Dechlorination of hexachlorobenzene by nano zero-valent iron/activated carbon composite: iron loading, kinetics and pathway

Abstract

Nano zero-valent iron (nZVI) and activated carbon composites were synthesized and employed to remove hexachlorobenzene (HCB). Methods of impregnation and adsorption were employed to firstly load iron salt onto activated carbon, and thus loaded iron was then reduced to form zero valent iron. Results indicate that the method of iron loading significantly impacted the amount of iron and HCB dechlorination. In general, nZVI/activated carbon composite synthesized using an adsorption method showed considerably higher removal efficiency and dechlorination capability. More than 80% of the HCB was dechlorinated after 48 h. The dechlorination of HCB followed a stepwise pathway: HCB (hexachlorobenzene) → PCB (pentachlorobenzene) → 1,2,4,5 TeCB (tetrachlorobenzene) → TriCB (trichlorobenzene) → 1,4 DCB (dichlorobenzene) → MCB (monochlorobenzene). Compared with activated carbon or nZVI, nZVI/activated carbon composite showed much higher HCB removal. Analysis of mass partition on solid and aqueous phases indicates that most of the HCB and its dechlorination intermediates were retained on solids probably due to the strong adsorption on activated carbon. In summary, activated carbon performs three important functions: it alleviated the limitation of nZVI agglomeration, contributed to HCB removal due to its own adsorption capability and increased the resistance of nZVI against pH change.

---

1. Introduction

Once considered to be an effective fungicide, rubber peptizing agent and wood preservative, hexachlorobenzene (HCB) is now banned for commercial production in many countries.¹ However, HCB still contaminates many sites and streams in China due to its persistence in the environment. Studies have related HCB to cases of breast cancer and various organ damages.² Along with other members from the halogenated organic compounds family, efficient and cost-effective remediation methods are required to treat the locations with HCB pollution.

Zero-valent iron was found to be an effective dehalogenation reagent due to its reducing power.³ In particular, studies have shown that nano zero-valent iron (nZVI) could provide even more rapid and a complete reduction of halogenated hydrocarbon.⁴ As a result, the use of nZVI for the in situ remediation of organic and inorganic pollution in soil and groundwater had gained significant interest over the last decade.⁵ Application of pure nZVI, however, was limited by the strong tendency of material to agglomerate, low wettability with organic phase and high density difference to water. All these characteristics contributed to the low mobility and loss of reactivity of nZVI in aqueous streams.⁶ Many studies employed additives (e.g. anionic polyelectrolytes and surfactants) to improve mobility and create a stable nZVI suspension.⁷ In addition to additives, depositing nZVI on carrier materials has been extensively explored as another way to overcome the limitations of nZVI and extend its environmental applicability. Silica,⁸ alginate bead,⁹ carbon black,¹⁰ zeolite,¹¹ carbon nanotube,¹² and activated carbon¹³ are among the various materials investigated as carriers.

In this study, activated carbon was chosen as the carrying material. Activated carbon is renowned for its high surface area and pore volume. The pore structure of activated carbon provides ample room for nZVI. Bleyl stated that the pore system of activated carbon provides a foundation for the formation of finely dispersed nano-iron cluster.¹⁴ Moreover, activated carbon is very effective for the adsorption of organics, which creates an environment of enriched organics around nZVI particles, and thus it may accelerate the destruction of pollutants. This is especially beneficial in the environment, where the concentration of pollutant is low, as often observed in the case of a groundwater plume. Xu et al. believed that sorption-assisted dechlorination significantly increased the destruction of pollutant in their study for the removal of 2,4-dichlorophenol using Pb/Fe nano particles dispersed on a multiwalled carbon nanotube support.¹²

However, there are numerous ways to deposit nZVI on supporting materials. In general, the production of ZVI-supported materials was divided into two steps: Fe(II) or Fe(III) was first adsorbed or impregnated onto material, and then reduced to form nZVI. Chemical reduction by NaBH₄ of Fe(II) or Fe(III) on the carrier is among the most popular ways because of its simple operation, especially at lab scale. For carbon-based carriers, there are many reports for the carbothermal synthesis of nZVI-carbon composite using the reductive power of elemental carbon.^15,16 Tseng et al. combined chemical reduction with high temperature calcinations (700 °C) under N₂ atmosphere for the synthesis of activated carbon/zero valent iron composite for the purpose of TCE removal.¹⁷ Their results showed that calcination helped to further disperse ZVI particles in activated carbon, and a higher TCE removal was achieved.

