Effect of NiFe₂O₄ on PCDF byproducts formation during thermal degradation of decachlorobiphenyl

Abstract

Formation of 2,3,7,8-PCDD/Fs during the catalytic decomposition of decachlorobiphenyl over spinel NiFe₂O₄ was evaluated at 300 °C and compared with that over Fe₂O₃ and in a catalyst-free system. The total concentration of 2,3,7,8-PCDFs, dominated by OCDF and then 2,3,7,8-HpCDF, were significantly formed, being 6–53 times higher than the total concentration of 2,3,7,8-PCDD in all three reaction systems studied. Time-dependent distribution profile of 2,3,7,8-PCDF indicated that the degradation and formation of PCDF occurred concurrently. The initial hydrodechlorination of OCDF preferentially occurred at the 1- and 9-position than at the 4- and 6-position (thus, higher amounts of 1,2,3,4,6,7,8-HpCDF were generated relative to 1,2,3,4,7,8,9-HpCDF), followed by further major hydrodechlorination pathway of 1,2,3,4,6,7,8-HpCDF → 1,2,3,6,7,8-HxCDF → 1,2,3,7,8-PeCDF → 2,3,7,8-TCDF. Relative to the catalyst-free reaction system, the concentration and total toxic equivalent values of 2,3,7,8-PCDFs decreased by 24.1–99.7% and 86–98.4%, respectively, in the NiFe₂O₄ catalyst system, while initial increases by 744.5% and 426.3%, respectively, followed by reduction up to 81.6% and 90.7%, respectively, were observed in the Fe₂O₃ catalyst system. Current findings indicate that NiFe₂O₄ hindered PCDF formation, which was attributed to its chemical structural stability and high activity towards degradation of biphenyl-like constituents in PCBs.

---

1. Introduction

Polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs) are extremely toxic and persistent organic pollutants (POPs). The toxicity levels of PCDD/Fs primarily depend on the amount and position of the chlorine substituents. PCDD/Fs with chlorine atoms substituted at the 2,3,7,8-positions (7 PCDDs and 10 PCDFs) are believed to be the most toxic congeners.¹ Because of their extreme toxicities, strict control of PCDD/F emissions in the environment is necessary. PCDD/Fs are naturally produced in almost all thermal processes in the presence of a chlorine source. Processes include general combustion, metallurgical-related reactions, and municipal waste incineration.^2–5 The two fundamental PCDD/F formation mechanisms proceed via a precursor pathway and a de novo pathway.^6–10 In the precursor pathway, PCDDs and PCDFs are formed by the Eley–Rideal and/or Langmuir–Hinshelwood mechanisms in heterogeneous surface-mediated processes.^8,11,12 The Eley–Rideal mechanism involves the combination of a surface-bound precursor with a gas-phase precursor to form PCDD/Fs, whereas the Langmuir–Hinshelwood mechanism involves two surface-adsorbed precursors.¹³ Typical and widely studied PCDD/F precursors are chlorophenols and chlorobenzenes.^13–15 However, biphenyl-like precursors, which contain more than one benzene ring, also play an important role in PCDF formation. Fullana et al.¹⁶ demonstrated that biphenyl and fluorine can be easily converted into PCDFs at 300 °C. Similar results were obtained in other studies that examined biphenyl-like compounds as precursors.^17,18 Zhao et al.¹⁹ reported that considerable amounts of PCDFs were formed when sediments containing polychlorinated biphenyls (PCBs) were subjected to thermal processes.

PCBs are a class of compounds composed of 209 congeners, and they are categorized as POPs under the Stockholm Convention because of their high levels of toxicity and bioaccumulation. Since the 1930s, they have been used in a variety of industrial applications, e.g., as dielectric fluids in electrical transformers and plasticizers in paints.²⁰ Although the production of PCBs has been banned since the mid-1970s, large amounts of PCBs have been released into the environment because of their historical widespread use. Accidental exposure to PCBs still occurs owing to their presence in old transformers and capacitors.²¹ Controlling PCB emissions in the environment is therefore a matter of public concern. Directing efforts towards the environmentally sound management and disposal of waste PCBs by 2028 is one of the aims as established under the Stockholm Convention. However, an important criterion for assessing PCB decomposition technology according to the Stockholm Convention is to ensure that the selected process does not potentially create further toxic byproducts, especially PCDD/Fs.²²

A number of methods for the disposal of POPs have been developed, including incineration,²³ microwave decomposition,²⁴ mechanochemical decomposition,^25,26 and catalytic degradation.^27,28 Thermal catalytic degradation of POPs using metal oxides under relatively mild conditions is attracting increasing attention because of the relatively low cost, high catalytic activity and thermal stability, and ease of preparation of high surface area materials.^29–32 Fe₂O₃ has been widely investigated for the decomposition of chlorinated aromatics.^33–36 For example, Khaleel et al.³⁴ investigated the activity of α-Fe₂O₃ in the oxidative decomposition of chlorobenzene, whereas Jia et al.³³ used micro/nano α-Fe₂O₃ to degrade hexachlorobenzene. However, some studies have indicated that α-Fe₂O₃ also possessed catalytic properties towards the promotion of the formation of PCDD/Fs. Thus, the development of a catalyst system, possessing high stability and activity towards degradation reactions, as well as the capabilities of reduced PCDD/Fs formation is challenging.

