Electrochimica Acta 92 (2013) 20–30

Contents lists available at SciVerse ScienceDirect

Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta

Oxidation of phenol and the adsorption of breakdown products using a graphite

adsorbent with electrochemical regeneration
S.N. Hussain a,∗ , E.P.L. Roberts a,1 , H.M.A. Asghar a , A.K. Campen b , N.W. Brown b
a
School of Chemical Engineering and Analytical Science, The University of Manchester, Manchester M13 9PL, UK
b
Arvia Technology Limited, Daresbury Innovation Centre, Keckwick Lane, Daresbury, Cheshire WA4 4FS, UK

a r t i c l e i n f o a b s t r a c t

Article history: The process of adsorption using an unexpanded graphite intercalation compound (GIC) adsorbent with
Received 16 August 2012 electrochemical regeneration has been investigated for the removal and oxidation of phenol in waste-
Received in revised form 3 January 2013 water. During electrochemical regeneration of the adsorbent, the formation of intermediate oxidation
Accepted 3 January 2013
products is critical because these could affect the toxicity of the treated effluent. The present study is
Available online 11 January 2013
concerned with the formation of breakdown products released in the treated water during electrochem-
ical regeneration of GIC adsorbents, using phenol as a model pollutant. Batch studies were carried out
Keywords:
on a small laboratory scale electrochemical cell and flow-through treatment studies were carried out
Breakdown products
Adsorption
in a pilot scale circulating fluidised bed reactor. In batch studies, the main breakdown products formed
Electrochemical regeneration were benzoquinone and catechol, with small amounts of hydroquinone. It was found that hydroquinone
Electrochemical oxidation was oxidized to benzoquinone in the presence of the GIC adsorbent. In the continuous process, studied
Phenol in a batch recycle mode, benzoquinone and maleic acid were the main breakdown products detected.
Graphite intercalation compound This investigation has revealed that there are similarities in the breakdown products formed during
electrochemical regeneration of GIC adsorbents with those observed during the indirect electrochemi-
cal oxidation of phenol in an electrochemical cell. However much lower concentrations of breakdown
products can be achieved with the process of adsorption/electrochemical regeneration than with indirect
oxidation.
© 2013 Elsevier Ltd. All rights reserved.

1. Introduction of activated carbon remain a key issue. For environmental and eco-
nomic reasons, disposal of loaded adsorbents is increasingly a less
A number of processes have been developed for the removal attractive option. Thermal regeneration techniques are widely used
of organic pollutants from wastewater including chemical [1] and but these are characterized by 5–10% material losses, high energy
electrochemical oxidation [2], coagulation and flocculation [3], costs and off-site regeneration which results in high transporta-
membrane separation [4], biological treatment [5] and adsorption tion cost [8]. Various alternative regeneration techniques have been
[6]. Adsorption has become increasingly important in recent years investigated, including chemical/solvent regeneration [9], micro-
due to its high efficiency for the removal of non-biodegradable con- bial regeneration [10], wet air oxidation [11], ultrasonic [12] and
taminants. Activated carbon is considered to be the most effective electrochemical regeneration [13].
material for the removal of low concentrations of organics from Electrochemical regeneration of activated carbons has been
wastewaters [7]. The economics and feasibility of the adsorption investigated as an alternative to thermal regeneration [14]. It has
process on a large scale employed for the treatment of wastewater a number of potential advantages compared to thermal regener-
depends upon the disposal or regeneration and reuse of the acti- ation, including in situ regeneration, minimal adsorbent losses,
vated carbon. However, problems associated with the regeneration high regeneration efficiencies and suitability for use in small and
medium sized treatment units [15]. Nevertheless, the high porosity
of activated carbon and its low electrical conductivity lead to long
regeneration times and high energy consumption [16]. In addition,
∗ Corresponding author. Present address: Institute of Chemical Engineering the intra-particle diffusion usually controls the rate of adsorption
& Technology, University of the Punjab, Quaid-e-Azam Campus, Lahore 54590, and desorption of organics from activated carbons which require
Pakistan. Tel.: +92 3004239260; fax: +92 42 99231159. long adsorption times [14]. These issues associated with the regen-
E-mail addresses: hussain [email protected] (S.N. Hussain),
[email protected] (E.P.L. Roberts).
eration of activated carbon have been addressed by the use of a
1
Present address: Department of Chemical and Petroleum Engineering, Univer- non-porous graphite based adsorbent (an unexpanded graphite
sity of Calgary, Calgary AB, T2N 1N4, Canada. Tel. +1 403 220 4466. intercalation compound, GIC), to remove low concentrations