This study assesses the effectiveness of HCB removal by activated carbon supported nZVI. The effects of dosage and pH, HCB removal kinetics, and the dechlorination pathway were investigated. Chemical reduction by NaBH₄ was employed to reduce iron salt to zero valent iron. The effects of iron loading by wet impregnation and adsorption on the morphology of synthesized nZVI, and their effects on HCB dechlorination were also studied.

2. Materials and methods

2.1 Chemicals

Lignite-based granular activated carbon, sodium borohydride, ethanol and n-hexane were purchased from Sinopharm Chemical (Shanghai, China). Hexachlorobenzene, ferrous chloride tetrahydrate and chlorobenzene standard solutions were obtained from Aladdin Industrial Co. (Shanghai, China). All chemicals were of analytical grade and used as received. All the reagents were prepared with deionized water.

2.2 Preparation of nZVI/activated carbon

Activated carbon of 88–150 μm in size was used as the nZVI carrier. There were two steps in making the nZVI/activated carbon composite. First, ferrous salts were loaded onto activated carbon by wet impregnation or adsorption process. Loaded iron was then reduced to Fe⁰ by NaBH₄ according to the following reaction (1):¹⁸

2Fe²⁺ + BH₄⁻ + 2H₂O = 2Fe⁰ + BO₂²⁻ + 2H₂ + 4H⁺ (1)

For wet impregnation, 2 g of activated carbon was added to 50 mL of ferrous chloride solution (concentration of 1 mol L⁻¹). The resulting slurry was shaken for 24 h on a rotary shaker at room temperature (25 °C), and then put into an evaporator at 100 °C for 5 h. For nZVI/activated carbon composite synthesis, 2 g of iron-impregnated activated carbon was slowly added to 50 mL of 1 mol L⁻¹ NaBH₄. After 3 h of reaction, nZVI/activated carbon was separated by filtration from liquid and thoroughly washed with deionized water to remove loose nZVI, dried and stored in a N₂ purged container. The as-synthesized nZVI/activated carbon was named as nZVI/AC-I.

For adsorption, 2 g of activated carbon was added to 20 mL of 1 mol L⁻¹ of ferrous chloride solution. The mixture was shaken for 24 h on a rotary shaker at 25 °C, and 15 mL of ethanol and 15 mL of water were then added to the mixture. Subsequently, 50 mL of 1 mol L⁻¹ NaBH₄ was added dropwise to the mixture with vigorous stirring. The resulting particles were separated and named as nZVI/AC-A.

For comparison, nZVI was synthesized using the modified NaBH₄ reduction method.¹⁹ 21.36 g of ferrous chloride was added to a mixture of 96 mL of ethanol and 24 mL of water. The solution was slowly added to 400 mL of 1 mol L⁻¹ NaBH₄, and then was shaken for 2 h on a rotary shaker at 25 °C. Subsequently, the particles were separated, washed, dried and stored in a N₂ purged container.

2.3 Characterization of nZVI/AC

Surface area and pore size distribution were determined by nitrogen adsorption–desorption isotherms at a temperature of 77 K (−196 °C) by a 3H-2000PS4 unit (Beishide Instrument, China). Total pore volume was estimated from the adsorbed amount at a relative pressure of 0.99. Brunauer–Emmett–Teller (BET) and Barrett–Joyner–Halenda (BJH) equations were used to calculate surface area and pore size distribution, respectively.

Scanning electron microscope (SEM) S4800 (Hitachi, Japan) and high resolution transmission electron microscope (HR-TEM) Tecnai G20 (FEI, USA) were used to observe the morphology and size of the composites. Infrared spectra of the nZVI, nZVI/activated carbon were obtained by a fourier transform infrared spectrometer (FTIR) Magna-IR 750 (Nicolet, USA) in the diffuse reflectance mode with KBr pellets. Spectra were generated from the mid-infrared range of 4000–500 cm⁻¹ with 256 scans per spectrum at a spectral resolution of 4 cm⁻¹.

2.4 Effect of dosage on HCB removal

10, 30, 50, 80, 160 and 300 mg of nZVI, activated carbon or nZVI/AC composite was added to vials, each containing 20 mL of 200 μg L⁻¹ HCB solution. Initial pH was adjusted to 5.6. The mixtures were then shaken for 8 h at 150 rpm and 25 °C. Solid particles were separated from liquid by filtration; a 0.45 μm membrane filter was used for all the filtrations. A preliminary test indicated that the retention of organics by membrane is negligible. Filtrate was extracted for 0.5 h with 20 mL of n-hexane and analyzed for HCB.

