Valorization of human hair as methylene blue dye adsorbents

2.1 Preparation and characterization of human hair adsorbents

Human hair was cut to a size of approximately 2 cm. Then it was soaked in distilled water, and boiled for 15 min to remove impurities. This step was repeated until the washed water becomes clear. Then the sample was oven-dried at 110°C for 24 h. After that, the dried sample was immersed in 0.1 m HNO₃ for 24 h, and then it was washed repeatedly using hot water until there is no further change in the solution pH. Similar step was repeated for the basic-treated human hair adsorbent using 0.1 m KOH. No noticeable change in the shape or form of human hair was observed upon the treatment strategies. The three adsorbents, i.e. untreated, acid-treated, and basic-treated were designated as UH, AH and BH, respectively. All adsorbents were dried prior to use.

Human hair was characterized for its thermal behavior using a Perkin-Elmer thermogravimetric analyzer (TGA 7). The measurement was carried out under a nitrogen flow of 20 ml/min at a heating rate of 10°C/min. All adsorbents were analyzed for specific surface area, surface functional groups, and elemental composition. The measurement of single-point specific surface area was performed by Brunauer-Emmett-Teller method using a Micromeritics surface analyzer (Gemini 2360). The presence of surface functional groups was determined using a Shimadzu spectrometer (IRTracer-100). The energy dispersive spectrometry (EDS) was performed to obtain the elemental composition of human hair using a Hitachi field emission scanning electron microscope coupled with an EDS detector (SU8020). The measurements were performed twice and the average data were reported.

2.2 Adsorption studies

Thirty milligrams of the adsorbent was added in a series of conical flasks containing 30 ml of MB solution with concentrations between 2 and 50 mg/l. The solution pH was not adjusted, and was measured as pH 5.7±0.1. The control solution was prepared for reference. The mixture was allowed to equilibrate at room temperature for 72 h. After that, the supernatant was taken out for concentration measurement using a Dynamica visible spectrophotometer (Halo Vis –10) at a wavelength of 564 nm. The calibration was determined as a.u.=0.0176×concentration, with R² = 0.991 for MB concentrations between 2 and 25 mg/l. The capacity of adsorption at equilibrium, q_e (mg/g) was calculated as q_e = (C_o–C_e)×(V/m), where C_o and C_e (mg/l) are the initial concentration and concentration at equilibrium, respectively, m (g) is the adsorbent mass, and V (l) is the volume of solution. The removal efficiency (%) was calculated as (C_o–C_e)×100/C_o. The effect of concentration was used to screen the removal performance by the three adsorbents; the resulting data were used to select the one used for the subsequent parameters evaluation which is described as follows.

The effect of pH was evaluated at MB concentrations of 5 and 10 mg/l, in which the solution pH was adjusted between pH 2 and 10 using 0.1 m NaOH or 0.1 m HCl. The concentrations were selected based on the equilibrium curve that demonstrates appreciable adsorption affinity. The control solution with the same adjusted pH was prepared for reference. The mixture was allowed to reach the equilibrium for 72 h, after which the concentration and pH at the equilibrium were measured.

Three concentrations were selected for the rate of adsorption and the effect of temperature based on appreciable adsorption affinity and maximum capacity. Thirty milligrams of human hair adsorbent was brought into intimate contact with a series of 30 ml MB solution of known concentrations. The supernatant was withdrawn from the solution at the preset time intervals for the concentration measurement. The adsorption capacity at time t, q_t was calculated as q_t = (C_o–C_t)×(V/m), where C_t (mg/l) is the measured MB concentration at time t. The adsorption of MB at varying temperatures was performed by adjusting the water bath temperature at 35°C, 45°C and 55°C. The control solution at the same tested temperature was employed for reference. Other experimental steps are identical as described for the effect of concentration.

3.1 Characteristics of human hair adsorbents

Figure 1 displays the thermal profile of human hair. The initial weight loss of 7% from 33°C to 130°C is due to the release of the physically adsorbed water. The sharp peaks at 249°C and 345°C with 50% weight loss could be attributed to the denaturation of keratin polypeptide chains that make up the morphologies and structures of human hair [21]. Denaturation is a process whereby the proteins lose their original structure, become more simplified, and start to degrade with increasing temperature. The degradation of carbonic chains takes place at a temperature range of 345°C–470°C [22], while a complete degradation of keratin and hair structures was observed at 700°C with residual of 20%.

Figure 1:

Thermogravimetric curve of human hair.

Table 1 shows the characteristics of human hair adsorbents. The adsorbents exhibit similar yield of 92±1% with pH varying between 4.95 and 5.44. The basic-treated human hair adsorbent (BH) displays a higher specific surface area of 3.51 m²/g upon treatment with KOH, while a small increase was observed for HNO₃-treated one (AH). It is suggested that the basic treatment could be effective to flush the impurities (probably trace oils) on the hair surface, thus exposing the hair structures and also the functional groups. This could be supported by a slight decrease in yield, and an increase in oxygen content (EDS data) for BH.

