfactory (AV drop ≥ 20 mg of KOH/g). That area covered 45 to 60 min for time and 60 to 90°C temperature. On the other hand, the plot demonstrates that a contact time of less than 45 min is not interesting at any level of temperature.

Optimization of the AV drop

Using the Response Optimizer Function, further evaluation has been carried out in order to determine the optimal conditions of the process with respect to the acid value reference. Those conditions are then reported in Table 4 and plotted in Fig. 6.

For the aforementioned reasons, 26.08 mg of KOH/g was taken as the targeted response and 20 mg of KOH/g as the minimum value. The system has generated a double response prediction with an ideal desirability value of 1.000 with 95% of confidence.

Table 4

Optimized solution of the model
Solution	t/min	T/°C	d (%)	AV drop Fit	Standard Error fit	Confidence Index	Desirability
1	60	90	2	27.18	0.659	95%	1
2	53.8	90	2	26.08	0.659	95%	1

Verification and validation of the model

For a model to be validated, the predicted responses must be close to the experimental ones. A scatterplot has been drawn to verify the relationship between AV drop (experimental) and AV drop (Predicted) values. The results are shown in Fig. 7. The comparison between these results shows that we have an excellent estimate of the model since almost all the points are along the first bisector line. It can therefore be adapted for the interpretation of tests.

Studies of some parameters on the adsorption of FFA onto the Silica/smectite Composite

In order to extend the study on the adsorption of FFAs from palm oil, a detailed evaluation of the influence of the parameters contact time, temperature and dosage of the adsorbent has been carried out on palm oil with 30 mg/KOH as initial acid value.

Effect of temperature and adsorbent dosage on the removal of FFAs

The plots of the FFAs removal as a function of temperature on one hand and as a function of adsorbent dosage on the other hand are presented in Fig. 8.

Figure 8(a) presents the influence of the temperature on the FFA adsorption assessed from the treated palm oil. It shows that the rate of adsorption of FFAs increases with the treatment temperature. Indeed, the higher the temperature from 60–90 min, the lower is the viscosity of the palm oil, thus promoting greater mobility of the molecules of FFA. Simultaneously, there would also be activation of an increasing number of adsorption sites when the temperature rises (Ahmad et al. 2009; Baptiste et al. 2013, 2020). The process would be endothermic because it is favored by the increase in temperature (Yuney et al. 2020). Figure 8(b) indicates a perceptible increase in the adsorbed quantity of FFAs with the dosage from 1 to 2% after which the variation is almost stable. That initial rapidly increasing tendency in the rate of absorption of FFA when the adsorbent dose increases would be the result of an abundance of the active sites available to act with the molecules of FFA. This behaviour is no longer significant above 2% and can be explained by the fact that, beyond this dose, practically all the FFAs are already retained. Some authors share the idea of the formation of a bulk which on the one hand would obstruct the other active sites of the composite and on the other hand, would limit the mobility of FFA in the oil/adsorbent mixture (Baptiste et al. 2020). It, therefore, seems judicious for economic purposes to consider 2% for yields around 90%.

Kinetics studies

• Effect of contact time

In order to further understand the mechanism of the adsorption of FFAs onto the silica/smectite based composite. Kinetic studies involving pseudo-first-order, pseudo-second-order and intraparticle diffusion model in non-linear regression have been done and the results are given in Fig. 9.

Figure 9 shows that the amount of adsorbed FFAs increases with contact time until it reaches a steady state. Indeed, from 0 to 60 min of oil/adsorbent contact, a rapid increase in the quantity adsorbed is noted due to the high concentration of active sites at the surface of the adsorbent used. After 60 min, the rate becomes slower and tends towards equilibrium, which corresponds to about 27 mg of KOH/g (about 90% of FFAs adsorbed). The steady-state trend is explained by the reduction of sites and the potential diffusion of adsorbate molecules into the pores of the adsorbent, and therefore this rate ceases to develop when almost all the active sites are occupied. Similar results were obtained by Susemo et al. in the adsorption of FFAs from sardine oil by attapulgite and bentonite (Suseno et al. 2013). The effectiveness of the composite is mostly explained by the presence of pores and mesopores in its exfoliated structure. A similar explanation has been reported in the literature for the adsorption of free fatty acids from rice husk on magnesium silicate (Clowutimon et al. 2011). The kinetics constants are given in Table 5.

