Effective in situ remediation of groundwater requires the successful delivery of reactive iron particles through sand. However, the agglomeration of nano zero-valent iron (NZVI) particles limits the migration distance, which inhibits their usefulness. In the study described herein, NZVI supported by mesoporous silica microspheres covered with FeOOH (SiO2@FeOOH@Fe) was synthesized, and its mobility was demonstrated on the basis of transport in porous media. Degradation of decabromodiphenyl ether (BDE209) was more efficient by SiO2@FeOOH@Fe than by ‘bare’ NZVI. Breakthrough curves and mass recovery showed the mobility of SiO2@FeOOH@Fe in granular media was better than that of bare NZVI. It increased greatly in the presence of natural organic matter (NOM) and decreased when high Ca2+ and Mg2+ concentrations were encountered. Analysis of the transport data on the basis of filtration theory showed diffusion to be the main mechanism for particle removal in silicon sand. Increasing the NOM may decrease agglomeration of the grains of sand, which has a positive effect on the mobility of SiO2@FeOOH@Fe. Presumably, increasing the concentrations of Ca2+ and Mg2+ compresses the diffuse double layer of SiO2@FeOOH@Fe, resulting in a reduction of mobility.
The widespread accumulation of halogenated organic compounds (HOCs) such as chlorinated organic compounds and brominated organic compounds in groundwater and in soil is a serious environmental concern (Qiu & Fang 2010). Due to their high hydrophobicity, persistence and bioaccumulative property, HOCs (e.g., polychlorinated biphenyls) are often retained in soil and rock matrix and are very slowly released. Thus, they represent a potential threat to drinking water supplies (Johnson et al. 2009; He et al. 2010).
Recent published studies have explored the direct injection of nano zero-valent iron (NZVI) into aquifers which have potential for in situ degradation of organic pollutants. In situ injection of NZVI may provide some advantages over traditional technologies: many pollutants such as organic halogenated hydrocarbons, nitrates, heavy metals, insecticides and dyes can be effectively degraded by NZVI (Shu 2007; Li et al. 2007; Geng et al. 2009; Saad et al. 2010), and NZVI in some circumstances may reach contaminated zones in deep aquifers or areas where excavation or permeable reactive barriers are not applicable (Johnson et al. 2009). Although several studies have claimed good results in using NZVI for in situ remediation, others have shown that NZVI particles without any modification cluster easily and then form particles with diameters of micron, or millimetre (or larger) scale due to Van der Waals and magnetic attraction forces (Vecchia et al. 2009; He et al. 2009). The agglomeration reduces the mobility and reactivity of NZVI, which is then not deliverable to the targeted source zones.
Extensive effort has gone into reducing aggregation of NZVI. A common method is the use of surfactants, starch and cellulose to refine the surface of NZVI particles. He et al. used carboxymethyl cellulose (CMC) as a stabilizer to modify Fe and investigated transport of CMC-stabilized ZVI nanoparticles (He et al. 2009). The CMC–Fe nanoparticles were readily delivered into various saturated porous media including soils. Another method is to immobilize the NZVI onto carriers that are able to move readily through the subsurface without aggregation of the iron. However, investigations into the mobility of carrier-attached NZVI have been limited (Choi et al. 2009; Jiemvarangkul et al. 2011).
Recently, NZVI immobilized on mesoporous silica microspheres (SiO2@FeOOH@Fe) was synthesized by our team, and the ability of the immobilized NZVI to degrade BDE209 was shown to be greater than that of NZVI (Qiu et al. 2011). As a carrier, SiO2 not only prevents aggregation of NZVI but also may improve mobility of the materials. To explore that possibility, we compared mobility of SiO2@FeOOH@Fe with that of ‘bare’ NZVI. Column experiments were used to evaluate the effects of divalent ions Mg2+ and Ca2+ and natural organic matter (NOM) on the mobility of NZVI immobilized on porous media, and the experimental transport data using the classical filtration theory were simulated.
MATERIALS AND METHODS
Ferrous sulfate (FeSO4·7H2O, 99% minimum), sodium borohydride (NaBH4, 98% minimum), polyethylene glycol and ethanol (99.7%) were supplied by Tianjin Damao Chemical Agent Co. (Tianjin, China). Tetraethoxysilane and dodecylamine were purchased from Aladdin Reagent Co. (Shanghai, China). Methanol (HPLC grade) was supplied by Tianjin Kermel Chemical Reagent Co.