0013-4686/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.electacta.2013.01.020
 S.N. Hussain et al. / Electrochimica Acta 92 (2013) 20–30 21

of coloured, toxic and non-biodegradable organic pollutants from

wastewaters [16,17]. This was followed by the development of an
innovative process in which the organic contaminants in water are
adsorbed onto the surface of a GIC adsorbent and then its electro-
chemical regeneration is carried out in a simple electrochemical cell
with low cell potentials associated with the high electrical conduc-
tivity of the adsorbent [18]. It was shown that the adsorbent can be
rapidly electrochemically regenerated with low energy consump-
tion [16,17]. However, these adsorbent materials have low adsorp-
tive capacity due to their non-porous character. The adsorption
kinetics are however much quicker than those of porous adsor-
bents, with equilibrium achieved within 30–60 min [16,17,19].
These GIC adsorbents have been shown to be effective in
removing a wide range of organic contaminants from wastewater
such as natural organic matter, pesticides (atrazine and metalde-
hyde), chlorinated species (chlorinated phenols, dichloromethane,
trichloroethylene), oils and solvents, recalcitrant COD and organic
dyes [16,17,19].
An ideal electrochemical regeneration process should mineral-
ize the adsorbed pollutants completely to CO2 and H2 O thereby
converting the pollutants into less harmful products. However, Fig. 1. SEM image of GIC particles showing smooth and flat morphology at 100×
magnification.
depending upon the mechanism of the regeneration process, the
contaminants may not be mineralized completely and therefore
some of the adsorbed organics might be converted into undesir- porosimetry showed that there are no pores and therefore no inter-
able breakdown products [20]. Electrochemical regeneration of GIC nal surface area. The mean particle diameter as determined by
adsorbents is achieved through anodic oxidation of the adsorbed Mastersizer-2000 (Malvern Instruments, UK) was 484 ␮m (range
species in which each particle of adsorbent behaves as an anode 50–850 ␮m). The Brunauer Emmett Teller (BET) surface area of
within the electrochemical cell. The regenerated adsorbent, includ- NyexTM 1000 was determined by nitrogen adsorption and found to
ing any water trapped in the bed of adsorbent, is transferred be around 1.0 m2 g−1 .
without additional treatment to the next adsorption cycle. This
means that no secondary waste is being generated in the process. 2.2. Electrochemical regeneration using the batch
However, partially oxidized organics could be released into the electrochemical cell
treated water causing contamination with potentially more toxic
compounds. For investigations of the formation of breakdown products,
Previous studies have suggested that breakdown products batch adsorption was followed by adsorbent/liquid separation and
remain adsorbed on the GIC during electrochemical regeneration electrochemical regeneration in a simple batch cell, as shown in
[19,21]. However, there have been no studies of the adsorp- Fig. 2. The anode current feeder of the cell was a dimensionally sta-
tion characteristics of typical breakdown products. Furthermore, ble anode composed of an iridium/ruthenium mixed metal oxide
a recent study of the treatment of phenol by adsorption onto GIC coated titanium plate (supplied by Electrode Products Technology
adsorbent and electrochemical regeneration showed that oxidation Ltd., UK). The cathode was a perforated stainless steel plate, with
of adsorbed phenol occurs in parallel with indirect oxidation of phe- 1 mm thickness and 3 mm diameter perforations giving an open
nol in solution [22]. In this study, the oxidation characteristics of area of 33%. The anodic and cathodic compartments were sepa-
the adsorbed phenol, and the adsorption behaviour of the break- rated by a microporous Daramic 350 membrane (Daramic Gmbh,
down products are investigated. The implications for wastewater Germany), a high molecular weight polyethylene ribbed sheet con-
treatment process design are considered and investigated using a taining amorphous silica. The role of the membrane is to prevent the
flow through adsorption/electrochemical regeneration reactor. adsorbent bed coming into contact with the cathode short circuit-
ing the electrochemical cell, and to restrict the transport of the salts
2. Experimental in the catholyte into the treated water. The active area of each elec-
trode was 25 cm2 , with anode and cathode compartments being
2.1. Materials 10 mm deep. The cell was constructed from grey PVC using sol-
vent adhesive construction. The DC power source was supplied by
Phenol was chosen as a model pollutant for adsorption Thurlby Thandar Instruments (PL 320QMD, UK).
with electrochemical regeneration of GIC adsorbents. Analytical A mass of 15 g of the GIC adsorbent was added to 100 mL of
grade phenol was supplied as a white crystalline powder by 100 mg L−1 phenol solution in a 250 mL volumetric flask and mixed
Sigma–Aldrich (UK). Analytical grades of other organic chemi- for 30 min (to achieve equilibrium [19]) on a UNIMAX 1010 shaker
cals including 1,4-benzoquinone, hydroquinone, catechol, maleic (Heidolph, UK) at 400 rpm. After adsorption the contents of the flask
and oxalic acids, and inorganic chemicals such as NaCl and HCl were vacuum filtered using a VWR-413 grade filter paper and the
were also supplied by Sigma–Aldrich. All solutions were made concentration of phenol in the filtrate was determined (see below).
up using ultrapure water obtained from a laboratory purification The phenol loaded filtered adsorbent was pressed into the
unit (MilliporeTM Elix 5, Germany). The conductivity of the water anode compartment of the cell to make an adsorbent bed 10 mm
was <0.2 ␮S cm−1 at 25 ◦ C and the total organic carbon content thick. No electrolyte was added to the anode compartment in order
was <30 ppb. The adsorbent used in this work was an unexpanded to avoid the associated cost and the contamination of the treated
GIC, NyexTM 1000 supplied by Arvia Technology Ltd., in the form of water. In addition, it was not necessary to add electrolyte to the
graphitic flakes. The flakes were observed to be flat with a smooth anode compartment because of the high electrical conductivity
morphology as shown in Fig. 1. The specification for NyexTM 1000 of the GIC adsorbent. On the cathode side, 0.3% NaCl solution
provided by Arvia shows that it has 98% carbon content. Mercury acidified with HCl (to pH, 2 ± 1) was added to the same level as
22 S.N. Hussain et al. / Electrochimica Acta 92 (2013) 20–30

