Introduction

Bisphenol A (BPA) which has the chemical name 2,2-(4,4-dihydroxy phenyl) propane is an estrogen-like endocrine disrupting chemical (EEDC) with two unsaturated phenolic rings (Jafari et al. 2009). It is an important industrial chemical widely used in polycarbonated plastic as well as epoxy resin, both of which are very practical materials often present in food containers (Staples et al. 1998). The effluents from polycarbonate factories have been found to contain high amounts of BPA around 100 mg/L (Yeo and Kang 2006; Ivanov et al. 2016). BPA either enters the environment during the factory processing and thereby pollutes the rivers and groundwater, sediments in the soil and accumulates there or it can be discharged as a waste product into surface and drinking water (Inoue et al. 2008; Santhi et al. 2012).

BPA-containing wastewater can be a source of contamination in aquatic environments and BPA is hazardous to aquatic systems since it threatens the life of aquaculture through disruption of endocrine glands (Kuramitz et al. 2001; Soares et al. 2008). Low doses of BPA have been reported to cause hypertension, proliferation of prostate cancer cells, heart disease, diabetes, and abnormalities in liver enzymes in humans (Vandenberg et al. 2007; Hugo et al. 2008).

Therefore, the removal of BPA from aquatic environments is of great significance and several methods have been applied for the treatment of water containing BPA, including physical/chemical adsorption (Joseph et al. 2011; Fan et al. 2011), ultrafiltration (Heo et al. 2012), photoelectrocatalytic (Brugnera et al. 2010), biological reduction (Chen et al. 2006), ultraviolet light irradiation (Irmak et al. 2005), catalyzed oxidation method (Coleman et al. 2005) and separation by membrane process (Wu et al. 2010). Since most of these methods have certain disadvantages such as high cost, the formation of hazardous by-product, low efficiency, etc., adsorption process has been found to be very effective for the removal of phenolic compounds (Ahmaruzzaman 2008; Rastegari et al. 2016). Many adsorbents including lignin (Han et al. 2012), activated carbon (Nakanishi et al. 2002), carbon nanotubes (Pan et al. 2008), chitosan (Dehghani et al. 2016) have been investigated for the removal of BPA.

Recently, nanoparticles have proven to be good adsorbents for the removal phenolic compounds from contaminated water since they possess large surface area. Iron oxide has attracted attention in adsorption process because they are environmental friendly (Zou et al. 2016). In fact iron oxides especially magnetite nanoparticles have been used as adsorbents to remove phenolic compounds from water (Shirsath and Shrivastava 2015). Magnetite is blessed with outstanding sorption capacity, separation property, small size and low toxicity (Wang et al. 2016). Magnetite composites have been reportedly used for the removal of BPA to facilitate the separation of adsorbents from solutions after adsorption (Zhang et al. 2014; Park et al. 2015; Balci and Erkurt 2017). Magnetite has also been applied as a catalyst for catalytic degradation of BPA (Tang et al. 2011; Du et al. 2016). In this work, the efficiency of removal of BPA from water by magnetite through adsorption process was studied. The factors affecting the adsorption process were also investigated to determine the optimal conditions for the sorption of BPA onto magnetite nanoparticles.

Experimental procedures

The materials used for this research are of analytical grade and include bisphenol A (99% Purity from Sigma Aldrich)), ferrous chloride, ferric chloride, acetone and sodium hydroxide was obtained from Chemical store in University of Ilorin, Nigeria, supplied by Merck.

Reverse co-precipitation method was employed for the synthesis of the magnetite nanoparticles. This was done by dissolving 21.26 g (0.02 mol) of FeCl3.6H2O and 7.95 g (0.01 mol) of FeCl2.4H2O in 20 ml deionized water at 25 °C for 20 min. 8 mol/L of alkaline solution was prepared by adding 13.74 g of NaOH to 42 mL of deionized water and stirred for 30 min at 25 °C. Magnetic nanoparticles were then precipitated by adding the iron salts solution drop wisely into the NaOH solution at 80 °C with continuous stirring. The co-precipitated samples were then washed with deionized water and acetone to remove the NaCl (Hamid et al. 2012).

The synthesized material was then characterized with appropriate instrumentations. Phase composition of the material was analyzed using X-ray diffractometer. The mean crystallite size of the samples was calculated by XRD line-broadening technique using Scherrer formula. Scanning electron microscopy (SEM) was used to determine the surface morphology of the sample, X-ray fluorescence (XRF) examined the elemental composition and Fourier transform infrared (FTIR) was used to determine the functional groups present.

The magnetite particles were then applied as an adsorbent to remove BPA from aqueous solutions using the batch adsorption process. This was done by introducing a specific mass of the magnetite (0.1 g) to different concentrations of BPA (10, 20, 50, 75 ppm). The solutions were then placed in the rotary shaker for 45 min at 30 °C. The concentrations of BPA left in the solutions after removing the magnetite were then determined using the UV spectrophotometer. The quantity of BPA adsorbed onto magnetite was then estimated using Eq. 1:

$$q_{\text{e}} = \frac{{C_{\text{o}} - C_{\text{e}} }}{M}{\text{XV,}}$$
(1)

where C0 is the initial sorbate concentration (mg/L), Ce the equilibrium adsorbate concentration (mg/L), V is the volume of solution (L) and M is the mass of the adsorbent (g).

