Simultaneous utilization of soju industrial waste for silica production and its residue ash as effective cationic dye adsorbent

2.1 Materials and reagents

Soju industrial waste was collected from local soju industry in Iksan-Si (South Korea), and all other chemicals used in the experiments were of analytical grade. MB was purchased from Sigma-Aldrich, Co. (St. Louis, MO, USA).

2.2 Extraction of silica and preparation of RA adsorbent

SIW was air dried, milled, and passed through a 60 mesh sieve to collect powder of size <0.250 mm. The obtained powder was washed several times in water to remove aqueous soluble substances and oven dried at 60°C for 24 h. The dried SIW powder (50 g) was burned to ash using a programmable muffle furnace – at 250°C for 1 h, followed by 350°C for 1 h, and finally at 550°C for 10 h at a rate of 5°C/min – to remove all of the hydrocarbons, and the resultant white ash was used for silica extraction (28). The chemical composition of SIW-A is presented in Table 1.

SIW-A (5 g) was dispersed in 0.5 m NaOH aqueous solution and heated at 110°C for 4 h under continuous stirring. The insoluble residue was collected by filtration through Micro Filter Holder Assembly. The transparent filtrate of sodium silicate was allowed to cool (26°C) and titrated with 10% H₂SO₄ to pH 7. The solution was stirred for another 48 h and allowed to stand for 72 h. The silica was quickly frozen by liquid nitrogen and dried overnight using a vacuum freeze dryer. The dried product was stored in a vacuum desiccator for further characterization. After extraction, the RA was washed with distilled water until the supernatant reached pH 7 and dried overnight at 120°C. RA was stored in a desiccator and used for adsorption experiments.

2.3 Characterization of synthesized materials

The chemical analysis of SIW-A and RA was performed by X-ray fluorescence using a PANalytical SuperQ 50 spectrometer. Fourier transform infrared (FTIR) spectroscopy (Perkin-Elmer FTIR spectrophotometer, CA, USA) was used to identify the surface functional groups. The surface morphology and fundamental physical properties were observed by field emission scanning electron microscopy (SEM; Hitachi S-4700, Tokyo, Japan). The surface area of the samples were determined using the Brunauer-Emmett-Teller (BET) equation and the total pore volume was estimated to be the liquid volume of nitrogen at a relative pressure of 0.98 (ASAP 2010, Micromeritics, USA). The structure of the silica was observed by biological transmission electron microscopy (TEM) (H-7650, HITACHI, Japan). Diffractograms following X-ray diffraction (XRD) of the silica were obtained using a Cu Kα incident beam (k=0.1546 nm), monochromated by a nickel filtering wave at a tube voltage of 40 kV and tube current of 30 mA.

2.4 Batch adsorption experiments with adsorbent RA

Equilibrium studies were carried out at 30°C with different initial dye concentrations (50 mg/l, 100 mg/l, 200 mg/l, 300 mg/l, 400 mg/l) and adsorbent dosage of 1 g/l, and agitated at 100 rpm for 330 min. The concentration of MB in solutions before and after adsorption experiments were determined by a UV-vis spectrometer at 668 nm. The effect of pH (2, 4, 6, 8, 10, 12) was determined by batch experiments, and pH of the solution was adjusted using 1N HCl or 1N NaOH. The following equation (Eq. [1]) was used to calculate the amount of MB adsorbed at equilibrium q_e (mg/g):

where C_o and C_e (mg/l) are the initial and equilibrium concentrations of MB, respectively, V (l) is the volume of the solution, and W (g) is the mass of RA used. The percentage removal of MB was calculated according to Eq. [2]:

The kinetic experiments were carried out at 30°C similar to equilibrium studies, by measuring the concentration of MB left in supernatant solutions at predetermined time intervals. The amount of adsorption at time t, q_t (mg/g), was calculated by Eq. [3]:

where C_o and C_t are the concentrations of the dye in liquid phase at initial and any time, respectively.