In recent years, many efforts have been made to develop multicomponent catalysts for the decomposition of chemical pollutants because of the remarkable multifunctionality that they can afford over single metal oxides.^37–40 Spinel oxides, as a type of complex oxides, with the chemical formula AB₂O₄, have normal and inverse structures. In normal spinels, A is generally a divalent cation occupying tetrahedral sites and B is a trivalent cation occupying octahedral sites. In inverse spinels, half of the B cations occupy tetrahedral sites. Numerous studies have reported that spinels, as multicomponent materials, display high levels of activity towards the degradation of chlorinated aromatics.^41–44 Cesteros et al.⁴¹ showed that the hydrodechlorination of 1,2,4-trichlorobenzene to benzene proceeded at a considerably faster rate over a spinel-supported nickel material (Ni/NiAl₂O₄) when compared with that over an aluminum-supported nickel material (Ni/Al₂O₃) at 150–250 °C. Moreover, Fan et al.⁴² reported higher degradation efficiencies of hexachlorobenzene and PCDD over CuAl₂O₄ at 250 °C when compared with those achieved by the corresponding metallic copper and copper oxide materials. However, to the best of our knowledge, there are few studies on the evaluation of PCDD/F formation during the thermal catalytic degradation of chlorinated aromatics, especially PCBs, using spinel oxides as catalyst.

In a former study, as-prepared NiFe₂O₄ spinel was demonstrated as a highly efficient catalyst for the decomposition of decachlorobiphenyl (CB-209), achieving a final dechlorination efficiency of 95.9% at 300 °C after 1 h.⁴⁵ The identified degradation products included less-chlorinated biphenyls, hydroxyl compounds, chlorobenzenes (CBzs), and organic acids. In this study, we further evaluated PCDD/F formation during the thermal degradation of CB-209 over as-prepared NiFe₂O₄. The results were compared with those obtained using a Fe₂O₃-based catalyst system and a catalyst-free system. The levels of the formed 2,3,7,8-PCDFs and 2,3,7,8-PCDDs were compared in all three reaction systems studied, and the time-dependent distribution of 2,3,7,8-PCDFs were discussed. The effect of NiFe₂O₄ on the concentration and toxic equivalent (TEQ) of the produced 2,3,7,8-PCDFs was also assessed.

2. Materials and methods

2.1 Catalyst preparation

NiFe₂O₄ was prepared by coprecipitation according to the procedure described in a former study.⁴⁵ In a typical process, precipitation was achieved by drop-wise addition of ammonium hydroxide to a solution of nickel and iron nitrate with continuous stirring at room temperature until the pH reached 8. When precipitation was complete, the obtained suspension was stirred for 1 h. The collected precipitate was dried at 105 °C overnight in an oven, and then calcined at 400 °C in air for 3 h. The resulting NiFe₂O₄ sample was ground and stored in a desiccator before use. The prepared NiFe₂O₄ exhibited irregular morphology and magnetic property (presented in Fig. S1 and S2 in the ESI†). A former study⁴⁵ reported that all diffraction peaks of the wide angle XRD patterns of NiFe₂O₄ sample corresponded to the NiFe₂O₄ phase (JCPDS 01-086-2267). This catalyst exhibited a relatively high surface area of 41.0 m² g⁻¹ with an average pore size of 11.5 nm and a total pore volume of 1.22 × 10⁻¹ cc g⁻¹.

All reagents used above were purchased from Beijing Chemical Reagents Company (Beijing, China); moreover, analytical grade Fe₂O₃ was also purchased from Beijing Chemical Reagents Company (Beijing, China).

2.2 Thermal reactions

Reactions were conducted in sealed glass ampoules (approximately 1.5 mL). A hexane solution containing 1 mg CB-209 (2004 nmol) was added to a glass ampoule, and the mixture was dried under a nitrogen atmosphere at room temperature. Then, 50 mg of either NiFe₂O₄ or Fe₂O₃ was added and the ampoule was heated at 300 °C for 30–120 min. The same procedure was used for the blank reaction, which was conducted in the absence of a catalyst. All the experiments were performed in triplicate to ensure the repeatability of the results.

2.3 PCDD/F analysis

The PCDD/Fs were analyzed mainly using the United States Environmental Protection Agency Method 23 modifications. After the thermal reaction, the ampoule was cooled to room temperature and crushed. The fragments were spiked with known amounts of ¹³C₁₂-labeled PCDD/F internal standards (Wellington Laboratories, Guelph, Canada) and the sample was extracted. The extract was cleaned using a basic alumina column. The final solution was concentrated to 20 μL, and then spiked with known amounts of ¹³C₁₂-labeled injection standards (Wellington Laboratories, Guelph, Canada), which were used to quantify the internal standard recoveries. The analytical recoveries of ¹³C₁₂-labeled PCDD/F internal standards were 86–106%, which met the analytical method requirement of PCDD/Fs.^4,46 PCDD/F analysis was performed using an Agilent 6890 high-resolution gas chromatograph (Agilent Technologies Inc., Santa Clara, CA, USA) equipped with a DB-5 MS column (J&W, Agilent Technologies Inc.) and a Waters Autospec Ultima high-resolution mass spectrometer (Waters Corp., Milford, MA, USA). The detailed instrumental information on the analysis of PCDD/Fs is described in the ESI.† Laboratory method blanks were run with every batch of samples, and no interferences were observed.