The experimental data from the sorption studies were modeled using the Langmuir, Freundlich and Temkin isotherms. The Langmuir which predicts monolayer coverage of the adsorbate molecules on the outer surface of the adsorbent has the linear form as shown in Eq. 2 (Adekola et al. 2014):

$$\frac{{C_{\text{e}} }}{{q_{\text{e}} }} = \frac{{C_{\text{e}} }}{{q_{\text{m}} }} + \frac{1}{{q_{\text{m}} K_{\text{L}} }},$$
(2)

where qm (mg/g) is maximum monolayer coverage related to adsorption capacity of the adsorbent while kL (L/mg) is Langmuir constant related to rate of adsorption. The constants are calculated from the slope and the intercept of the plot of Ce/qe against Ce.

The Freundlich isotherm describes multilayer adsorption on heterogeneous surface of adsorbent (Adekola et al. 2014). Equation 3 is the linear form of Freundlich isotherm:

$$Lnq_{e} = {\text{ LnK}}_{f} + 1/n{\text{ LnCe,}}$$
(3)

where Kf (mg/g) is the Freundlich constant indicating adsorption capacity and n is the adsorption intensity. Both constants are deduced from the slope and the intercept of the plot of Lnqe vs LnCe.

Temkin isotherm can be used to determine the interaction between the adsorbate molecules and adsorbent and the linear form is presented in Eq. 4:

$$q_{\text{e}} = B{\text{ Ln}}A \, + B{\text{ Ln}}C_{\text{e}} ,$$
(4)

where A (L/g) is the Temkin constant related to maximum binding energy and B (J/mol) is the Temkin constant related to heat of sorption which are determined from the plot of qe vs LnCe.

The pseudo-first-order kinetics, pseudo-second-order kinetics and Elovich models were used to study the kinetics of the adsorption process.

Results and discussion

Characterization

The result of the XRF analysis of the synthesized material is presented in Table 1. The result which revealed the elemental compositions of the sample indicated that Fe3O4 is predominantly present in the sample with a percentage of 94.4% along with other metal oxides in trace amount as impurities. This suggested that the synthesized material is indeed an oxide of iron.

Table 1 XRF result of the synthesized magnetite
Full size table

The results of SEM, XRD and FTIR are presented in Fig. 1a, b, c, respectively. XRD is used to know the diffraction pattern for particles size. The mineralogical composition of the magnetite was determined by the X-ray diffraction analyses for the qualitative evaluation of the common and predominant phases within the sample.

Fig. 1
figure 1

a XRD image of synthesized magnetite (Fe3O4). b The SEM image of synthesized of magnetite (Fe3O4). c The FTIR image of synthesized magnetite (Fe3O4)

Full size image

The XRD patterns of synthesized magnetite has characteristic peaks of magnetite which appeared at 29.5, 36.0 43.5 57.5, 62.5 corresponding to the (220), (311), (422), (511) and (440) crystal planes of a pure Fe3O4 with a spinal structure (JCPDS file PDF no. 65-3107). Characteristic peaks of impurities were absent and this result revealed that the synthesized material is magnetite (Li et al. 2008).

The SEM image revealed the particles are agglomerated and the surface is coarse with pores which indicated the suitability of the material for adsorbent. The agglomeration could be due to forces such as quantum size-effects.

From the FTIR spectra of the synthesized magnetite has an absorbance peak at 565.14 cm−1 which is attributed to Fe–O bond as a result of the formation of ferrite phase. The peak at 3423.65 cm−1 corresponds to –OH vibration which could be due to the presence of moisture in KBr used for the analysis (Shirsath and Shrivastava 2015).

Effect of pH and adsorbent dose on adsorption

The effect of the BPA solution pH and amount of magnetite used are shown in Fig. 2a, b, respectively. The studies were carried out with 50 ppm of BPA solution for 60 min. The adsorption process was found to be dependent on the pH of the BPA solution (Fig. 2a). It can be seen that the there was slight increase in the quantity of BPA adsorbed over the acidic pH range. The quantity adsorbed falls from pH 6 to pH 8 and continues to decrease over the basic pH range. This is mostly likely due to the pKa value of BPA and the fact that at basic pH range, the hydroxyl ions compete with BPA molecules for active sites on the magnetite nanoparticles (Xiao et al. 2012; Jin et al. 2015). Low adsorption capacity observed that lower and higher pH values could also be attributed to the fact that magnetite nanoparticles are quite unstable at higher and lower pH. Equilibrium adsorption was also observed with adsorbent dose of 0.1 g (Fig. 2b).