2.5 Effect of pH on HCB removal

nZVI and activated carbon or nZVI/AC, 160 mg in mass, was added to 20 mL of 200 μg L⁻¹ HCB solution. The pH of the solution was adjusted to 3, 5, 7, 9 and 11 using 0.5 mol L⁻¹ HCl or NaOH. The mixtures were then shaken for 8 h at 150 rpm, 25 °C. Solids were separated from liquid through filtration and filtrate was extracted for HCB analysis.

2.6 HCB removal kinetics

nZVI and activated carbon or nZVI/AC, 160 mg in mass, was added to 20 mL of 200 μg L⁻¹ HCB solution. The initial pH of the solution was adjusted to 5.6. The mixtures were shaken for 0.5, 1, 4, 8, 24 and 48 h. At the end of each time interval, solids were separated from liquid by filtration. 10 mL of filtrate was extracted with 20 mL of n-hexane for 0.5 h to separate residual HCB or dechlorination products from water.

In addition, to determine the mass of organics retained by nZVI/AC-I or nZVI/AC-A solids, nZVI/AC-I or nZVI/AC-A was mixed with 20 mL of n-hexane for 4 h to extract the adsorbed organics. HCB and its potential dechlorination products in the extract were analyzed.

All the abovementioned tests were conducted in triplicate.

2.7 Chemical analysis

HCB and its degradation chlorobenzene products were analyzed with an Agilent 7890A-5975C gas chromatography-mass spectrometry (Agilent Technologies, USA) equipped with an HP-5MS capillary column. 1 μL of the extraction was automatically injected in splitless mode. The temperatures of injector and detector were set at 320 °C and 350 °C, respectively. Separation was controlled with a temperature program that started at an oven temperature of 40 °C, held for 5 min, then ramped at 20 °C min⁻¹ to 200 °C, and then ramped at 5 °C min⁻¹ to 250 °C, and then held for 2 min.

Iron content of activated carbon or nZVI/AC composite was determined by a digestion method. 0.1 g of sample was added to 30 mL of 3 mol L⁻¹ HCl. The mixture was shaken for 2 h, and the digestion solution was diluted and analyzed for iron by an AA-6300C atomic absorption spectrophotometer unit (Purkinje General, China) with flame atomization.

3. Results and discussion

3.1 Characterization of nZVI/AC

Fig. 1a–f are the illustrations of SEM and TEM images of activated carbon, nZVI, nZVI/AC-I and nZVI/AC-A. Activated carbon shows an uneven and porous surface. The nZVI without support severely agglomerated into clusters (Fig. 1b); nZVI was much more dispersed when activated carbon was used as the supporting material. However, the structure of nZVI differed with different iron loading methods. As shown in Fig. 1c, with wet impregnation, the resultant nZVI was more randomly shaped, whereas for adsorption, nZVI was generally in a rod-like structure.

Fig. 1 SEM images (magnitude 50.0 k) of (a) activated carbon, (b) nZVI, (c) nZVI/AC-I, (d) nZVI/AC-A; TEM images of (e) nZVI/AC-I, (f) nZVI/AC-A.

TEM images correspond well with images from SEM. The dark clusters in these images are identified to be iron particles. Fig. 1e reveals that nZVI/AC-I composite consists of roughly spherical iron particles, whereas those of nZVI/AC-A were mostly in rod shape; particles from nZVI/AC-I are about 50 nm in diameter. However, adsorption produced relatively smaller particles with diameters around 20 nm. The difference in the size may be due to the difference in the morphology of loaded iron. In impregnation, water was evaporated, and iron salts solidified and precipitated as iron hydroxides, which were later reduced. While in adsorption, iron was most likely adsorbed as hydrated or complexed ions; ethanol could also be playing a role in limiting the size. Jiu et al. believed that the presence of ethanol could first increase the amount of iron loading.²⁰ Ferrous ions were bound to the hydroxyl groups in ethanol whereas the nonpolar part of the ethanol could more easily attach to the nonpolar surface of the activated carbon. Moreover, ethanol formed a finite structure surrounding the iron ions, and the size of nanoparticles created is thus more limited.