3. Results and discussion

3.1 Concentrations and congener patterns of PCDD/Fs

The total concentration of the 2,3,7,8-PCDD/F congeners detected in the three reaction systems, after reaction at 300 °C for 30 min, and in the pure CB-209 standard is shown in Table 1. In the CB-209 standard, only trace amounts of OCDF and OCDD were detected. When subjected to heating in the absence of a catalyst (blank reaction), higher amounts of OCDF and newly formed 1,2,3,4,6,7,8-HpCDF and 1,2,3,4,7,8,9-HpCDF were detected; moreover, high concentrations of 2,3,7,8-PCDFs (2.80 × 10³ ng g⁻¹) were observed. This is consistent with previous reports,⁴⁷ which show that conversion of PCB into PCDF occurs in air at temperatures as low as 300 °C in the absence of catalysts. Similar congeners were also identified in the NiFe₂O₄ catalyst reaction system. However, ∑2,3,7,8-PCDD/F concentration was very low at 674 ng g⁻¹. The significant decreases by 76.0% in the ∑2,3,7,8-PCDF concentrations in the NiFe₂O₄ reaction system compared with those in the CB-209 blank reaction system indicated that NiFe₂O₄ catalyst has a strong negative effect on the PCDF formation. In contrast, 1,2,3,4,7,8-HxCDF, 1,2,3,6,7,8-HxCDF, 2,3,4,6,7,8-HxCDF, and 1,2,3,7,8,9-HxCDF were additionally produced when Fe₂O₃ catalyst was used in the thermal degradation process of CB-209. The differences in the congeners between the NiFe₂O₄ and Fe₂O₃ reaction systems suggest that the type of catalyst involved in the PCB degradation system influenced the distribution of the formed PCDD/F congeners. The ∑2,3,7,8-PCDF concentration achieved in the Fe₂O₃ reaction system was 2.36 × 10⁴ ng g⁻¹ that is eight times higher than that obtained in the CB-209 blank reaction system, thus indicating that Fe₂O₃ has a positive effect on PCDF formation under these reaction conditions. These results are consistent with previously published studies.^11,48 Low concentration levels of 2,3,7,8-substituted PCDD were determined in all three reaction systems, i.e., 462, 48.5, and 446 ng g⁻¹. Moreover, only OCDD was detected as the 2,3,7,8-substituted PCDD congener in these reaction systems. The ∑2,3,7,8-PCDF concentration was 6–53 times higher than the ∑2,3,7,8-PCDD concentration in the three reaction systems, suggesting the preferential formation of 2,3,7,8-PCDFs over 2,3,7,8-PCDDs when CB-209 is subjected to a thermal process. As reported, polyaromatic hydrocarbons with biphenyl-like structures produce high amounts of PCDF and low amounts of PCDD.¹⁶ Similar results were obtained in the oxidative process of pesticide α-cypermethrin at 300–650 °C,¹⁷ thermal treatment of model waste incinerator fly ash in the presence of oxygenated-PAH (polyaromatic hydrocarbon) at 250–350 °C,¹⁸ and thermal treatment of sediments containing polychlorinated biphenyls.¹⁹ The characteristic mid-ring structure of PCDD consists of a hexagonal ring bonded to two oxygen atoms, whereas that of PCDF consists of a pentagonal ring bonded to one oxygen atom. The biphenyl-like compounds can easily be converted into PCDFs, possibly due to the insertion of an oxygen directly without breaking the C–C bridge bond of biphenyl.⁴⁹

Table 1 Average concentration levels of PCDD/Fs in decachlorobiphenyl (CB-209) standard and three reaction systems following heating for 30 min^a

Pollutants	CB-209 standard	CB-209 blank reaction	CB-209 reaction with NiFe2O4	CB-209 reaction with Fe2O3
a Concentration levels are reported as ng g^−1 CB-209. n.d.: not detected.
2,3,7,8-TCDF	n.d.	n.d.	n.d.	n.d.
1,2,3,7,8-PeCDF	n.d.	n.d.	n.d.	n.d.
2,3,4,7,8-PeCDF	n.d.	n.d.	n.d.	n.d.
1,2,3,4,7,8-HxCDF	n.d.	n.d.	n.d.	2.55
1,2,3,6,7,8-HxCDF	n.d.	n.d.	n.d.	4.75
2,3,4,6,7,8-HxCDF	n.d.	n.d.	n.d.	2.55
1,2,3,7,8,9-HxCDF	n.d.	n.d.	n.d.	4.15
1,2,3,4,6,7,8-HpCDF	n.d.	539	70.4	2.45 × 10^3
1,2,3,4,7,8,9-HpCDF	n.d.	21.1	n.d.	81.9
OCDF	10.6	2.24 × 10^3	604	2.11 × 10^4
2,3,7,8-TCDD	n.d.	n.d.	n.d.	n.d.
1,2,3,7,8-PeCDD	n.d.	n.d.	n.d.	n.d.
1,2,3,4,7,8-HxCDD	n.d.	n.d.	n.d.	n.d.
1,2,3,6,7,8-HxCDD	n.d.	n.d.	n.d.	n.d.
1,2,3,7,8,9-HxCDD	n.d.	n.d.	n.d.	n.d.
1,2,3,4,6,7,8-HpCDD	n.d.	n.d.	n.d.	n.d.
OCDD	27.9	462	48.5	446
∑2,3,7,8-PCDDs	27.9	462	48.5	446
∑2,3,7,8-PCDFs	10.6	2.80 × 10^3	674	2.36 × 10^4
∑WHO-TEQ	1.20 × 10^−2	6.41	0.90	33.2