The infrared penetrates the outer surface of hair, known as cuticle that is primarily composed of cystine, and other amino acids which are not usually found in alpha helical polypeptides [21]. Cystine is an oxidized dimer of amino acid. It contains a disulfide bond (-S-S-) that binds two cystine molecules together, carboxyl group (-CO-OH), and amino group (-NH₂).

From the Fourier transform infrared analysis, all adsorbents exhibit similar spectra signifying the presence of identical surface functional groups. The broad and medium intensity band ranging from 3600 to 3000 cm⁻¹ corresponds to the stretches of carboxylic acid (-CO-OH), phenolic (-OH), and amino acid (-NH₂) groups. The peaks located at 1640 cm⁻¹ (Amide I), 1520 cm⁻¹ (Amide II), 1240 cm⁻¹ (Amide III), and 900 cm⁻¹ (Amide V) are the characteristics of amino acids, while peaks at 1040 cm⁻¹, 1080 cm⁻¹, and 1120 cm⁻¹ are associated with distinct sulfur groups of cystine. The presence of these functional groups is in agreement with the typical chemical constituents of human hair. The variation in peak intensity between the treated human hair-based adsorbents and the untreated one (UH) was not observed.

3.2 Adsorption equilibrium

Figure 2 shows the effect of initial concentration of MB dye on the adsorption performance of human hair as adsorbents. Generally, the adsorption capacity increased with increasing concentration. All adsorbents exhibit a concentration-dependent adsorption due to concentration gradient that provides the driving force to suppress the adsorbent mass transfer resistance for adsorption to take place. To a point of surface saturation, the active sites can no longer accommodate the dye molecules, and consequently the removal capacity begins to level off. The adsorption capacity was found to be in the order of BH>UH>AH. This could be explained by the characteristics of BH as displayed in Table 1. BH shows a maximum removal efficiency of 80% (q_e = 10 mg/g) at C_o = 13 mg/l, while a higher capacity of 14 mg/g was obtained at C_o = 42 mg/l with a decreased efficiency of 33%. The decrease in the removal efficiency is linked with the increase in residual concentration upon equilibrium due to the decrease of vacant sites with increasing concentration. Similar pattern was also demonstrated by AH and UH, with maximum removal efficiency of 22% (q_e = 3.6 mg/g, C_o = 17 mg/l) and 30% (q_e = 3.8 mg/g, C_o = 13 mg/l), respectively.

Figure 2:

Effect of concentration on methylene blue removal by human hair adsorbents (bars: adsorption capacity; lines: removal efficiency).

Figure 3 shows the equilibrium curves of MB adsorption onto human hair adsorbents. The concave downward pattern indicates a favorable MB adsorption onto human hair adsorbents. The maximum adsorption capacity of BH was recorded as 13.5 mg/g, while AH and UH shared a similar capacity of 3.4 mg/g. BH exhibits a comparable removal performance with orange peel (surface area = 1.17 m²/g, Q_m = 18.3 mg/g) [16] and Prosopis sicigera L. wood (surface area = 121 m²/g, Q_m = 25.1 mg/g) [17].

Figure 3:

Equilibrium adsorption of methylene blue onto human hair adsorbents (lines were predicted by, solid: S-shaped isotherm; dashed: Langmuir isotherm).

The adsorption data were fitted into two models, namely Langmuir [Eq. (1)] [23] and S-shaped isotherm [(Eq. (2)], as follows,

where Q_m (mg/g) is the maximum adsorption capacity, b (l/mg) is the adsorption affinity, and m_e (mg/l) is the midpoint equilibrium concentration. The isotherm constants were solved by nonlinear regression using Solver-add in MS Excel for the least sum-of-squared error (SSE) that yields the optimum coefficient of determination, R². Table 2 summarizes the values of isotherm constants. The adsorption data fitted well into S-shaped isotherm. This is true for all adsorbents studied. It implies that the adsorption did not follow an ordinary Langmuir isotherm that describes a monolayer adsorption onto homogeneous adsorbent surface [23]. Rather, the S-shaped isotherm suggests a cooperative adsorption, whereby the already adsorbed dye molecules formed a layer to attract the mobile molecules in the bulk solution to be captured as well [24]. It is evident from Figure 3, where the uptake capacity is initially small at low equilibrium concentration to a point where the capacity displays a sudden steep increase with a small increase in concentration which renders a maximum value of adsorption capacity.

