Adsorption of thallium(I) on rutile nano-titanium dioxide and environmental implications

Materials

Analytical reagent (AR)-grade RNTD was purchased from Aladdin Reagents Co., Ltd. (Shanghai, China). Thallium(I) nitrate (99.5%, metal basis) was purchased from Alfa Aesar, A Johnson Matthey Company. AR-grade NaClO₄, HClO₄, and NaOH were obtained from Guangzhou Chemical Reagent Co., Ltd. (China). All experiments were performed using Milli-Q water (Millipore Corp.).

Batch adsorption experiments

All the batch adsorption experiments were conducted in a rotary shaker (25 ± 0.2 °C, 200 rpm) using a 50 mL PP centrifuge tube. The pH was adjusted to the desired value by using 0.1 mol/L HClO₄ or NaOH. Stock solution of TlNO₃ (100 mg/L) was prepared with deionized water.

First, the influence of RNTD dosage on the adsorption of Tl(I) was studied. For this purpose, the concentration of RNTD was set in the range 0.5 to 5 g/L at desired intervals, using 0.1 mol/L NaClO₄ as the background, to ensure constant ionic strength with pH 7.0 ± 0.3. To investigate the effect of ionic strength on Tl(I) adsorption, an appropriate amount of NaClO₄ was added with the concentration ranging from 0.05 to 3 mol/L at desired intervals. The influence of pH on Tl(I) adsorption was studied in the range of 2 to 11. For isotherm studies, the temperature was set at 298 K, 303 K and 313 K, respectively, and the initial concentration of Tl(I) varied from 0.02 to 20 mg/L. The solutions obtained were filtered rapidly through a 0.45 mm membrane after 300 min, and Tl concentration was measured immediately by adsorption thermodynamics experiments. To evaluate the adsorption kinetics, samples were collected at desired intervals in the range of 0 to 300 min. The concentration of Tl(I) was determined by inductively coupled plasma–mass spectrometry (ICP-MS; NexION, PerkinElmer US).

The adsorption amount and adsorption rate are calculated as follows:

(1)${\displaystyle \text{Adsorption rate}:\alpha \text{\%}=\frac{C0-Cs}{C0}\times 100\text{\%}}$ (2)${\displaystyle \text{Adsorption amount}:Qe=\frac{\left(C0-Cs\right)V}{m}}$

Where, C₀—initial concentration (mg/L);

C_s—remaining concentration of Tl after adsorption (mg/L);

V—solution volume (L);

m—total mass of adsorbent added (g).

Characterization

The morphologies of RNTD were observed by a JSM-7001F scanning electron microscopy (SEM, JOEL, Japan) integrated with an energy-dispersive X-ray spectrometer (Oxford Instruments, UK). To examine surface properties, the FTIR (https://www.sciencedirect.com/topics/chemistry/absorption-spectrum) of TiO₂ nanomaterials were analyzed by using a Bruker (TENSOR27, Germany) FTIR spectrometer in the frequency range of 400–4,000 cm⁻¹, operating with a spectral resolution of 2 cm⁻¹. The crystal phase of the sample was visualized by X-ray powder diffraction (XRD) using an X’Pert Pro Diffractometer (PANanalytical, EA Almelo, The Netherlands) X-ray diffractometry operating at 100 kV and 40 mA, using Cu Ka radiation at a scan speed of 4°/min.

Effect of RNTD dosage

The adsorption effect of adsorbent on Tl is due to the existence of active adsorption sites in the adsorbent. Under a constant initial Tl concentration, increase of the dosage of adsorbent will enhance the contact site, thereby increasing the adsorption amount of adsorbent on Tl (Shen et al., 2009). However, a decrease in adsorption amount with increased dosage was observed (Fig. 1), and the adsorption rate increased from 55.4% to 92% as the dosage increased from 0.5 to 2.5 g/L. It did not continue to rise afterwards due to further addition of the adsorbent will cause difficulty in dispersion, which can hinder Tl adsorption. According to the literature, the mass of TiO₂ particles could increase the sedimentation aggregation (He & Zhang, 2003).

Effect of initial thallium concentration

Initial concentration of Tl in natural systems will also influence Tl adsorption behavior (Zhang et al., 2017). As shown in Fig. 2, at 0.02 to 10 mg/L, as the initial concentration increases, the adsorption rate shows a tendency to decrease unevenly from 70%. From 10 to 20 mg/L, the adsorption rate no obvious change with the initial concentration. Obviously, without changing the amount of adsorbent, if initial Tl concentration in the water is higher, the nano TiO₂ can adsorb more Tl. The adsorption rate no longer decreases at initial Tl concentration of 10 mg/L. From the perspective of cost savings, the Tl concentration of 10 mg/L would be most suitable for adsorption by rutile nano TiO₂, which deserves highest environmental benefits.