---

The PCDD/F congener distribution profiles were used as fingerprints to identify the formation of PCDD/Fs during thermal degradation of CB-209. The distribution profiles of the three reaction systems after reaction at 300 °C for 30 min are shown in Fig. 1. OCDF and 1,2,3,4,6,7,8-HpCDF fractions predominated in CB-209 blank reaction, representing 68.8% and 16.5% of total 2,3,7,8-substituted PCDD/F, respectively. In contrast, OCDD concentration only constituted 14.2% of the total 2,3,7,8-substituted PCDD/F concentration. Similar trends were observed in the NiFe₂O₄ and Fe₂O₃ reaction systems, i.e., the congener concentration profile decreased as follows: OCDF (83.2% and 87.5%, respectively) > 1,2,3,4,6,7,8-HpCDF (10% and 10.1%, respectively) > OCDD (6.7% and 1.9%, respectively). The PCDD/F congener profiles more clearly indicated 2,3,78-PCDFs, in particular OCDF, were formed more easily than 2,3,7,8-PCDDs when CB-209 was subjected to a thermal process. The current congener profile differed from that reported by Morita et al., in which tetra- and penta-CDFs were observed as the predominant congeners in the oxidation of Aroclor 1248.⁴⁷ The chlorine substitution pattern of the treated PCBs possibly has a significant influence on the type of congeners formed.

Fig. 1 PCDD/F congener profiles in the three reaction systems at 30 min.

Two general PCDD/F formation mechanisms have been proposed: the precursor process and the de novo process.⁵⁰ The precursor pathway involves the formation of PCDD/Fs from precursors with similar structural features, e.g., chlorophenols, chlorobenzenes, and PCBs. PCDF-to-PCDD ratios (R_DF/DD) of lower than one typically indicate that the PCDD/Fs are formed from precursors and are observed in large-scale thermal processes, including municipal solid waste incinerators and combustion of industrial wood waste.^51,52 However, despite the precursor pathway adopted in our study, the amount of PCDFs formed was considerably larger than the amount of PCDDs formed, i.e., R_DF/DD > 1. Large R_DF/DD ratios were also reported when PCDD/Fs were formed using biphenyl-like compounds as precursors.^16,17,19,49 Evaluation of the formation pathways of PCDD/Fs should therefore not only take into account the R_DF/DD value, but also the reaction conditions and type of matrix.

3.2 WHO-TEQ level and contribution of PCDD/F congeners

Toxic equivalent factors (TEF) obtained from the 2005 World Health Organisation (WHO) report (herein, denoted as WHO-TEF₂₀₀₅) were used to calculate the WHO-TEQs of the PCDD/Fs formed in the samples. For congeners with concentrations below the limit of detection, a value of 0 was assigned for the TEQ calculations. The total values of WHO-TEQ (∑WHO-TEQ) of the PCDD/Fs calculated for the different systems following reaction for 30 min are given in Table 1. Notably, ∑WHO-TEQ of the NiFe₂O₄ reaction system (i.e., 0.9 ng g⁻¹) was 7.1 times lower than that of the CB-209 blank reaction system (i.e., 6.41 ng g⁻¹). The result showed that the use of a NiFe₂O₄ catalyst could efficiently reduce the toxicity of the intermediate products during the pyrolysis of PCBs. In contrast, the Fe₂O₃ reaction system displayed a ∑WHO-TEQ value of 33.2 ng g⁻¹, which was 5.2 times higher than that of the CB-209 blank reaction system, indicating that the Fe₂O₃ catalyst increased the toxicity of the intermediate products during pyrolysis of PCBs. The relationship of ∑WHO-TEQ levels in the three systems was consistent with that of ∑2,3,7,8-PCDD/F concentrations.

The TEQ congener patterns of PCDD/Fs in the three reaction systems after reaction for 30 min are shown in Fig. 2. The abundance sequences were similar for the three reaction systems. In the CB-209 blank reaction, 1,2,3,4,6,7,8-HpCDF contributed to 84.2% of the total TEQ, followed by OCDF (10.7%), and OCDD (2.03%). The same trend was observed in the NiFe₂O₄ and Fe₂O₃ reaction systems, i.e., 1,2,3,4,6,7,8-HpCDF (78.4% and 73.8%, respectively), followed by OCDF (20% and 19.1%, respectively), and OCDD (1.6% and 0.4%, respectively). The TEQ patterns are relatively different from the PCDD/F congener profiles, which was caused by the different WHO-TEF₂₀₀₅ values of different congeners. The TEQ congener patterns of PCDD/Fs in all three reaction systems indicated that the toxicity of the thermal reaction of CB-209 mainly originated from the toxicity contribution of 2,3,7,8-PCDFs, which was 48.2–247 times higher than that of 2,3,7,8-PCDDs.

Fig. 2 TEQ percentage contribution of PCDD/F congeners in the three reaction systems at 30 min.

3.3 Time-dependent concentration and WHO-TEQ level of PCDFs

Examining whether the PCDD/Fs byproducts will disappear as a function of reaction time is intriguing. 2,3,7,8-PCDFs were identified as the significant contributors towards the formation of PCDD/Fs during thermal degradation of PCBs. Hence, a series of experiments were performed on the different reaction systems to investigate the effect of reaction time on the formation of this type of congener, as shown in Table 2. The concentration of ∑2,3,7,8-PCDD/F in the CB-209 blank reaction system increased rapidly from 2.80 × 10³ to 7.23 × 10⁴ ng g⁻¹ within 30–120 min of heating. In contrast, the concentration of ∑2,3,7,8-PCDF in the NiFe₂O₄ reaction system decreased from 674 to 241 ng g⁻¹ by the end of the 120 min heating process. Accordingly, when compared with the CB-209 blank reaction system, the ∑2,3,7,8-PCDF of the NiFe₂O₄ reaction system decreased from 24.1% to 99.7% with the increase of reaction times. In contrast, the concentration of ∑2,3,7,8-PCDF in Fe₂O₃ reaction system was determined at 2.36 × 10⁴ ng g⁻¹ with the increase by 745% at 30 min, followed by a reduction up to 1.29 × 10³ ng g⁻¹ with the decrease by 81.6% at 120 min. The above results indicated that both decomposition and formation of PCDF occurred concurrently. Thus, during the thermal catalytic degradation of PCBs, the reaction time needs to be carefully regulated to control the release of PCDFs. In NiFe₂O₄ reaction system, the destruction was significantly predominant in the whole period of reaction. The promotion effect of Fe₂O₃ played a dominant role in PCDF formation at the early stages of the reaction, and then the destruction activity dominated the later stages of the reaction possibly with the decreasing amounts of precursors. However, as mentioned above, the concentration of ∑2,3,7,8-PCDF in the NiFe₂O₄ reaction system was always significantly lower than that in the Fe₂O₃ reaction system, suggesting that NiFe₂O₄ is a more environment friendly catalyst for PCB pyrolysis.

