w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 0 9 1 – 4 1 0 0

Avai lab le a t www.sc iencedi rec t .com

journa l homepage : www.e lsev ie r . com/ loca te /wat res

Potential for activated persulfate degradation of BTEXcontamination

Chenju Liang*, Chiu-Fen Huang, Yan-Jyun Chen

Department of Environmental Engineering, National Chung Hsing University, 250 Kuo-kuang Road, Taichung 402, Taiwan

a r t i c l e i n f o

Article history:

Received 21 March 2008

Received in revised form

17 June 2008

Accepted 19 June 2008

Available online 29 June 2008

Keywords:

In situ chemical oxidation

Sulfate radical

Gasoline hydrocarbons

Citric acid

Activation

* Corresponding author. Tel./fax: þ886 4 228E-mail address: cliang@dragon.nchu.edu

0043-1354/$ – see front matter ª 2008 Elsevidoi:10.1016/j.watres.2008.06.022

a b s t r a c t

The present study focused on evaluation of activated persulfate (PS) anion (S2O82�) oxidative

degradation of benzene, toluene, ethylbenzene, and xylene (constituents of gasoline and

known collectively as BTEX) contamination. The results indicated that BTEX were effec-

tively oxidized by PS in aqueous and soil slurry systems at 20 �C. PS can be activated ther-

mally, or chemically activated with Fe2þ to form the sulfate radical (SO4��) with a redox

potential of 2.4 V. The degradation rate constants of BTEX were found to increase with

increased persulfate concentrations. For two PS/BTEX molar ratios of 20/1 and 100/1 exper-

iments, the observed aqueous phase BTEX degradation half-lives ranged from 3.0 to 23.1

days and 1.5 to 20.3 days in aqueous and soil slurry systems, respectively. In the interest

of accelerating contaminant degradation, Fe2þ and chelated Fe2þ activated persulfate

oxidations were investigated. For all iron activation experiments, BTEX and persulfate

degradations appear to occur almost instantaneously and result in partial BTEX removals.

It is speculated that the incomplete degradation reaction may be due to the cannibalization

of SO4�� in the presence of excess Fe2þ. Furthermore, the effects of various chelating agents

including, hydroxylpropyl-b-cyclodextrin (HPCD), ethylenediaminetetraacetic acid (EDTA),

and citric acid (CA) on maintaining available Fe2þ and activating PS for the degradation of

benzene were studied. The results indicated that HPCD and EDTA may be less susceptible

to chelated Fe2þ. In contrast, CA is a more suitable chelating agent in the iron activated per-

sulfate system and with a PS/CA/Fe2þ/B molar ratio of 20/5/5/1 benzene can be completely

degraded within a 70-min period.

ª 2008 Elsevier Ltd. All rights reserved.

1. Introduction fraction of gasoline (Pawlowski, 1998). The presence of BTEX

The United States Environmental Protection Agency (U.S. EPA)

estimates that 35% of the U.S.’s gasoline and diesel fuel under-

ground storage tanks (USTs) are leaking. Approximately 40%

of these leaking USTs likely have resulted in soil and ground-

water contaminations. Note that this information was

reported in a compilation by Pawlowski (1998). Gasoline

hydrocarbons including benzene, toluene, ethylbenzene, and

xylenes (known collectively as BTEX) are volatile and water-

soluble constituents that comprise 50% of the water-soluble

56610..tw (C. Liang).er Ltd. All rights reserved

in groundwater can create a hazard to public health and the

environment. The U.S. EPA National Primary Drinking Water

Regulations Maximum Contaminant Levels for BTEX in

drinking water are 0.005, 1.0, 0.7, and 10.0 mg l�1, respectively.

Limitations of conventional groundwater cleanup technolo-

gies (e.g., pump-and-treat) have made in situ chemical oxida-

tion (ISCO) an attractive remediation alternative. Among all

current commonly used ISCO oxidants, sodium persulfate

(Na2S2O8) is a relatively new oxidant which exhibits some

potential advantages for hazardous waste site remediation.

.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 0 9 1 – 4 1 0 04092

The persulfate (PS) anion (S2O82�) with a high redox potential

(E0¼ 2.01 V) (Eq. (1); Latimer, 1952) can be thermally or chem-

ically activated (see Eqs. (2) and (3), respectively) (House, 1962;

Kolthoff and Miller, 1951) to form the sulfate radical (SO4��),

a stronger oxidant (E0¼ 2.4 V) (Eq. (4); Huie et al., 1991).

S2O2�8 þ 2e�/2SO2�

4 E0 ¼ 2:01 V (1)

Thermal activation : S2O2�8 þ heat/2SO�4

� (2)

Chemical activation : S2O2�8 þ Fe2þ/SO�4

� þ Fe3þ þ SO2�4 (3)

SO�4� þ e�/SO2�

4 E0 ¼ 2:40 V (4)

Persulfate itself is a strong oxidant; however, it usually

requires higher reaction activation energy than other ISCO

oxidants such as permanganate. Liang et al. (2003) reported

that an activation energy of 98 kJ/mol is required for persul-

fate oxidation of trichloroethylene (TCE), as opposed to

35 kJ/mol required for the reaction between potassium

permanganate and TCE (Huang et al., 1999). It can be seen

that even though persulfate has a higher redox potential

(E0¼ 2.01 V) than permanganate (E0¼ 1.68 V), chemical

degradation of TCE with persulfate without supplemental

activation is relatively slow. Thus, persulfate may be thermo-

dynamically stable in the subsurface. Moreover, the solubility

of Na2S2O8 is high (730 g l�1) and persulfate is not significantly

involved in sorption reactions, suggesting that the natural soil

oxidant demand is low when persulfate is used for ISCO appli-

cation (Huling and Pivetz, 2006).