Fig. 2. Schematic diagram of a batch electrochemical cell used for regeneration showing front and side views.

the adsorbent bed on other side of the membrane (as shown in concentrations of 100 mL of adsorbate solution with 20 g of adsor-
Fig. 2). The formation of hydrogen on the cathode produces alkali bent in 250 mL volumetric flasks. Mixing was carried out using
that necessitates pH adjustment to maintain acid conditions in a magnetic stirrer at 700 min−1 for 30 min at room temperature.
the cathode compartment in order to avoid the deterioration of This was observed to be the time required to attain equilibrium
Daramic 350 membrane. The power supply was connected to during kinetic studies of phenol adsorption [19]. Considering the
the anode and cathode of the cell and a DC current of 500 mA non-porous nature of the GIC adsorbents, the same equilibrium
(corresponding to a current density 20 mA cm−2 , suitable for elec- time (30 min) was assumed to be applicable for the study of the
trochemical regeneration of GIC [16]) was applied across the cell adsorption isotherms of the breakdown products. Initial and final
for regeneration times in the range of 2–15 min. This gave a charge samples were taken during each experiment and these were
passed in the range of 4–30 C g−1 , based on 15 g of adsorbent. There analyzed as described below. The adsorbent loading, q (mg g−1 ),
was no flow in the cell and the only mixing was caused by the gas was determined from the initial and final concentrations as:
bubbles generated at the electrodes. (Ci − Cf )
After regeneration, the catholyte was gently poured from the q= V (1)
m
cell and all the contents of anode compartment, without any addi-
tional treatment, were transferred to a 250 mL flask containing a where Ci and Cf are the initial and final concentrations (mg L−1 ) of
fresh 100 mL batch of 100 mg L−1 phenol solution. Re-adsorption adsorbate, V is the volume (L) of solution and m is the mass (g) of
was carried out for 30 min on a Unimax shaker at 400 rpm. After adsorbent used.
re-adsorption, the flasks contents were vacuum filtered, and the
filtrate was analyzed to determine its phenol content and break 2.3.2. Multicomponent adsorption
down products using HPLC as described below. Multicomponent adsorption of phenol with its aromatic break-
In order to compare the behaviour of electrochemical oxidation down products was investigated by mixing a range of known
of phenol in a similar electrochemical reactor, in the absence of the masses of adsorbent with 100 mL of a mixed solution containing
adsorbent, the reactor used has been described in detail elsewhere known concentrations of phenol, benzoquinone, hydroquinone and
[21,22], and the same anode was used as in the batch cell discussed catechol using a magnetic stirrer at 700 min−1 for 30 min at room
above. In this case, 500 mL of phenol solution (85 mg L−1 ) con- temperature. Initial and final (after adsorption) samples were fil-
taining a supporting electrolyte (0.3% (w/v) Na2 SO4 ) was treated tered through a 0.45 ␮m syringe filter and analyzed as described
by electrochemical oxidation at a current density of 10 mA cm−2 . below.
Samples of solution were taken at regular intervals during electro-
chemical treatment, but in all other respects the electrochemical 2.4. Continuous adsorption with electrochemical regeneration
regeneration procedure was the same as that in the presence of the
GIC adsorbent. The reactor shown schematically in Fig. 3 was used to perform
continuous (recirculation) adsorption and electrochemical regen-
2.3. Adsorption studies eration in separate zones in the same unit for the batch treatment
of a given volume of phenol solution. The reactor design has been
2.3.1. Adsorption isotherms described in detail by Mohammed et al. [21] who also evaluated and
Adsorption isotherms were determined for phenol and its modelled the reactor performance. The reactor was constructed
intermediate oxidation products by mixing a range of known from clear polycarbonate and the internal dimensions of the unit
 S.N. Hussain et al. / Electrochimica Acta 92 (2013) 20–30 23