Table 2 PCDF concentration levels in different reaction systems as a function of reaction time^a

	CB-209 blank reaction			NiFe2O4 reaction system			Fe2O3 reaction system
	30 min	60 min	120 min	30 min	60 min	120 min	30 min	60 min	120 min
a Concentration levels are reported as ng g^−1 CB-209. n.d.: not detected.
2,3,7,8-TCDF	n.d.	n.d.	n.d.	n.d.	19.6	n.d.	n.d.	n.d.	n.d.
1,2,3,7,8-PeCDF	n.d.	n.d.	n.d.	n.d.	12.4	n.d.	n.d.	n.d.	n.d.
1,2,3,4,7,8-HxCDF	n.d.	4.25	7.25	n.d.	8.45	4.70	2.55	n.d.	n.d.
1,2,3,6,7,8-HxCDF	n.d.	27.9	42.8	n.d.	13.0	3.80	4.75	n.d.	n.d.
2,3,4,6,7,8-HxCDF	n.d.	7.00	6.90	n.d.	5.45	4.5	2.55	n.d.	n.d.
1,2,3,7,8,9-HxCDF	n.d.	11.2	33.5	n.d.	9.65	5.55	4.15	n.d.	n.d.
1,2,3,4,6,7,8-HpCDF	539	6.62 × 10^3	1.21 × 10^4	70.4	80.4	42.2	2.45 × 10^3	998	918
1,2,3,4,7,8,9-HpCDF	21.1	250	446	n.d.	12.0	17.7	81.9	62.5	37.4
OCDF	2.24 × 10^3	2.63 × 10^4	5.97 × 10^4	604	206	163	2.11 × 10^4	1.60 × 10^4	1.23 × 10^4
∑2,3,7,8-PCDFs	2.80 × 10^3	3.32 × 10^4	7.23 × 10^4	674	367	241	2.36 × 10^4	1.71 × 10^4	1.33 × 10^4
∑WHO-TEQ (2378-PCDFs)	6.27	81.8	152	0.885	6.96	2.50	33.0	15.4	13.3

---

During the thermal degradation of CB-209 in the three reaction systems for different reaction times, 2,3,7,8-OCDF to -TCDF products were detected at varied concentration levels, as shown in Table 2. Such distribution behavior of the 2,3,7,8-PCDFs indicated that successive hydrodechlorination reactions did occur. Regardless of the reaction system, within a reaction heating time of 30–120 min, OCDF was the dominant congener: 79.2–82.6%, 56.1–89.5%, and 89.2–93.8% in the CB-209 blank, NiFe₂O₄, and Fe₂O₃ reaction systems, respectively. 1,2,3,4,6,7,8-HpCDF was the second dominant congener in all three reaction systems. The 2,3,7,8-HpCDF/OCDF ratio was constantly at ∼1 : 4 in the blank reaction system, and varied in the range of 1 : 2–1 : 9 in the NiFe₂O₄ reaction system, and in the range of 1 : 8–1 : 15 in the Fe₂O₃ reaction system. Moreover, it is interesting to note that the concentration of the 1,2,3,4,6,7,8-HpCDF (up to 1.21 × 10⁴ ng g⁻¹) was always significantly higher than that of 1,2,3,4,7,8,9-HpCDF isomer (up to 446 ng g⁻¹). The higher concentration of 1,2,3,4,6,7,8-HpCDF relative to that of 1,2,3,4,7,8,9-HpCDF suggested that the first OCDF hydrodechlorination step preferentially occurred at the 1- and 9-position than in the 6- and 4-position. A selective dechlorination of the 1- and 9-position of OCDF was also observed by Ballschmiter et al. during UV-photolysis⁵³ and by Weber et al. during pyrolysis of PCDD/Fs on selected fly ash.⁵⁴ The driving force for the fast dechlorination of the 1- and 9-position in OCDF was possibly caused by the steric crowding of the two chlorine atoms in 1- and 9-position.⁵⁵ The steric crowding caused the molecule to twist out of planarity to relieve the crowding, which resulted in the instability of chlorine atoms in 1- and 9-position of OCDF. When compared with 2,3,7,8-substituted HpCDFs, particularly OCDF, 2,3,7,8-substituted HxCDFs were detected at relatively lower concentration levels in all three reaction systems. Nevertheless, 1,2,3,6,7,8-HxCDF was identified as the major isomer among the three 2,3,7,8-substituted HxCDF isomers detected. Similar to 2,3,7,8-substituted HxCDFs, 2,3,7,8-substituted PeCDF and 2,3,7,8-substituted TCDF were detected in relatively low concentration levels in the NiFe₂O₄ reaction system, whereas no 2,3,7,8-substituted isomers were available for these two types of compounds. The major hydrodechlorination pathways were determined by identifying the major intermediate products at each stage. In conclusion, it can be concluded that the predominant hydrodechlorination pathway of the OCDF formed during the thermal degradation of CB-209 to its daughter 2,3,7,8-PCDFs is as follows: OCDF → 1,2,3,4,6,7,8-HpCDF → 1,2,3,6,7,8-HxCDF → 1,2,3,7,8-PeCDF → 2,3,7,8-TCDF (Scheme 1), which was consistent with main hydrodechlorination route of the OCDF on an alumina support.⁵⁶ The further dechlorination of OCDF might relate with thermal stability of the 2,3,7,8-PCDFs and the electronic requirement of each dechlorination reaction pathway.⁵³