Although persulfate without additional activation will not

appreciably oxidize most organic contaminants including

BTEX, persulfate may oxidize BTEX over a longer period of

time at ambient temperatures. Because changing the magni-

tude of thermal activation temperature is one of the ways to

control the generation of sulfate radicals, a slow reaction

rate would be expected for the persulfate oxidation of BTEX

at ambient temperatures (e.g., 20 �C). Persulfate activation

with iron requires an activation energy of 12 kcal mol�1 (Ford-

ham and Williams, 1951), which is lower than the value of

33.5 kcal mol�1 required for thermal activation (Kolthoff and

Miller, 1951). The iron activated persulfate may be a more

rapid way of degrading contaminants. However, it has been

reported by Liang et al. (2004a) that even though ferrous ion

activated persulfate can quickly generate sulfate radicals for

the destruction of TCE, incomplete TCE destruction would

occur almost instantaneously and then the reaction stalled.

It was hypothesized that either destruction of sulfate radical

in the presence of excess Fe2þ or the rapid conversion of all

Fe2þ to Fe3þ limited the ultimate oxidizing capability (see Eq.

(5); Gupta and Gupta, 1981).

SO�4� þ Fe2þ/Fe3þ þ SO2�

4 k ¼ 4:6� 109 M�1 s�1 (5)

A means of maintaining a suitable amount of Fe2þ to reduce

the competition for sulfate radical between the target organic

contaminant and excess Fe2þ is the use of a complexing agent

for controlling the concentration of Fe2þ in solution (Liang

et al., 2004b; Crimi and Taylor, 2007). Liang et al. (2004b)

evaluated four types of complexing agents on maintaining

chelated ferrous ion for persulfate activation to degrade TCE

in an aqueous system and demonstrated that when citric

acid (CA) chelated ferrous ion was used as an activator,

complete TCE degradations were achieved after 24 h while

others resulted in only partial degradations. Furthermore,

a Fe2þ/CA molar ratio of 5/1 was suggested to be the lowest

acceptable ratio to maintain sufficient quantities of Fe2þ in

the persulfate activation system. Crimi and Taylor (2007)

applied two Fe2þ/CA molar ratios (i.e., 5/1 and 2/1) in degrad-

ing simulated contaminated groundwater with BTEX. They

reported that >99% of BTEX was destroyed within 3 weeks.

The results of these studies demonstrated potential persulfate

reactivities in the destruction of organic contaminants. There-

fore, in this study, we evaluated (1) the thermally activated

persulfate oxidation of each BTEX compound in both aqueous

and soil slurry systems at 20 �C and (2) iron and chelated iron

activated persulfate oxidation reactivities in an aqueous

system. In addition to citric acid chelate, alternative chelating

agents such as ethylenediaminetetraacetic acid (EDTA) and

hydroxypropyl-b-cyclodextrin (HPCD) were also evaluated.

2. Materials and methods

2.1. Chemicals

Water used for preparation of BTEX contaminated solutions

was purified by a Millipore reverse osmosis (RO) purification

system. The chemicals used were purchased from the

following sources: benzene (C6H6, min. 99.7%), carbon disul-

fide (CS2, 99.9%), and sodium thiosulfate (Na2S2O3$5H2O,

min. 99.5%) from Riedel-deHaen; toluene (C6H5CH3, ACS) and

potassium iodide (KI, min. 99.5%), from UNION Taiwan; acetic

acid (CH3CO2H, min. 99.8%), ethyl benzene (C6H5C2H5, min.

98%) and m-xylene (C6H4(CH3)2, min. 98%) from Fluca; sodium

persulfate (Na2S2O8, min. 99.0%) and citric acid (HOC(COOH)

(CH2COOH)2, >99.5%) from Merck; ferrous sulfate (FeS-

O4$7H2O, 100.8%) from J. T. Baker; HP-b-CD (C51H88O38, 97%)

from Acros Organics; EDTA ((HO2CCH2)2NCH2CH2N(CH2-

CO2H)2, w99%) from Sigma–Aldrich.

Soil materials used in soil slurry experiments were

obtained from farm land located in southern Taiwan and

collected from a layer located approximately within 30–

100 cm below the ground surface. The grain size distribution

of the soil was determined in accordance with ASTM (1998)

Standard D442(63) and revealed that the soil consisted of

29.9% sand and 70.1% of silt/clay. The results of a partial char-

acterization of the soil are reported in Table 1. The silt soil was

air-dried and sieved (i.e., passed sieve #10 and retained on

sieve #200) prior to use.

2.2. Experimental setup

The experimental procedure was in accordance with the

method described by Liang et al. (2003). BTEX degradation

experiments were conducted individually for B, T, E, and X

on both aqueous solutions and soil slurries.

Aqueous test solutions were prepared at an initial concen-

tration of 1 mM (B: 78; T: 92; E: 106; X: 106 mg l�1) in a 2 l

Table 1 – Properties of soil

Parameters Value References

pH 7.1 NIEA S410.61C

Total chromate

oxidizable matter (%)

1.64 Page (1982)

Total carbon (%) 1.67

Total organic carbon (%) 1.51

Total inorganic carbon (%) 0.06

Cation exchange

capacity (meq per 100 g)

18.8 NIEA S202.60A

Iron (Fe) (mg kg�1) 31100 NIEA S321.63B

Manganese (Mn) (mg kg�1) 242 NIEA S321.63B

Copper (Cu) (mg kg�1) 27 NIEA S321.63B

Zinc (Zn) (mg kg�1) 147 NIEA S321.63B

Note: National Institute of Environmental Analysis (NIEA) methods

established by the government of Taiwan.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 0 9 1 – 4 1 0 0 4093

borosilicate reservoir (Schott Puran) equipped with a Teflon

stopper and valved bottom outlet by adding pure BTEX directly

into RO water. The contaminant solution was mixed overnight

in a constant temperature chamber at 20 �C. For the thermally

activated persulfate experiments in the aqueous system at

20 �C, a predetermined amount of sodium persulfate was added

to each contaminant reservoir and mixed for approximately

1 min before filling a series of 40 ml EPA amber glass vials.