Table 5

Kinetics constants of adsorption of FFAs onto Silica/smectite composite
Kinetic models	Equations	Constants notations	Constant values
Pseudo-1st order		K₁	0.0176
		Qe	0.531
	${Q}_{t={Q}_{e}(1-{Exp}^{(-}}{K}_{1}t)$	R²	0.867
		RMSD	3.188
Pseudo-2nd order		K₂	0.009
		Qe	1.151
	${Q}_{t}=\frac{{K}_{2{Q}_{e}^{2}t}}{1+{K}_{2}{Q}_{e}t}$	R²	0.996
		RMSD	0.555
Intraparticle Diffusion		K_ind	0.0986
	${Q}_{t}= {K}_{id}{t}^{1/2}+ \gamma$	$\gamma$	0.0023
		R²	0.9643
		RMSD	1.650

K ₁ (min^− 1), K₂ (mg^− 1g.min), and k_id (mg/gmin^0.5) are the kinetic constant of the 1st order, 2nd order, and intra-particle diffusion kinetic constant respectively. $\varvec{\gamma }$ (mg/g) is associated with the boundary layer thickness, t (min) is the time; Q_e and Q_t (mg/g) represent the amounts of FFAs adsorbed at equilibrium and at time t respectively.

The most fitted model is the pseudo-2nd order which exhibited the highest determination constant R² (0.9958) and the lowest Root Mean Square Deviation (RMSD) of 0.556 compared to other models. This model suggests an adsorption mechanism with two steps including mostly hydrogen bonds and electrostatic interactions. The proposed mechanism is given in Fig. 8. It starts with a speedy step corresponding to the external diffusion of FFAs from the oil to the surface of the adsorbent due to the gradient of concentrations between the oil and adsorbent surface. The following step is a steady one that can be ascribed to the interaction between the adsorbate and the adsorbent particles (Nwabanne et al. 2018; Baptiste et al. 2020). However, the determination constant corresponding to intraparticle diffusion is R² = 0.9643. That value is bigger enough to involve that some molecules can also perfectly fit the intraparticle diffusion model as the porous structure of the composite is also favorable for that (Mbah et al. 2005; Sabah et al. 2007).

• Proposed mechanism of the adsorption of FFA onto Silica/smectite composite

Based on the FTIR analysis of the composite which previously evidenced the presence of residual silanol (Si-OH) groups on the adsorbent surface (Kepdieu et al. 2023). The adsorption mechanism is weak interactions such as hydrogen bonding between the hydroxyl groups both found on the FFA molecules (RCOOH) and the silanols groups (Si-OH) on the composite surface. Moreover, palm oil is an acid substance (Afane ZE et al. 1997; Domonhédo et al. 2018), it therefore gives to protonate the oxygen atoms of Si-O-Si and Si-OH groups found on the composite surface before the formation of electrostatic interactions. The protonated groups (Si-OH⁺-Si and Si-OH₂⁺) and other FFAs molecules interact following the illustrative scheme given in Fig. 10. Another mechanism tendency may involve the ionic interaction between the polar carbonyl groups of FFA and the hydrophilic silanol (≡ Si-OH) and siloxane (≡ Si-O-Si≡) groups which may drop the surface tension at the interface of FFA and silica. This will then higher affinity and an increased adsorptive potential of adsorbent to FFA. (Kheang et al. 2006; Parida et al. 2006; Siwińska-Stefańska et al. 2008).

Comparison of the results obtained in this study with other published works.

Table 6 displays a comparative value of the bleaching capacity of adsorbents used in this work with respect to those of other materials available in the literature.

