This is a previous version of the article published in Chemosphere. 2017, 169: 361-368. doi:10.1016/j.chemosphere.2016.11.074

Study of the presence of PCDDs/PCDFs on zero-valent iron nanoparticles

Blanca Calderon*a,b, Lisa Lundina, Ignacio Aracilb, Andres Fullanab

a Department of Chemistry, Umeå University, SE-901 87 Umeå Sweden.

b Permanent address: Chemical Engineering Department, University of Alicante, San Vicente del Raspeig Road, s/n

03690 San Vicente del Raspeig, Alicante, Spain.

* Corresponding author: Tel. +34 965 903 400 X.1116; Fax: + 34 965 903 826; email address: blanca.calderon@ua.es

ABSTRACT

Studies show that nanoscale zero-valent iron (nZVI) particles enhance the formation of

chlorinated compounds such as polychlorinated dioxins and furans (PCDD/Fs) during thermal

processes. However, it is unclear whether nZVI acts as a catalyst for the formation of these

compounds or contains impurities, such as PCDD/Fs, within its structure. We analyzed the

presence of PCDD/Fs in nZVI particles synthesized through various production methods to

elucidate this uncertainty. None of the 2,3,7,8-substituted congeners were found in the

commercially-produced nZVI, but they were present in the laboratory-synthesized nZVI

produced through the borohydride method, particularly in particles synthesized from iron (III)

chloride rather than from iron sulfate. Total PCDD/F WHO-TEQ concentrations of up to 35

pg/g were observed in nZVI particles, with hepta- and octa-chlorinated congeners being the

most abundant. The reagents used in the borohydride method were also analyzed, and our

findings suggest that FeCl3 effectively contains PCDD/Fs at concentrations that could explain

the concentrations observed in the nZVI product. Both FeCl3 and nZVI showed a similar

PCDD/F patterns with slight differences. These results suggest that PCDD/Fs might transfer

from FeCl3 to nZVI during the production method, and thus, care should be taken when

employing certain nZVI for environmental remediation.

KEYWORDS: toxicity; PCDD/F; nanoparticles; nZVI; iron chloride

1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

1. INTRODUCTION

Nanoscale zero-valent iron (nZVI) has proven effective for removing a wide range of

environmental contaminants including chlorinated compounds, heavy metals and nitrates

(Cundy, et al., 2008; Yan, et al., 2013). Due to its nanometric size, it presents a large overall

surface area and numerous active sites, thereby remarkably enhancing the efficiency of

contaminant degradation (Boparai, et al., 2011). Consequently, nZVI is considered a promising

solution for contaminant remediation of soils, groundwater and wastewater, and is being applied

in pilot plant and field tests in both underground water and soil throughout Europe, the United

States of America (the U.S.), and Asia (Karn, et al., 2009).

The synthesis of nZVI is commonly performed using the wet chemical reduction of iron (II) or

(III) with sodium borohydride (see equations 1 and 2) because this reduction can occur at

ambient temperature and pressure (Elliott and Zhang, 2001; Zhang, 2003). Nevertheless, this

method is very costly for large production of nanoparticles due to the high cost of sodium

borohydride (Li, et al., 2009) and the concurrent production of large quantities of hydrogen

(Crane and Scott, 2012). As a result, nowadays, large-scale production of commercial

nanoparticles occurs mainly by using high temperature reduction of oxides in a hydrogen

atmosphere (He and Zhao, 2007). Other alternative methods such as grinding, ultrasound or

electrochemical methods are currently being developed (Han, et al., 2015; Stefaniuk, et al.,

2016; Wiesner, et al., 2006).

4 Fe3+¿+3 BH 4−¿+9 H 2O →4 Fe0 ↓+3 H 2B O3

−¿+12 H+¿+6 H2↑¿

¿¿ ¿ (1)

Fe(H2 O)62+¿+2BH 4

−¿→Fe 0↓+2B( OH )3+7 H 2↑¿ ¿ (2)

Chlorinated organic compounds have received much attention from laboratory scale and field

test researchers studying contaminant degradation by nZVI, due to their widespread presence in

contaminated sites [13, 14]. Studies show that nZVI is effective at degrading a large number of

these compounds, such as trichloroethylene, polychlorinated biphenyls (PCBs), and

2

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

23

organochlorinated pesticides (Li, et al., 2006; Varanasi, et al., 2007). The mechanism for

degradation is the adsorption of the contaminants into the nZVI surface, followed by their

reductive dehalogenation by electron donation (Yan, et al., 2013) (see equations 3 and 4):

Fe0 → Fe2+¿+2 e−¿¿¿ E0 = -0.44 V (3)

RCl+2e−¿+H +¿→ RH+Cl−¿¿¿¿ E0 = 0.5-1.5 V at pH = 7 (4)

Polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs), which are also chlorinated

organic compounds, have caused special concern. PCDD/Fs are persistent, highly toxic, organic

pollutants, especially those with chlorine atoms present at positions 2, 3, 7, and 8 of the

compound. Long term exposure to PCDD/Fs causes immunological, neurological and

reproductive effects, among others, and 2,3,7,8-TCDD is classified as carcinogenic by the U.S.

Environmental Protection Agency (Vallejo, et al., 2015). Given the nature of PCDD/Fs, the

Stockholm Convention established in 2001 that they should irreversibly be transformed or

destroyed to mitigate their ubiquitous presence in the environment. A few studies have

demonstrated effective methods for degrading PCDD/Fs in gas streams (Shu, et al., 2008), fly

ash handling systems (Weber, et al., 2002) and aqueous solutions (Kim, et al., 2008) with

metallic iron nanoparticles or related compounds.