The distribution of the solute molecules between the solid phase and liquid phase can be understood by the adsorption isotherms. The isotherm data were fitted to the Freundlich and Langmuir isotherms to find a suitable model that can be used to optimize the design of an adsorption system (29). The Langmuir isotherm is represented by the following linear equation (30):

where C_e (mg/g) and q_e (mg/g) are the equilibrium concentration and the amount of dye adsorbed per unit mass of adsorbent, respectively. Q_o and b are the Langmuir constant related to adsorption capacity and rate of adsorption, respectively. The linear plot C_e/q_e (specific adsorption) versus C_e (equilibrium concentration) shows a straight line with slope 1/Q_o, which indicates that the adsorption follows the Langmuir isotherm.

The Freundlich isotherm model was derived assuming the adsorbent to have a heterogeneous surface with nonuniform distribution of heat of sorption while interacting between dye molecules. The Freundlich adsorption isotherm can be expressed as:

where q_e is the amount of dye (mg) adsorbed per unit of adsorbent at equilibrium (mg/g), C_e is the equilibrium concentration of dye in solution (mg/l), K_f is the Freundlich constant expressing the capacity of adsorption, and n indicates the intensity of the adsorption (31).

To perceive and comprehend the nature of the mechanism of the adsorption process, two widely used kinetic models such as pseudo-first-order and pseudo-second-order were applied to estimate the rate constants, initial adsorption rates, and adsorption capacities of the RA.

The linear form of pseudo-first-order kinetic model of Lagergren is given as follows (32):

where q_eand q_t are the amount of MB adsorbed (mg/g) at equilibrium and at time t (h), respectively, and k₁ is the rate constant of first-order adsorption (h^-1).

The data were modeled using the following equation (5) as described by (33, 34):

where k₂ (g/mg h) is the pseudo-second-order rate constant and q_e (mg/g) was calculated from the plot between t/q_t versus t.

3.1 Characterization of silica

The XRD pattern of nanosilica is shown in Figure 2A. A strong broad peak centered at 22° (2θ) confirms the amorphous nature of silica. Due to a three-step thermochemical process, the crystalline particles transform into amorphous silica, which is indicated by broad peak. The FTIR spectrum of nanosilica is presented in Figure 2B. A large broad band at 3440 cm^-1 is assigned to the presence of the OH stretching frequency of the silanol group (35). A strong intense band at 1100 cm^-1 is associated to the siloxane Si-O-Si vibration of the molecules (36). Amorphous silica exhibits a relatively strong peak around at 804 cm^-1, and the peak at 475 cm^-1 is due to Si-O bending mode of vibration (37).

Figure 2:

X-ray diffraction (XRD) and Fourier transform infrared (FTIR) spectra of nanosilica obtained from soju industrial bio-residue ash (SIW-A).

The N₂ adsorption-desorption isotherm at -195.8°C (Figure 3A) was carried out for vacuum freeze dried silica powder. The results showed that the isotherm is of type IV, indicating that the samples were mesoporous (i.e. they possess pore diameters in the range of 2–50 nm) in nature. The hysteresis loop is of type H1, indicating the presence of agglomerates with cylindrical pores that have narrow pore distribution (38, 39). The pore size distribution was analyzed, and the average pore width was found to be 34 nm in the insert of Figure 3A. The Barret-Joyner-Halenda (BJH) cumulative surface area of silica powder is 235.6 m²/g, and the BET surface area is 175.5 m²/g. The BJH method has a relatively higher area compared to BET analysis, indicating that the material is of a highly porous nature. The surface areas were similar to that of commercial silica and that reported for silica nanoparticles from different agricultural wastes such as rice straw and wheat straw (26, 40).

Figure 3:

Characterization of synthesized nanosilica. (A) N₂ adsorption-desorption isotherm at 77 K, and pore size distribution from adsorption isotherm using the Barret-Joyner-Halenda (BJH) method of nanosilica, (B) field emission scanning electron microscopy (FE-SEM) micrograph showing aggregation of nanosilica particles due to H-bonding and chemical composition of nanosilica on energy dispersive spectroscopy (EDS) profile, and (C) transmission electron microscopy (TEM) micrograph showing individual sphere-shaped silica nanoparticles around 14–18 nm.