Fig. 3. Schematic diagram of the equipment used for continuous adsorption and electrochemical regeneration. (a) Schematic diagram of the continuous unit with auxiliary
equipments, and (b) schematic diagram of the electrochemical regeneration zone illustrating a cross section through line X–X in (a).

were 35 cm wide, 2.2 cm deep and 147 cm tall. The adsorption zone composed of an iridium/ruthenium mixed metal oxide coated tita-
consisted of two symmetrical side zones, each with a rectangular nium plate (supplied by Electrode Products Technology Ltd., UK)
cross section of 10 cm by 2.2 cm. The regeneration section was com- and the cathode was a perforated stainless steel plate. The area of
posed of an anode and cathode separated by a membrane and with each electrode was 12 cm wide by 60 cm in height. A microporous
a bed of GIC particles flowing downward in the anode compart- polyethylene membrane (Daramic GmbH, Germany) separated the
ment. The anode current feeder was a dimensionally stable anode GIC adsorbent from the cathode. In order to generate mixing and to
24 S.N. Hussain et al. / Electrochimica Acta 92 (2013) 20–30

Table 1 Separation of the organic acids such as maleic and oxalic acid
Operating parameters for continuous adsorption and electrochemical regeneration
was carried out using a Hi-Plex-H column [24]. The mobile phase
unit.
was 0.1% trifluoroacetic acid in ultrapure water at a flow rate of
Operating parameters Values 0.6 mL min−1 and the oven temperature was maintained at 60 ◦ C. A
Solution flow rate 380 mL min−1 UV detector was used out at a wavelength of 210 nm. The detection
Solution concentration 500 mg L−1 limits of maleic and oxalic acid were 0.15 and 0.18 mg L−1 , respec-
Current passed 7 mA cm−2 tively. Total organic carbon (TOC) of samples was determined using
Operating time 21 h
a TOC analyzer, VCSH/CSN , (Shimadzu).
Air flow rate 4.8 L min−1
Mode of operation Recirculation
3. Results and discussion