Soil slurry test mixtures were prepared by adding 10 g of

soil into a 40 ml bottle prior to filling with 38 ml contaminant

solution. The resulting soil slurries were shaken on a recip-

rocal shaker (IKA HS 260 control) at 160 rpm with the chamber

at 20 �C for a period of 48 h to ensure the equilibrium of

contaminants. Thereafter, shaking was stopped and the vials

were allowed to equilibrate for an additional 24 h. Before the

start of tests, 3 ml of supernatant were removed from each

vial and discarded and then 3 ml of concentrated persulfate

stock solution were added into each bottle to reach the desired

persulfate concentrations (i.e., 20 and 100 mM).

The iron activated persulfate experimental procedure was

similar to that reported by Liang et al. (2007a). Contaminant

solution was prepared in a 1.3 l heavy-wall plain pressure

reaction flask (ACE Glass) that was placed in the tempera-

ture-controlled chamber at 20 �C and the top of the flask was

covered with a flat Teflon reaction head that was sealed

with a stainless steel clamp. Ferrous ion or chelated iron

was added to the flask during the preparation of the BTEX

contaminant solutions. At each sampling time, a 5 ml aliquot

of solution was withdrawn using a gas-tight syringe through

a sampling port for analysis of BTEX and persulfate. Control

tests in the absence of persulfate were carried out in parallel.

2.3. Analysis

The aqueous phase BTEX concentrations from both the

aqueous and soil slurry systems were analyzed by high pres-

sure liquid chromatography (Agilent 1100 HPLC) with a UV

detector. A reversed-phase ZORBAX Eclipse XDB-C18

(4.6� 150 mm� 5 mm) column was used. The mobile phase

was acetonitrile–water (70:30, v/v) with a flow rate at

1.00 mL/min and the effluent was monitored at 254 nm.

The soil slurry samples were separated into aqueous and

soil phases by centrifuging for 20 min at 1500 rpm. The

aqueous phase solution was decanted and an aliquot was

analyzed by HPLC, as described above. For measuring soil

sorbed BTEX concentrations, 1 g of soil was transferred to

a 5 ml amber glass bottle for extraction using 2 ml of carbon

disulfide. The extraction procedure included mixing for

5 min on a vortex shaker (Thermolyne Type 65800), and then

equilibration for 5 min. The carbon disulfide extractant was

analyzed by an Agilent 6890 N gas chromatograph equipped

with flame ionization detector and an Agilent 7683B autosam-

pler using an Agilent HP-5 fused silica capillary column

(30 m� 0.32 mm� 0.25 mm). Nitrogen was used as the carrier

gas at a flow rate of 15 ml min�1. The temperatures for

column, injector, and detector were 45 (isothermal), 220, and

250 �C, respectively.

Persulfate anion was determined by iodometric titration

with sodium thiosulfate (Kolthoff and Stenger, 1947). The pH

was measured by SUNTEX TS-100 pH meter. Total organic

carbon (TOC) analysis for soil samples was performed using

a TOC analyzer (O.I. Analytical SOLID M.). Metals were

measured using a Perkin Elmer Analyst 100 flame atomic

absorption spectrophotometer. Selected experiments were

duplicated or in triplicate and the data obtained were aver-

aged. The error bars in all figures represent �1 standard devi-

ation from the mean of triplicate data.

3. Results and discussion

3.1. Persulfate oxidation of BTEX in the aqueous system

As shown in Fig. 1, the overall kinetic data indicate that the

degradation rate of BTEX can be characterized by a pseudo-

first-order reaction kinetic model, as evidenced by a high

correlation coefficient (R2> 0.95) (data presented in Table 2).

The degradation rate constants (kobs) of BTEX were found to

increase with increased persulfate concentrations. For two

PS/BTEX molar ratios of 20/1 and 100/1 experiments, half-lives

(t1/2) of BTEX degradations were in the range of 3.0–23.1 days.

Huang et al. (2005) reported that t1/2 values for BTEX degrada-

tions with persulfate ranged from 3.3 to 5.2 days in an exper-

imental system where 1 g l�1 Na2S2O8 was employed to

degrade 59 volatile organic compounds (VOCs), with initial

BTEX concentrations in the range of 77–156 mg l�1 at 20 �C.

Moreover, Crimi and Taylor (2007) evaluated persulfate oxida-

tion for simulated groundwater contaminated with BTEX and

reported BTEX reduction ranging from 54% to 92% during

a reaction period of 3 weeks in an experimental system,

with a PS/BTEX molar ratio of 20/1 and pH¼ 11 at ambient

temperature. These results reveal the capability of persulfate

oxidation of BTEX even at ambient temperatures (e.g., 20 �C).

Among four BTEX compounds, benzene was most resistant

to persulfate oxidation at 20 �C. For example, in the PS/BTEX

molar ratio of 100/1 experiment, a kobs value of 0.095 day�1

for benzene is approximately 2–3 fold less than other

compounds (e.g., kobs¼ 0.23 day�1 for toluene). These observa-

tions are in agreement with conclusions made by Huang et al.

