Interaction of engineered nanomaterials with hydrophobic organic pollutants

For this study, phenanthrene, naphthalene, dichloromethane (CH₂Cl₂), aldrin (C₁₂H₈Cl₆) and TiO₂ (Aeroxide^® P25) were purchased from Sigma-Aldrich (USA). Sodium chloride (NaCl) was purchased from Fisher Scientific (USA). Pure MWCNTs and MWCNT-COOH were purchased from Sun Innovation Inc. (Fremont, California, USA). All the materials were used as received for all the experiments. The physical properties of the MWCNTS are listed in table 1 [23–25]. The purity of the ENMs and the concentration of the trace inorganics were discussed in our previous study [26, 27]. The physical properties of the pollutants selected for this study are listed in table 2. Laser doppler velocimetry in conjunction with phase analysis light scattering (Malvern Zetasizer Nano, Malvern Instruments Ltd) was used to measure the zeta potential of TiO₂ and MWCNTs at selected pH values.

Table 2.  Properties of organics pollutants tested.

Compound	Chemical structure	Mole weight (g mol−1)	Boiling point (°C)	Water solubility (mg l−1)	Log Kow
Naphthalene		128	218	31.7	3.5
Phenanthrene		178	340	1.29	4.45
Aldrin		364	145	0.003	6.5

Stock solutions of phenanthrene and aldrin were prepared in octanol with concentrations ranging from 0.1 to 10.0 g l⁻¹. A naphthalene standard solution of 10 mg l⁻¹ was prepared in CH₂Cl₂. Each PAH solution of 0–20 mg l⁻¹ was prepared with 50 ml octanol and then added to 900 ml DI water that contained 0–20 mg l⁻¹ of each nanomaterial in a 2 l flask. A measured amount of 1 M NaCl solution was added to the flask according to each experimental condition. A known weight of ENMs was placed in a 125 ml umber glass vial with a measured concentration of PAH in distilled water. The solution pH was adjusted using either 1 M NaOH or HCl to the preset pH. A minimum of headspace was left to avoid PAH losses in the vapor phase. The flasks were then sealed with a rubber cock, placed on a stir plate and kept under gentle stirring (i.e., the vortex was less than 1 cm) for five days to reach equilibrium. A glass tube was inserted through the rubber cock to take out samples in the water layer without disturbing the octanol layer. To avoid chemical or photodegradation, the tests were conducted under anaerobic conditions in the absence of light by wrapping the vials with aluminum foil. Samples (30 ml) were taken from the water layer using a 60 ml plastic syringe 3 h after stirring had stopped. Each compound of PAH in the samples was quantified by GC/MS (Agilent GC/MS 6890/5973) equipped with a split/splitless injector. The octanol phase was also analyzed by GC/MS. A Supel-Q™ PLOT Capillary GC Column (Supelco, 30 m × 0.32 mm, average thickness 15 μm) was used. The sample injection volume was 1 μl and the injector temperature was 300 °C. The oven temperature was initially held at 35 °C for 4 min and was then raised at 15 °C min⁻¹ up to 185 °C, with a second heating rate at 10 °C min⁻¹ up to 230 °C , and then a final heating rate at 25 °C min⁻¹ up to 310 °C, where it was held for 2 min.

The samples collected from the aqueous samples were centrifuged using a Sorvall Legend T+ centrifuge (Thermo Scientific), and this was followed by microfiltration. Sodium chloride was dissolved to form a concentration of 10 g l⁻¹ in the filtrate to assist the extraction by increasing the ionic strength to promote the transfer of hydrophobic pollutants to the organic phase. The sample and 20 ml dimethyl chloride were added into a reparatory funnel, and samples were shaken for 10 min. The organic phase was decanted, and the extraction process was repeated with the same volume of dimethyl chloride before the GC/MS analysis. For the GC/MS analysis, the 30 ml aqueous sample was extracted with 30 ml of CH₂Cl₂ using a 125 ml separatory funnel. The mixture of the aqueous sample and CH₂Cl₂ was then taken and transferred to a 50 ml conical centrifuge tube. The tube was under nitrogen flow until all the extracted solutions were evaporated. After the evaporation, 2 ml of a mixture of 50:50 CH₂Cl₂ and acetone was added to the tube and used to rinse the wall of the tube well. The solution in the tube after rinsing was analyzed by GC/MS to quantify PAHs.

In addition, the PAHs adsorbed on the nanomaterials were determined with TGA (PerkinElmer Pyrus 7) combined with evolved analysis using an integrated GC/MS system (PerkinElmer Clarus 600) (TGA/GC/MS). For the analysis, nanomaterials were collected by filtration using a glass fiber filter (Whatman Glass Microfibre Filter, GE Healthcare Life Sciences, Pittsburgh, Pennsylvania, USA) or centrifugation. The collected ENMs were dried in a conventional oven at 35 °C for 8 h. The mass fraction of the adsorbed PAHs on nanomaterials was analyzed using TGA (PerkinElmer Pyrus 7) under nitrogen flow.