Column breakthrough experiments
Synthesis of NZVI by immobilization on mesoporous silica microspheres (SiO2@FeOOH@Fe) has been described previously (Qiu et al. 2011), and the overall process of synthesizing SiO2@FeOOH@Fe is shown schematically in Figure 1.
Silicon sand (Hainan, China) was selected for column breakthrough experiments. The sand is composed of mineral quartz (>99.8%, silicon dioxide, SiO2). The particle size is in the range of 30–50 mesh and its color is white. The sand was rinsed with deionized (DI) water. Organic impurities were removed by soaking the rinsed sand in hydrogen peroxide (5%) for 3 hours, rinsing the sand again with DI water and then soaking in hydrochloric acid (12 M HCl) overnight. The sand was then thoroughly rinsed with DI water and air-dried. Column breakthrough experiments were carried out in water-saturated silicon sand that was prepared as described above and packed in a vertical glass column 20 cm in length and 2.5 cm in inner diameter. A small nylon sieve (80 mesh) at the bottom of the column prevented loss of the sand. For each experiment, the column was packed wet (porosity = 34.5%) and fed a background ion (Mg2+ or Ca2+) or NOM solution with a dual channel peristaltic pump (HL-2, Chunding, Shanghai, China) for 2 hours to remove background turbidity and to provide a uniform collector surface charge. For the transport of NZVI or SiO2@FeOOH@Fe, a 4 mL solution was introduced into the column. The flow rate was 12 mL/min. To prevent the NZVI or SiO2@FeOOH@Fe from sedimenting, the suspension was mixed at 200 rpm. After completing the injection, the same background solution was used to elute the materials in the column. The effluent samples were collected at selected time intervals. An aliquot of the NZVI effluent was digested with 1 M HCl for 2 hours. The total iron concentrations were determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES). Parallel experiments followed the same procedures.
Free sedimentation experiment
To assess the primary reason for the effect of NOM on the transport of SiO2@FeOOH@Fe through the silicon sand, a free sedimentation experiment was conducted with a spectrophotometer to first determine the effects of NOM on aggregation of SiO2@FeOOH@Fe. The SiO2@FeOOH@Fe suspension (2 g/L) at different concentration of NOM was prepared and sonicated for 5 minutes, while the absorbance was held continuously at 508 nm with a UV–vis light spectrophotometer.
Transmission electron microscopy (TEM) (Philips-Tecnai 10 electron microscope, Philips, Amsterdam, The Netherlands) was used to observe the size and morphology of particles. Specific surface area and pore characteristics were measured with an accelerated surface area and porosimetry analyzer (ASAP 2020M, Micromeritics Instrument Corp., Norcross, GA, USA). Particle iron content was determined by ICP-AES (IRIS Intrepid II XSP, Thermo Elemental, Waltham, MA, USA).
RESULTS AND DISCUSSION
Characterization of synthetic particles
TEM images of SiO2@FeOOH@Fe and bare NZVI are shown in Figure 2. As can be seen in Figure 2(a), the synthesized nanoparticles (NZVI) were spherical, with grain size ranging from about 20 nm to 60 nm. The chain-like structure may have resulted from both the static magnetism and surface tension. The several large particles indicated on the image might have resulted from agglomeration of the nanoparticles due to their magnetic properties (Phenrat et al. 2008; Lin et al. 2010). Mesoporous silica microspheres loaded with NZVI are shown in Figure 2(b). A number of tiny particles of less than 100 nm are seen on the surface of the microspheres. These tiny particles are NZVI particles. The specific surface area of mesoporous silica microspheres loaded with NZVI was 383.477 m2/g and greater than that of bare NZVI (35 m2/g). The ICP-AES results indicated that Fe-loading capacity of the microspheres accounts for 18.01% of the microsphere mass. SiO2@FeOOH@Fe is characterized in detail in our previous publication (Qiu et al. 2011).