Effect of ionic strength

Another important factor affecting Tl(I) concentration in natural systems is ionic strength (Liu et al., 2011). As shown in Fig. 3, with an increase of NaClO₄ concentration, the adsorption rate gradually increased from 33.6% to 86.7% and then tended to reach equilibrium. The adsorption of Tl mainly depends on the surface negative charge of the adsorbent. Increasing NaClO₄ concentration is favorable for adsorption, but excessive concentrations reduce adsorption rate. When NaClO₄ concentration is >1.5 mol/L, the adsorption rate tends to become constant (Vilar, Botelho & Rui, 2005). At this time, the adsorption increases with the increase of the ionic strength or is not sensitive to the ion concentration. It is likely that the inner surface complex (ISSC) is formed between RNTD and Tl, the chemical bond between RNTD and Tl is generally a coordination covalent bond.

Effect of pH

The effect of the pH value of the solution on the adsorption is mainly achieved by changing the charge carried by the adsorbent and the adsorbate, thereby affecting the electrostatic interaction between the adsorbent and the adsorbate (Wu, Joo & Lee, 2005; Wu et al., 2004). As shown in Fig. 4, the pH of the solution exerted a significant role on the adsorption of Tl(I). The adsorption percentage rised from 28.9% to 60.2% when the pH increased from 2 to 11. No obvious change in the adsorption occurred from pH 2 to pH 5. The adsorption amount increased slowly in the pH range from 2 to 5, but very significantly in the pH range from 5 to 9, followed by a plateau at pH 9 to 11. In aqueous solution, Tl⁺ is the predominant species in the pH range of 2–11. The elevated adsorption amount with pH can be owing to pH_pzc of the RNTD adsorbent (Govender et al., 2007; Mahamuni et al., 1999; Vayssieres et al., 2001). At pH values of 2∼5, which are lower than pH_pzc, the surface charge of the RNTD was positive. It facilitated great electrostatic repulsion between Tl⁺ and positively-charged RNTD, which inhibited adsorption of Tl(I) ions. At pH values (6∼9) higher than pH_pzc, the surface charge of the TNTs turned negative, which promoted the adsorption of Tl⁺, due to the electrostatic force of attraction between Tl(I) ions and the surface. For further increase in pH, approximately 60% removal efficiency was achieved when pH was above 9, suggesting a fairy good adsorption performance of RNTD.

Adsorption kinetics

As shown in Fig. 5, the adsorption amount rapidly increased from 2.8 to 3.3 mg/g in the first 5–15 min but did not change significantly further on, indicating a rapid adsorption of Tl(I) by RNTD. The adsorption process was complete within 15 min, The rapid uptake was mainly due to the large amount of active sites on the surface of the RNTD.

In order to better understand the behavior of Tl(I) adsorption and possible controlling mechanism, the kinetics experimental data were fitted with the commonly used kinetic models, such as the pseudo first-order, pseudo second-order, and Elovich models (Liu et al., 2014; Pu et al., 2013).

The pseudo-first-order kinetic model equation is expressed by: (3)${\displaystyle \text{lg}\left(Qe-Qt\right)=lgQe-\left(k1t\right)∕2.303}$

Where Q_e is the adsorption amount (mg/g); Q_t is the adsorption amount (mg/g) at time t; and k ₁ is the first-order adsorption rate constant (g/mg min).

The pseudo-second-order kinetic model equation is expressed as follows: (4)${\displaystyle t∕Qt=1∕k2Q{e}^2+t∕Qe}$ Where Q_e is the adsorption amount (mg/g); Q_t is the adsorption amount (mg/g); k ₂ is the secondary adsorption rate constant (g/mg min).

Elovich dynamic model is expressed by: (5)${\displaystyle Qt=ke\times lnt+C}$ Where Q_t is the adsorption amount (mg/g) at time t; k_e and C are constants; t is the adsorption time (min); and k _e is related to the adsorption efficiency. The greater the k_e value, the higher is the adsorption efficiency.

As listed in Table 1, the kinetic results are fitted by the pseudo-second-order model very well, with a high correlation coefficient (R² = 0.9969). This suggests that the rate-controlling step for the adsorption was the initial diffusion of metal ions from the solution and their subsequent interaction with the -ONa/-OH groups on the RNTD surface, and the subsequent interaction between the -ONa/-OH groups of RNTD and the metal ions.

