1 Introduction

The presence of industrial dyes in wastewater is a very serious environment problem because of their highly toxicity and carcinogenicity. They are known as potential hazards to living organism. Dyes have complex aromatic structures and synthetic origins that make them stable to biodegradation, oxidizing agents, and photodegradation [1]. The undesirability of the presence of dyes in aqueous prevents the passage of light through water, which leads to the reduction of photosynthesis activity and consequently decreases the dissolved oxygen in the water and spoils the ecosystem [1, 2]. Among these dyes, crystal violet (CV), a triphenylmethane dye, is immensely used in the dyeing of textile, the printing of paper, the coloring of leather and paint industry, and then it enters into the aquatic systems that were one of the major causes of water pollution [3]. CV is also well known for its cause of certain serious problems such as mutagenic, teratogenic, and mitotic poisoning nature. A number of methods aimed at removing CV from the wastewaters have been studied such as coagulation, adsorption, flocculation, reverse osmosis, biological methods, and membrane filtration [4]. Amid them, adsorption is one of the most effective techniques thanks to its advantages such as great ease of operation, selectivity, sensitivity, low cost, and low energetic requirements [5]. Nowadays, many kinds of composites and advanced materials have been developed for adsorption and adsorption integrated degradation of hazardous contaminants including CV in water [6, 7]. However, their disadvantages can be multi-step synthesis procedure, using toxic chemicals for material production, and eco-unfriendliness. Therefore, utilization of green and sustainable adsorbents is highly expected in the field.

The use of biomass as an adsorbent has been gained much attention. Indeed, many publications have revealed the use of various biomass adsorbents for the efficiency treatment of dyes from textile wastewaters such as rice husk ash [8], grapefruit peel [9], ginger waste [10], tomato plant root [11], date palm fiber [12], teak leaf powder [12], and waste black tea [13]. Considering advantages of these adsorbents, they are mainly derived from eco-friendly, low-cost, safe, and sustainable biowastes or agricultural by-products. As a result, the conversion of these biowastes into effective adsorbents is not only in accordance with the purpose of alleviating environmental pollution but also promotes the value of wastes for value-added products.

Water hyacinth (Eichhornia crassipes) is an aquatic weed plant and one of the most typical invasive species. They can be found ubiquitously in rivers, lakes, ponds, canals, and agricultural fields. Because of their very strong adaptability, reproduction, and spreading powers, water hyacinth has posed an alarming threat to indigenous plants, water quality, soil, as well as biodiversity [14]. To mitigate the harmful effect of this species, a new protocol of transforming water hyacinth into biosorbents is worth considered. Indeed, water hyacinth can be used as a bioadsorbent to adsorb various industrial dyes such as CV, methylene blue, Congo red, and Vitoria blue from aqueous media [15]. Some parts of water hyacinth plant such as roots and leaves can be used as effective materials for removing dyes from the industrial wastewaters. A number of publications investigated the use of the water hyacinth-based biosorbents, biocomposites, and bio-based photocatalysts to improve the dye removal efficiency [16,17,18]. However, the production of these adsorbents often prolonged and required toxic chemicals, leading to the possibility of secondary pollution. Moreover, the adsorption data in these works were analyzed in the form of univariate, which can lead to statistical mistakes and misunderstanding of synergistic interactions among adsorption parameters (e.g., concentration, dosage, pH, time, temperature, etc.). Difference from multivariate experiments and univariate experiments endures step-by-step investigations, resulting in the fact that a large number of trials is required. In many cases, optimum values rely completely on the results of all trials. As such, it is important to develop optimization approaches to aid the treatment process efficiently and reduce the number of trials.

Response surface methodology (RSM) is considered a common and easy-to-use tool to predict the processes in chemistry in order to reduce the number of experimental runs and produce high-quality products [19]. RSM describes the relationship between several independent variables and various dependent variables [20]. For instance, Eddebbagh et al. [21] reported a procedure based on the central composite design combined with response surface methodology to optimize the conditions of several parameters such as the activation temperature, activation time, and chemical impregnation. They found that the removal of methylene blue at optimized process conditions was 90.5%. In a previous study conducted by Moughaoui et al. [22], the optimum adsorption of methylene blue and malachite green was investigated with sugarcane bagasse adsorbent using RSM. Their work showed a good correlation between the predicted factors of the RSM model and the obtained experimental values. Therefore, the central composite rotatable design combined with response surface methodology can be a useful tool for optimization of dyes removal from wastewaters.