Table 6

Comparison with other results on FFA adsorption
Adsorbent	Optimum parameters Temperature/contact time/Adsorbent/oil ratio/solution	FFA adsorbed (mg /g of adsorbent)	Reference
Coffee Husk Ash	30°C /1%.w)/5.5 h	1966	(Vanida Chairgulprasert 2018)
Silica/smectite composite	95°C/1h/2%. W	617 (27 mg of KOH/g of oil)	Current work
Strong base anion exchange resin	Ambient/8 h/ 0.4%	589.5	(Mhadmhan et al. 2023)
Magnesium silicate-derived rice husk	50°C /1h / 2%.w	185	(Clowutimon et al. 2011)
Mesoporous silica functionalized (mSiO₂-NH₂)	70°C//1.5h / 5%	147 (528 µmol/g)	(Ahn and Kwak 2020)
Commercial Magnesium silicate	50°C/ 1h/ 4%.w	70	(Clowutimon et al. 2011)
Montmorillonite	25°C / 1h/x%.w	46.20	(Bayrak 2006)
Mesoporous silica functionalized (mSiO₂-OH)	70°C//1.5h/5%	28 (101µmol/g )	(Ahn and Kwak 2020)

The adsorption of Free Fatty Acids (FFAs) onto a silica/smectite based composite can be well undertood by Response Surface Methodolgy. The main parameters affecting the process such as time, temperature, and adsorbent dosage were determined by Pareto Charts of Standardized effects. It showed that time is the more influential and significant parameter in the process, followed by adsorbent dosage and then temperature with respective coefficients of 4.158, 1.271, and 1.231. The temperature was effective in the process with corresponding contour and surface plots allowed us to know the sets of factors’ levels that lead to good results (remaining Acid value ≤ 10 mg of KOH/g). The optimization of the model confirms that more than 27 mg of KOH/g, i.e. around 90% of the FFA were adsorbed. The adsorption process well fitted with second order with a determination coefficient of 0.996 and a value of RMSD (0.555). It, therefore, displayed a two steps mechanism with the transfer of FFA molecules from the oil to the surface of the composite followed by the interaction between these molecules and the surface of the composite surface. The Box-Behnken Design in Response Surface Methodology is therefore convenient and suitable to explain the influence of the studied factors. The mechanism of elimination of FFAs in oil by a low cost and environmentally friendly adsorbent silica/smectite based composites was also detailed.

Ethics Approval

Not applicable: Authors declare no research involving human participants and/or animals was conducted.

Consent to Participate

All the authors are voluntarily participating for the submission of this research work.

Consent to publication

The authors confirm that this manuscript has not been submited or published previously to any other journal and give full consent for publication of the research work.

Funding Statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflicts of Interest

All authors certify that they have no affiliations with or involvement in any organization or entity with any financial interest or non-financial interest in the subject matter discussed in this manuscript.

Data Availability

The [xls] data used to support the findings of this study are available from the corresponding author upon request.

Authors Contributions

Jean Marie Kepdieu: Investigation, Roles/Writing - original draft, Data curation. Njiomou Djangang Chantale: Conceptualization, Methodology; Writing - review & editing, Visualization: Validation; Supervision. Gustave Tchanang: Writing - review & editing. Jacques Romain Njimou: Investigation, Writing - review & editing. Sanda Andrada Maicaneanu and Chedly Tizaoui: Methodology, Writing - review & editing.

Acknowledgments

The authors are grateful to Dr Jacques Richard Mache for providing the smectite clay.