The S-shaped isotherm satisfactorily predicts the maximum adsorption capacity from the experimental data (Table 2). Despite having an identical maximum capacity, AH reaches the same point at a higher equilibrium concentration compared to UH. It could be due to the protonated surface of AH (pH = 5.0, pH_e = 5.2) that somewhat hinders and drives away the positively charged MB molecules at lower concentration, leading to a lower removal efficiency. However, this was somewhat overcome as the concentration increases because of the increase in the interaction probabilities.

The adsorption of MB onto human hair could be driven by surface functional groups. During KOH treatment, oil covering the hair surface is removed, thus exposing the surface functional groups and thereby increasing the specific surface area. The main component in hair is cystine, while the probable functional groups for adsorption are amine (-NH₂), phenolic (-OH), and lactonic ( = O). Amine has one lone pair electron, while phenolic and lactonic possess two lone pair electrons. Hence, they are all electron donating groups that are prone to release electron to the oxidizing agent or electron withdrawer/acceptor (MB). However, based on the electronegativity rule, oxygen is more electronegative than nitrogen that implies a stronger hold of electrons by oxygen. Hence, it is presumed that the adsorption of methylene is much easier on the amine groups compared to the oxygen groups. The following mechanisms can be proposed for MB adsorption onto human hair.

Figure 4 shows the effect of solution pH on MB adsorption onto basic-treated human hair adsorbent (BH). For the two concentrations studied, a similar pattern of pH-dependent adsorption was demonstrated, in which the adsorption of MB increased with increasing solution pH. The trend is more pronounced for C_o = 10 mg/l. This could be attributed to the isoelectric point (pI) of human hair which is in the range of 2.45–3.17 [25]. The solution pH affects the adsorbent surface charge through the dissociation of functional groups, and the degree of ionization of dye molecule. The surface charge of human hair is positive when the solution pH is below pI due to the protonated surface by hydrogen ions, while the surface becomes negatively charged when the solution pH is greater than pI. As a result, the electrostatic interaction/attraction between the deprotonated surface and the positively charged MB molecules increased with increasing solution pH [26]. However, extra caution should be taken as the increase of solution pH could in some way affect the wavelength and density of the solution, thus nullifying the concentration value (e.g., for pH 10, before pH adjustment C_o = 10 mg/l, after pH adjustment C_o = 18 mg/l).

Figure 4:

Effect of solution pH on the removal of methylene blue by basic-treated human hair adsorbent.

3.3 Rate of adsorption

Figure 5 displays the rate of MB adsorption by BH at three selected concentrations, i.e. 5, 13, and 25 mg/l. A rapid adsorption was observed in the first 2.5 h, after which the removal capacity starts to level-off at different values depending on the concentrations used. The initial fast removal is due to high concentration gradient and unoccupied active sites on the adsorbent surface. As the MB molecules start to lodge on the active sites, the remaining molecules in the bulk solution become lesser thus decreasing the rate of adsorption (dC_t/dt→0).

Figure 5:

Rate of methylene blue adsorption by basic-treated human hair adsorbent (Lines were predicted by pseudo-first-order and S-shaped kinetics models).

The quasi-equilibrium was observed at around 6 h, with q_t = 13.5, 8, and 3.5 mg/g with descending concentrations. Later at t>17 h, a jump in MB adsorption was detected for concentrations 13 and 25 mg/l, thus creating an additional adsorption isotherm. The q_e values obtained after 17 h are in agreement with the data shown in Figures 2 and 3. From Figure 5, it can be generalized that the actual equilibrium for higher concentrations could be attained at t>17 h, while a shorter period may be sufficient for lower concentrations. This also supports the hypothesis of cooperative adsorption as described by the S-shaped isotherm earlier.

The kinetics data were evaluated using the pseudo-first-order [Eq. (5)] [27], pseudo-second-order [Eq. (6)] [28], and intraparticle diffusion [Eq. (7)] [29] models for the rate-controlling steps. The kinetics models are given as follows,

where q_t (mg/g) is the amount of MB adsorbed at time t (h), k₁ (h⁻¹), k₂ (g/mg·h), and K_p (mg/g·h^0.5) are the rate constants for the pseudo-first-order, pseudo-second-order and intraparticle diffusion models, respectively, and C (mg/g) is the intercept that gives an idea about the boundary layer thickness. A generic model, namely S-shaped kinetics [Eq. (8)] was also introduced to represent an increase in the adsorption capacity at t>17 h for C_o = 13 and 25 mg/l. It is expressed as,

where q_e (mg/g) is the equilibrium capacity, k_S (h⁻¹) is the rate constant, and m_t (h) is the midpoint of time. All kinetics constants were solved using Solver add-in of Microsoft (MS) Excel, and the values are summarized in Table 3.