The objectives of this research, therefore, involve (i) to apply central composite rotatable design (CCRD) combined with RSM in order to get the convenient adsorption conditions and optimization modeling for the maximizing adsorption capacity of water hyacinth powder (WHP) for CV removal from aqueous solution; (ii) to understand the effects of four independent variables such as the initial pH solution, the contact time, the adsorbent dosage, and the initial concentration of CV and their interactions on the adsorption capacity of WHP for CV; (iii) to study the adsorption isotherms and kinetic models of CV adsorption on WHP.

2 Materials and Methods

2.1 Materials

Crystal violet (C25H30N3Cl, Sigma – USA), sodium hydroxide (NaOH, Merck-Germany), and hydrochloric acid (HCl, Xilong Scientific Co., Ltd, China) were used without further purification. Stock solution (1000 mg/L) was prepared by dissolving accurately the weight quantity of the CV (1.0 g) in distilled water. Experimental CV solutions of different concentrations were prepared by the dilution of a stock solution. The initial solution pH was adjusted with HCl 0.1 M and NaOH 0.1 M solution.

2.2 Preparation of Biosorbent

Water hyacinth plants, which were used in this study, were collected from Saigon River, Ho Chi Minh City, Vietnam. The water hyacinth was washed with distilled water to remove dust and dried at 60 °C for 24 h. After, the dried plants were sieved with particle size lower than 0.15 mm and stored in a desiccator until used.

2.3 Biosorption Studies

WHP biosorption studies (Fig. 1) were performed in batch using the colored solution of CV, wherein dye volume and the stirring speed were maintained at 25 mL and at 500 rpm, respectively, in all the experimental runs. In this experiment, the effect of variables such as the initial pH solution (2–9), the contact time (10–240 min), the adsorbent dosage (0.1–3.0 g), and the initial concentration of CV (100–600 mg/L) was studied. The amount of dye adsorbed by the biosorbent and the percent removal dye from the solution were determined based on the difference between the initial dye concentration and the dye concentration in the following Eqs. 1 and 2:

$$q=\frac{\left({C}_{0}-C\right)\times V}{m}$$
(1)
$$\% \, \text{removal}=\frac{\left({C}_{0}-C\right)}{{C}_{0}}\times 100$$
(2)

where q is the amount of dye taken up by the biosorbent (mg/g), C0 is the initial concentration of dye (mg/L), C is the concentration of dye in the supernatant (mg/L), V is the volume of the solution (L), and m is the mass of biosorbent (g).

Fig. 1
figure 1

Illustrative preparation and crystal violet adsorption performance of biosorbent powder derived from water hyacinth species

Full size image

2.4 Experimental Design

The estimation of the optimum conditions of CV removal from aqueous solution using WHP adsorbent depends on the factors such as the initial pH solution, the contact time, the adsorbent dosage, and the initial dye concentration. RSM and CCRD were used to investigate the effects of the four independent variables on response surface for optimization of adsorption conditions [23]. In this work, a total of 52 experiments were employed to evaluate the individual and interactive effects of the four main independent parameters on the CV adsorption efficiency. The maximizing adsorption capacity of WHP for CV removal from aqueous solution was carried out as a response Y of the system under four process parameters, namely the initial pH solution: 2–8, the contact time: 60–180 min, the adsorbent dosage: 1–3 g/L, the initial concentration of CV: 400–800 mg/L. The variables and levels of the design model are given in Table 1. Using this design, the experimental data were fitted according to Eq. 3 as a quadratic polynomial equation including individual and cross-effects of variables.

$$\begin{aligned} Y & {\kern 1pt} = \beta_{0} + \, \beta_{1} X_{1} + \, \beta_{2} X_{1}^{2} + \, \beta_{3} X_{2} + \, \beta_{4} X_{2}^{2} \\ & \quad + \, \beta_{5} X_{3} + \, \beta_{6} X_{3}^{2} + \, \beta_{7} X_{4} + \, \beta_{8} X_{4}^{2} \\ & \quad + \, \beta_{9} X_{1} X_{2} + \, \beta_{10} X_{1} X_{3} + \, \beta_{11} X_{1} X_{4} \\ & \quad + \beta_{12} X_{2} X_{3} + \, \beta_{13} X_{2} X_{4} + \, \beta_{14} X_{3} X_{4} \\ \end{aligned}$$
(3)

where β0, β1, β2, …, β14 are regression coefficients, Y is the maximizing adsorption capacity (response), X1 is the initial pH solution, X2 is the contact time, X3 is the adsorbent dose, and X4 is the initial concentration of CV. Multiple regression analysis, response surface plots, and statistical analyses were performed using JMP 10 software (SAS Institute, North Carolina, USA).