Mobility of SiO2@FeOOH@Fe and NZVI
Breakthrough profiles of SiO2@FeOOH@Fe and NZVI are shown in Figure 3. Nearly 92.3% of the SiO2@FeOOH@Fe particles were eluted from the silicon sand column under the specified conditions. The effluent Fe concentration from the SiO2@FeOOH@Fe reached a plateau at ∼8 pore volume (PV). In contrast, no significant elution of bare NZVI was observed, as reported previously (Zhan et al. 2008; Tratnyek & Johnson 2006). Tufenkji et al. developed a model (Tufenkji–Elimelech model) that predicts optimal particle sizes between 200 and 1000 nm for ZVI particles in typical groundwater flow conditions (Tufenkji & Elimelech 2004a; Zheng et al. 2008). Although the particle size of fresh NZVI was on the order of 20–60 nm, the particles aggregate rapidly (within a few minutes) in water, resulting in micro-, millimetre- or even larger-scale aggregates in the absence of an effective stabilizer, thereby leading to a decline in mobility. So the bare NZVI has very little mobility in porous media, and it is difficult to transport more than 1 g/L through porous media without pressure pulsing and very high flow velocities (compared to typical regional groundwater flow velocities).
Diffusion (ηD), interception (ηI) and gravitational sedimentation (ηG) were calculated to be 0.008275, 0.000152 and , respectively. Thus, diffusion was identified as the main mechanism responsible for the loss of nanoparticles. Because of the marked dispersibility and suspension property of SiO2@FeOOH@Fe, the losses resulting from interception (ηI) and gravitational sedimentation (ηG) were limited.
Effect of NOM concentrations on SiO2@FeOOH@Fe mobility
Recovery of SiO2@FeOOH@Fe in column experiments performed at different effective PVs with NOM at 0, 50 or 100 mg/L is shown in Figure 4. The mobility of SiO2@FeOOH@Fe increased with increases in NOM concentration. Two possible explanations for this phenomenon were explored. The first is that the sorption of NOM onto the surface of SiO2@FeOOH@Fe particles may decrease aggregation of SiO2@FeOOH@Fe. The second is that sorption of NOM onto the surface of sand particles may decrease the attachment of SiO2@FeOOH@Fe to the surface by combined electrostatic and steric stabilization effects (Johnson et al. 2009).
Effects of cations calcium and magnesium on SiO2@FeOOH@Fe mobility
In groundwater, the concentrations of divalent cations such as Ca2+ and Mg2+ are typically 0.1–2 mM. To explore the effects of Ca2+ and Mg2+ on the mobility of SiO2@FeOOH@Fe, the column experiment was carried out with high cation concentrations. The effects of Ca2+ and Mg2+ at various concentrations on the mobility of SiO2@FeOOH@Fe are shown in Figure 6. Although the maximum peak remained at ∼6 PV, recovery of Fe decreased from 93.77 to 87.36% when the Mg2+ concentration was increased from 0 to 40 mM. A similar trend was seen with Ca2+: recovery of Fe decreased from 93.77 to 85.98%. According to the Derjaguin–Landau–Verwey–Overbeek theory, increasing the ionic strength would alter the particle surface charge and compress the diffused double layer, resulting in decreased colloidal stability. So, when the Ca2+ and Mg2+ concentrations were increased, the repulsive energy barrier between the SiO2@FeOOH@Fe particles or between the collectors and SiO2@FeOOH@Fe particles decreased, resulting in a reduction of mobility.
In summary, we developed a new approach to immobilization of NZVI in mesoporous silica microspheres to enhance the mobility of NZVI, and demonstrated that reactive SiO2@FeOOH@Fe particles move much more easily than bare NZVI through model soils by a column experiment. This mobility increased greatly in the presence of NOM and decreased with high Ca2+ and Mg2+ concentrations. Analysis of the transport data showed diffusion to be the main mechanism for particle removal in silicon sand. Increasing the NOM may decrease agglomeration of the gains of sand, which has a positive effect on the mobility of SiO2@FeOOH@Fe. Presumably, increasing the concentrations of Ca2+ and Mg2+ compresses the diffuse double layer of SiO2@FeOOH@Fe, resulting in a reduction of mobility. Results of this work indicate that SiO2@FeOOH@Fe nanoparticles have the potential for delivery into porous media to become an effective reactive material for in situ groundwater remediation.
This research was supported by the National Natural Science Foundation of China (no. 41471259) and Guangdong Technology Research Centre for Ecological Management and Remediation of Urban Water Systems (2012gczxA005). The authors are grateful to all of the study participants, and thankful for the financial support of the Guangdong Technology Research Centre for Ecological Management and Remediation of Water Systems.