Abdulsalam M, Man HC, Abidin ZZ, et al (2020) Decolorization of Palm Oil Mill Effluent by Klebsiella Pneumonia ABZ11: Remediation Efficacy and Statistical Optimization of Treatment Conditions. Front Microbiol 11:675. https://doi.org/10.3389/fmicb.2020.00675
Afane ZE, Lando G, BIYITI BI ESSAM L, et al (1997) Les vapeurs de l’huile de palme bouillante, un acide broncho-irritant. Médecine d’Afrique Noire 44:604–607
Ahmad A, Chan C, Abd Shukor S, Mashitah M (2009) Adsorption kinetics and thermodynamics of β-carotene on silica-based adsorbent. Chemical Engineering Journal 148:378–384. https://doi.org/10.1016/j.cej.2008.09.011
Ahn Y, Kwak S-Y (2020) Functional mesoporous silica with controlled pore size for selective adsorption of free fatty acid and chlorophyll. Microporous and Mesoporous Materials 306:110410. https://doi.org/10.1016/j.micromeso.2020.110410
Akinola F, Oguntibeju O, Adisa A, Owojuyigbe O (2010) Physico-chemical properties of palm oil from different palm oil local factories in Nigeria. Journal of Food, Agriculture & Environment 8:264–269
Annadurai G, Sheeja RY (1998) Use of Box-Behnken design of experiments for the adsorption of verofix red using biopolymer. Bioprocess Engineering 18:463. https://doi.org/10.1007/s004490050472
Arenas LT, Lima EC, dos Santos AA, et al (2007) Use of statistical design of experiments to evaluate the sorption capacity of 1,4-diazoniabicycle[2.2.2]octane/silica chloride for Cr(VI) adsorption. Colloids and Surfaces A: Physicochemical and Engineering Aspects 297:240–248. https://doi.org/10.1016/j.colsurfa.2006.10.050
Atinafu DG, Bedemo B (2011) Estimation of total free fatty acid and cholesterol content in some commercial edible oils in Ethiopia, Bahir DAR. Journal of Cereals and Oil seeds 2:71–76. https://doi.org/DOI: 10.5897/JCO11.025
Ayu Putranti MLT, Wirawan SK, Bendiyasa IM (2018) Adsorption of Free Fatty Acid (FFA) in Low-Grade Cooking Oil Used Activated Natural Zeolite as Adsorbent. IOP Conf Ser: Mater Sci Eng 299:012085. https://doi.org/10.1088/1757-899X/299/1/012085
Azeman NH, Yusof NA, Othman AI (2015) Detection of free fatty acid in crude palm oil. Asian Journal of Chemistry 27:1569. http://dx.doi.org/10.14233/ajchem.2015.17810
Baptiste BJ, Esther N, Mirela P, Richard K (2013) Adsorption isotherm and kinetics modeling of carotene and free fatty acids adsorption from palm oil onto montmorillonite. Int J Biosci 15–24. https://doi.org/10.12692/ijb/3.3.15-24
Baptiste BMJ, Daniele BK, Charlène EM, et al (2020) Adsorption mechanisms of pigments and free fatty acids in the discoloration of shea butter and palm oil by an acid-activated Cameroonian smectite. Scientific African 9:e00498. https://doi.org/DOI 10.1002/ejlt.200501149
Bayrak Y (2006) Application of Langmuir isotherm to saturated fatty acid adsorption. Microporous and Mesoporous Materials 87:203–206. https://doi.org/10.1016/j.micromeso.2005.08.009
Bilici Baskan M, Pala A (2010) A statistical experiment design approach for arsenic removal by coagulation process using aluminum sulfate. Desalination 254:42–48. https://doi.org/10.1016/j.desal.2009.12.016
Burton G, Hook J, Butler J, Hellwig R (1988) Effect of temperature, daylength, and solar radiation on production of coastal bermudagrass. Agronomy Journal 80:557–560. https://doi.org/10.2134/agronj1988.00021962008000040002x
Che Man Y, Haryati T, Ghazali HM, Asbi B (1999) Composition and thermal profile of crude palm oil and its products. Journal of the American oil chemists’ society 76:237–242. https://doi.org/10.1007/s11746-999-0224-y
Clowutimon W, Kitchaiya P, Assawasaengrat P (2011) Adsorption of free fatty acid from crude palm oil on magnesium silicate derived from rice husk. Engineering Journal 15:15–26. https://doi.org/10.4186/ej.2011.15.3.15
Dbik A, El Messaoudi N, Bentahar S, et al (2022) Optimization of methylene blue adsorption on agricultural solid waste using box–behnken design (BBD) combined with response surface methodology (RSM) modeling. Biointerface Res Appl Chem 12:4567–4583
Domonhédo H, Cros D, Nodichao L, et al (2018) Enjeux et amélioration de la réduction de l’acidité dans les fruits mûrs du palmier à huile, Elaeis guineensis Jacq.(synthèse bibliographique). https://doi.org/10.25518/1780-4507.16228
Ebongue GFN, Koona P, Nouy B, et al (2008) Identification of oil palm breeding lines producing oils with low acid values. European journal of lipid science and technology 110:505–509. https://doi.org/10.1002/ejlt.200700263