Scheme 1 Main hydrodechlorination pathway of the formed 2,3,7,8-PCDF in the NiFe₂O₄ reaction system.

Table 2 shows the ∑WHO-TEQ of 2,3,7,8-PCDFs as a function of reaction time for all three reaction systems. Similar trends were observed in all three reaction systems. Relative to the CB-209 blank reaction system, the ∑WHO-TEQ value of the NiFe₂O₄ reaction system decreased by 86% after 30 min and subsequently by 98.4% after 120 min. The significant reduction of TEQ further indicated the negative effect of NiFe₂O₄ on PCDF formation. In contrast, the ∑WHO-TEQ value of the Fe₂O₃ reaction system increased significantly by 426% after 30 min, and then decreased by 90.7% after 120 min. However, NiFe₂O₄ offers better prospects as a catalyst for the disposal of PCBs based on its inhibitory effects towards the formation of 2,3,7,8-PCDFs within an affordable time frame and time cost.

3.4 Origin of PCDF formation inhibition by NiFe₂O₄

It has been reported that the addition of CaO and urea to a PCDD/F formation system significantly decreases PCDD/Fs emissions.^57–59 This was due to the basicity of the added materials. Similarly, the amount of PCDFs formed was significantly lower in the NiFe₂O₄ reaction system when compared with that in the CB-209 blank reaction system, although NiFe₂O₄ is not basic. The mass balance of CB-209 during the thermal degradation was analyzed by calculation of chlorine atom distribution in the degraded products. Considering the NiFe₂O₄ reaction system as an example, the chlorine atom distribution percentage after degradation of 1 mg CB-209 using 50 mg NiFe₂O₄ at 300 °C for 60 min is shown in Table S1 in the ESI.† The chlorine percentage in 2,3,7,8-PCDD/Fs was only 7.80 × 10⁻⁵% due to the inhibition of PCDF formation by NiFe₂O₄. The percentage of chlorine in the detected PCBs, CBzs and inorganic Cl⁻ were 5.69%, 5.07%, and 74.0%, respectively. The relevant detailed information can be seen in the ESI.†

The inhibition of PCDF formation in the presence of NiFe₂O₄ can be potentially explained as follows. NiFe₂O₄ displays high activities towards the decomposition of biphenyl-like components in CB-209 and daughter PCB products. As previously reported,⁴⁵ NiFe₂O₄ can weakly adsorb bonded oxygen species and surface oxygen such as O₂⁻, O⁻, and OH. The surfaces of metal oxides, especially transition metal oxides, have centers which can supply electrons to the adsorbed oxygen molecules. The active O₂⁻ and O⁻ might be generated from stepwise electron gain by the adsorbed oxygen O₂ on the surface of NiFe₂O₄.⁶⁰ At the early stages of the reaction, the PCB molecules adsorb onto the NiFe₂O₄ surface (which has hydroxyl and metallic sites) through formation of partial hydroxyl bonding and Ni²⁺–Cl or Fe³⁺–Cl interactions, as shown in Scheme 2a. This step is followed by attack of the biphenyl-like components by O₂⁻ and O⁻, resulting in formation of chlorobenzenes and chlorophenols. However, these products are structurally less similar to PCDFs than PCBs; thus, it is anticipated that it will be more difficult to form PCDFs from them than PCBs. Furthermore, highly reactive O₂⁻˙ could be produced during thermal reaction of CB-209 over NiFe₂O₄, and could oxidize PCBs and intermediates of chlorobenzenes and chlorophenols to form micromolecules, such as formic, acetic, and propionic acids, which would hinder PCDF formation. The other possible reason for the relatively low PCDF concentration produced in the NiFe₂O₄ reaction system is related to the spinel structure of NiFe₂O₄. According to previous reports,^19,61 the proposed formation pathway of PCDFs from PCB precursors over NiFe₂O₄ involves the successive steps: adsorption → oxidation → condensation → desorption, as shown in Scheme 2b. It has been reported that the intermediate reduction of iron ions plays an important role in the formation of PCDF from chlorinated aromatics in ferric oxide reaction systems.¹¹ The chemical stability of the spinel structure does not allow for facile metal valence changes from Ni²⁺ and/or Fe³⁺ in NiFe₂O₄ to Ni⁺ and/or Fe²⁺, potentially suppressing PCDF formation. The new and efficient catalytic technique explored in this study might provide a reference method for the disposal of the wastes with PCBs and structurally similar chlorinated aromatics under relatively low temperature.

Scheme 2 (a): Decomposition of CB-209 over NiFe₂O₄ and (b): formation pathway of PCDF using PCB as precursor.