Nevertheless, the high reactivity of nZVI, which makes it effective for environmental

remediation, could have undesirable effects on living organisms, as evaluated in some recent

studies (Crane and Scott, 2012). In addition to their direct effects on cells and organisms,

nanoparticles possess high surface area and strong adsorbent capacity, and could serve as

transport vectors of other contaminants, such as heavy metals and organic substances (Brar, et

al., 2010; Karn, et al., 2009). Furthermore, some studies have shown that iron, including nZVI,

can increase the production of PCDD/Fs under certain conditions. In a recent study by our team,

the addition of nZVI during thermal degradation of PCDD/Fs resulted in decreased amounts of

PCDD/Fs released into the gas phase, and a remarkable increase of PCDD/F content in the soil

(Lundin, et al., 2013). Also, a study by Font et al. showed that combustion of polyvinyl chloride

3

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

45

(PVC) in the presence of nZVI led to increased PCDD/F emissions up to a factor of 104 (Font, et

al., 2010). These studies demonstrate that contaminant degradation treatment using iron

nanoparticles can cause increased production of PCDD/F content in some way. It is still unclear

which mechanisms lead to the formation of PCDD/Fs with use of iron nanoparticles because no

investigation of this phenomenon has been found in the literature. Moreover, it is still unclear

whether the PCDD/Fs come from the nanoparticles, or whether nZVI itself participates in the

formation of PCDD/Fs under certain specific conditions.

This work is aimed to investigate the possible presence of PCDD/Fs in nZVI synthesized in the

lab with varying starting materials and commercially-available nZVI purchased from different

companies. Commercially-obtained nZVI, presumably produced through the dry chemical

reduction method, and lab-synthesized nZVI, using the wet chemical reduction method were

evaluated. In the case of synthesized nZVI, different raw materials were studied to determine

whether differences in PCDD/F content in nZVI depend on the raw materials used, and which

mechanisms (if any) lead to their formation. This investigation is highly novel and could

provide a better understanding of the possible toxic effects of nZVI, which is being widely

applied for the remediation of contaminants.

2. MATERIALS AND METHODS

2.1. Chemical reagents

Chemical reagents used to synthesize nZVI were selected from various sources and differed in

quality. Specifically, five different kinds of iron (III) chloride (FeCl3·6H2O), three kinds of iron

sulfate (FeSO4·7H2O), and two kinds of sodium borohydrate (NaBH4) were used.

Commercially-produced nZVI was purchased from two different companies, Nano Iron s.r.o.,

and Sigma Aldrich. Table 1 shows the specific properties and the origins of the reagents used

for nZVI synthesis, and for the commercially-available nanoparticles purchased for the work.

The rest of the chemicals used for PCDD/Fs analysis (sulfuric acid (H2SO4), n-hexane,

4

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

67

tetradecane, methanol, dichloromethane, etc.) were of analytical grade, and all aqueous

solutions were prepared using high quality purified water (Millipore).

Table 1. Properties of the reagents used for the nZVI synthesis, and of the commercial nZVI

Reagent Company Reference Purity Form Price (€/100g) Abbreviation

FeCl3·6H2O

VWR 24.208.260 ≥ 99% Lumps 10.1 FeCl3-VWRAlfa Aesar A16231.36 > 96% Lumps 3.8 FeCl3-AA-IAlfa Aesar 12497 97-102% Lump 9.6 FeCl3-AA-II

Sigma-Aldrich F2877 ≥ 98% Chunks 18.1 FeCl3-SA-I

Sigma-Aldrich 236489 97% Chunk 13.1 FeCl3-SA-II

FeSO4·7H2O

VWR 24.244.298 99.5–105.0% Powder 4.5 FeSO4-VWR

Alfa Aesar 14498.3 ≥ 99% Crystalline 9.4 FeSO4-AASigma-Aldrich F7002 ≥ 99% Powder 13.8 FeSO4-SA

NaBH4

Alfa Aesar 38788 ≥ 97.1% Caplets 69.9 NaBH4-AASigma-Aldrich 452882 ≥ 98% Powder 59.6 NaBH4-SA

nZVI

Sigma Aldrich 746843 99.5% Powder (35-

45 nm) 1076.0 nZVI-SA

Nanoiron, s.r.o.

NANOFER 25

14-18% of Fe in water

Aqueous suspension

(nZVI size < 50 nm)

6.1 nZVI-NI

2.2. Synthesis of nZVI

The following wet chemical reduction method, known as the borohydride reduction method but

with slight modifications, was used to synthesize nZVI. Briefly, 1.5 molar excess of an aqueous

solution of selected NaBH4 was added slowly to a solution of FeCl3 or FeSO4 in water

(depending on the case) to produce the nanoparticles. After that, the nanoparticles were

separated from the water by decantation, with the aid of a magnet, and then washed four times

with distilled water to remove the remaining borohydride. The synthesis of nZVI was performed

using both ferrous and ferric iron, using different qualities of FeCl3 and FeSO4, respectively.

PCDD/Fs extraction and analysis were performed immediately after nZVI synthesis.

2.3. PCDD/Fs analysis

5

104

105

106

107

108

109

110

111

112

113

114

115

116

117

89

The FeCl3, FeSO4 and NaBH4 were dissolved in 250 mL of 0.5 M H2SO4. Then, the solutions

were extracted with n-hexane by three sequential extractions with 100 mL each, and the extracts

were combined and concentrated to generate 40 l of tetradecane which was then cleaned

according to a method by Liljelind et al. (Liljelind, et al., 2003).