Thermal analysis provides an important insight into the thermodynamic and mechanical properties of materials heated under a controlled atmosphere and their interaction with probe molecules. Thermogravimetric analysis (TGA) is a thermal analysis method in which air-dried ENM samples are heated under a controlled aerobic or anaerobic environment, and the changes in the weight of the samples as a function of the temperature is recorded. This weight change data provide useful information about the physical and chemical properties of the samples. However, TGA cannot identify what material is emitted at a specific temperature. The analysis of gases evolved during a TGA experiment by gas chromatography/mass spectrometry provides us with a means of identifying the compound or groups of compounds evolved during a specific weight-loss event. Hence, TGA integrated with gas chromatography and mass spectroscopy (TGA/GC/MS) improves our ability to understand a sample's physical properties and chemical composition. Coupling a TGA to a GC/MS allows us to detect, trap and isolate low levels of materials sorbed to ENMs from complex mixtures, and to identify their components. The nanomaterials filtered from the sample water of our batch adsorption tests are air dried. A measured mass of (∼10 mg) was put on a PerkinElmer's TGA (Pyrus 1) and the evolved gas containing PAHs was transferred through a heated transfer capillary tube that was kept under partial vacuum to the sampling loop (5 ml) integrated with a Clarus 600 GC/MS. The system was controlled by Turbo Mass software.

For GC-MS analysis, an Agilent DB-5ms capillary column (30 m × 0.250 mm, film thickness 0.25 um) was used. The oven temperature was initially held at 35 °C for 4 min and was then raised at 15 °C min⁻¹ up to 185 °C, with a second heating rate at 10 °C min⁻¹ up to 230 °C, and then a final heating rate at 25 °C min⁻¹ up to 310 °C, where it was held for 2 min. The injection volume was 1 μl and the injector temperature was 300 °C. The EPC was run with the splitless mode, and the column flow rate of helium was 1 ml min⁻¹. The average velocity in the column was 37 cm s⁻¹ and for MS acquisition a secondary ion mass spectrometry (SIMS) was used. Measured masses of pure PAH were placed in a TGA pan for qualitative calibration of the TGA/GC/MS response as a function of evolved gas PAH concentration. To quantify the naphthalene in MWCNTs, an internal standard was used. Scheme 1 shows a simple experimental flow for studying the interaction of PAHs and nanomaterials in water. The detailed experimental procedure is provided in the supporting information (scheme S1).

Zoom In Zoom Out Reset image size

Figure 1 shows the zeta potential values (mV) as a function of pH in dispersions of pristine and acid-functionalized CNTs and TiO₂. Pristine MWCNTs were positively charged from pH 3 to pH 8. Above pH 8, the surface of pure MWCNTs was negatively charged. The acid functional groups had a higher electron negativity of oxygen than neat CNTs and polarized for oxygen. The presence of the hydroxyl group helped to stabilize the COOH-CNT suspensions [28], increasing their dispersion in water, which can increase their ion exchange capacity [29]. Figure 1 shows that the points of zero charges (pH_PZC) for pristine-CNT and COOH-CNT were 7.9 and 1.7, respectively. For MWCNT-COOH, the zeta potential reached a low plateau at a pH of 2.7, which can be explained by the fact that COOH-CNT is completely deprotonated above this pH [30–33]. These results are in good agreement with previous studies, which reported that the surface of MWCNTs is negatively charged after functionalization with acid or noncovalent treatments due to the formation of carboxyl functional groups [30–32]. The different zeta potential values of the CNTs are directly correlated with the ability of CNTs to stay in suspension and form stable colloidal solutions, and to facilitate the sorption mechanisms [34]. The increased quantity of oxygen groups present on the surface lattice of CNTs enhances the attraction between the oxygen functional groups and the water molecules, overcoming the CNTs' gravitational force and leading to the formation of stable suspensions [28]. Based on its colloidal stability, COOH-CNT is a better adsorbent than pristine-CNT. Pristine-CNT also has deprotonated groups, which result in decreased electronegativity at higher pHs. The attachment of hydroxyl ions onto the hydrophilic surface or residual oxygen groups during the synthesis process may also result in a negative zeta potential for untreated pristine-CNTs [33]. In the case of TiO₂, the particles have negative zeta potential values above a pH of 6 and at alkaline conditions, and the pH_PZC for commercial P25 TiO₂ (a 25%:75% mixture of rutile and anatase) was found between 6.2, which is in good agreement with previous studies [35–37].