(2005) who demonstrated that VOCs with higher degradation

rates in the presence of persulfate generally contain

0.1

0.2

0.4

0.6

0.81.0

0 10 15 20 25 30 350.1

0.2

0.4

0.6

0.81.0

PS/BTEX M.R. = 0/1 PS/BTEX M.R. = 20/1 PS/BTEX M.R. = 100/1

C/C

OC

/CO

Reaction time (days) Reaction time (days)

0.1

0.2

0.4

0.6

0.81.0

PS/BTEX M.R. = 20/1 PS/BTEX M.R. = 100/1P

S C

/CO

Reaction time (days)

Benzene Toluene

Ethylbenzene Xylene

Persulfate

Initial pH (BTEX solution) ~ 5.8 Final pH for PS/BTEX M.R.= 0/1, 20/1, 100/1 ~ 5.9, 2.9, and 2.2, respectively

5

0 10 15 20 25 30 355

0 10 15 20 25 30 355

a

c

e

d

b

Fig. 1 – Aqueous BTEX and persulfate degradations as a function of PS/contaminant molar ratios at 20 8C in the aqueous

system. [B], [T], [E], [X] [ 1 mM, [S2O82L] [ 20, 100 mM. Persulfate data are the average of each BTEX degradation data.

Table 2 – Kinetic data of the 20 8C thermal persulfate oxidation of BTEX in aqueous and soil slurry systems

PS/BTEX M.R. Pseudo-first-order BTEX degradation rate constant (kobs)� 102 day�1 (t1/2, days) (R2)

Benzene Toluene Ethylbenzene Xylene PSa

(1) Aqueous system

20/1 3.0� 0.1, (23.1) (0.99) 4.4� 0.1, (15.8) (0.99) 3.8� 0.1, (18.2) (0.99) 4.7� 0.1, (14.7) (0.99) 0.9� 0.1, (77.0) (>0.95)

100/1 9.5� 0.2, (7.3) (0.99) 23.2� 0.9, (3.0) (0.99) 14.5� 1.3, (4.8) (0.95) 21.9� 0.9, (3.2) (0.99) 0.6� 0.0, (115.5) (>0.97)

(2) Soil slurry system – aqueous phase

20/1 3.4� 0.4, (20.3) (0.91) 3.9� 0.4, (17.8) (0.93) 5.9� 0.5, (11.7) (0.96) 6.1� 0.6, (11.3) (0.93) 3.9� 0.3, (17.8) (>0.93)

100/1 46.1� 7.9, (1.5) (0.90) 33.5� 4.0, (2.1) (0.96) 27.6� 2.1, (2.5) (0.98) 38.1� 4.1, (1.8) (0.96) 1.9� 0.2, (36.5) (>0.87)

(3) Soil slurry system – soil sorbed phase

20/1 6.6� 1.5, (10.5) (0.76) 12.7� 1.1, (5.5) (0.97) 12.5� 1.4, (5.5) (0.94) 7.1� 1.1, (9.7) (0.89) –

100/1 36.4� 5.7, (1.9) (0.91) 38.0� 4.0, (1.8) (0.94) 31.6� 4.6, (2.2) (0.92) 41.7� 7.5, (1.7) (0.89) –

M.R.¼molar ratio.

a The reported kobs for persulfate is the average of four sets of data obtained from each BTEX degradation experiment.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 0 9 1 – 4 1 0 04094

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 0 9 1 – 4 1 0 0 4095

carbon–carbon double bonds or are aromatic ring compounds

with substituted functional groups. On the other hand, persul-

fate anion appears to be very stable at 20 �C with half-lives of

several months in aqueous phase. Changing the magnitude of

thermal activation is one of the ways to control the generation

of sulfate radicals for destroying organic contaminants (Liang

et al., 2007b). Therefore, as demonstrated in Fig. 1, 20 �C may

serve as an effective thermal activation temperature in the

persulfate oxidation system for BTEX destruction over rela-

tively longer period of times and higher persulfate concentra-

tions could increase the BTEX degradation rates.

3.2. Persulfate oxidation of BTEX in the soil slurrysystem

As shown in Fig. 2, in the soil slurry system BTEX can be

degraded in both the aqueous phase and the soil sorbed

phase. All of the BTEX kinetic plots indicate that the rates of

BTEX degradation are faster during the early stage of oxida-

tion reaction followed by a much slower degradation rate.

Moreover, the observed degradation rate constants in the

0.1

1

10

100

0 10 15 20 25 30 350.1

1

10

100

PS/BTEX M.R. = 0/1PS/BTEX M.R. = 20/1PS/BTEX M.R. = 100/1

mg/

Lm

g/L

Reaction time (days)

0 10 150.1

0.2

0.4

0.6

0.81.0

PS/BTEX M.R. = 2PS/BTEX M.R. = 1P

S C

/CO

Reaction tim

Persulf

Benzene

Ethylbenzene

5

5

a

c

e

Fig. 2 – Aqueous BTEX and persulfate degradations as a function

[B], [T], [E], [X] [ 1 mM, [S2O82L] [ 20, 100 mM. Persulfate data ar

soil slurry system are mostly higher than those observed

previously in the aqueous system (see Table 2). One possible

cause of the increase in reaction rate constant can be attrib-

uted to native soil minerals such as iron that might activate

persulfate to generate more SO4�� for the initial quick destruc-

tion of contaminants. Persulfate anions are degraded at rela-

tively fast rates in the presence of soil, as shown in the

comparison of Figs. 1(e) and 2(e), and also in Table 2. These

observations are supported by explanation presented by Liang

et al. (2003) who postulated that soil organics bound to Fe2þ

generates persulfate activation. Furthermore, it is possible

that contaminant degradation intermediates and soil organics

compete for SO4�� and some reduction in reaction rates may

reflect these competition reactions.