The FTIR spectrum of activated carbon, nZVI, nZVI/AC-I and nZVI/AC-A provide a further confirmation of nZVI loading onto activated carbon. As shown in Fig. 2, broad bands at around 3400 cm⁻¹ in activated carbon and nZVI/AC composites were attributed to O–H stretching, most likely due to H₂O or C–OH.²¹ Other characteristic bands include C=C at 1600 cm⁻¹ and C–OH at 1130 cm⁻¹.²² The strong vibration at 1355 cm⁻¹ and 1100 cm⁻¹ for nZVI/AC-A was probably due to hydroxyl groups stretching caused by ethanol used during sample preparation. Bands around 2000 cm⁻¹ and at less than 900 cm⁻¹ observed in nZVI, nZVI/AC-I and nZVI/AC-A were attributed partly to iron oxides. These bands were weaker for nZVI/AC-I and nZVI/AC-A, indicating that less oxidation occurred for activated carbon-supported Fe⁰.²³

Fig. 2 FTIR spectra of (a) activated carbon, (b) nZVI, (c) nZVI/AC-I, (d) nZVI/AC-A.

Fig. 3 shows the plot of the incremental pore volume vs. pore width for activated carbon, nZVI/AC-I and nZVI/AC-A. Table 1 lists the BET surface area and total pore volume. There are obvious drops in both surface area and pore volume after the activated carbon was loaded with nZVI. This was mainly attributed to the fact that the pores were occupied by nZVI. According to Fig. 3, decrease in the pore volume mostly occurred with pores that are less than 50 Å in diameter, indicating that these pores may be blocked by nZVI and are no longer accessible.

Fig. 3 Pore size distribution of activated carbon, nZVI/AC-I and nZVI/AC-A.

Table 1 BET surface area, total pore volume and iron content of activated carbon, nZVI/AC-I and nZVI/AC-A

Samples	BET surface area (m^2 g^−1)	Total pore volume (mL g^−1)	Iron content (mg g^−1)
Activated carbon	708 ± 14	0.44 ± 0.05	0.30 ± 0.05
nZVI/AC-I	597 ± 10	0.36 ± 0.04	61.08 ± 4.23
nZVI/AC-A	502 ± 12	0.32 ± 0.02	217.38 ± 5.12

---

Table 1 is a summary of the surface area and total pore volume from pore size analysis. It also includes the iron contents for the samples. After the synthesis, iron content increased from about 0.30 mg g⁻¹ for activated carbon to 61.08 and 217.08 mg g⁻¹ for nZVI/AC-I and nZVI/AC-A respectively.

Through characterization, two nZVI/AC composites showed differences both in nZVI morphology and iron content. The iron loading process had significant effect on the amount of elemental iron and the characteristics of thus created nZVI.

The difference in pollutant removal by these composites was subsequently tested on HCB.

3.2 Effect of dosage on HCB removal

Fig. 4 presents the effect of dosages of activated carbon, nZVI, nZVI/AC-I and nZVI/AC-A on HCB removal. In all the cases, HCB removal increased with the increase of dosage, until it started to level off after the dosage reached 8 g L⁻¹. Thus, it can be concluded that if all other conditions are kept same, there will be more reactive sites available for reaction at a higher dosage, and thus there would be higher HCB removal. In addition, HCB varied with different materials. For instance, at 8 g L⁻¹ dosage, activated carbon, nZVI, nZVI/AC-I and nZVI/AC-A removed 61%, 40%, 85% and 95% of the initial HCB, respectively. In general, at the same dosage, the removal efficiency followed the order of nZVI/AC-A > nZVI/AC-I > activated carbon > nZVI. It appears that combining nZVI with activated carbon showed a synergistic effect. nZVI/AC-A showed the highest HCB removal capability probably because of its high nZVI content. With more nZVI particles, there will be more surface sites for the reactions.

Fig. 4 Effect of dosage on HCB removal by activated carbon, nZVI, nZVI/AC-I and nZVI/AC-A.