Zoom In Zoom Out Reset image size

The TGA weight-loss profiles of neat naphthalene, the glass filter used to recover MWCNTs and the pristine and carboxyl-functionalized MWCNTs as a function of temperature are presented in figure 2. Since the boiling point of naphthalene is 218 °C, the total weight loss of pure naphthalene from the TGA pan occurred at around 230 °C. The glass filter used to separate the nanomaterials was very stable (no weight loss) within the temperature range for TGA analysis. The MWCNTs and MWCNT-COOH exhibited a small weight loss above 500 °C. The TGA profile of the naphthalene adsorbed MWCNT and MWCNT-COOH samples showed weight loss around 390 °C to 400 °C, following a similar trend to that demonstrated by the neat MWCNTs and MWCNT-COOH. Significant weight loss was observed around 470 °C and 620 °C for the naphthalene-adsorbed MWCNTs and MWCNT-COOH, respectively. The delay in the weight loss accounting for the release of naphthalene indicates the strong adsorbtion of naphthalene to the ENMs.

Zoom In Zoom Out Reset image size

In addition to the naphthalene adsorption on MWCNTs, the adsorption of phenanthrene to nanostructured-TiO₂ was investigated. Figure 3 shows the TGA profile for dried or wet TiO₂ before and after phenanthrene adsorption. As seen in figure 3, TiO₂ had a high capacity for phenanthrene adsorption and released it during a thermal heating step. Since wet TiO₂ contains water, its weight loss was larger than that of dried TiO₂. After phenanthrene adsorption, both dried and wet TiO₂ showed a similar trend to the TGA results for TiO₂ before phenanthrene adsorption. The phenanthrene-adsorbed wet TiO₂ sample lost more weight than the dried one. Interestingly, the TiO₂ samples adsorbed more water and phenanthrene in the absence of light illumination. This might have resulted from TiO₂ photocatalysis. Adsorbed phenanthrene on the surface of TiO₂ starts to be decomposed by reactive oxygen species generated by TiO₂ photocatalytic reaction. The formed reaction intermediates during TiO₂ photocatalysis may interrupt phenanthrene adsorption due to competition between the phenanthrene and reaction intermediates on the surface of TiO₂ to occupy active sites of TiO₂. In addition, the surface chemistry of TiO₂ might be changed by TiO₂ photocatalysis in a way that lowers phenanthrene adsorption. Moreover, TiO₂ becomes hydrophilic in the presence of light [38]. The lowered hydrophobicity of TiO₂ reduced its interaction with the hydrophobic phenanthrene, resulting in decreased phenanthrene adsorption. Similar results were reported in a previous study where the adsorption of contaminants by TiO₂ in the presence of UV light was lower than in the absence of light [39].

Zoom In Zoom Out Reset image size

Based on the GC/MS analysis, the K_ow of phenanthrene in the presence of TiO₂ was calculated via equation (1):

$$\begin{eqnarray}&&{K}_{{\rm{o}}{\rm{w}}}\;=\;\displaystyle \frac{{C}_{{\rm{o}}}}{{C}_{{\rm{w}}}},\end{eqnarray} \tag{ 1 }$$

where C_o is the centration of phenanthrene in octanol and C_w is the concentration of phenanthrene in water. The calculated log K_ow are listed in table 3. As seen in table 3, the values of K_ow were proportional to the solution pHs, while K_ow decreased with an increase in mixing times. More phenanthrene goes into the water phase with longer mixing times. Compared to the log K_ow (4.46) of phenanthrene under normal conditions [40], in the presence of TiO₂ the calculated log K_ow became smaller, indicating that TiO₂ helps phenanthrene to become more partitioned in water. Moreover, the K_ow increased slightly at pH 3, but decreased at pH 10. It might be that the environmental conditions around TiO₂ change due to the photocatalytic activity of TiO₂ at different solution pHs.

In addition to phenanthrene adsorption on TiO₂, adsorption of the hydrophobic insecticide aldrin on TiO₂, pristine MWCNTs and MWCNT-COOH was studied by GC/MS analysis. Figure 4 shows the results for aldrin adsorption on TiO₂ and MWCNTs, which were calculated via liquid phase analysis using GC/MS. As seen in figure 4, the adsorbed aldrin on all TiO₂, pristine MWCNTs and MWCNT-COOH was proportional to the equilibrium concentration of aldrin in experimental systems. Interestingly, MWCNTs adsorbed a smaller amount of aldrin than TiO₂, while MWCNT-COOH showed a higher adsorption of aldrin than TiO₂ and pristine MWCNTs. This might have been caused by their different physicochemical properties (e.g., Brunauer, Emmett and Teller (BET) surface areas, sizes and surface charges).