The removal of soil sorbed BTEX in the soil slurry system

was studied by analyzing the residual contaminant concen-

trations on the soil phase. BTEX were concurrently removed

in both aqueous and sorbed phases, as shown in Figs. 2 and

3, respectively. Analogous to the results of the aqueous

system, sorbed BTEX removal rates generally increase with

higher concentrations of persulfate (see Table 2). Additionally,

0 10 15 20 25 30 35Reaction time (days)

20 25 30 35

0/100/1

e (days)

Initial pH (BTEX solution) ~ 6.9 Final pH for PS/BTEX M.R.= 0/1, 20/1, 100/1 ~ 6.8, 4.0, and 2.7, respectively

ate

Toluene

Xylene

5

d

b

of PS/BTEX molar ratios at 20 8C in the soil slurry system.

e the average of each BTEX degradation data.

0.1

1

10

100

0 5 10 15 20 25 30 350.1

1

10

100

0 5 10 15 20 25 30 35

mg/

kgPS/BTEX M.R. = 0/1PS/BTEX M.R. = 20/1PS/BTEX M.R. = 100/1

mg/

kg

Reaction time (days) Reaction time (days)

Benzene Toluene

Ethylbenzene Xylene

a b

c d

Fig. 3 – Sorbed BTEX degradations as a function of PS/BTEX molar ratios at 20 8C in the soil slurry system. [B], [T], [E],

[X] [ 1 mM, [S2O82L] [ 20, 100 mM.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 0 9 1 – 4 1 0 04096

more complete removals of sorbed BTEX were achieved, as

compared to the aqueous phases (see Fig. 2). Considering

a soil slurry system with hydrophobic organic contaminants,

the contaminants preferentially adsorb to soil organic matter,

and in addition are not significantly adsorbed to minerals

(Chiou and Kile, 2000). Therefore, we speculate that the

mass removal of BTEX from soils could be possibly caused

by two mechanisms including degradation and desorption.

First, persulfate or sulfate radicals may directly oxidize sorbed

contaminants. Secondly, because the contaminant sorption

occurs predominantly by partition into the soil organic

matter, once soil organics are destroyed by persulfate, it

may result in desorption of sorbed BTEX and subsequently

oxidation in the aqueous phase. As a result, the data shown

in Fig. 3 present the net results of these two BTEX removal

mechanisms. It should be noted that the average distribution

percentages of BTEX to soil vs. aqueous solution are 9.7% vs.

75.4%, 15.1% vs. 65.5%, 30.7% vs. 51.2, and 34.6 vs. 47.3%,

respectively, in control experiments without the presence of

persulfate. It can be seen that overall extraction recoveries

ranged from 81% to 85% and some losses of contaminant

mass may have occurred due to possible biodegradation

during a 72-h soil slurry preparation period. Further investiga-

tion of soil biological activity would be required to clarify this

issue.

3.3. Iron activated persulfate oxidation of BTEX in theaqueous system

As demonstrated in previous sections, a lower rate of sulfate

radical production, as occurs during thermal persulfate acti-

vation at 20 �C, could still result in BTEX destructions with

half-lives of days to weeks depending on the strength of per-

sulfate used. Furthermore, in the interest of accelerating

contaminant degradation, generation of sulfate radicals by

Fe2þ and chelated Fe2þ activated persulfate, as described in

Eq. (3), was investigated. Fig. 4 shows the influence of persul-

fate concentrations on Fe2þ activated persulfate oxidation of

BTEX. The influence of Fe2þ concentrations is shown in

Fig. 5. Increases of persulfate concentrations at a fixed level

of Fe2þ/BTEX molar ratio (i.e., 5/1) or increases of Fe2þ concen-

trations at a fixed level of PS/BTEX molar ratio (i.e., 20/1)

resulted in increases in both BTEX and persulfate degrada-

tions. In all Fe2þ activation experiments, BTEX and persulfate

were quickly degraded and after the second sampling period

(i.e., 7 min for BTEX experiments) little further degradation

was observed. These results are similar to those previously

reported by Liang et al. (2004a), where Fe2þ activated persul-

fate oxidation of TCE appears to occur almost instantaneously

and then stall. It is speculated that the halt of reaction may be

due to the cannibalization of sulfate radicals in the presence

of excess Fe2þ in accordance with Eq. (5).

The molar ratios of DFe2þ/DPS for all BTEX degradations at

each fixed PS/Fe2þ/BTEX molar ratio are generally smaller

than the stoichiometric ratio of 2, in accordance with the

following equation by combining the reactions of Eqs. (3)

and (5).