Table 1 Coded units used for the variable parameters for the statistical analysis
Full size table

2.5 Equilibrium and Kinetic Models

In this study, the classical models of Langmuir and Freundlich were used in order to describe the equilibrium between biosorption capacity and sorbate concentration at a constant temperature. The nonlinear pseudo-first-order and pseudo-second-order models were applied to study the adsorption kinetic [24]. The equations are presented in Table 2. where qe is the equilibrium dye uptake per g of the biosorbent (mg/g), qt is the amount of dye adsorbed at time t (mg/g), KL is the Langmuir constant (L/mg), KF is the Freundlich constant (mg/g), n is a constant dependent on the nature of adsorbate and temperature, and Qo is the adsorption capacity (mol/L).

Table 2 Biosorption equilibrium and kinetic models
Full size table

3 Results and Discussion

3.1 Characterization of Water Hyacinth Powder

The FTIR spectrum of WHP before (red line) and after (black line) adsorption of CV was shown in Fig. 2. The presence of a peak at 3427 cm−1 indicates the O − H stretching. The peak at 2932 cm−1 can be ascribed to the stretching vibration of the C − H bonds. The peaks were observed at 1640 cm−1, 1318 cm−1, and 1041 cm−1, which were associated with C = O, C − O, and sulfonic groups, respectively. After biosorption, the appearance of a peak at 1588 cm−1 reflected the C-N group, and the disappearance of the peak at 1318 cm−1 corresponded to the –CO group. Furthermore, the characteristic –OH band at 3427 cm−1 shifted to 3446 cm−1 confirmed the participation of these groups in the dye uptake. Similar results were also reported by Kumar et al. [10] and Saha et al. [25]. The results of functional groups on water hyacinth powder by FTIR analysis as a main characterization in this work were also in accordance with the previous work [26, 27].

Fig. 2
figure 2

FTIR spectra of WHP before adsorption (red line) and after adsorption (black line) of CV dye

Full size image

The textual structure examination of WHP particles can be observed before and after the biosorption of CV from the SEM (Fig. 3). Before dye uptake, the PR-WH particles show porous and irregular surfaces (Fig. 3a). The pores were partially filled after the adsorption of CV, showing the CV adsorbed on the surface of WHP (Fig. 3b). This finding indicates that the adsorption of CV molecules over WHP occurred.

Fig. 3
figure 3

SEM images of WHP before (a) and after (b) adsorption of CV dye

Full size image

3.2 Response Surface Methodology

The influence of experimental conditions with four variable parameters, namely the initial pH solution, the contact time (min), the adsorbent dosage (mg/g), and the initial concentration of CV (mg/L) on the CV adsorption, was studied using RSM. The total of 52 experiments is shown in Table 3. The relationship between variables and response can be represented in Eq. 4.

$$\begin{aligned} Y\left( {q_{e} } \right) & = 180.33588 \, + \, 6.1529438 \, X_{1} \\ &\quad + \, 3.4459979 \, X_{2} {-} \, 7.319777 \, X_{3} \\ &\quad + \, 3.5297188 \, X_{4} + \, 2.1051781 \, X_{3} X_{4} \\ &\quad - \, 3.375615 \, X_{2}^{2} {-} \, 2.900884 \, X_{3}^{2} - 3.298121 \, X_{4}^{2} \\ \end{aligned}$$
(4)
Table 3 The observed (Y) and the predicted values of the response
Full size table

The diagnostic plot of the model is shown in Fig. 4. Accordingly, the coefficient of linear regression R2 of 0.92, which is higher than 0.80, indicating the fitness of polynomial models used for describing the effect of variables on the response would be more accurate and 8% of the variations are not considered [28,29,30]. The results indicated that there was an agreement between the experimental and the prediction in adsorption capacity of WHP.

Fig. 4
figure 4

Actual and predicted plot of CV adsorption capacity (qe)