SEM in combination with energy dispersive spectroscopy (EDS) was used to investigate the morphological property of individual particles and the percentage of the elements present. SEM images and EDS of silica are given in Figure 3B. The agglomeration of the particles could be due to hydrogen bonding formed by the -OH groups on the surface of the particle (41). The results from SEM with EDS suggested that spherical particles consist mainly of silica (99%) (Figure 3B). TEM analysis further confirmed the morphology of synthesized nanosilica as spherical in shape, and it exhibits the agglomerated particle with uniform size in the range of 14–17 nm (Figure 3C). The presence of the nano-sized sphere-shaped particles leads to a highly porous structure.

3.2 Characterization of RA adsorbent before and after adsorption

The BET surface area, total pore volume, and average pore diameter of RA are found to be 411.4 m²/g, 0.456 cm³/g, and 4.019 nm, respectively. The average pore diameter of the RA indicates that the adsorbent obtained is in the mesoporous region according to IUPAC, 1972 (42). The development of a high surface area and porosity of the adsorbent is associated with gasification reactions occurring during the physiochemical process (43).

X-ray fluorescence analysis done for mineral content in SIW-A and RA is presented in (Table 1). The SIW-A contains 72% SiO₂ with some impurities (K, Al, Fe, Mg, Ca). After silica extraction, SiO₂ concentration was largely reduced compared to other elemental composition in the remaining RA. The results are consistent with previous studies which reported a major decrease in the SiO₂ concentration after silica extraction (37). The FTIR spectra of three samples (SIW and RA before and after adsorption) are presented in Figure 4A. The broad peaks from 3800 cm^-1 to 3036 cm^-1 indicate the stretching of O-H and N-H functional groups. The peaks at 1739 cm^-1 and 1639 cm^-1 in SIW represent the stretching of C=O functional groups. In addition, stretching of C=C and benzene groups were observed at 1508 cm^-1 and 1581 cm^-1. A peak at 1438 cm^-1 indicates the bend of C-O from inorganic carbonate compounds. The N-O stretch was found at 1381 cm^-1, whereas 1059 cm^-1 and 558 cm^-1 band widths had some molecules containing S=O bonds and silica Si-O-Si bend, respectively. The S=O and Si-O-Si bandwidths were present in RA but not in SIW. The peaks at 3402 cm^-1, 1608 cm^-1, 1059 cm^-1, and 558 cm^-1 are shifted to 3389 cm^-1, 1597 cm^-1, 1044 cm^-1, and 547 cm^-1, respectively, suggesting the involvement of different functional groups on the adsorption of MB. The surface morphology of RA has been characterized by SEM (Figure 4B). The representative micrograph showed the presence of heterogeneous pores, providing a good possibility for dyes to be trapped and adsorbed. However, after adsorption, a significant change was observed on the surface of the adsorbent, which indicates that the dye molecules entered the pores and interacted with RA.

Figure 4:

Characterization of adsorbent before and after adsorption of dye. (A) Fourier transform infrared (FTIR) spectra of soju industrial bio-residue (SIW) and residual ash (RA) before and after adsorption of methylene blue (MB) and (B) scanning electron microscopy (SEM) micrographs of RA before and after adsorption of MB.

3.3 Factors effecting adsorption process

3.3.1 Effect of pH

The initial pH of the dye solution is an important parameter to determine the adsorption capacity of the adsorbent (4). The effect of pH on MB adsorption onto RA is shown in Figure 5A. The adsorption capacity was low at pH 2 and 4, due to the presence of high H+ ions that compete with the cationic ions of dye. Moreover, at lower pH (2 and 4), electrostatic attraction becomes weakened as the carboxylate groups on the polymeric chains may convert to the carboxylic groups, and subsequently adsorption capacity was decreased. The MB sorption rate was increased at pH 6 and remained almost constant up to pH 12. The negatively charged absorption sites were generated in the adsorbent when the pH increased from 6 to 10, which enhanced the positively charged MB cations by means of electrostatic forces of attraction (4). This trend show resemblance with adsorption of MB onto Ficus carica bast (44) and yellow passion fruit peel (45). Thus, the adsorption behavior of RA at various pH values suggests that it can be potentially applied in a wide pH range.