S2O2�8 þ 2Fe2þ/2Fe3þ þ 2SO2�

4 (6)

The higher initial Fe2þ concentrations resulted in increased

ratios of DFe2þ/DPS that are closer to 2. Smaller ratios suggest

less consumptive/scavenging of Fe2þ by sulfate radicals (see

Eq. (5)); in other words, sulfate radicals generated are mostly

used for contaminant destruction, and may result in higher

contaminant degradation efficiency (e.g., lower persulfate

0.2

0.4

0.6

0.81.0

0.2

0.4

0.6

0.81.0

0 20 40 60 80 0 20 40 60 80

C/C

oPS/Fe2+/BTEX = 0/0/1PS/Fe2+/BTEX = 0/5/1PS/Fe2+/BTEX = 5/5/1PS/Fe2+/BTEX = 20/5/1PS/Fe2+/BTEX = 100/5/1

C/C

O

Reaction time (mins) Reaction time (mins)

0 20 40 60 80

0.2

0.4

0.6

0.81.0

PS/Fe2+/BTEX M.R. = 5/5/1PS/Fe2+/BTEX M.R. = 20/5/1PS/Fe2+/BTEX M.R. = 100/5/1

PS

C/C

O

Reaction time (mins)

Initial pH (BTEX solution) ~ 5.6 Final pH for 3 S2O8

2- conc. tested ~ 2.4

Benzene Toluene

Ethylbenzene Xylene

Persulfate

ba

c d

e

Fig. 4 – Influence of persulfate concentrations on ferrous ion activated persulfate oxidation of BTEX at 20 8C in the aqueous

system. [B], [T], [E], [X] [ 1 mM, [Fe2D] [ 5 mM, [S2O82L] [ 5, 20, 100 mM. Persulfate data are the average of each BTEX

degradation data.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 0 9 1 – 4 1 0 0 4097

consumption per unit mass of contaminant degraded). As can

be seen from the DPS/DBTEX data, the higher DFe2þ/DPS ratio

resulted in the higher DPS/DBTEX ratio (i.e., more persulfate

consumption). As presented in Table 3, for example: by

comparing the PS/Fe2þ/BTEX M.R.¼ 20/1/1 vs. 20/20/1 with

DFe2þ/DPS ratios of 1.46 vs. 1.86 correspond to DPS/DB ratios

of 2.78 vs. 14.99. However, the influence of persulfate on the

DPS/DBTEX ratio under a fixed Fe2þ concentration is less

significant.

The Fe2þ activated persulfate oxidation is significantly

affected by the cannibalization of sulfate radical in the pres-

ence of free Fe2þ and results in partial BTEX degradations. It

has been demonstrated that chelating agents can regulate

Fe2þ in solution and be available to activate the persulfate

reaction (Liang et al., 2004b, 2007a; Crimi and Taylor, 2007).

Therefore, the effects of various chelating agents (i.e., HPCD,

EDTA, and CA) and chelate/Fe2þ molar ratios on maintaining

available Fe2þ for persulfate activation were investigated to

evaluate the degradation of benzene, which is the BTEX

compound most resistant to persulfate oxidation. As shown

in Fig. 6, it can be seen that benzene was only completely

degraded (i.e., near 100% degradation) in the case of the citric

acid chelated iron experiment with a PS/CA/Fe2þ/B molar ratio

of 20/5/5/1 with a 70-min period. HPCD and EDTA seem less

able to maintain available Fe2þ in the persulfate system.

However, the fate and long-term effect of chelating agents

were not studied and their presence within soil media may

behave differently. Liang et al. (2003) has demonstrated that

citric acid is a more suitable chelating agent in the iron

activated persulfate system when used to treat TCE contami-

nation. In contrast to EDTA, the rates of formation of iron–

citrate complexes are not as sensitive to the concentration

of competing metals (Fujii et al., 2008) and hence may be

more stable. HPCD complexed iron has been applied in acti-

vating peroxide (i.e., modified Fenton reaction) in treating

contaminants such as PAHs (Lindsey et al., 2003) and to acti-

vate persulfate in treating TCE and tetrachloroethylene (Liang

et al., 2007a). There are two major roles of HPCD: enhancing

organics solubilities and simultaneously complexed iron acti-

vator (Rossi and Rossi, 2004; Liang et al., 2007a). However, the

0.2

0.4

0.6

0.81.0

0.4

0.6

0.81.0

C/C

o

PS/Fe2+/BTEX = 0/0/1PS/Fe2+/BTEX = 0/5/1PS/Fe2+/BTEX = 20/1/1PS/Fe2+/BTEX = 20/5/1PS/Fe2+/BTEX = 20/20/1

C/C

O

0 20 40 60 80

0.2

0 20 40 60 80

Reaction time (mins) Reaction time (mins)

Benzene Toluene

Ethylbenzene Xylene

a b

c d

0 20 40 60 80

0.2

0.4

0.6

0.81.0

PS

C/C

O

Reaction time (mins)

PS/Fe2+/BTEX M.R. = 20/1/1 PS/Fe2+/BTEX M.R. = 20/5/1 PS/Fe2+/BTEX M.R. = 20/20/1

Persulfate

Initial pH (BTEX solution) ~ 5.6 Final pH for 3 Fe2+ conc. tested ~ 2.1 to 2.6

(high to low Fe2+)

e

Fig. 5 – Influence of ferrous ion concentrations on ferrous ion activated persulfate oxidation of BTEX at 20 8C in the aqueous

system. [B], [T], [E], [X] [ 1 mM, [Fe2D] [ 1, 5, 20 mM, [S2O82L] [ 20 mM. Persulfate data are the average of each BTEX

degradation data.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 0 9 1 – 4 1 0 04098

effectiveness of iron complexation by HPCD in treating

benzene appears low.

In general, the chelated iron activated persulfate resulted

in DFe2þ/DPS molar ratios less than those obtained in the

unchelated iron activated persulfate experiments. The lower

ratios indicate that scavenging of sulfate radical by Fe2þmight

be well controlled. In the HPCD experiment, a lower DFe2þ/DPS

due to increases of HPCD concentrations resulted in the

increase of DPS/DB values (i.e., 9.59–28.31, see Table 3), which

reveal significant persulfate consumption. An explanation

may be that HPCD, a solubility enhancing chemical, can

contain contaminant inside the cavity structure of HPCD,

thereby preventing benzene from oxidation and the elevated

persulfate consumption may be mainly caused by oxidation

of HPCD, instead of benzene, by persulfate or sulfate radicals.