Full size image

The analysis of variance (ANOVA) for the response surface quadratic model is given in Table 4. The ANOVA results indicated that a model with a large F-value of 30.6818 and p-value < 0.0001 is more significant. When the p-value < 0.01 is typically viewed as an indication, the model was adequate to explain the correlation between the factors and the corresponding response [31]. The lack of fit was significant with a F-value of 44.1986 and a p-value < 0.0001. The probability of F-values which are fewer than 0.05 indicated that the model terms are significant [32]. It is obvious that the p-value is found to be less than 0.005 for all four individual parameters such as the initial pH solution (X1), the contact time (X2), the adsorbent dosage (X3), and the initial concentration of CV (X4) under linear conditions (Table 5). The model contained a six two-way interaction (the initial pH solution x the contact time, the initial pH solution x the adsorbent dosage, the contact time x the adsorbent dosage, the initial pH solution x the initial concentration of CV, the contact time x the initial concentration of CV, and the adsorbent dosage x the initial concentration of CV). Only the cross-interaction between the adsorbent dosage and the initial dye concentration was significant while all the rest of these interactions did not have a significant effect on the determination of the CV adsorption since the p-value was more than 0.05. However, the quadratic effects of X22, X32, and X42 were all signed with a p-value < 0.001 showing that the variables had a great influence on the CV adsorption capacity of the WHP. The 3-D response surface plots used for the predicted responses are shown in Fig. 5 that confirming the presence of strong interaction between each pair variable. The interaction of the initial pH solution with the contact time, the adsorbent dosage, and the initial CV concentration on the CV adsorption capacity response is shown in Figs. 4a, b, and c. The results indicated that the CV adsorption capacity of the WHP increased with increasing pH from 4 to 6 and slightly increased from pH 6 to 8. It seems that the maximum adsorption was found to be at pH 6. This result could be explained as due to the electrostatic repulsion interaction between the negatively charged WHP particles and the positively charged dye molecule [33]. Thus, as the H+ ions concentration in the solution decreased, the numbers of negative charges increased which results in an increase in the percentage removal of CV. Similar observations have been reported by Saeed et al. [9].

Table 4 Analysis of variances for optimization of CV adsorption capacity (qe) using the JMP software
Full size table
Table 5 Estimated regression coefficients and their significance
Full size table
Fig. 5
figure 5

The 3D plots showing the effect of a the initial pH solution and the contact time, b the initial pH solution and the adsorbent dosage, c the initial pH solution and the initial dye concentration, d the contact time and the adsorbent dosage, e the contact time and the initial dye concentration, f the adsorbent dosage and the initial dye concentration on the CV adsorption capacity

Full size image

The effect of the contact time with the initial pH solution, the adsorbent dosage, and the initial CV concentration are visualized in Figs. 5a, d, and e. From these data, it is revealed that the dye CV was rapidly adsorbed at the beginning of the experiment until 120 min due to a large number of vacant sites on the adsorbent particles surface available for the adsorption of CV and then relatively stable from 120 to 150 min. The interaction of the adsorbent dosage with other independent process variables is presented in Figs. 5b, d, and f. It can be observed that the adsorption efficiency increased with the increase in the adsorbent dosage from 1 to 2 g/L and then remained almost constant with a further increase in the adsorbent dosage. These results could be due to the increase in available surface area and consequently more active sites for efficient adsorption [34].

The optimum conditions for the initial pH solution, the contact time, the adsorbent dosage, and the initial concentration of CV were found to be 6.246, 125.698 min, 1.382 g/L, and 615.865 mg/L, respectively. At these conditions, the maximum adsorption of the CV dye was 180.336 mg/g with a desirability of 0.600. Under the optimal condition of parameters, the corresponding experimental value of the CV adsorption was 180.071 mg/g, indicating that there was conformity between the experiment results with the optimized values.

3.3 Adsorption Isotherms

In order to investigate the adsorption isotherms of the CV by the water hyacinth powder at different temperatures, two nonlinear isotherm models have been used: Langmuir and Freundlich isotherm models. The isotherm parameters are summarized in Table 6. Data shown in Figs. 5a and b represented the three various equilibrium isotherm models for the adsorption of CV on WHP. By comparing the R2 obtained from Langmuir and Freundlich isotherms (R2 > 0.97), it can be suggested that the Langmuir isotherm was more favored by the adsorption process at various temperatures. This indicated that the uptakes occur by monolayer adsorption onto a homogenous surface with a finite number of identical sites. The maximum adsorption capacity qmax was 181.818 mg/g, 192.308 mg/g, and 196.078 mg/g at 30 °C, 40 °C, and 50 °C, respectively. According to Table 6, obtained values of separation factor or equilibrium constant RL were 0.101, 0.111, and 0.105 at 30 °C, 40 °C, and 50 °C, respectively. These findings indicated that the favorability of the adsorption process [35].