Fig. 2
figure 2

a Effect of pH on the sorption of BPA onto nFe3O4. b Effect of adsorbent dose on the sorption of BPA onto nFe3O4

Full size image

Adsorption isotherms

The amount of BPA adsorbed increased with increase in concentration up to an optimum concentration of 50 ppm where the amount adsorbed was observed to be 4.511 mg/g (Fig. 3a). The rapid adsorption at the initial stage was most likely due to the large number of available active sites and high driving force, both of which made BPA molecules transfer rapidly the surfaces and pores of magnetite particles. Shortly afterwards, a drop in the quantity of BPA adsorbed was observed, which could be ascribed to the fewer available adsorption sites and the decrease in the driving force (Jiang et al. 2015). The optimum concentration of 50 ppm was used for further experiment.

Fig. 3
figure 3

a Effect of concentration on the sorption of BPA onto nFe3O4. b Langmuir isotherm plot for the sorption of BPA onto nFe3O4. c Freundlich isotherm plot for the sorption of BPA onto nFe3O4. d Temkin isotherm plot for the sorption of BPA onto nFe3O4

Full size image

The isotherm plots for the adsorption process are shown in Fig. 3b–d, while isotherm parameters calculated from the intercept and slope of the plots are shown in Table 2. The extent of the fitness of data with an isotherm model is determined by the regression value (R2) of the plot. The experimental data fit best the Langmuir isotherm model with R2 value of 0.957 which suggests monolayer coverage of the adsorbent surface. The Langmuir constant RL value of 0.179 indicates that adsorption process is favorable since it less than 1 (Radu et al. 2015). The Freundlich constant ‘n’ further suggests that the adsorption is favorable. The Temkin isotherm plot has the least regression value of 0.862.

Table 2 Isotherm parameters for the sorption of BPA onto Fe3O4
Full size table

Adsorption kinetics

The uptake of the BPA molecules by the magnetite particles was found to be very rapid within the first few minutes and the uptake reached equilibrium after 60 min (Fig. 4a). The rapid uptake of BPA can be due to the availability large amounts of adsorption sites in the adsorbent, as the contact time prolongs, there was reduction in the available adsorption sites (Zeng et al., 2006).

Fig. 4
figure 4

a Effect of time on the sorption of BPA onto nFe3O4. b Pseudo-first-order kinetic plot. c Pseudo-second-order kinetic plot. d Elovich kinetic model plot

Full size image

The pseudo-first-order, pseudo-second-order and Elovich kinetics plots are presented in Fig. 4b, c, d, respectively, while the parameters calculated from the intercept and slope of the plots are shown Table 3. The experimental data fit the pseudo-second-order kinetics best with R2 value of 0.980. In fact, the experimental quantity adsorbed was found to be very close to the calculated quantity adsorbed and this revealed that the process can best be described by pseudo-second-order kinetics model. This finding is similar to those reported earlier in the literature (Dehghani et al. 2016).

Table 3 Kinetics parameters
Full size table

Thermodynamic studies

From Fig. 5a, the effect of temperature on the adsorption process can be seen. The temperatures studied are 30, 35, 40, 45, 50, 55 and 60 °C. The highest adsorption of BPA onto magnetite was observed at 30 °C after which the quantity adsorbed decreased slightly. This shows that the process is exothermic and favorable with lower temperature (Liu et al. 2009).

Fig. 5
figure 5

a Effect of temperature on the sorption of BPA onto nFe3O4. b Thermodynamic plot

Full size image

The plot of the thermodynamic studies is shown in Fig. 5b and the values of the thermodynamic parameters are presented in Table 4. The thermodynamic parameters give indications about internal energy changes during the adsorption process. From the results, the enthalpy change ΔH was found to be – 26,396.95 Jmol−1. The negative value of ΔH indicated that the process was exothermic and this was evident in the decrease of adsorption efficiency with increase in temperature. The Gibbs free energy ΔG was found to be negative at lower temperatures which revealed that at these temperature the adsorption process is feasible and spontaneous. Similar findings have been reported by Radu et al. (2015).

Table 4 Thermodynamic parameters
Full size table

Conclusions

The removal of BPA from water by magnetite through adsorption process was studied. The magnetite was synthesized using reverse co-precipitation method and fully characterized. The XRF results revealed that iron oxide is the major component of the synthesized material and both XRD and FTIR results confirmed the sample to be nano-magnetite. Various physicochemical parameters affecting the adsorption of BPA using magnetite were studied as well. The optimum time for the adsorption process was found to be 60 min at pH of 6, adsorbent dose of 0.2 g and 50 ppm of BPA. The adsorption data fit the Langmuir adsorption isotherm best with a regression value of 0.957. The RL value was 0.179 which revealed that the process is favorable. The Freundlich constant n which was 1.901 also revealed that the adsorption is normal and favorable. The data were in agreement with the pseudo-second-order kinetics with regression value of 0.98 and the experimental quantity adsorbed is close to the calculated quantity adsorbed. From the thermodynamic studies, the process was found to be exothermic and the Gibb’s free energy value which was negative showed that the adsorption was spontaneous. The synthesized magnetite therefore offers great potential for the remediation of bisphenol A-contaminated media.