Foucaud Y, Lainé J, Filippov LO, et al (2021) Adsorption mechanisms of fatty acids on fluorite unraveled by infrared spectroscopy and first-principles calculations. Journal of Colloid and Interface Science 583:692–703. https://doi.org/10.1016/j.jcis.2020.09.062
Gadiraju TV, Patel Y, Gaziano JM, Djoussé L (2015) Fried food consumption and cardiovascular health: a review of current evidence. Nutrients 7:8424–8430. https://doi.org/10.3390/nu7105404
Gato S, Jayasuriya N, Roberts P (2007) Temperature and rainfall thresholds for base use urban water demand modelling. Journal of hydrology 337:364–376. https://doi.org/10.1016/j.jhydrol.2007.02.014
Guo Q, Gao S, Sun Y, et al (2016) Antioxidant efficacy of rosemary ethanol extract in palm oil during frying and accelerated storage. Industrial Crops and Products 94:82–88. https://doi.org/10.1016/j.indcrop.2016.08.032
Ismaila A, Saimon NN, Jusoh M, Zakaria ZY (2017) Adsorption of free fatty acid in biodiesel from palm fatty acid distillate using KOH-activated starch. CHEMICAL ENGINEERING 56:. https://doi.org/DOI: 10.3303/CET1756104
Japir AA-W, Salimon J, Derawi D, et al (2017) Physicochemical characteristics of high free fatty acid crude palm oil. Ocl 24:D506. https://doi.org/10.1051/ocl/2017033
Jawad AH, Malek NNA, Abdulhameed AS, Razuan R (2020) Synthesis of Magnetic Chitosan-Fly Ash/Fe3O4 Composite for Adsorption of Reactive Orange 16 Dye: Optimization by Box–Behnken Design. J Polym Environ 28:1068–1082. https://doi.org/10.1007/s10924-020-01669-z
Kepdieu JM, Djangang CN, Njimou JR, et al (2023) Low-cost pathways to synthesize silica-smectite clay-based composites. Silicon 1–12. https://doi.org/10.1007/s12633-023-02569-2
Kheang LS, Foon CS, May CY, Ngan MA (2006) A study of residual oils recovered from spent bleaching earth: their characteristics and applications. American Journal of Applied Sciences 3:2063–2067
Khuri AI, Mukhopadhyay S (2010) Response surface methodology. WIREs Comp Stat 2:128–149. https://doi.org/10.1002/wics.73
Kumar N, Das D (2017) Alkali treatment on nettle fibers part II: design of experiment and desirability function approach to study enhancement of tensile properties. The Journal of the Textile Institute 108:1468–1475. https://doi.org/10.1080/00405000.2016.1257347
Kurtulbaş E, Toprakçı İ, Pekel AG, Şahin S (2021) A model study for decolorization reasons: β-carotene removal and its kinetics and thermodynamics behaviors. Biomass Conversion and Biorefinery 1–7. https://doi.org/10.1007/s13399-021-01644-1
Mbah JBB, Kamga R, Nguetnkam JP, Fanni J (2005) Adsorption of pigments and free fatty acids from shea butter on activated Cameroonian clays. European Journal of Lipid Science and Technology 107:387–394. https://doi.org/10.1002/ejlt.200501149
Medeiros Vicentini-Polette C, Rodolfo Ramos P, Bernardo Gonçalves C, Lopes De Oliveira A (2021) Determination of free fatty acids in crude vegetable oil samples obtained by high-pressure processes. Food Chemistry: X 12:100166. https://doi.org/10.1016/j.fochx.2021.100166
Merabet S, Bouzaza A, Bouhelassa M, Wolbert D (2009) Modélisation et optimisation de la photodégradation du 4-méthylphénol dans un réacteur à recirculation en présence d’UV/ZnO. Revue des sciences de l’eau 22:565–573. https://doi.org/10.7202/038331ar
Mhadmhan S, Yoosuk B, Chareonteraboon B, et al (2023) Elimination of free fatty acid from palm oil by adsorption process using a strong base anion exchange resin. Separation and Purification Technology 310:123211. https://doi.org/10.1016/j.seppur.2023.123211
Myers RH, Khuri AI, Carter WH (1989) Response Surface Methodology: 1966–l988. Technometrics 31:137–157. https://doi.org/10.1080/00401706.1989.10488509
Nwabanne JT, Onu CE, Nwankwoukwu OC (2018) Equilibrium, kinetics and thermodynamics of the bleaching of palm oil using activated nando clay. Journal of Engineering Research and Reports 1:1–13. https://doi.org/DOI: 10.9734/JERR/2018/42699
Ojewumi ME, Ehinmowo AB, Obanla OR, et al (2021) Comparative analysis on the bleaching of crude palm oil using activated groundnut hull, snail shell and rice husk. Heliyon 7:e07747. https://doi.org/10.1016/j.heliyon.2021.e07747
Parida SK, Dash S, Patel S, Mishra BK (2006) Adsorption of organic molecules on silica surface. Advances in Colloid and Interface Science 121:77–110. https://doi.org/10.1016/j.cis.2006.05.028