3.3 Effect of initial pH

Fig. 5 shows the effect of initial pH on HCB removal. Solution pH was adjusted from 2 to 11 before adding adsorbents. According to Fig. 5, pH had a significant effect only on HCB removal by nZVI. HCB removal was favorable at acidic pH. This may be because the dissolution of the passivating (oxy)hydroxide layer on nZVI was facilitated at low pH, which increased the reactivity. Moreover, at higher pH (>9), Fe(II) and (III) hydroxides were formed, which could have resulted in hydroxide covering the Fe⁰ surfaces and hampering the electron transfer.²⁴ In contrast, statistical analysis (single factor t-test) revealed that variations in HCB removal by activated carbon, nZVI/AC-I and nZVI/AC-A were not significant at α = 0.05 level (p-value>0.05). Depositing of nZVI on activated carbon weakened the effect of pH, and the materials were more robust against pH change. In regards to activated carbon, it removed HCB mainly by adsorption. Godino-Salido et al. believed that the adsorption of aromatics on activated carbon consists of a plane to plane interaction between the aromatic moieties and the arene centres at the graphite sheets.²⁵ Therefore, the pH of solution may not have a significant effect on this mechanism.

Fig. 5 Effect of pH on HCB removal by activated carbon, nZVI, nZVI/AC-I and nZVI/AC-A.

3.4 Adsorption kinetics

The effect of the reaction time on HCB removal is shown in Fig. 6. For activated carbon, the removal efficiency increased from 26% at 0.5 h to around 45% at 8 h, then reached a plateau of around 65%. For nZVI/AC-I and nZVI/AC-A, the removal efficiency leveled off at 80% and 90%, respectively, only after 0.5 h of contact. Removal by nZVI was considerably lower; it never reached beyond 30% even after 48 hours.

Fig. 6 Effect of contact time on HCB removal by activated carbon, nZVI, nZVI/AC-I and nZVI/AC-A.

Kinetic data were fitted to pseudo-first- and second-order kinetic eqn (2) and (3), respectively, as follows:

Pseudo-first order:

Pseudo-second order:

where q_t and q_e are the adsorption at time t and equilibrium, and k₁ and k₂ are first- and second-order adsorption rate constants.

Table 2 lists the kinetic parameters of first- and second-order reactions. It is noted that in all the cases, the correlation coefficient (R²) for the pseudo-first-order model is less than 0.90, which indicates a bad correlation. The application of the pseudo-second-order model leads to a much better coefficient of around 0.98–0.99. Moreover, the calculated values of q_e are in much better agreement with experimental data. Thus, the second-order kinetic model is more suitable to depict the HCB removal, suggesting that the process involves a chemical reaction mechanism

Table 2 Parameters for kinetics of HCB removal by activated carbon, nZVI, nZVI-I and nZVI-A

Samples	Pseudo-first order				Pseudo-second order
	qe,exp^a (mg g^−1)	qe,cal^b (mg g^−1)	k1 (h^−1)	R^2	qe,cal^b (mg g^−1)	k2 (g (mg h)^−1)	R^2	qe,exp^1 (mg g^−1)
a Experimental data.b Calculated from model.
Activated carbon	9.85	20.49	0.32	0.89	9.28	0.024	0.99	9.85
nZVI	7.98	4.57	0.08	0.88	8.13	0.057	0.99	7.98
nZVI/AC-I	18.34	1.52	0.06	0.77	18.34	0.25	0.99	18.34
nZVI/AC-A	21.55	3.27	0.08	0.86	21.59	0.13	0.98	21.55

---

Reactions of dechlorination and oxidation of zero-valent iron to Fe(II) or Fe(III) (eqn (4)–(8)) could occur as follows, when nZVI came into contact with chlorinated organics:³¹

Fe⁰ + RCl + H⁺ → Fe²⁺ + RH + Cl⁻ (4)

2Fe⁰ + O₂ + H₂O → 4OH⁻ + 2Fe²⁺ (5)

2Fe²⁺ + RCl + H⁺ → 2Fe³⁺ + RH + Cl⁻ (6)

Fe³⁺ + 3OH⁻ → Fe(OH)₃ (7)

Fe²⁺ + HO˙→ Fe³⁺ + (OH)⁻ (8)

The nZVI/AC-A manifested the highest HCB removal capacity, with calculated adsorption capacity at an equilibrium of 21.55 mg g⁻¹, which is more than double compared to that of the activated carbon or nZVI. In addition, second-order reaction rates (k₂) were one magnitude higher for nZVI/AC-I and nZVI/AC-A, indicating that the combination of activated carbon and nZVI also improved the rates of removal.

Table 3 compares the HCB reaction rates reported in the peer-reviewed journals to the results obtained from this research. nZVI/AC-A manifested a relatively faster reduction than kaolin, activated carbon, nanoscale Fe, Pd/Fe or Cu/Fe particles, and a slower reduction than Ag/Fe particles indicating that the materials synthesized could be quite competitive.