In short, 13C-labelled internal standards were added to the extract. Apart from the additional use

of 1,2,3,4,6-pentachlorodibenzofuran (1,2,3,4,6-PeCDF) and 1,2,3,7,8,9-

hexachlorodibenzofuran (1,2,3,7,8,9-HxCDF) with the labeled standards and the use of super-

activated carbon AX21-carbon, subsequent treatment and analysis of PCDD/F in these samples

followed the EU standard methods 1948:1-3. Samples were purified prior to PCDD/F analysis

using a multilayer silica column, a super alumina column, and a column comprising a pre-

washed mixture of AX21-carbon and Celite (7.9:92.1 w/w). The concentrations of PCDD/F

congeners were determined using gas chromatography-high resolution mass spectrometry (GC-

HRMS), equipped with a DB5 column and a Waters AutoSpec ULTIMA NT 2000D high

resolution mass spectrometer. Recovery of the internal standard was well within the limits set

by the EU standard methods 1948:1-3.

The accuracy and precision of the analysis were checked by measuring blanks and certified

standards. The analytes were quantified by the isotope dilution technique using 13C-labeled

internal and recovery standards. The World Health Organization 2005 Toxic Equivalency

Factors (WHO2005-TEF) were applied to obtain the toxicity of the samples in terms of PCDD/F

content. All analyses were performed in triplicate, and the standard deviation of all

measurements was determined and is represented as error bars in all figures.

3. RESULTS AND DISCUSSION

3.1. PCDD/Fs in nZVI particles

6

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

1011

Four different kinds of nanoparticles, two commercially-available and two lab-synthesized

through the borohydride method, were analyzed for PCDD/F content. Table 2 shows the results

of the concentrations of the 2,3,7,8-substituted congeners and Table S1 in Supporting

Information the concentrations of the total homologues of PCDDs and PCDFs. No PCDD/Fs

were detected in either of the commercially-produced nZVI particles, except for MoCDF and

DiCDD in the nZVI from Sigma Aldrich. Of the two lab-synthesized nZVI particles, the ones

produced using FeCl3 (nZVI-B-ferric) contained about 100 times more PCDD/Fs than those

produced using FeSO4 (nZVI-B-ferrous). More explicitly, the nZVI-B-ferric contained high

amounts of the highly-substituted PCDD/Fs, i.e. the hepta- and octa-chlorinated congeners

which are more abundant, with concentrations of 850, 830 and 3500 pg g-1 for HpCDD, HpCDF

and OCDF respectively. Two noteworthy findings were that the PCDD/Fs present in the nZVI-

B-ferrous particles were also the highly-chlorinated hepta- and octa-chlorinated congeners, and

that no other degrees of chlorination were detected except for the mono-chlorinated congeners.

The nZVI-B-ferric particles possessed approximately 5-fold more PCDFs than PCDDs.

No 2,3,7,8-substituted congeners were found in the commercially-produced nZVI particles. The

lab-synthesized nZVI-B-ferrous particles only contained 1,2,3,4,6,7,8-HpCDF and OCDF,

while nZVI-B-ferric particles contained all the 2,3,7,8-substituted PCDF congeners. The nZVI-

B-ferric particles possessed lower concentrations of 2,3,7,8-substituted PCDD than PCDF, but

neither 2,3,7,8-TCDD nor 1,2,3,4,7,8-HxCDD were detected.

However, since nZVI-B-ferric particles also contained significant concentrations of the most

toxic homologues, such as PeCDD, TCDD and PeCDF, their WHO toxic equivalent (WHO–

TEQ) concentration reached a value of 35 pg g-1. In the case of nZVI-B-ferrous particles, again

only 1,2,3,4,6,7,8-HpCDF and OCDF were detected, thus giving the material a lower WHO-

TEQ concentration than the measurement detection limit (0.1 pg g-1).

Consequently, we concluded that the commercially-produced nZVI particles possess a

negligible PCDD/F content, while lab-synthesized nZVI particles produced using the

7

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

1213

borohydride method, particularly the ones synthesized with iron chloride, possess significant

PCDD/F content. The possible origin of these persistent compounds will be further assessed in

subsequent sections.

Table 2. Concentrations of 2,3,7,8-substituted congeners in commercially-produced and lab-

synthesized nZVI. Concentrations are in pg g-1 sample. The detection limit of the analysis was

0.5 pg g-1

Commercially-produced Lab-synthesizednZVI-SA nZVI-NI nZVI-B-ferric nZVI-B-ferrous

2-MoCDD n.d. - n.d. n.d.23-DiCDD n.d. - n.d. n.d.237-TriCDD n.d. n.d. 1±0.2 n.d.2378-TCDD n.d. n.d. n.d. n.d.12378-PeCDD n.d. n.d. 3±0.7 n.d.123478-HxCDD n.d. n.d. n.d. n.d.123678-HxCDD n.d. n.d. 4±0.1 n.d.123789-HxCDD n.d. n.d. 5±1 n.d.1234678-HpCDD n.d. 1±0.2 19±3.1 n.d.OCDD n.d. n.d. 78±14 n.d.2-MoCDF 15±1.5 n.d. 4±0.8 n.d.28-DiCDF n.d. n.d. n.d. n.d.248-TriCDF n.d. n.d. 3±0.6 n.d.2378-TCDF n.d. n.d. 8±1 n.d.12378-PeCDF n.d. n.d. 23±3.7 n.d.23478-PeCDF n.d. n.d. 13±2.2 n.d.123478-HxCDF n.d. n.d. 97±12 n.d.123678-HxCDF n.d. n.d. 51±2.3 n.d.123789-HxCDF n.d. n.d. 17±2.6 n.d.234678-HxCDF n.d. n.d. 25±6.1 n.d.1234678-HpCDF n.d. n.d. 460±96 7±11234789-HpCDF n.d. n.d. 75±7.0 n.d.OCDF n.d. n.d. 3500±370 43±3.5

WHO2005-TEQ 0.0 0.01±0.002 35±5 0.1±0.01

- : could not be quantified due to disturbances during GC/HRMS analysis.

n.d.: under detection limit (<0.5 pg g-1)nZVI-SA: commercially-produced nanoparticles purchased from

Sigma-Aldrich (their characteristics can be viewed in Table 1).

nZVI-NI: commercially-produced nanoparticles purchased from Nano Iron, s.r.o.

nZVI-B-ferric: nZVI synthesized through the borohydride method using the reactants FeCl3-SA-I and

NaBH4-AA as raw materials.