Piñeiro G, Perelman S, Guerschman JP, Paruelo JM (2008) How to evaluate models: observed vs. predicted or predicted vs. observed? Ecological modelling 216:316–322. https://doi.org/10.1016/j.ecolmodel.2008.05.006
Pohndorf R, Cadaval Jr T, Pinto L (2016) Kinetics and thermodynamics adsorption of carotenoids and chlorophylls in rice bran oil bleaching. Journal of Food Engineering 185:9–16. https://doi.org/10.1016/j.jfoodeng.2016.03.028
Ranganath M, Vipin H (2015) Surface roughness prediction model for CNC turning of EN-8 steel using response surface methodology. International Journal of Emerging Technology and Advanced Engineering 5:135–143
Resende RF, Leal PVB, Pereira DH, et al (2020) Removal of fatty acid by natural and modified bentonites: Elucidation of adsorption mechanism. Colloids and Surfaces A: Physicochemical and Engineering Aspects 605:125340. https://doi.org/10.1016/j.colsurfa.2020.125340
Sabah E, Cinar M, Çelik M (2007) Decolorization of vegetable oils: Adsorption mechanism of β-carotene on acid-activated sepiolite. Food Chemistry 100:1661–1668. https://doi.org/10.1016/j.foodchem.2005.12.052
Sarkar O, Mohan SV (2017) Pre-aeration of food waste to augment acidogenic process at higher organic load: Valorizing biohydrogen, volatile fatty acids and biohythane. Bioresource technology 242:68–76. https://doi.org/10.1016/j.biortech.2017.05.053
Siwińska-Stefańska K, Walkowiak J, Krysztafkiewicz A, Jesionowski T (2008) Polymer adsorption on the surface of highly dispersed silica. Applied surface science 254:3591–3600. https://doi.org/10.1016/j.apsusc.2007.11.046
Sukiran MA, Kheang LS, Bakar NA, May CY (2011) Production and characterization of bio-char from the pyrolysis of empty fruit bunches. American Journal of Applied Sciences 8:984
Suseno SH, Izaki AF, Suptijah P, Jacoeb AM (2013) Kinetic Study of Free Fatty Acid Adsorption Using Adsorbent in Sardine (Sardinella sp.) Oil Refining
Susilowati E, Hasan A, Syarif A (2019a) Free fatty acid reduction in a waste cooking oil as a raw material for biodiesel with activated coal ash adsorbent. IOP Publishing, p 012035
Susilowati E, Hasan A, Syarif A (2019b) Free fatty acid reduction in a waste cooking oil as a raw material for biodiesel with activated coal ash adsorbent. IOP Publishing, p 012035
Syamila M, Gedi M, Briars R, et al (2019) Effect of temperature, oxygen and light on the degradation of β-carotene, lutein and α-tocopherol in spray-dried spinach juice powder during storage. Food Chemistry 284:188–197. https://doi.org/10.1016/j.foodchem.2019.01.055
Tan CX, Gun Hean C, Hamzah H, Ghazali HM (2018) Optimization of ultrasound‐assisted aqueous extraction to produce virgin avocado oil with low free fatty acids. Journal of Food Process Engineering 41:e12656. https://doi.org/10.1111/jfpe.12656
Turan NG, Elevli S, Mesci B (2011) Adsorption of copper and zinc ions on illite: Determination of the optimal conditions by the statistical design of experiments. Applied Clay Science 52:392–399. https://doi.org/10.1016/j.clay.2011.04.010
Vanida Chairgulprasert (2018) Removal of Free Fatty Acid from Used Palm Oil by Coffee Husk Ash. Science & Technology Asia 23:19. https://doi.org/10.14456/SCITECHASIA.2018.18
Woumfo D, Kamga R, Figueras F, Njopwouo D (2007) Acid activation and bleaching capacity of some Cameroonian smectite soil clays. Applied Clay Science 37:149–156. https://doi.org/10.1016/j.clay.2006.12.008
Yuney K, Oladipo AA, Gazi M, Younis DZ (2020) CuO coated olive cake nanocomposites for rapid phenol removal and effective discoloration of high strength olive mill wastewater. Chemosphere 253:126703
Zhang Z, Zheng H (2009) Optimization for decolorization of azo dye acid green 20 by ultrasound and H2O2 using response surface methodology. Journal of Hazardous Materials 172:1388–1393
Zolgharnein J, Shariatmanesh T, Asanjarani N, Zolanvari A (2015) Doehlert design as optimization approach for the removal of Pb(II) from aqueous solution by Catalpa Speciosa tree leaves: adsorption characterization. Desalination and Water Treatment 53:430–445. https://doi.org/10.1080/19443994.2013.853625
Zulfiqar U, Subhani T, Husain SW (2016) Synthesis and characterization of silica nanoparticles from clay. Journal of Asian Ceramic Societies 4:91–96. https://doi.org/10.1016/j.jascer.2015.12.001