Table 3 Comparison of HCB reaction rates with other materials

Material description	Reaction rate constants	References
Kaolin	0.0047	26
Nanoscale Fe	0.14	27
Nanoscale Pd/Fe particles	0.23	27
Nanoscale Cu/Fe particles	0.056	28
Nano Ni/Fe particles	0.065	29
Micro Ag/Fe particles	0.452	30
Activated carbon	0.024	This research
nZVI/AC-I	0.13	This research
nZVI/AC-A	0.25	This research

---

In addition, the concentration of iron in the solution was monitored throughout the kinetics test. Overall, the leaching of iron is not significant because the iron concentration never reached beyond 0.08 mg L⁻¹ in all situations. It appears that Fe(II) or Fe(III) produced during the dechlorination process was retained by activated carbon, and there is not much leaching of iron.

3.5 Pathway of HCB dechlorination reaction

Because dechlorination reactions may occur during HCB removal, in particular with nZVI/AC composites, potential degradation products of HCB were monitored. Fig. 7 and 8 show the trends of various dechlorination by-products (in percentage of initial HCB) against time for nZVI/AC-I and nZVI/AC-A, respectively. Intermediates detected include monochlorobenzene (MCB), isomers of dichlorobenzene (DCB), trichlorobenzene (TriCB), tetrachlorobenzene (TeCB) and pentachlorobenzene (PCB). The detection of various intermediates in solution proved that reductive dechlorination reaction by the loss of chloride atoms occurred with nZVI/activated carbon. The concentrations of by-products in the solution with activated carbon were below the detection limit, which may indicate that dechlorination by activated carbon was negligible.

Fig. 7 Changes of HCB and its dechlorination by-products vs. time with nZVI/AC-I in solution.

Fig. 8 Change of HCB and its dechlorination by-products vs. time with nZVI/AC-A in solution.

The mass balance analysis of chlorinated aromatics based on the amount of by-products detected in solution was poor because of the fact that dechlorination intermediates may be retained on the activated carbon surface and not released in the solution. On an average, after 48 h of contact with nZVI/AC-A and nZVI/AC-I, only 35% and 18% of the initial HCB mass was detected in solution, respectively, indicating the strong adsorption by activated carbon. Among them, about 70–80% is HCB, and the remaining part is composed of various dechlorination by-products. In particular, in the case of nZVI/AC-A, concentrations of the by-products peaked at 72 h of contact, and then gradually decreased until almost no by-product was detected in solution after 120 h, as shown in Fig. 8. Dechlorination by-products were taken up by nZVI/AC-A after a long contact time. This means that it could provide better removal of HCB.

Combining the mass from the solution and that retained on the solids, Fig. 9 and 10 show trends of HCB and dechlorination intermediates in total. The mass balance analysis showed that around 90% of the mass was accounted. The remaining 10% may be strongly adsorbed on activated carbon and was not extractable by n-hexane. According to the mass analysis, after 2 h, more than 50% of the initial HCB was dechlorinated with nZVI/AC-I. After 72 h, HCB accounted for about 20% of the total mass. nZVI/AC-A was able to dechlorinate more than 70% of HCB after 2 h, and only 9.5% of the initial HCB still existed after 72 h; a greater amount of HCB was dechlorinated with nZVI/AC-A. This may also be attributed to the higher iron content of nZVI/AC-A.

Fig. 9 Change of HCB and its dechlorination by-products vs. time with nZVI/AC-I in total.

Fig. 10 Change of HCB and its dechlorination by-products vs. time with nZVI/AC-A in total.

It is generally believed that the higher chlorination tended to higher toxicity.³² Dechlorination by-products are generally less toxic and more biodegradable. Dechlorination showed by both nZVI/AC composites suggests that these products can be used for the remediation of HCB in the environment.

The nZVI/AC-I and nZVI/AC-A showed much similarity with regards to the dechlorination pathway, as shown in Fig. 9 and 10. DCB and TeCB were the main by-products. 1,4-DCB and 1,2,4,5-TeCB were detected to be the main species of TeCB and DCB in this research. 1,4-DCB and 1,2,4,5-TeCB both manifest symmetric structures and are more stable. Studies have shown that chloride atoms at the first and third carbon atoms have lower electron density, and are much more susceptible to nucleophilic substitution.³³ Nie and Liu also mentioned 1,2,4,5-TeCB being the main by-products of PCB dechlorination in their study.³⁴ However, due to the low concentration of TriCB, no distinction among TriCB isomers was made.