8

168

169

170

171

172

173

174

175

176

177

178

179

180

1415

nZVI-B-ferrous: nZVI synthesized through the borohydride method using the reactants FeSO4-AA and

NaBH4-AA as raw materials.

3.2. Determination of the origin of PCDD/F in the nZVI

Secondly, we analyzed the raw materials commonly used to synthesize nZVI for laboratory and

field applications for PCDD/Fs to determine the origin of the PCDD/Fs detected in the lab-

synthesized nZVI particles. Synthesis of nZVI particles with the borohydride method can be

performed using either ferrous or ferric iron (see equations 3 and 4). Studies of the

environmental remediation of contaminants with nZVI use iron sulfate as a source of ferrous

iron and iron chloride for ferric iron for nZVI particle synthesis. NaBH4 is the reducing agent

for nZVI particle synthesis in a majority of studies (Fu, et al., 2014). Thus, we evaluated

FeCl3·6H2O, FeSO4·7H2O, and NaBH4 from various sources for PCDD/Fs content. Table 3

shows the results of the PCDD/F 2,3,7,8-substituted congeners analyses for iron chlorides and

iron sulfates, Table S2 in the Supporting Information their concentration of the total

homologues of PCDD/Fs and Table S3, shows the PCDD/F analysis for NaBH4.

Table 3. Concentrations of 2,3,7,8-substituted congeners in reagents used for nZVI particle

synthesis using the wet chemical reduction method. Concentrations are in pg g-1 sample. The

detection limit of the analysis was 0.5 pg g-1

FeCl3-VWR

FeCl3-AA-I

FeCl3-AA-II

FeCl3-SA-I

FeCl3-SA-II

FeSO4-VWR

FeSO4-AA

FeSO4-SA

2-MoCDD - - - - - - - -23-DiCDD - - n.d. n.d. n.d. - - -237-TriCDD n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.2378-TCDD n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.12378-PeCDD n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.123478-HxCDD n.d. n.d. n.d. 0.5±0.06 n.d. n.d. n.d. n.d.123678-HxCDD n.d. n.d. n.d. 1±0.1 n.d. n.d. n.d. n.d.123789-HxCDD n.d. n.d. n.d. 0.5±0.04 n.d. n.d. n.d. n.d.1234678-HpCDD 2±0.2 3±0.5 n.d. 5±0.5 1±0.1 1±0.1 n.d. n.d.OCDD 8±2 14±2 1±0.1 26±3.5 2±0.6 3±0.7 1±0.1 n.d.

9

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

1617

FeCl3-VWR

FeCl3-AA-I

FeCl3-AA-II

FeCl3-SA-I

FeCl3-SA-II

FeSO4-VWR

FeSO4-AA

FeSO4-SA

2-MoCDF 84±29 37±8 0.5±0.02 17±3.2 2±0.4 n.d. n.d. n.d.28-DiCDF n.d. n.d. n.d. 1±0.1 n.d. n.d. n.d. n.d.248-TriCDF n.d. 1±0.1 n.d. 1±0.2 n.d. n.d. n.d. n.d.2378-TCDF n.d. n.d n.d. 1±0.2 n.d. n.d. n.d. n.d.12378-PeCDF n.d. n.d. n.d. 1±0.01 3±0.9 n.d. n.d. n.d.23478-PeCDF n.d. n.d. n.d. 1±0.1 n.d. n.d. n.d. n.d.123478-HxCDF n.d. n.d. n.d. 7±1.2 n.d. n.d. n.d. n.d.123678-HxCDF n.d. n.d. n.d. 1±0.1 n.d. n.d. n.d. n.d.123789-HxCDF 1±0.1 1±0.1 1±0.1 1±0.1 1. ±0.1 0.6±0.2 1±0.2 n.d.234678-HxCDF n.d. n.d. n.d. 1±0.1 n.d. n.d. n.d. n.d.1234678-HpCDF n.d. 3±0.3 1±0.2 23±2.9 n.d. n.d. n.d. n.d.1234789-HpCDF n.d. 1±0.2 n.d. 6±0.4 n.d. n.d. n.d. n.d.OCDF 2±0.1 216±21 1±0.5 740±74 4±1 2±0.3 1±0.2 n.d.

WHO-TEQ 0.1±0.02 0.2±0.02 0.1±0.02 2.1±0.3 0.1±0.04 0.1±0.02 0.1±0.03 0.0

n.d.: under detection limit (<0.5 pg g-1)

- : could not be quantified due to disturbances during GC/HRMS analysis

As shown in Table S2 and S3 in the Supporting Information, the iron chlorides contained

PCDD/Fs, while the iron sulfates and the NaBH4 were practically free of them. The FeCl3-SA-I

and FeCl3-AA-I samples had the highest PCDD/Fs content, followed by FeCl3-VWR. In

general, PCDFs were present in higher concentrations than PCDD, and the highly-substituted

PCDD/Fs were more abundant than less substituted ones, except for MoCDD/F, which was

present at remarkable concentrations.