From Table 3, the kinetics data for t<17 h were well fitted into the pseudo-first-order model with reasonably high R², and a good approximation of q_e (or quasi q_e for C_o = 13 and 25 mg/l) to the experimental data. The applicability of this model for the stated adsorption period indicates that the external diffusion is a significant step [4]. Also, the rate constant, k₁ increased with increasing concentration, signifying a faster adsorption at a higher concentration, and therefore the favorability of adsorption as the concentration increases. This could be seen from the change in the gradients of q_t vs. t as shown in Figure 5. Depending on the availability of active sites, the adsorbent can accommodate a higher adsorbate concentration for a higher removal capacity. However, the adsorption rate is likely to slow down due to the inherent repulsion between the neighboring adsorbate molecules, and the competition for the active sites. Because of that, k₁ may as well decline at higher concentration (for C_o = 25 mg/l).

For the whole adsorption period, the pseudo-first-order model could only satisfy the adsorption data of C_o = 5 mg/l. Hence, a generic S-shaped kinetics model was used to describe a sudden increase in the adsorption after quasi-equilibrium at t>17 h for C_o = 13 and 25 mg/l. From Table 3, this model satisfactorily fitted the remaining data with a good prediction of q_e as that of the experimental data (Figure 5). In addition, the values of rate constant, k_S for both concentrations are comparable, but somewhat lower compared to that of k₁. It indicates that the cooperative adsorption onto the already adsorbed dye molecules occurs at a same rate, which is much slower than the initial adsorption by the role of surface functional groups. The possible mechanism for the cooperative adsorption is π–π interaction among the aromatic rings of MB molecules. The aromatic (benzene) ring enables the π-electrons to delocalize, thus prompting a weak attractive force towards its neighboring molecule [30]. The contact may become more effective because the already adsorbed dye molecules had lost their cationic character as a result of anchoring onto the surface functional groups at quasi-equilibrium.

Figure 6 displays the intraparticle diffusion plots. In general, all lines are not linear through origin, where each line can be split into two or three connecting lines of different slopes. Because of that, the intraparticle diffusion is not the only rate-limiting step that controls the adsorption [5]. From Figure 6, the first part of the line indicates the initial stage of adsorption, and reflects the thickness of boundary layer, while the second and the third parts represent the effect of intraparticle diffusion (adsorption) and cooperative adsorption, respectively. Hence, the rate-limiting step (slowest step) could be film diffusion to overcome the adsorbent mass transfer resistant. The time taken to attain the quasi-equilibrium (t<17 h), in general, also infers that the adsorption process is possibly film diffusion-controlled [31]. Therefore, the adsorption of MB onto basic-treated human hair could be based on the following mechanisms: (i) film/external diffusion (slowest step), (ii) intraparticle diffusion (adsorption), and (iii) cooperative adsorption.

Figure 6:

Intraparticle diffusion plots.

3.4 Adsorption thermodynamics

Figure 7 shows the removal of MB dye onto basic-treated adsorbent at different temperatures. Generally, the adsorption slightly increased with increasing temperature, indicating that the process is temperature-dependent and endothermic. The thermodynamics properties, i.e. enthalpy of adsorption (ΔH), Gibbs energy (ΔG), and entropy (ΔS) were calculated using the Gibbs expression as follows,

Figure 7:

Effect of temperature on methylene blue removal by basic-treated human hair adsorbent.

where, K = C_e,a/C_e is the equilibrium constant, C_e,a (mg/l) is the amount of dye adsorbed, T (K) is the absolute temperature of solution, and R = 8.314 J/mol·K is the gas constant. The respective thermodynamics data are summarized in Table 4.

From Table 4, the adsorption of MB onto BH is exothermic at lower concentration, and becomes more endothermic as the concentration increases. The higher the concentration the greater the energy is absorbed during the adsorption. This is in accordance with the adsorption behavior that increased with increasing temperature. Likewise, the adsorption system demonstrates an increasing disorder or randomness because of the accumulation of dye molecules on the surface as the concentration increases. It is shown by the positive ∆S (∆S = 18.1 J/mol·K, C_o = 20 mg/l), which also signifies that the process (or system) is feasible [32]. In general, the negative values of ∆G indicate that the MB adsorption onto BH is spontaneous. However, the spontaneity of the process decreased with increasing concentration. For example, the ∆G increased from –3.1×10⁻³ to –4.6×10⁻² J/mol with increasing concentration from 5 to 20 mg/l at 35°C. This could be partly due to the nature of cooperative adsorption, whereby a concentration gradient is required as the driving force for subsequent increase in capacity after reaching the quasi-equilibrium. Nevertheless, the spontaneity of the process slightly improved as temperature increases to 55°C at C_o = 20 mg/l (∆G decreased from –4.6×10⁻² to –8.2×10⁻² J/mol), signifying the positive effect of temperature as the driving force for adsorption at higher concentration [32, 33].