Zoom In Zoom Out Reset image size

Moreover, the interaction of naphthalene and MWCNTs in water was studied. As seen in figure 5, the concentration of naphthalene in both the solid and liquid phases increased with an increase of the concentration of spiked naphthalene. This indicates that the MWCNTs adsorbed naphthalene effectively in addition to aiding naphthalene dissolution in water. The naphthalene concentrations for pristine MWCNTs in both phases increased linearly, but for MWCNT-COOH, it seemed that the adsorption reached equilibrium in the solid phase, even though more naphthalene dissolved in water. Based on the results of the GC/MS analysis, the nanoparticle–water partition coefficient (K_d) for MWCNTs at different concentrations of naphthalene was calculated using equation (2). The calculated coefficients are listed in table 4.

$$\begin{eqnarray}&&{K}_{{\rm{d}}}\;=\;\displaystyle \frac{{C}_{{\rm{p}}}}{{C}_{{\rm{w}}}},\end{eqnarray} \tag{ 2 }$$

where C_p is the centration of naphthalene in MWCNTs and C_w is the concentration of naphthalene in water. At a higher concentration of naphthalene, log K_d decreased slightly for MWCNTs, indicating that more naphthalene went to the water phase. However, for MWCNT-COOH, the log K_d was smaller than that for bare MWCNT, showing less adsorption of naphthalene. This may have resulted from the different hydrophobicities of bare and functionalized MWCNT. Similar results were reported in previous studies, namely that adsorption of PAHs decreased significantly after the functionalization of MWCNTs [41–44].

Zoom In Zoom Out Reset image size

In addition, the adsorption isotherm was determined using 48 h of equilibrium time. The experimental data for chlorinated pesticide, aldrin and naphthalene adsorption to TiO₂, MWCNT and MWCNT-COOH were fit with the Freundlich isotherm ($\mathrm{log}\;q=\;\mathrm{log}\;{K}_{{\rm{F}}}+\;1/n\mathrm{log}\;{C}_{{\rm{w}}})$, as shown in figures 4 and 5. The correlation coefficients vary from 0.95 to 0.98. For TiO₂, pristine and COOH-functionalized MWCNT the constants K_F for aldrin adsorption were 19.1, 13.8 and 21.9, respectively, and the n values were 0.56, 0.47 and 0.61, respectively. The average particle size of the Degussa P25 nanoparticles, 20 nm, is comparable with the OD of MWCNT (10–20 nm), affording a high surface area. Suspended TiO₂ and MWCNT-COOH have a zeta potential of about −29 mV in the experimental conditions of pH = 7, which accounts for the higher electrostatic attraction of these pollutants compared with MWCNT.

As the temperature of the nanomaterial sample obtained from the batch equilibrium study continued to increase, the evolved gas from TGA was transferred via a heated line and analyzed using GC/MS to determine the quantity of PAHs adsorbed on the nanomaterials for batch equilibrium study. Vapor phase samples were pulled to the GC/MS by a vacuum pump and trapped in the injection port. The MS response was calibrated by putting multiple masses of naphthalene on the TGA pan and assessing the MS responses. The data set obtained by TGA/GC/MS analysis is provided in figure S1 in the supporting information. Figure 6 shows the concentration of naphthalene in MWCNTs and MWCNT-COOH obtained by direct solid phase analysis with TGA/GC/MS and by calculation based on liquid phase analysis with GC/MS. As seen in figure 6, the results of TGA/GC/MS analysis for solid phase was compared to calculated amount of naphthalene in nanomaterials based on the GC/MS analysis for liquid phase analysis. The adsorption capacity, q, of both pristine and COOH-functionalized MWCNT for naphthalene was linearly proportional to the concentration of spiked naphthalene in a batch three-phase system, as shown in figures 6(a) and (b), respectively. The adsorption capacity estimates based on solid phase analysis using TGA/GC/MS were 6% to 19% higher than the measurements made using liquid samples. To investigate whether there is a significant difference between liquid phase analysis with GC/MS and solid phase analysis with TGA/GC/MS in quantifying the adsorbed amount of naphthalene, statistical analysis was conducted. Further investigation proved that there is no significant difference between the two analytical methods at 95% confidence. Moreover, the difference in the adsorption capacity between pristine and COOH-functionalized MWCNT for naphthalene adsorption was not significant (figures 6(c) and (d)).

Zoom In Zoom Out Reset image size

The views expressed in this article are those of the authors and do not reflect the official policy or position of the Unites State Environmental Protection Agency. Mention of trade names, products, or services does not convey official EPA approval, endorsement, or recommendation. This manuscript has been subjected to the Agency's review and has been approved for publication.