Table 6 Langmuir and Freundlich constants at different temperatures
Full size table

To assess the maximum adsorption capacity obtained in this work, a comparison study has been conducted with emphasis on the biosorbents used, adsorption temperature, and capacity. According to Table 7, water hyacinth powder can adsorb crystal violet dye with very high qm values of 181.18–196.08 mg/g at various temperatures. These outcomes were so far higher than those obtained by Ficus religiosa leaves powder and Daucus carota pomace powder (2.4–27 mg/g), date palm leaves powder (36.74 mg/g). coffee husks (12.24 mg/g), and anatolian black pine (12.36 mg/g). The water hyacinth powder in this work had higher adsorption performance than modified biosorbents such as charred rice husk (62.85 mg/g) and xanthated rice husk (90 mg/g), but slighly lower than citric acid-modified red seaweed (217.38 mg/g) or centaurea stem powder (476.8 mg/g). The water hyacinth biosorbents also gave good adsorption capacity results in comparison with Ananas comosus (pineapple) leaf powder, red seaweed powder, and cottonwood seeds. To sum up, it can be concluded that the water hyacinth powder can be a competitive biosorbent for the removal of crystal violet from water.

Table 7 The comparative adsorption performance of various biosorbents for removal of crystal violet dye
Full size table

3.4 Adsorption Kinetics

In order to study the mechanism of CV adsorption by the WHP at various concentrations of CV, two kinetic models were considered: the pseudo-first-order model proposed by Lagergren and the pseudo-second-order model. Table 8 summarizes the kinetic parameters of the fitted models. Figures 6c and d indicates that the plots of nonlinear pseudo-second-order model showed a higher correlation coefficient R2 for all initial dye concentrations than that of the pseudo-first-order model. Moreover, the qe increased with the increase in initial concentration of CV, while the nonlinear pseudo-second-order rate constant K2 decreased with the increase in initial dye concentration. These results revealed that the pseudo-second-order model is appropriate only to describe the adsorption kinetics of CV by WHP adsorbent and it implies that chemisorption mechanics took place during the adsorption process. This is in accordance with the results obtained by Cheruiyot [49] using coffee husks for the adsorption of toxic crystal violet from an aqueous solution [49].

Table 8 Comparison of kinetic parameters for different initial concentrations CV at 30 °C
Full size table
Fig. 6
figure 6

The adsorption isotherm a Langmuir adsorption isotherm and b Freundlich adsorption isotherm for the adsorption of CV on water hyacinth powder and the adsorption kinetic, c pseudo-first-order plots, and d pseudo-second-order plots for the adsorption of CV on WHP

Full size image

From the best-fitting results obtained by the pseudo-second-order model, it can be suggested that the adsorption is controlled by chemisorption phenomenon, which surface functional groups are responsible for the interaction with crystal violet molecules. Indeed, the possible interaction between the CV and WHP can be elaborated as follows. WHP adsorbent possess a range of surface functional groups such as hydroxyl and carbonyl groups. These surface functional groups can interact with amine groups of CV through hydrogen interaction. Such interactions have been reported in the previous works, which enhances the adsorption performance of the biomass adsorbent [51]. Moreover, the water hyacinth powder biosorbent was not reusable as it was reflecting the chemical bonding mechanism between the dye molecules and the adsorbent sites. After the adsorption, the loss of surface functional groups led to the significant reduction of reuse performance. As a result, the recycle of water hyacinth powder biosorbent was not successful.

4 Conclusion

In the present work, the water hyacinth powder has been successfully produced via a facile procedure and was found with many surface functional groups. This biosorbent was used as an efficient adsorbent for crystal violet adsorption from the aqueous solution. The central composite rotatable design combined with response surface methodology has been adopted as a powerful tool for optimization of process parameters to maximize the dye adsorption capacity. It was observed that individual factors such as the initial pH solution, the contact time, the adsorption dosage, and the initial concentration of CV played an important role in the CV adsorption capacity of water hyacinth powder. Under optimum operating points, giving maximum CV adsorption (qe = 180.336 mg/g) was determined as follows: the initial pH solution, 6.246; the contact time, 125.698 min; the adsorbent dosage, 1.382 g/L; and the initial dye concentration, 615.865 mg/L. The tested values were in very high correlation with the predicted model values of the response variable, suggesting high validation and reliability of proposed quadratic model. The adsorption isotherms studies showed that the equilibrium process could be well described by Langmuir model, revealing the monolayer adsorption of dye on water hyacinth powder. The kinetic adsorption also suggested the best fitting of pseudo-second-order kinetic equation, which might confirm the control of surface functional groups toward adsorption process. With high maximum adsorption capacity (181.18–196.08 mg/g) at various temperatures, our findings recommend that water hyacinth powder can be an attractive candidate for removing crystal violet from the wastewater.