circulate the adsorbent continuously, air was injected at the bottom 3.1. Breakdown products in the batch electrochemical cell
of both adsorption zones through 14 injection nozzles. The total
flow rate of air was adjusted to ca. 5 L min−1 so that the injected Previous studies of the electrochemical oxidation of phenol have
air entrained the adsorbent from the bottom of the regeneration shown that the main breakdown products are hydroquinone, cat-
zone into the adsorption zones and a uniform bed movement was echol, benzoquinone and carboxylic acids (typically maleic and
achieved. The effluent to be treated was injected close to the bottom oxalic acid) [25,26]. In addition, chlorinated phenols can be formed
of the adsorption zones. At the top of the adsorption zones, the air due to indirect oxidation by active chlorine [27]. Similar breakdown
bubbles disengaged and the water flowed into a settlement zone, in products have been detected during electrochemical regeneration
which the loaded adsorbent particles settled out from the upward of phenol loaded GIC [22]. In this study we focus on the breakdown
flowing treated effluent. The high density adsorbent particles set- products obtained from the adsorbed phenol on the GIC surface, by
tled in the centre of the unit, forming a moving packed bed between oxidizing a bed of the filtered adsorbent. The breakdown products
the anode and the membrane in the regeneration chamber. were recovered by mixing the adsorbent with a phenol solution, so
To start up the continuous unit, it was necessary to develop a the concentrations will depend on the adsorption characteristics of
bed of GIC adsorbent in the regeneration zone. The reactor was the breakdown products. Analysis of the phenol concentrations also
filled with clean water from a feed tank using a peristaltic pump. provides information on the regeneration efficiency. It was found
A solution of sodium chloride (0.3%, w/v) at pH 1 (adjusted using that the regeneration efficiency was greater than 90% for regener-
HCl) from a separate tank was circulated through the cathode com- ation times of 2–15 min. However, analysis of regeneration carried
partment as shown in Fig. 3. A mass of 2.5 kg of GIC adsorbent was out over several cycles indicated that regeneration times of 15 min
added to the cell. It was set initially in motion by injecting a moder- or greater are required in order to achieve 100% regeneration.
ate air flow to the outer nozzles located near the walls of the reactor The main breakdown products detected were benzoquinone
in order to have adsorbent circulation in the two adsorption zones. and catechol, with only small amounts of hydroquinone (see
At this stage, no air was injected at the injection points close to Fig. 4a). Almost no carboxylic acids (maleic, fumaric or oxalic acid)
the regeneration bed. After the build up of the bed in the regenera- were detected. The high surface area of the packed bed electrode
tion zone, air was supplied to the injection nozzles close to the bed and the pre-adsorption step should ensure that the pattern of
so that the bed also started to move. These conditions were then breakdown products detected was associated with oxidation on
maintained to achieve a steady state bed movement. the GIC rather than at the current feeder. However, to confirm that
The adsorbent was cleaned for 3 h by flowing tap water through this was the case, oxidation of a phenol solution was carried out in
the cell and providing DC power to the electrochemical cell at the absence of the adsorbent, in order to determine the breakdown
a current density of 7 mA cm−2 . The flow rate of water was products formed at the mixed metal oxide coated titanium anode.
200 mL min−1 . After the adsorbent had been cleaned, the water sup- In this case, the main aromatic breakdown products were benzo-
ply was turned off and the feed was switched to a second tank which quinone and hydroquinone, with lower concentrations of catechol
was filled with 25 L of 500 mg L−1 of phenol solution. It was allowed (Fig. 4b). In addition, significant concentrations of carboxylic acids,
to flow through the reactor in a recirculation mode, i.e. the outlet mainly maleic acid, were detected (data not shown), consistent
from the reactor was returned to the feed tank, under the operating with results reported for a RuO2 coated titanium electrode [28]. It is
conditions given in Table 1. not possible to quantitatively compare the concentrations shown in
Samples were taken from the feed tank and the reactor out- Fig. 4a and b directly, since the amount of phenol oxidized in each
let every 30 min. The pH of the catholyte was checked at regular case is different and the effects of adsorption would need to be
intervals and was maintained at around 1–2. accounted for. However, the relative concentrations of the various
breakdown products obtained suggests that although the mecha-
2.5. Analysis nism of electrochemical oxidation of phenol on GIC is similar to that
on other active electrodes, the relative rates of the various reactions
Phenol and its oxidation products were analyzed using high per- is significantly different.
formance liquid chromatography (HPLC, Varian Prostar). This was Considering the well-established phenol electrochemical oxida-
carried out by determining the retention times of the standard com- tion mechanism [25,26,28]:
pounds of phenol and its oxidation products. All aqueous samples C2 H5 OH + H2 O → C6 H4 (OH)2 + 2e− + 2H+  C6 H4 (O)2 + 2e− + 2H+
were filtered through a 0.45 ␮m filter (Phenomenex Ltd.) prior to
phenol hydroquinone p-benzoquinone (2)
their injection into the respective column.
catechol o-benzoquinone
Separation of the aromatic compounds such as phenol, 1,4-
benzoquinone, hydroquinone and catechol was performed using with subsequent oxidation of benzoquinone to carboxylic acids.
a C18 column (Phenomenex, Macclesfield, UK). The mobile phase The data shown in Fig. 4a suggest that the rate of oxidation of
was 50% methanol (HPLC grade) in ultrapure water at a flow rate hydroquinone on GIC is rapid while the rate of oxidation of cat-
of 0.75 mL min−1 . The oven temperature was maintained at 25 ◦ C. echol is relatively slow. The low concentrations of carboxylic acids
A UV detector was used at wavelengths of 210 and 254 nm [23]. detected may indicate several possibilities: (i) there was insuffi-
The detection limits of phenol, benzoquinone, hydroquinone and cient oxidation to generate organic acids; (ii) the rate of oxidation
catechol were 0.03, 0.01, 0.02 and 0.015 mg L−1 respectively. of the aromatics to carboxylic acids is slow on GIC; (iii) the rate of
 S.N. Hussain et al. / Electrochimica Acta 92 (2013) 20–30 25

8 12
(a) (b)

10

-1

-1
c / mg L

c / mg L
4 6

0 0
0 5 10 15 20 0 20 40 60 80
Regeneraon me / min Treatment me / min

Fig. 4. Breakdown products detected after: (a) electrochemical regeneration of a phenol loaded GIC adsorbent (adsorbent dose: 15 g in 100 mL−1 solution, initial phenol
concentration 100 mg L−1 ), and (b) oxidation of 500 mL of solution containing 85 mg L−1 phenol and 0.3% (w/v) in a batch electrochemical cell (with no adsorbent) at a mixed
metal oxide dimensionally stable anode. A current of 0.5 A was used in both cases, corresponding to a current density: 10 mA cm−2 . () Benzoquinone; () catechol; (䊉)
hydroquinone.