4. Conclusions

The concentrations and congener distribution patterns of 2,3,7,8-PCDD/Fs produced during the decomposition of decachlorobiphenyl (CB-209) over NiFe₂O₄ were investigated and compared with those obtained with Fe₂O₃ catalyst and a catalyst-free system. In all three degradation systems, preferential generation of 2,3,7,8-PCDFs over 2,3,7,8-PCDDs occurred during the degradation of CB-209, suggesting the easier conversion of PCBs into PCDFs. Within the studied reaction period of 30–120 min, OCDF was the dominant congener (56.1–93.8%), followed by 2,3,7,8-substituted HpCDF. Time-dependent formation profile of the 2,3,7,8-PCDFs indicated that decomposition and formation of PCDF occurred simultaneously. The main hydrodechlorination pathway of the formed 2,3,7,8-PCDF is as follows: OCDF → 1,2,3,4,6,7,8-HpCDF → 1,2,3,6,7,8-HxCDF → 1,2,3,7,8-PeCDF → 2,3,7,8-TCDF. The first hydrodechlorination step of OCDF occurred preferentially at the 1- and 9-position than at the 4- and 6-position. Relative to the CB-209 blank reaction, within a reaction time of 30–120 min, the concentration and ∑WHO-TEQ of 2,3,7,8-PCDF decreased by 24.1–99.7% and 86–98.4%, respectively, in NiFe₂O₄ reaction system. However, in the Fe₂O₃ reaction system, the concentration and ∑WHO-TEQ of 2,3,7,8-PCDF increased significantly by 745% and 426%, respectively, followed by reduction to 81.6% and 90.7%, respectively. These results indicated that as-prepared NiFe₂O₄ had a steadily negative effect on PCDF formation. This negative effect might be attributed to its high activity in the destruction of biphenyl-like structure of PCBs as well as its chemical structural stability.

Acknowledgements

This study was supported by the Chinese Academy of Sciences (CAS) (Grant no. KZCX2-YW-QN407), the National 863 Program (2012AA062803), and the National Natural Science Foundation of China (21377147, 21177141, 21321004), and the Youth Innovation Promotion Association, CAS.