The monitoring of dechlorination by-products showed a clear stepwise dechlorination pattern for both nZVI/AC-I and nZVI/AC-A. There was a general decrease in HCB with the increase of contact time. PCB peaked after 2 h of reaction; then slowly decreased until the end of the experiment. TeCB became the predominant species after 10 h and reached to the apex at around 48 h. TriCB and DCB were both observed after 8 h. DCB replaced TeCB to become the predominant species after 72 h. It appears that the dechlorination of TriCB to DCB was rapid, and the accumulation of TriCB did not occur. Therefore, TriCB never dominated and was stable around 10% after 48 h. Moreover, the percentage of MCB was relatively low throughout the whole test, meaning that it may have taken a considerably longer time for HCB to degrade to MCB.

In summary, the dechlorination results agreed well with several studies of HCB reduction by nZVI.³⁵ All these studies reported stepwise dechlorination. In addition, the percentage of less chlorinated by-products (e.g. DCB and MCB) over initial total HCB was higher with nZVI/AC-A than nZVI/AC-I. After 120 h, DCB and MCB accounted for 40% of the HCB with nZVI/AC-A. This number is 25% for nZVI/AC-I. nZVI/AC-A was more effective in dechlorination.

The dechlorination by-products and pathways in the stepwise reductive dechlorination of HCB with nZVI/AC are as follows: the first product was PCB. PCB quickly dechlorinated to form tetrachlorobenzene, which includes three isomers. 1,2,4,5-TeCB is found to be the main product. TriCB was the next step and was further converted to 1,4-DCB. There was also a small amount of MCB detected. Therefore, the main pathway proposed was HCB → PCB → 1,2,4,5-TeCB → TriCB → 1,4-DCB → MCB.

4. Conclusions

nZVI/AC particles were prepared by first loading iron salts iron onto activated carbon, and then nZVI was synthesized through a chemical reduction. The performance for HCB removal was investigated and compared with that of the activated carbon and nZVI. The iron loading method played a significant role in the morphology of nZVI and the extent of iron content. With the help of ethanol, the adsorption method was able to produce nZVI with rod-like structure, and thus created an nZVI/activated carbon composite that contained much higher iron content.

Batch experiments showed that nZVI/AC composite resulted in the enhancement of HCB removal and resistance against pH change as compared to the activated carbon or nZVI alone. In addition, HCB removal by nZVI/AC fits well with second-order reaction consumptions.

Dechlorination of HCB was proven to play a significant part in HCB removal. The majority of HCB was dechlorinated and converted into an array of intermediate chlorobenzene products with TeCB and DCB as the predominant species. In particular, nZVI/AC composite with iron adsorption as the starting step showed advantages in both HCB removal efficiency and dechlorination capability. Removal efficiency reached as high as 95% for HCB, and 90% of the removed HCB was dechlorinated after 72 h of reaction.

Overall, combining nZVI with activated carbon was able to take advantage of the adsorption capacity of activated carbon, and the dechlorination reactivity of nZVI, and the resultant material showed a promising behavior to be a candidate for HCB remediation.

Acknowledgements

This work was supported by Shanghai Natural Science Foundation (14ZR1428900), National Natural Science Foundation of China (51078233), Shanghai Committee of Science and Technology (13230502300) and Tianjin Natural Science Foundation (13JCYBJC18600).