As for the other two cases, the DPS/DB values are either

decreased or remained nearly the same when the concentra-

tion of chelating agents increased and may indicate that the

chelating capability of EDTA and CA are more resistant to per-

sulfate oxidation.

4. Conclusions

In this study, thermally activated persulfate oxidation of BTEX

at 20 �C was investigated in both aqueous and soil slurry

systems. The results show that persulfate oxidation of BTEX

at ambient temperatures (e.g., 20 �C) may serve as an effective

method in treating BTEX contamination. The observed kobs

values for the PS/BTEX molar ratios of 20/1 and 100/1 ranged

from 3� 10�2 to 23.2� 10�2 day�1 in the aqueous system and

from 3.4� 10�2 to 46.1� 10�2 day�1 in the soil slurry system.

The kobs was found to increase with the increased PS concen-

tration. Also, soil sorbed BTEX degradation exhibited kobs

values ranging from 6.6� 10�2 to 41.7� 10�2 day�1 in the soil

slurry system. Among BTEX compounds, benzene was most

resistant to persulfate oxidation. Fe2þ activated persulfate

oxidation is significantly affected by the cannibalization of

SO4�� in the presence of free Fe2þ (unchelated) and results in

partial BTEX degradations. The citric acid chelated Fe2þ was

demonstrated to be a suitable chelating agent in regulating

Table 3 – Kinetic data of the iron activated persulfate oxidation of BTEX in the aqueous system

Objectives PS/Fe2þ/BTEX or PS/chelate/Fe2þ/B M.R. DFe2þ/DPS for BTEXa or for Bb DPS/DBTEXa or DPS/DBb

Influence of PS

(Fe2þ ¼ 5 mM)

5/5/1 1.86, 1.85, 1.88, 1.89 7.49, 11.23, 7.02, 7.03

20/5/1 1.73, 1.75, 1.77, 1.76 7.22, 9.08, 5.81, 6.52

100/5/1 1.70, 1.76, 1.73, 1.75 7.85, 5.76, 4.21, 4.51

Influence of Fe2þ

(PS¼ 20 mM)

20/1/1 1.46, 1.45, 1.57, 1.62 2.78, 3.28, 2.41, 2.64

20/5/1 1.73, 1.75, 1.77, 1.76 7.22, 9.08, 5.81, 6.52

20/20/1 1.86, 1.84, 1.89, 1.90 14.99, 22.32, 12.89, 14.49

Influence of HPCD

(PS¼ 20 mM, Fe2þ ¼ 5 mM)

20/1/5/1 1.14 9.59

20/5/5/1 1.04 12.91

20/25/5/1 0.60 28.31

Influence of EDTA

(PS¼ 20 mM, Fe2þ ¼ 5 mM)

20/1/5/1 0.69 12.19

20/5/5/1 1.10 8.23

20/25/5/1 1.09 9.23

Influence of CA

(PS¼ 20 mM, Fe2þ ¼ 5 mM)

20/0.25/5/1 1.27 6.49

20/1/5/1 0.96 6.79

20/5/5/1 0.67 8.04

M.R.¼molar ratio.

a DFe2þ/DPS for BTEX or DPS/DBTEX represents the molar ratio of the Fe2þ, PS, BTEX consumed or degraded which are the average of data

obtained between 20 and 70 min.

b DFe2þ/DPS for B or DPS/DB represents the molar ratio of the Fe2þ, PS, B consumed or degraded at 70 min.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 0 9 1 – 4 1 0 0 4099

available Fe2þ for persulfate activation. In the case of PS/CA/

Fe2þ/B molar ratio of 20/5/5/1, the results show complete

benzene removal, while the case of PS/Fe2þ/B molar ratio of

20/5/1 shows nearly 40% removal. The results discussed here

demonstrated the BTEX degradation capability by 20 �C

thermal persulfate activation in both aqueous and soil slurry

systems. In the interest of accelerating degradation rate, citric

acid chelated Fe2þ activated persulfate can be a good alterna-

tive in the treatment of BTEX contamination.

0 200 20 40 60 80

0.2

0.4

0.6

0.81.0

0.2

0.4

0.6

0.81.0

PS/EDTA/Fe2+

PS/EDTA/Fe2+

PS/EDTA/Fe2+

PS/EDTA/Fe2+

Reaction

PS/HPCD/Fe2+/B = 20/0/5/1

PS/HPCD/Fe2+/B = 20/1/5/1

PS/HPCD/Fe2+/B = 20/5/5/1

PS/HPCD/Fe2+/B = 20/25/5/1

PS

C/C

o

Reaction time (min)

Ben

zene

C/C

o

Initial pH for B solution ~ 5.6

Final pH ~ 2.1 to 2.4

(high to low HPCD)

Final pH ~ 1.7 to 2.(high to low ED

HPCD EDa b

Fig. 6 – Influence of chelate concentrations on ferrous ion activa

system. [B] [ 1 mM, [Fe2D] [ 1, 5, 20 mM, [S2O82L] [ 20 mM.

Acknowledgement

This study was partially funded by the National Science

Council (NSC) of Taiwan under project number of 96-2622-E-

005-009-CC3. The authors acknowledge Jack Miano and Nihar

Mohanty, Environmental Engineers of Massachusetts Depart-

ment of Environmental Protection for much appreciated

discussions.