oxidation of these carboxylic acids on GIC is relatively fast; or (iv) 3.3. Hydroquinone oxidation
oxidation of adsorbed benzoquinone led to other organic acids not
studied here, such as succinic acid, malonic acid or acetic acid. During the adsorption studies with hydroquinone, some ben-
zoquinone was also observed, as shown in Fig. 7. It is likely
that the GIC, which is an oxidized form of graphite, might have
3.2. Adsorption isotherms oxidized the hydroquinone or catalyzed the oxidation of the hydro-
quinone by dissolved oxygen. In order to demonstrate this effect,
The objective of developing adsorption isotherms for phenol and two experiments were performed whereby 100 mL of an aque-
its possible breakdown products was to evaluate the adsorptive ous solution containing 200 mg L−1 hydroquinone was stirred for
capacities of these species on the GIC adsorbent. The results reveal 30 min in two separate volumetric flasks. In one of these flasks,
that two different types of adsorption isotherms were observed 20 g of GIC was added while no adsorbent was added to the
for aromatic compounds and carboxylic acids (Figs. 5 and 6). other flask. After 30 min, samples from both the flasks were ana-
The adsorption of phenol and its aromatic intermediates (ben- lyzed for hydroquinone and any other oxidation product. For the
zoquinone, hydroquinone and catechol) were described by the flask without adsorbent, the initial and final samples were exactly
L-shaped (type I) isotherms [29]. These isotherms increased sharply the same and no oxidation products were observed. When the
at low concentrations in comparison to their increase at high con- GIC was present, 14 mg L−1 of benzoquinone was also detected in
centrations due to the availability of readily accessible adsorption solution after 30 min, and the hydroquninone concentration was
sites under these conditions. With an increase in the concentration reduced to 167 mg L−1 . This demonstrates clearly that the pres-
of adsorbate, the GIC adsorbent may be approaching towards satu- ence of the GIC adsorbent leads to the oxidation of hydroquinone to
ration due to the less number of sites available for adsorption. The benzoquinone. The concentration of benzoquinone detected after
adsorptive capacities of phenol, hydroquinone, benzoquinone and adsorption appears to increase with the concentration of hydro-
catechol onto GIC adsorbent were all very similar (Fig. 5). However, quinone (Fig. 7). Since the same equilibrium time (30 min) was
the adsorptive capacities of the aromatic intermediates at low con- applied for the study of adsorption isotherm of hydroquinone at
centrations are relatively low compared to the adsorptive capacity its various initial concentrations, therefore it can be guessed from
of phenol (Fig. 5). Based on these adsorption isotherms it seems Fig. 7 that the rate of oxidation of hydroquinone in terms of the
likely that the pattern of aromatic breakdown products shown in rate of formation benzoquinone appears to increase with the con-
Fig. 4 reflects the distribution of breakdown products formed dur- centration of hydroquinone. Note that the experimental solutions
ing the electrochemical regeneration. used were not deoxygenated.
In contrast, isotherms for carboxylic acids appeared to follow Graphite intercalation compounds are known to be reactive so
the S-shaped (type III) form [29] indicating that adsorption of these the observation of hydroquinone oxidation is not surprising. For
compounds on GIC only takes place at high concentration (Fig. 6). example, Bertin et al. [30] studied esterification reactions in the
It also indicates that the GIC adsorbent has higher affinity for water presence of a GIC which served as an acid catalyst. Further work
than for the carboxylic acids at low concentrations. Canizares et al. is needed in order to determine the mechanism of hydroquinone
[24] found similar results for the adsorption of phenol and its oxi- oxidation.
dation products on granular activated carbon. They observed type Keeping in view the fact that some of the hydroquinone has
I adsorption isotherm for the aromatic compounds (phenol and been oxidized rather than adsorbed due to the formation of benzo-
hydroquinone) and type III adsorption isotherm for the carboxylic quinone, the actual loading of hydroquinone on the GIC adsorbent
acids (maleic and oxalic acids). will be lower than that calculated from the aqueous concentration
26 S.N. Hussain et al. / Electrochimica Acta 92 (2013) 20–30

0.25 0.35
Phenol BQ
0.30
0.20
0.25
0.15 0.20

0.10 0.15
0.10
0.05
0.05
-1
q / mg g
0.00 0.00
0 50 100 150 200 250 300 0 50 100 150 200

0.20 HQ 0.30 Catechol

0.25
0.15
0.20

0.10 0.15
0.10
0.05
0.05
0.00 0.00
0 50 100 150 200 250 300 0 50 100 150 200

-1
c / mg L
Fig. 5. Adsorption isotherms for phenol and its aromatic intermediate breakdown products on a GIC adsorbent, obtained at room temperature using a dose of 20 g of GIC
adsorbent in 100 mL of solution at a range initial concentrations.

of hydroquinone. In order to further illustrate, the adsorption when they are present in solution together along with phenol. Thus,
isotherm of hydroquinone has been corrected by adjusting the the multicomponent adsorption of these species with phenol was
value of q to account for the oxidation of hydroquinone to benzo- investigated. The results revealed that the decrease in concentra-
quinone (Fig. 8). This assumes that the adsorption of hydroquinone tion of phenol and catechol in solution with an increase of the
was not significantly affected by the presence of benzoquinone. adsorbent dose was similar up to a dose of 100 g L−1 (Fig. 9a). At
The adsorption capacity shown in Fig. 8 for hydroquinone on the higher doses, the concentration of catechol decreased below that
GIC can thus be estimated to be around 0.5–1.0 mg g−1 too high of phenol as shown in Fig. 10a. In contrast, the concentration of
due to the effect of oxidation to benzoquinone. benzoquinone increased with an increase of adsorbent dose due to
the oxidation of hydroquinone as discussed above. The concentra-
3.4. Multicomponent adsorption tion of hydroquinone decreased steeply with adsorbent dose due
to a combination of oxidation and adsorption (Fig. 9a).
As benzoquinone, hydroquinone and catechol were observed Detailed interpretation of this data in comparison with the sin-
to be the main intermediate products of phenol oxidation on GIC, gle component isotherms discussed previously is difficult because
it is also important to consider their adsorption behaviour on GIC of the continuous oxidation of hydroquinone. In order to eliminate