Notes and references

1. M. Tong and S. Yuan, J. Hazard. Mater., 2012, 229–230, 1.
2. J. F. Jarrell, A. Gocmen, D. Akyol and R. Brant, Reprod. Toxicol., 2002, 16, 65.
3. X. Ye, J. Jiang, Y. Yang and G. Gao, Chem. Eng. J., 2011, 173, 415.
4. Y. H. Shih, C. Y. Hsu and Y. F. Su, Sep. Purif. Technol., 2011, 76, 268.
5. S. Klimkova, M. Cernik, L. Lacinova, J. Filip, D. Jancik and R. Zboril, Chemosphere, 2011, 82, 1178.
6. K. D. Grieger, A. Fjordboge, N. B. Hartmann, E. Eriksson, P. L. Bjerg and A. Baun, J. Contam. Hydrol., 2010, 118, 165.
7. F. He, D. Zhao and C. Paul, Water Res., 2011, 44, 2360.
8. P. J. Dorathi and P. K. Kandasamy, J. Environ. Sci., 2012, 24, 765.
9. H. Kim, H. J. Hong, J. Jung, S. H. Kim and J. W. Wang, J. Hazard. Mater., 2010, 176, 1038.
10. L. B. Hoch, E. J. Mack, B. W. Hydustky, J. M. Hershman, I. M. Skluzacek and T. E. Mallouk, Environ. Sci. Technol., 2008, 42, 2600.
11. S. A. Kim, S. Kamala-Kannan, K. J. Lee, Y. J. Park, P. J. Shea, W. H. Lee, H. M. Kim and B. T. Oh, Chem. Eng. J., 2013, 217, 54.
12. J. Xu, X. Lv, J. Li, Y. Li, L. Shen, H. Zhou and X. Xu, J. Hazard. Mater., 2012, 225–226, 36.
13. M. C. Pereira, F. S. Coelho, C. C. Nascentes, J. D. Fabris, M. H. Araujo, K. Sapag, L. C. A. Oliveira and R. M. Lago, Chemosphere, 2010, 81, 7.
14. S. Bleyl, F. D. Kopinke and K. Mackenzie, Chem. Eng. J., 2012, 191, 588.
15. J. J. Zhan, B. Sunkara, J. J. Tang, Y. Q. Wang, J. B. He, G. L. McPherson and V. T. John, Ind. Eng. Chem. Res., 2011, 50, 13021.
16. H. Q. Sun, G. L. Zhou, S. Z. Liu, H. M. Ang, M. O. Tade and S. B. Wang, ACS Appl. Mater. Interfaces, 2012, 4, 6235–6241.
17. H. H. Tseng, J. G. Su and C. Liang, J. Hazard. Mater., 2011, 192, 500–506.
18. P. Jiemvarangkul, W. Zhang and H. Lien, Chem. Eng. J., 2011, 170, 482–491.
19. R. D. Rogers, A. H. Bond and C. B. Bauer, Sep. Sci. Technol., 1993, 28, 139.
20. J. T. Jiu, L. P. Li, Y. Ge, S. R. Zhang, Z. R. Hua and L. Nie, Chin. J. Inorg. Chem., 2001, 17, 361.
21. M. Mohapatra, H. Mohapatra, P. Singh, S. Anand and B. K. Mishra, Int. Trans. J. Eng., Manage., Appl. Sci. Technol., 2010, 2, 89–103.
22. V. I. Protilla, Am. Mineral., 1976, 61, 95.
23. A. L. Andrade, D. M. Souza, M. C. Pereira, J. D. Fabris and R. Z. Domingues, Ceramica, 2009, 55, 420.
24. G. Roy, P. D. de Donato, T. Görner and O. Barres, Water Res., 2003, 37, 4954.
25. M. L. Godino-Salido, R. López-Garzón, M. D. Gutiérrez-Valero, P. Arranz-Mascarós, M. Melguizo-Guijarro, M. D. L. de la Torre, V. Gómez-Serrano, M. Alexandre-Franco, D. Lozano-Castelló, D. Cazorla-Amorós and M. Domingo-García, Mater. Chem. Phys., 2014, 143, 1489.
26. S. Yuan, Z. Zheng, X. Z. Meng, J. Chen and L. Wang, Geoderma, 2010, 159, 165.
27. Y. H. Shih, Y. C. Chen, M. Y. Chen, Y. T. Tai and C. P. Tso, Colloids Surf., A, 2009, 332, 84–89.
28. N. Zhu, H. Luan, S. Yuan, J. Chen, X. Wu and L. Wang, J. Hazard. Mater., 2010, 76, 1101.
29. X. Xu, J. Wo, J. Zhang, Y. Wu and Y. Liu, Desalination, 2009, 242, 346.
30. X. Nie, J. Liu, X. Zeng and D. Yue, J. Environ. Sci., 2013, 25, 473.
31. B. Deng, D. R. Burniss and T. J. Campbell, Environ. Sci. Technol., 1999, 33, 2651.
32. A. Poland and J. C. Knutson, Annu. Rev. Pharmacol. Toxicol., 1982, 22, 517.
33. Y. Xu and W. X. Zhang, Ind. Eng. Chem. Res., 2000, 39, 2238.
34. X. Nie and J. G. Liu, J. Environ. Sci., 2013, 25, 473–478.
35. G. V. Lowry and K. M. Johnson, Environ. Sci. Technol., 2004, 38, 5208.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra06760f

---