References

1. M. Van den Berg, L. S. Birnbaum, M. Denison, M. De Vito, W. Farland, M. Feeley, H. Fiedler, H. Hakansson, A. Hanberg and L. Haws, Toxicol. Sci., 2006, 93, 223.
2. J. I. Baker and R. A. Hites, Environ. Sci. Technol., 2000, 34, 2879.
3. Z. Nie, M. Zheng, W. Liu, B. Zhang, G. Liu, G. Su, P. Lv and K. Xiao, Chemosphere, 2011, 85, 1707.
4. J. C. Hu, M. H. Zheng, Z. Q. Nie, W. B. Liu, G. R. Liu, B. Zhang and K. Xiao, Chemosphere, 2013, 90, 89.
5. V. Steck, H. Riedel and C. Marb, Gefahrstoffe–Reinhalt. Luft, 2007, 67, 471.
6. K. A. Tuppurainen, P. H. Ruokojärvi, A. H. Asikainen, M. Aatamila and J. Ruuskanen, Environ. Sci. Technol., 2000, 34, 4958.
7. K. Hell, L. Stieglitz and E. Dinjus, Environ. Sci. Technol., 2001, 35, 3892.
8. S. Lomnicki and B. Dellinger, J. Phys. Chem. A, 2003, 107, 4387.
9. K. Hell, L. Stieglitz, G. Zwick and R. Will, Organohalogen Compd., 1997, 31, 492.
10. X. Qu, H. Wang, Q. Zhang, X. Shi, F. Xu and W. Wang, Environ. Sci. Technol., 2009, 43, 4068.
11. S. Nganai, S. Lomnicki and B. Dellinger, Environ. Sci. Technol., 2009, 43, 368.
12. T. Lippert, A. Wokaun and D. Lenoir, Environ. Sci. Technol., 1991, 25, 1485.
13. S. L. Alderman, G. R. Farquar, E. D. Poliakoff and B. Dellinger, Environ. Sci. Technol., 2005, 39, 7396.
14. M. S. Milligan and E. R. Altwicker, Environ. Sci. Technol., 1995, 30, 225.
15. S. Lomnicki and B. Dellinger, Proc. Combust. Inst., 2002, 29, 2463.
16. A. Fullana, H. Nakka and S. Sidhu, Organohalogen Compd., 2004, 66, 1126.
17. S. L. Summoogum, D. Wojtalewicz, M. Altarawneh, J. C. Mackie, E. M. Kennedy and B. Z. Dlugogorski, Proc. Combust. Inst., 2013, 34, 3499.
18. Y. Hajizadeh, J. A. Onwudili and P. T. Williams, Chemosphere, 2011, 85, 1672.
19. L. Zhao, H. Hou, K. Shimoda, A. Terada and M. Hosomi, Chemosphere, 2012, 88, 1368.
20. D. A. Abramowicz, Crit. Rev. Biotechnol., 1990, 10, 241.
21. S. Miller, Environ. Sci. Technol., 1982, 16, 98A.
22. Stockholm Convention (SC), 2001, http://www.pops.int.
23. C. Lee and G. L. Huffman, Environ. Prog., 1989, 8, 190.
24. R. Salvador, B. Casal, M. Yates, M. Martín-Luengo and E. Ruiz-Hitzky, Appl. Clay Sci., 2002, 22, 103.
25. Q. Zhang, F. Saito, T. Ikoma, S. Tero-Kubota and K. Hatakeda, Environ. Sci. Technol., 2001, 35, 4933.
26. K. Zhang, J. Huang, H. Wang, G. Yu, B. Wang, S. Deng, J. Kano and Q. Zhang, RSC Adv., 2014, 4, 14719.
27. N. Matykiewiczová, J. Klánová and P. Klán, Environ. Sci. Technol., 2007, 41, 8308.
28. X. Ma, X. Feng, J. Guo, H. Cao, X. Suo, H. Sun and M. Zheng, Appl. Catal., B, 2014, 147, 666.
29. Y. Tanaka, Q. Zhang, F. Saito, T. Ikoma and S. Tero-Kubota, Chemosphere, 2005, 60, 939.
30. Y. F. Sun, M. Takaoka, N. Takeda, W. Wang, X. L. Zeng and T. L. Zhu, Chemosphere, 2012, 88, 895.
31. B. Yang, S. Deng, G. Yu, H. Zhang, J. Wu and Q. Zhuo, J. Hazard. Mater., 2011, 189, 76.
32. Y. Mitoma, N. Egashira and C. Simion, Chemosphere, 2009, 74, 968.
33. M. Jia, G. Su, M. Zheng, B. Zhang and S. Lin, J. Nanosci. Nanotechnol., 2011, 11, 2100.
34. A. Khaleel and A. Al-Nayli, Appl. Catal., B, 2008, 80, 176.
35. G. Su, Y. Liu, L. Huang, Y. Shi, A. Zhang, L. Zhang, W. Liu, L. Gao and M. Zheng, Chemosphere, 2012, 90, 103.
36. X. Ma, Q. Sun, X. Feng, X. He, J. Guo, H. Sun and H. Cao, Appl. Catal., A, 2013, 450, 143.
37. M. Martin-Luengo, H. Martins-Carvalho, J. Ladriere and P. Grange, Clay Miner., 1989, 24, 495.
38. R. T. Kumar, N. C. Sagaya Selvam, T. Adinaveen, L. J. Kennedy and J. J. Vijaya, React. Kinet., Mech. Catal., 2012, 106, 379.
39. A. H. Xu and C. L. Sun, Environ. Technol., 2012, 33, 1339.
40. K. Faungnawakij, N. Shimoda, T. Fukunaga, R. Kikuchi and K. Eguchi, Appl. Catal., B, 2009, 92, 341.
41. Y. Cesteros, P. Salagre, F. Medina and J. E. Sueiras, Appl. Catal., B, 2000, 25, 213.
42. Y. Fan, X. Lu, Y. Ni, H. Zhang, L. Zhao, J. Chen and C. Sun, Environ. Sci. Technol., 2010, 44, 3079.
43. Y. Fan, X. Lu, Y. Ni, H. Zhang, M. Zhu, Y. Li and J. Chen, Appl. Catal., B, 2011, 101, 606.
44. R. M. Freire, F. F. de Sousa, A. L. Pinheiro, E. Longhinotti, J. Mendes Filho, A. C. Oliveira, P. T. C. Freire, A. P. Ayala and A. C. Oliveira, Appl. Catal., A, 2009, 359, 165.
45. L. Huang, G. Su, A. Zhang, Y. Shi, C. Xia, H. Lu, L. Li, S. Liu and M. Zheng, J. Hazard. Mater., 2013, 261, 451.
46. Z. Nie, M. Zheng, G. Liu, W. Liu, P. Lv, B. Zhang, G. Su, L. Gao and K. Xiao, J. Hazard. Mater., 2012, 215, 259.
47. M. Morita, J. Nakagawa, K. Akiyama, S. Mimura and N. Isono, Bull. Environ. Contam. Toxicol., 1977, 18, 67.
48. S. Nganai, S. Lomnicki and B. Dellinger, Chemosphere, 2012, 88, 371.
49. J. Wilhelm, L. Stieglitz, E. Dinjus and R. Will, Chemosphere, 2001, 42, 797.
50. M. Altarawneh, B. Z. Dlugogorski, E. M. Kennedy and J. C. Mackie, Prog. Energy Combust. Sci., 2009, 35, 245.
51. E. R. Altwicker, J. Hazard. Mater., 1996, 47, 137.
52. R. Luijk, D. M. Akkerman, P. Slot, K. Olie and F. Kapteijn, Environ. Sci. Technol., 1994, 28, 312.
53. K. Ballschmiter, W. Zoller, H. Buchert and T. Class, Fresenius' Z. Anal. Chem., 1985, 322, 587.
54. R. Weber, K. Nagai, J. Nishino, H. Shiraishi, M. Ishida, T. Takasuga, K. Konndo and M. Hiraoka, Chemosphere, 2002, 46, 1247.
55. F. Hileman, J. Wehler, S. Gibson, R. Orth, M. Thompson and J. Wendling, Chemosphere, 1989, 18, 1039.
56. M. H. Schoonenboom, H. E. Zoetemeijer and K. Olie, Appl. Catal., B, 1995, 6, 11.
57. Y. F. Wang, L. C. Wang, L. T. Hsieh, H. W. Li, H. C. Jiang, Y. S. Lin and C. H. Tsai, Aerosol Air Qual. Res., 2012, 12, 191.
58. S. Y. Lu, T. Chen, J. H. Yan, X. D. Li, Y. Ni and K. F. Cen, J. Hazard. Mater., 2007, 147, 663.
59. P. Ruokojarvi, A. Asikainen, J. Ruuskanen, K. Tuppurainen, C. Mueller, P. Kilpinen and N. Yli-Keturi, J. Air Waste Manage. Assoc., 2001, 51, 422.
60. A. Bielański and J. Haber, Catal. Rev.: Sci. Eng., 1979, 19, 1.
61. R. Weber, S. Yoshida and K. Miwa, Environ. Sci. Technol., 2002, 36, 1839.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra02580f

---