Adsorption isotherms

With an increase in the initial Tl(I) concentration, the adsorption rate showed a trend of decrease from 70% to 50% (Fig. 6). The lowest adsorption amount was found at the reaction temperature of 313 K, while higher adsorption amounts were observed at lower temperatures, which indicates that the adsorption is exothermic. The adsorption amount of RNTD showed a near-linear dependent on the initial Tl concentration. The adsorption of Tl(I) was reduced by increase of temperature. Differences between adsorption isotherm become smaller at higher temperature. Langmuir and Freundlich adsorption models were used to fit the experimental data of Tl(I).

The equation for the Langmuir adsorption model is (6)${\displaystyle Ce∕Qe=Ce∕Qmax+1∕\left(Qmax\times b\right)}$ where Q_e is the adsorption amount (mg/g); Q_max is the maximum adsorption amount (mg/g); C_e is Tl concentration at adsorption equilibrium (mg/L); and b is the adsorption equilibrium constant (L/mg).

The equation for the Freundlich adsorption model is (7)${\displaystyle lnQe=lnk+\left(1∕n\right)lnCe}$ Where Q_e is the adsorption amount (mg/g); k and n are constants; and C_e is Tl concentration at adsorption equilibrium (mg/L);

Adsorption thermodynamics

The thermodynamic parameters provide in-depth information about the inherent energetic changes associated with adsorption. The standard enthalpy change (ΔH^o), standard entropy change (ΔS^o), and Gibbs free energy change (ΔG^o) can be calculated from the temperature-dependent adsorption data using the following equations: (8)${\displaystyle \Delta G=RTlnKs}$ (9)${\displaystyle \Delta G=\Delta H-T\Delta S}$ (10)${\displaystyle lnKs=-\frac{\Delta H}{RT}+\frac{\Delta S}{R}}$ where R is the gas constant; K_s is the equilibrium constant; and T is the absolute temperature.

ΔH^o and ΔS^o were obtained from the slope and intercept of the linear plot of ln K_s versus 1/T, respectively. The thermodynamic parameters calculated from Eqs. (5) to (7) are summarized in Table 2. The results showed that the adsorption amount of Tl(I) decreased with increasing temperature. The negative values of ΔH^o indicated that the adsorption of Tl was exothermic. This can be explained by the fact that the heat of adsorption of Tl(I) on RNTD exceeds the dehydration energy of the Tl(I) ions during adsorption. The negative values of ΔS^o showed that adsorption of Tl decreased the randomness of the liquid–solid system. The negative values of ΔG^o suggested that the adsorption of Tl was spontaneous.

FTIR analysis

Figure 7 exhibited the FTIR spectra of pure RNTD and RNTD after Tl adsorption. The peaks at ∼1,045 cm⁻¹ and the strong signal between 2,600∼3,600 cm⁻¹ were ascribed to hydroxyl bond O–H stretching (water adsorbed on TiO₂) (Tanzifi et al., 2018). The wide peak between 600 and 800 cm⁻¹ corresponded to Ti–O–Ti bonds was observed at both of the spectra. The weak absorption peak at ∼2,360 cm⁻¹ corresponds to the stretching vibration of TiOO-H. A strong peak attributed to ClO₄⁻ is seen at 1,085.8 cm⁻¹ . It can be due to the ion exchange between ClO₄⁻ on the surface of the RNTD after addition of NaClO₄ . A vibration peak due to physically adsorbed water appears at 1,637.5 cm⁻¹ (Yousefzadeh, Salarian & Kalal, 2018).

SEM and XRD analyses

As shown in the SEM images (Fig. 8), the surface of TiO₂ nano-particles before adsorption was covered by loose particles with diameter in nm, which implies a large specific surface area. After adsorption, due to the hydrophilicity of TiO₂, ultrafine nanoparticles moved randomly in the aqueous phase, diffused, and spontaneously agglomerated to produce secondary particle clumping. Thus, a dispersed system of micron-sized and nano-sized TiO₂ coexists in the solution, between the colloidal system and the coarsely dispersed system (Fig. 8).

The XRD patterns of pure RNTD and RNTD after Tl adsorption were presented in Fig. 9. Not obvious differences were observed between the two patterns. Anatase and rutile phases of TiO₂ were observed in both of the patterns at the 2θ angles of 27.48, 36.12, 39.24, 41.28, 44.08, 54.36, 56.68, 62.80, 64.12, 69.04, 69.80°. It matched well with the standard data for TiO₂ diffraction pattern (JCPDS89-4920) (Maleki et al., 2016).