Moreover, in the FeCl3-SA-I sample, all toxic PCDF isomers were present, resulting in a WHO-

TEQ concentration of 2.1 pg g-1. Although the toxicities of the other chlorides were not as high

as those of the FeCl3-SA-I sample, their values, which were lower than a WHO-TEQ

concentration of 0.2 pg g-1, should not be neglected. Since a 5:1 mass relationship (between the

iron salt reagent and nZVI particles, respectively) is used for nZVI particle synthesis, PCDD/F

transfer from the raw materials to the nZVI could generate high concentrations of PCDD/Fs in

the nZVI particles, thus producing significant toxicity values for the nZVI particles

10

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

1819

The iron sulfates had negligible concentrations of PCDD/Fs. Only 4 of the 2,3,7,8-substituted

congeners could be detected, 1,2,3,4,6,7,8-HpCDD, OCDD, 1,2,3,7,8,9-HxCDF and OCDF, all

of which were present at concentrations below 2 pg g -1. Consequently, they possessed low

WHO-TEQ concentration values (see Table 3). As expected, PCDD/F concentrations in the

sodium borohydride was very similar to those in the iron sulfates; no PCDD/F content was

detected.

PCDD/F content varied considerably from one chloride to another, as mentioned above. The

differences might be attributed to several factors: purity of the compounds, method of

production, etc. An observation of the properties of each compound (Table 1) reveals that

compounds possessing higher PCDD/F content are not necessarily less pure or less costly.

Indeed, FeCl3-SA-I had the highest PCDD/F content and toxicity, intermediate purity (>98%),

and the highest price. Moreover, FeCl3-VWR, which had a higher PCDD/F content than FeCl3-

SA-II and FeCl3-AA-II, also had a higher purity than the others (>99% compared to 97%).

Consequently, the variation between the samples in total PCDD/F content could be due to the

specific method of production of these compounds, and not simply to their purity.

According to the literature, iron (III) chloride solutions are prepared by dissolving iron or iron

ore in hydrochloric acid and oxidizing the resulting iron (II) chloride with chlorine:

Fe3O4(s) + 8 HCl(aq) → FeCl2(aq) + 2 FeCl3(aq) + 4 H2O (5)

2 FeCl2(aq) + Cl2(g) → 2 FeCl3(aq) (6)

In a continuous, closed-cycle process, iron (III) chloride solution is reduced with iron, and the

resulting iron (II) chloride solution is re-oxidized with chlorine. Solid iron (III) chloride

hexahydrate is then crystallized by cooling a hot concentrated solution (Elvers, et al., 1989).

Both the reagents and operating conditions used in all the steps of the iron (III) chloride

manufacturing process, starting with the raw materials, are likely to create a favorable

environment for PCDD/F formation at any stage. At the industrial level, iron ore is initially

11

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

2021

subjected to a sintering process that converts iron ore fines into larger agglomerates. A mixture

of fine ores, coke, lime or limestone, and iron-bearing residues is heated at high temperatures

and sintered into a porous feedstock acceptable for the blast furnace (Buekens, et al., 2001).

Since chlorine is unavoidable in the feed materials, PCDD/Fs are generated as byproducts of the

process, as previous studies have shown (Ooi and Lu, 2011; Sun, et al., 2016).

Ryan and Altwicker (Ryan and Altwicker, 2004) investigated the formation of PCDD/Fs with

mixtures of black carbon and three iron chloride types: iron (II) chloride tetrahydrate, iron (III)

chloride hexahydrate, and iron (III) oxychloride. They concluded that iron chlorides are

important chlorinating agents and promoters of low-temperature carbon gasification. The

predominant PCDD/F isomer was OCDF, as we have observed in this study. Kuzuhara et al.

studied the influence of several metal chlorides, including FeCl3·6H2O, on de novo PCDD/F

formation and concluded that iron chloride activity is important for the formation of PCDD/Fs

(Kuzuhara, et al., 2003). Lin et al. analyzed the influence of inorganic and organic flocculants

on PCDD/F formation during sewage sludge incineration (Lin, et al., 2015). They concluded

that the presence of poly-ferric chloride and polyaluminium chloride increased the formation of

PCDD/F during combustion experiments, suggesting that metals promote PCDD/F formation.

However, to our knowledge, none of these authors analyzed the PCDD/F content of the iron

compounds used in their studies. This analysis would have been useful for more accurately

determining whether iron compounds are catalysts for the formation of PCDD/Fs, or sources of

PCDD/Fs themselves. However, as the homologue patterns of PCDD/Fs changed as a function

of the reaction conditions and not only with the iron chloride concentration, it is more

reasonable to suggest that iron chloride could be involved in the de novo formation of PCDD/Fs

rather than being the only source of PCDD/Fs.

Accounting for all of the above mentioned findings, iron chloride may be responsible for

transporting PCDD/Fs into the iron nanoparticles. It is also expected that the higher the PCDD/F

content in the iron chloride, the higher the resulting PCDD/Fs concentration in the nZVI

particles.

12

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

2223

3.3. Transport and transformation of PCDD/Fs from iron chloride to nZVI during its synthesis

through the borohydride method.

During nZVI particle synthesis through the borohydride method, iron chloride is first diluted in

water. Consequently, any PCDD/Fs present in the iron chloride will also be present in the

aqueous solution. Sodium borohydride is slowly added to the ferric solution to produce the

nanoparticles, and the synthesis takes 30 to 60 minutes. When nZVI particles are formed,

PCDD/Fs present in the solution are most likely adsorbed into the formed particles. During

nZVI particle synthesis, PCDD/Fs may be chlorinated or dechlorinated, thereby changing the

homolog profile and isomer pattern. We compared the PCDD/F pattern in nZVI particles

synthesized from an iron chloride with a pre-determined PCDD/F content (nZVI-B-ferric) to the

expected PCDD/Fs pattern that would result if all PCDD/Fs coming from the iron chloride

source were attached to the nZVI particle surface during the synthesis (nZVI-calc). This

comparison was performed to assess changes in the isomer pattern or homolog profile due to

chlorination or dechlorination of PCDD/Fs during nZVI particle synthesis.