Download PDF

Journal Publication

published 26 Nov, 2023

Read the published version in Chemical Papers →

Reviewers agreed at journal
24 Aug, 2023
Reviewers invited by journal
24 Aug, 2023
Editor invited by journal
03 Aug, 2023
Editor assigned by journal
03 Aug, 2023
First submitted to journal
01 Aug, 2023

You are reading this latest preprint version

Absorptive Removal of Palm Oil Free Fatty Acids Onto Silica/smectite Composite: A Statistical Study using Box-Behnken Design in Response Surface Methodology

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Experimental

Adsorbent

Palm oil

Batch adsorption

Assessment of FFAs content of treated oil supernatant

Box-Behnken Design/Response Surface Methodology

Kinetic studies

Results and discussion

Response Surface Methodology

Experimental Design matrix and response

Mathematical modeling and Pareto charts of standardized effects

ANOVA and Henry Normal plot

Factorial plots

Optimization of the AV drop

Verification and validation of the model

Studies of some parameters on the adsorption of FFA onto the Silica/smectite Composite

Effect of temperature and adsorbent dosage on the removal of FFAs

Kinetics studies

• Effect of contact time

• Proposed mechanism of the adsorption of FFA onto Silica/smectite composite

Conclusions

Declarations

References

Status:

Journal Publication

Version 1