40 60 80 0 20 40 60 80

/B = 20/0/5/1

/B = 20/1/5/1/B = 20/5/5/1/B = 20/25/5/1

time (min)

PS/CA/Fe2+/B = 20/0/5/1

PS/CA/Fe2+/B = 20/0.2/5/1

PS/CA/Fe2+/B = 20/1/5/1

PS/CA/Fe2+/B = 20/5/5/1

Reaction time (mins)

3 TA)

Final pH ~ 1.8 to 2.4 (high to low CA)

TA CAc

ted persulfate oxidation of benzene at 20 8C in the aqueous

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 0 9 1 – 4 1 0 04100

r e f e r e n c e s

American Society for Testing and Materials (ASTM), 1998.Standard Test Method for Particle-size Analysis of Soils.ASTM. ASTM-International-D422–63.

Chiou, C.T., Kile, D.E., 2000. Contaminant Sorption by Soil and BedSediment. Fact Sheet-087-00. U.S. Department of the Interior,U.S. Geological Survey.

Crimi, M.L., Taylor, J., 2007. Experimental evaluation of catalyzedhydrogen peroxide and sodium persulfate for destruction ofBTEX contaminants. Soil Sediment Contam. 16, 29–45.

Fordham, J.W.L., Williams, H.L., 1951. The persulfate-iron (II)initiator system for free radical polymerizations. J. Am. Chem.Soc. 73, 4855–4859.

Fujii, M., Rose, A.L., Waite, T.D., Omura, T., 2008. Effect of divalentcationson the kinetics of Fe(III) complelxationby organic ligandsin natural waters. Geochim. Cosmochim. Acta 72, 1335–1349.

Gupta, S.S., Gupta, Y.K., 1981. Hydrogen ion dependence of theoxidation of iron (II) with peroxydisulfate in acid perchloratesolutions. Inorg. Chem. 20, 454–457.

House, D.A., 1962. Kinetics and mechanism of oxidations byperoxydisulfate. Chem. Rev. 62, 185–203.

Huang, K.-C., Hoag, G.E., Chheda, P., Woody, B.A., Dobbs, G.M.,1999. Kinetic study of oxidation of trichloroethylene bypotassium permanganate. Environ. Eng. Sci. 16, 265–274.

Huang, K.-C., Zhao, Z., Hoag, G.E., Dahmani, A., Block, P.A., 2005.Degradation of volatile organic compounds with thermallyactivated persulfate oxidation. Chemosphere 61, 551–560.

Huie, R.E., Clifton, C.L., Neta, P., 1991. Electron transfer reactionrates and equilibria of the carbonate and sulfate radical anions.Int. J. Radiat. Appl. Instrum. C. Radiat. Phys. Chem. 38, 477–481.

Huling, S.G., Pivetz, B.E., 28 July 2006. In-situ Chemical Oxidation.USEPA/600/R-06/072. USEPA.

Kolthoff, I.M., Stenger, V.A., 1947. Volumetric analysis. In:Titration Methods: Acid–base, Precipitation and ComplexReactions, second revised ed., vol. II. Interscience Publishers,Inc., New York.

Kolthoff, I.M., Miller, I.K., 1951. The chemistry of persulfate. I. Thekinetics and mechanism of the decomposition of the

persulfate ion in aqueous medium. J. Am. Chem. Soc. 73, 3055–3059.

Latimer, W.M., 1952. Oxidation Potentials. Prentice-Hall, Inc.,Englewood Cliffs, NJ.

Liang, C.J., Bruell, C.J., Marley, M.C., Sperry, K.L., 2003. Thermallyactivated persulfate oxidation of trichloroethylene (TCE) and1,1,1-trichloroethane (TCA) in aqueous systems and soilslurries. Soil Sediment Contam. 12, 207–228.

Liang, C., Bruell, C.J., Marley, M.C., Sperry, K.L., 2004a. Persulfateoxidation for in situ remediation of TCE. I. Activated by ferrousion with and without a persulfate–thiosulfate redox couple.Chemosphere 55, 1213–1223.

Liang, C., Bruell, C.J., Marley, M.C., Sperry, K.L., 2004b. Persulfateoxidation for in situ remediation of TCE. II. Activated bychelated ferrous ion. Chemosphere 55, 1225–1233.

Liang, C., Huang, C.-F., Mohanty, N., Lu, C.-J., Kurakalva, R.M.,2007a. Hydroxypropyl-beta-cyclodextrin-mediated iron-activated persulfate oxidation of trichloroethylene andtetrachloroethylene. Ind. Eng. Chem. Res. 46, 6466–6479.

Liang, C., Wang, Z.-S., Bruell, C.J., 2007b. Influence of pH onpersulfate oxidation of TCE at ambient temperatures.Chemosphere 66, 106–113.

Lindsey, M.E., Xu, G., Lu, J., Tarr, M.A., 2003. Enhanced Fentondegradation of hydrophobic organics by simultaneous ironand pollutant complexation with cyclodextrins. Sci. Tot.Environ. 307, 215–229.

NIEA, National Institute of Environmental Analysis,Environmental Analysis Laboratory. http://www.niea.gov.tw/,Executive Yuan, Taiwan, ROC.

Page, A.L., 1982. Methods of Soil Analysis – Part 2. Chemical andMicrobiological Properties, second ed. American Society ofAgronomy, Inc., Soil Science Society of America, Inc.,Madison, WI.

Pawlowski, M.H., 1998. Analytical and Field Test Methods forMeasuring BTEX Metabolite Occurrence and Transport inGroundwater. Oregon State University Corvallis Departmentof Chemistry, Defense Technical Information Center.ADA349678.

Rossi, L.I., Rossi, R.H.d., 2004. FeBr3–cyclodextrin complexes asefficient and chemoselective catalysts for sulfoxidationreactions. Appl. Catal. Gen. 267, 267–272.