0.25
Maleic acid 0.5
Oxalic acid

0.20
0.4

0.15
-1

-1

0.3
q / mg g

q / mg g

0.10
0.2

0.05 0.1

0.00 0.0
0 50 100 150 200 250 300 0 40 80 120 160 200
-1 -1
c / mg L c / mg L
Fig. 6. Adsorption isotherms of aliphatic carboxylic acids a GIC adsorbent, obtained at room temperature using a dose of 20 g of GIC adsorbent in 100 mL of solution at a
range initial concentrations.
 S.N. Hussain et al. / Electrochimica Acta 92 (2013) 20–30 27

25 the formation of benzoquinone, the competitive adsorption of a

mixture of phenol and benzoquinone was studied (Fig. 9b). The
adsorbent loadings obtained for phenol and benzoquinone were
20 calculated as a function of the equilibrium concentrations and
Concentration of BQ detected

compared to the single component isotherms (Fig. 10). The single

after adsorption / mg L-1

component isotherms have been plotted as a Freundlich isotherm

15 obtained from a non-linear least squared fit to the data plotted in
Fig. 6. For phenol, it appears that the presence of benzoquinone
does not affect the adsorption equilibrium (Fig. 10a). For benzo-
10
quinone, the presence of phenol leads to a slight decrease in the
equilibrium loading of around 15–20% (Fig. 10b).
5
It has been reported that in the adsorption of phenols on acti-
vated carbons, the role of the donor–acceptor mechanism involving
carbonyl oxygen groups of the activated carbon acting as an elec-
0 tron donor and the aromatic ring of the phenols as an acceptor is
0 50 100 150 200 250 300 important [31]. Based on the similar adsorption behaviour of phe-
Concentration of HQ before adsorption / mg L-1
nol and its aromatic breakdown products, it may be speculated that
the adsorption of these compounds on the GIC adsorbent occurs
Fig. 7. Formation of benzoquinone during adsorption of hydroquinone on GIC adsor- through a donor–acceptor mechanism.
bent at different initial concentrations of hydroquinone.

3.5. Continuous adsorption with electrochemical regeneration

0.35
An electrochemical reactor carrying out simultaneous adsorp-
0.30 tion and electrochemical regeneration using a GIC adsorbent has
recently been demonstrated for wastewater [21] and gas treatment
0.25 [32,33]. In this study, the concentrations of the breakdown prod-
ucts generated during the oxidation of phenol in this reactor are
0.20 evaluated for the first time. In this system, unlike the batch reactor,
q / mg g-1

a significant proportion of the phenol present is as dissolved phe-

0.15 nol in solution, with a relatively small fraction adsorbed on the GIC.
The breakdown products may thus arise due to both direct oxida-
0.10 tion of adsorbed phenol (and its breakdown products) on the GIC
and indirect oxidation of phenol (and its breakdown products) in
0.05 solution.
The electrochemical regeneration of GIC adsorbent loaded with
0.00 phenol in the continuous electrochemical reactor (Fig. 3) has simi-
0 50 100 150 200 250 300
larities with the studies carried out in batch electrochemical cells.
Concentration of HQ in liquid phase / mg L-1 In this context, the principle aromatic breakdown product observed
was benzoquinone, formed during the treatment of 25 L of water
Fig. 8. Adsorption isotherms of hydroquinone with () and (䊉) without corrected
concentration of hydroquinone in solution after adsorption on GIC adsorbent due
containing 500 mg L−1 of phenol circulated through the continu-
to the formation of benzoquinone. ous adsorption/electrochemical regeneration reactor operating at
a current density of 7 mA cm−2 . Benzoquinone was detected in the

100 100
Phenol (b)
(a) Phenol
90 BQ 90 BQ
HQ
80 Catechol 80

70 70

60 60
-1

-1
c / mg L

c / mg L

50 50

40 40

30 30

20 20

10 10

0 0
0 5 10 15 20 25 30 0 5 10 15 20 25
-1 -1
GIC dose / g (100 mL) GIC dose / g (100 mL)

Fig. 9. Filtrate phenol and aromatic intermediates concentrations after multicomponent adsorption on a GIC adsorbent with a range of doses: (a) 52 mg L−1 of phenol,
benzoquinone (BQ), hydroquinone (HQ) and catechol; (b) 52 mg L−1 of phenol and benzoquinone.
28 S.N. Hussain et al. / Electrochimica Acta 92 (2013) 20–30