FeCl3-SA-I (24.2 g) and NaBH4-AA (15.24 g) were used to synthesize nZVI particles (5 g).

Thus, PCDD/Fs were expected to be concentrated in the nZVI particles by a factor of 4.8.

13

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

2425

0.001

0.01

0.1

1

10C

(pm

ol/g

)

nZVI-calc nZVi-ferric

0.001

0.01

0.1

1

10

C (p

mol

/g)

nZVI-calc nZVI-B-ferric

Figure 1. Concentrations of 2,3,7,8-substituted congeners (a) and homolog concentrations of

mono- up to octa-chlorinated PCDD and PCDF (b) of nZVI-B-ferric and nZVI-calc. The nZVI-

B-ferric was synthesized using FeCl3-SA-I and NaBH4-AA.

14

a)

b)

286

287

288

289

290

2627

Table 4. WHO-TEQ concentration, total PCDD and PCDF content, and degree of chlorination

in nZVI-calc and nZVI-B-ferric.

nZVI-calc nZVI-B-ferricpg g-1 pmol g-1 pg g-1 pmol g-1

SUM PCDD 1680±300 3.95±0.72 1260±220 3.17±0.58SUM PCDF 4250±470 10.4±1.2 5390±680 13.0±1.7SUM PCDD/F 5930±770 14.3±2.9 6650±900 16±2.6WHO-TEQ 10.0±1 0.026±0.004 34.4±5.2 0.092±0.014Chlorine content in PCDD (pmol Cl g-1) 28±5.0 20±3.3Chlorine content in PCDF (pmol Cl g-1) 73±7.7 93±12Degree of Chlorination in PCDD1 (mol Cl mol-1 PCDD) 7.0±2.1 6.1±1.9

Average chlorine content in PCDF (mol Cl mol-1 PCDF) 7.0±1.7 7.1±1.9

1Degree of Chlorination = Σ(Homolog sum × No. of Cl)/ ΣPCDD

The results shown in Figure 1 and Table 4 reveal that PCDDs patterns were very similar for the

more highly chlorinated PCDDs in nZVI-calc particles and in nZVI-B-ferric particles, further

supporting the hypothesis that PCDD/Fs from the FeCl3 is transferred to the nZVI particles

during the synthesis. However, a larger difference in PCDF patterns between nZVI-calc and

nZVI-B-ferric particles was observed, which may imply a transformation between the different

congeners during synthesis. The total PCDD/F content in nZVI-B-ferric particles was a bit

higher than, but very similar to, that which was expected theoretically (nZVI-calc). Thus, we

could conclude that the synthesized nZVI particles probably adsorbed all, or almost all, of the

PCDD/Fs from the iron chloride. The difference between the actual and expected values was

lower than the standard deviation, hence our conclusion that both values were very similar.

Nevertheless, there was a remarkable change in total WHO-TEQ toxicity, since nZVI-B-ferric

particles were 3.5 times greater than the theoretical value. This change was attributed to a

change in the PCDD/F isomer pattern of the toxic congeners present in the nZVI particles. A

closer look at the 2,3,7,8-substituted congeners presented in Figure 1a shows that some of the

isomers that highly influence the WHO-TEQ concentration, such as 1,2,3,7,8-PCDD, 1,2,3,7,8-

15

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

2829

PeCDF, 1,2,3,4,7,8-HxCDF, etc, were present at higher concentrations in the nZVI-B-ferric

than expected.

The chlorine content was very similar in both cases (calculated and ferric nZVI particles), and

when accounting for the standard deviation of the values in Table 4, we could not attribute any

difference between those values to the average chlorine content. Thus, we conclude that during

nZVI synthesis, although the PCDD and PCDF contents are similar in the starting iron chloride

raw material and in the finished nZVI product, there is a change in the isomer pattern towards

formation of some of the most toxic congeners. The change in isomer pattern leads to a

significant increase in the WHO-TEQ concentration of the finished nZVI product.

3.4. Environmental implications

This study has shown that the use of iron chloride as a source of iron in nZVI particle synthesis

through the borohydride method can lead to the adsorption of PCDD/Fs onto the nZVI particles

during the synthesis procedure. We suggest that PCDD/Fs originate from the iron chloride raw

material due to the method of iron chloride production. Our results have remarkable

environmental implications since nZVI is used for environmental remediation of water or soils.

Under certain conditions, PCDD/Fs could be released from the nZVI particles and delivered to

drinking water or waste water. The PCDD/Fs could then end up in sewage sludge, thereby

increasing the PCDD/F content in the environment and resulting in negative effects. Such

transport of other contaminants, such as heavy metals and organic substances, has already been

observed in other studies (Brar, et al., 2010; Karn, et al., 2009) and could also occur for

PCDD/Fs. Thus, from an environmental perspective and with respect to PCDD/F content, we

recommend nZVI to be synthesized using iron sulfate as a starting material instead of iron

chloride, or that iron chloride is pre-treated to remove PCDD/Fs before use for nZVI synthesis.