Figure 5:

Isothermal study at different parameters. (A) Effect of initial pH on the adsorption of methylene blue (MB) onto residual ash (RA): adsorbent dosage=1 g/l; initial MB concentration=100 mg/l; temperature=30°C; agitation rate=100 rpm; volume of MB solution=100 ml, (B) removal percentage versus initial concentration (C) Langmuir isotherm for adsorption of MB on RA at temperature 30°C.

3.3.3 Effect of contact time and adsorption kinetics

There was a clear impact of contact time on the adsorption of MB onto RA as shown in Figure 6A. A rapid adsorption of dye molecules was observed in the first 30 min and gradually decreased with time until it reached equilibrium. The increased activity at the initial stage could be due to the availability of more reactive sites, and gradual occupancy of these sites reduced the interaction between MB and RA. The adsorption reached equilibrium at 210 min for an initial concentration of 50–100 mg/l and 280 min for an initial concentration of 200–400 mg/l. The time required to reach this equilibrium is the equilibrium time associated with these specific conditions. The amount of dye adsorbed at the equilibrium time was the maximum dye adsorption capacity of the adsorbent. The maximum adsorption capacity of RA was compared with different adsorbents for MB and reported in Table 3. In comparison, the results suggest that RA is an effective adsorbent for the removal of MB from aqueous solution.

Figure 6:

Kinetic studies at different parameters. (A) Effect of contact time and initial methylene blue (MB) concentration on the adsorption of MB onto residual ash (RA): adsorbent dosage=1 g/l; initial pH=7; temperature=30°C; agitation rate=100 rpm; volume of MB solution=100 ml, (B) pseudo-second-order kinetics for the adsorption of MB onto RA at temperature 30°C, (C) intraparticle diffusion plots under different initial MB concentration: adsorbent dosage=1 g/l; initial pH=7; temperature=30°C; agitation rate=100 rpm; volume of MB solution=100 ml.

Table 3

The comparison of maximum monolayer adsorption of methylene blue (MB) of various adsorbents investigated recently.

Adsorbents	MB adsorption capacity (mg/g)	References
Sodium chloride treated kapok fiber	105	(50)
Kenaf fiber char treated with HCl	18	(51)
NaOH modified rejected tea	242	(49)
Broad bean peels	192.72	(52)
Garlic peel	82.64	(7)
RA (SIW)	232.5	This study

RA, residual ash; SIW, soju industrial bio-residue.

The contact required for an adsorption process can be obtained through kinetic studies. The value of k₁ is the pseudo-first-order rate constant obtained from the plot of log (q_e-q_t) versus t, which gives a straight line. The first-order equation of Lagergren gives a straight line with poor regression coefficient (R²<0.83) and significantly lower values of equilibrium adsorption capacity q_e than the experimental q_e. As the values obtained indicated the inapplicability of this model, the data were modeled using pseudo-second-order kinetics.

The value of k₂ (g/mg h) is the pseudo-second-order rate constant, and we (mg/g) was calculated from the plot between t/q_t versus t using Eq. [7], and is represented in Figure 6B. The straight line provided by the plot is used to calculate the rate constant k₂, the correlation coefficient R², and the equilibrium adsorption capacity (q_e) (Table 4). Evidently, all of the kinetic parameters, including R² of the pseudo-second-order model, were >0.99, which is much higher than the pseudo-first-order. The experimental q_{e, exp} adsorption data were very close to the theoretical calculated q_{e, cal} and indicate that the pseudo-second-order model is the better fit and represents the adsorption process as being controlled by chemisorption. The results are consistent with a previous study reporting the chemisorptions of MB onto spent tea leaves (17).

Table 4

Pseudo-second-order kinetic parameters for the adsorption of different initial methylene blue (MB) on residual ash (RA).

Initial conc.	qe, exp (mg/g)	Pseudo-second-order kinetic model
		k2 (×103 g/[mg min])	qe, cal (mg/g)	R2
50	42.26	1.4	44.44	0.99
100	77.32	0.21	91.70	0.97
200	138.12	0.13	158.70	0.99
300	176.26	0.09	204.08	0.99
400	194.78	0.14	212.76	0.99