We should also consider the method of production of the nZVI when using them for

dechlorinating contaminant compounds. If the nZVI are produced from iron chloride using the

sodium borohydride method, the nZVI will likely adsorb PCDD/Fs onto its surface. Applying

16

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

3031

these nZVI particles as a specific treatment for removing any pollutant could lead to higher

concentrations of some chlorinated micropollutants in the environment than were present before

nZVI treatment. These increased concentrations of chlorinated micropollutants may lead to

unexplained or misunderstood results after any evaluations of nZVI treatments.

It is also noteworthy that this work demonstrated that some iron chlorides have a significantly

greater PCDD/F content than others which, to our knowledge, has not previously been shown in

the literature. Although the PCDD/F content in iron chlorides is not much higher than in other

substances, it must be taken into account, because iron chloride, which is a known coagulator in

waste water treatment, could be another source of PCDD/Fs in the environment. Iron chloride

may be one of other sources that partially contribute to an explanation for the occurrence of

PCDD/Fs in sewage sludge. PCDD/F levels encountered in sewage sludge samples as well as

total TEQ toxicity, and both homolog and isomer PCDD/F patterns, vary considerably

depending on several factors including the origin of the sludge (rural/urban/industrial areas) and

waste water treatment (Balasubramani and Rifai, 2015). As a matter of interest, Pereira and

Kuch studied the content of PCDD/Fs, PCBs, and metals in three sewage sludge samples

(Pereira and Kuch, 2005). Their results show a strong correlation between the iron content in

each sample (9.1, 19.7 and 60.1 g kg-1) and the PCDD/F WHO-TEQ values obtained (2.3, 26.9

and 128.5 pg WHO-TEQ g-1). Obviously, this finding does not prove that PCDD/Fs come from

use of ferric chloride as a coagulant in waste water treatment, but in any case we think it would

be interesting to study this in further detail.

4. CONCLUSIONS

This study showed that, under certain conditions, PCDD/Fs can appear in nZVI synthesized in

the laboratory through the borohydride method when iron (III) chloride is the ferric source. We

propose that iron (III) chloride, used as a reagent in the process, is the source of the PCDD/Fs

present in nZVI particles. Iron (III) chloride can contain impurities, including PCDD/Fs, that

17

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

3233

may have formed during its manufacture. Among the PCDD/F isomers encountered, hepta- and

octa-chlorinated PCDD/Fs were more abundant both in ferric chloride and nZVI particles, with

much higher concentrations of PCDFs than PCDDs. There were slight differences in the

patterns between ferric chloride and nZVI particles that influenced their respective total TEQ

values.

None of the toxic 2,3,7,8-substituted PCDD/F isomers were detected in the commercially-

produced nZVI particles. Thus, if lab-synthesized nZVI particles produced through the

borohydride method using iron (III) chloride are to be used for environmental remediation of

contaminants, the likely presence of PCDD/Fs in nZVI should be meaningfully considered, and

a preliminary cleaning treatment should be performed before these nZVI are used. Although the

maximum PCDD/F WHO-TEQ concentration observed in nZVI, (35 pg g-1) was not especially

high, these values must be acknowledged since nZVI have proven to be very efficient transport

vectors of contaminants that are adsorbed onto their surface.

Finally, we recommend further investigation of, but not limited to, the following: (i) whether the

potential risk that PCDD/Fs are transferred into the environment from nZVI exists; (ii) broader

spectra of ferric chloride samples to analyze their PCDD/F content; and (iii) whether the

presence of PCDD/Fs in ferric chloride is relevant when applying this compound during

processes such as waste water treatment. Any correlation between PCDD/F and iron content in

sewage sludge samples relative to the various sources of iron (such as ferric chloride, ferric

sulfate, etc.) could provide insight for the latter recommendation in particular.

ACKNOWLEDGEMENTS

The study was performed at the Environmental Chemistry Laboratory, Umeå University and the

PCDD/F analysis were done at the Trace Analysis Platform, Umeå University. We would like to

acknowledge the University of Alicante (EEBB-UA2013) for financing this study and Per

Liljelind for his valuable help with the PCDD/F analysis.

REFERENCES

18

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

3435

References

Balasubramani, A., Rifai, H.S., 2015. Occurrence and distribution of polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans (PCDD/Fs) in industrial and domestic sewage sludge. Environ. Sci. Pollut. Res. 22, 14801-14808.

Boparai, H.K., Joseph, M., O'Carroll, D.M., 2011. Kinetics and thermodynamics of cadmium ion removal by adsorption onto nano zerovalent iron particles. J. Hazard. Mater. 186, 458-465.

Brar, S.K., Verma, M., Tyagi, R.D., Surampalli, R.Y., 2010. Engineered nanoparticles in wastewater and wastewater sludge – Evidence and impacts. Waste Manage. 30, 504-520.

Buekens, A., Stieglitz, L., Hell, K., Huang, H., Segers, P., 2001. Dioxins from thermal and metallurgical processes: Recent studies for the iron and steel industry. Chemosphere. 42, 729-735.

Crane, R.A., Scott, T.B., 2012. Nanoscale zero-valent iron: Future prospects for an emerging water treatment technology. J. Hazard. Mater. 211-212, 112-125.

Cundy, A.B., Hopkinson, L., Whitby, R.L.D., 2008. Use of iron-based technologies in contaminated land and groundwater remediation: A review. Sci. Total Environ. 400, 42-51.

Elliott, D.W., Zhang, W.-., 2001. Field assessment of nanoscale bimetallic particles for groundwater treatment. Environ. Sci. Technol. 35, 4922-4926.

Elvers, B., Hawkins, S., Ravenscroft, M., Schulz, G., 1989. Ullmann's Encyclopedia of Industrial Chemistry. A, 14, 5th ed.

Font, R., Gálvez, A., Moltó, J., Fullana, A., Aracil, I., 2010. Formation of polychlorinated compounds in the combustion of PVC with iron nanoparticles. Chemosphere. 78, 152-159.

Fu, F., Dionysiou, D.D., Liu, H., 2014. The use of zero-valent iron for groundwater remediation and wastewater treatment: A review. J. Hazard. Mater. 267, 194-205.

Han, Y., Yang, M.D.Y., Zhang, W., Yan, W., 2015. Optimizing synthesis conditions of nanoscale zero-valent iron (nZVI) through aqueous reactivity assessment. Front. Environ. Sci. Eng.

He, F., Zhao, D., 2007. Manipulating the size and dispersibility of zerovalent iron nanoparticles by use of carboxymethyl cellulose stabilizers. Environ. Sci. Technol. 41, 6216-6221.

Karn, B., Kuiken, T., Otto, M., 2009. Nanotechnology and in situ remediation: A review of the benefits and potential risks. Environ. Health Perspect. 117, 1823-1831.

Kim, J.-., Tratnyek, P.G., Chang, Y.-., 2008. Rapid dechlorination of polychlorinated dibenzo-p-dioxins by bimetallic and nanosized zerovalent iron. Environ. Sci. Technol. 42, 4106-4112.

Kuzuhara, S., Sato, H., Kasai, E., Nakamura, T., 2003. Influence of metallic chlorides on the formation of PCDD/Fs during low-temperature oxidation of carbon. Environ. Sci. Technol. 37, 2431-2435.

Li, S., Yan, W., Zhang, W.-., 2009. Solvent-free production of nanoscale zero-valent iron (nZVI) with precision milling. Green Chem. 11, 1618-1626.

19

388

389390391

392393

394395

396397398

399400

401402

403404

405406

407408

409410

411412413

414415

416417

418419

420421422

423424

3637

Li, X.-., Elliott, D.W., Zhang, W.-., 2006. Zero-valent iron nanoparticles for abatement of environmental pollutants: Materials and engineering aspects. Crit. Rev. Solid Statae Mater Sci. 31, 111-122.

Liljelind, P., Söderström, G., Hedman, B., Karlsson, S., Lundin, L., Marklund, S., 2003. Method for multiresidue determination of halogenated aromatics and PAHs in combustion-related samples. Environ. Sci. Technol. 37, 3680-3686.

Lin, X., Li, X., Lu, S., Wang, F., Chen, T., Yan, J., 2015. Influence of organic and inorganic flocculants on the formation of PCDD/Fs during sewage sludge incineration. Environ. Sci. Pollut. Res. 22, 14629-14636.

Lundin, L., Moltó, J., Fullana, A., 2013. Low temperature thermal degradation of PCDD/Fs in soil using nanosized particles of zerovalent iron and CaO. Chemosphere. 91, 740-744.

Ooi, T.C., Lu, L., 2011. Formation and mitigation of PCDD/Fs in iron ore sintering. Chemosphere. 85, 291-299.

Pereira, M.D.S., Kuch, B., 2005. Heavy metals, PCDD/F and PCB in sewage sludge samples from two wastewater treatment facilities in Rio de Janeiro State, Brazil. Chemosphere. 60, 844-853.

Ryan, S.P., Altwicker, E.R., 2004. Understanding the role of iron chlorides in the de novo synthesis of polychlorinated dibenzo-p-dioxins/dibenzofurans. Environ. Sci. Technol. 38, 1708-1717.

Shu, H.C., Jhy, W.Y., Hung, M.C., Li, Y.H., Kai, H.C., Moo, B.C., 2008. PCDD/F adsorption and destruction in the flue gas streams of MWI and MSP via Cu and Fe catalysts supported on carbon. Environ. Sci. Technol. 42, 5727-5733.

Stefaniuk, M., Oleszczuk, P., Ok, Y.S., 2016. Review on nano zerovalent iron (nZVI): From synthesis to environmental applications. Chem. Eng. J. 287, 618-632.

Sun, Y., Liu, L., Fu, X., Zhu, T., Buekens, A., Yang, X., Wang, Q., 2016. Mechanism of unintentionally produced persistent organic pollutant formation in iron ore sintering. J. Hazard. Mater. 306, 41-49.

Vallejo, M., Fresnedo San Román, M., Ortiz, I., Irabien, A., 2015. Overview of the PCDD/Fs degradation potential and formation risk in the application of advanced oxidation processes (AOPs) to wastewater treatment. Chemosphere. 118, 44-56.

Varanasi, P., Fullana, A., Sidhu, S., 2007. Remediation of PCB contaminated soils using iron nano-particles. Chemosphere. 66, 1031-1038.

Weber, R., Nagai, K., Nishino, J., Shiraishi, H., Ishida, M., Takasuga, T., Konndo, K., Hiraoka, M., 2002. Effects of selected metal oxides on the dechlorination and destruction of PCDD and PCDF. Chemosphere. 46, 1247-1253.

Wiesner, M.R., Lowry, G.V., Alvarez, P., Dionysiou, D., Biswas, P., 2006. Assessing the risks of manufactured nanomaterials. Environ. Sci. Technol. 40, 4336-4345.

Yan, W., Lien, H.-., Koel, B.E., Zhang, W.-., 2013. Iron nanoparticles for environmental clean-up: Recent developments and future outlook. Environ. Sci. Process. Impacts. 15, 63-77.

20

425426427

428429430

431432433

434435

436437

438439440

441442443

444445446

447448

449450451

452453454

455456

457458459

460461

462463

3839

Zhang, W.-., 2003. Nanoscale iron particles for environmental remediation: An overview. J. Nanopart. Res. 5, 323-332.

21

464465

466

4041