80

70
Oxalic acid
60 Maleic acid

50

-1
c / mg L
40

30

20

10

0 2 4 6 8 10 12 14 16 18 20 22
Time / h

Fig. 12. Concentrations of maleic and oxalic acids in the water outlet during the
treatment of 25 L of water containing an initial concentration of 500 mg L−1 phe-
nol circulated through the continuous adsorption and electrochemical regeneration
(shown in Fig. 3) operating at a current density of 7 mA cm−2 .

outlet stream after 15 h of operation (Fig. 11). The concentration of

benzoquinone increased over this time interval and the maximum
concentration of benzoquinone generated was 56 mg L−1 after 21 h.
This accumulation of benzoquinone indicated that the formation
rate was greater than the rate of degradation. This suggests that the
oxidation of benzoquinone can be considered as the rate limiting
step for the oxidation of phenol into carboxylic acids. The other non-
chlorinated aromatic intermediate identified was hydroquinone,
present in very low concentrations (data not shown). As has already
Fig. 10. Adsorption isotherm data for: (a) phenol in the presence of benzoquinone,
been discussed, hydroquinone is the first product formed by the
and (b) hydroquinone in the presence of phenol, on a GIC adsorbent. The lines show
the Freundlich isotherms obtained from a non-linear least squared error fit to the oxidation of phenol. The presence of very low concentrations of
data in Fig. 5, while the data points are based on the data shown in Fig. 9b. hydroquinone suggests that its oxidation to benzoquinone is rela-
tively fast.
feed tank as well as in outlet stream (Fig. 11). The concentration In addition to the formation of these aromatic intermediates,
of benzoquinone was slightly higher in the outlet compared to its maleic and oxalic acid were also observed during the continuous
concentration in the feed tank as shown in Fig. 11. This suggests process as shown in Fig. 12. Relatively low concentrations of maleic
that benzoquinone was being generated in the reactor through- acid were found in comparison to the oxalic acid whose concentra-
out the experiment. Conversely, the concentration of phenol in the tion increased with time, indicating that the oxidation of maleic
tank was higher than its concentration in the outlet, indicating that acid is relatively fast.
phenol removal/oxidation was occurring in the reactor. The con- In order to further analyze the continuous process of adsorption
centration of phenol was eliminated completely from the water with electrochemical regeneration, it was assumed that the benzo-
quinone was the main aromatic intermediate generated during the
continuous electrochemical regeneration for the phenol solution.
60 The percentage of phenol degraded at any time (t) and the percent-
500
age of aromatics formed at any time (t) relative to the amount of
50 phenol removed during the continuous process were calculated as
400 [34]:
40
-1

(phenol)0 − (phenol)t
[Phenol] / mg L

-1

%(Phenol converted)t = × 100 (3)

[BQ] / mg L

300 (phenol)0
Phenol (outlet) 30
Phenol (feed tank) [aromatics]t
200 BQ (outlet) 20 %(Aromatics)t = × 100 (4)
BQ (feed tank) (phenol)0 − (phenol)t

10 where [aromatics]t is the concentration of aromatic intermediates

100
(benzoquinone) in mmol dm−3 present at time (t). (Phenol)0 and
0 (phenol)t are the concentrations of phenol at the start of the process
0 and at any time t in mmol dm−3 .
0 2 4 6 8 10 12 14 16 18 20 22 Fig. 13 shows that around 10% of the phenol was converted to
Time / h aromatics after 12 h. It is also clear from the data shown in Fig. 11,
that 50% of the total benzoquinone formed was produced in the
Fig. 11. Concentration of phenol and benzoquinone during the treatment of 25 L of
first 3 h suggesting that the rate of formation of benzoquinone was
water containing an initial concentration of 500 mg L−1 phenol circulated through
the continuous adsorption and electrochemical regeneration (shown in Fig. 3) oper- high in the first 3–4 h. Similarly the rate of removal of phenol was
ating at a current density of 7 mA cm−2 . higher in the first few hours as shown in Fig. 11. In the first 12 h,
 S.N. Hussain et al. / Electrochimica Acta 92 (2013) 20–30 29

regeneration. At long times, the concentration of organic carbon

% [Aromatics] (based on benzoquinone)
28 100 (TOC) reached a steady state, with only 64% removal achieved after
24 21 h of treatment. Of the total TOC removal achieved after 21 h,
80 95% of this was achieved in the first 12 h. Based on the amounts of

% [Phenol] converted
20 known breakdown products in solution after 21 h, the correspond-
% Aromatics (outlet)
60 ing value of TOC may be estimated around 40% of the TOC actually
% Aromatics (feed tank)
16 present in solution after 21 h. The remaining TOC may be due to the
% Phenol converted (outlet)
% Phenol converted (feed tank) formation of other organics acids (e.g. malonic, succinic, or acetic
12 40
acid) polymers and/or chlorinated breakdown products. Consider-
ing the low concentrations of organic acids (maleic or oxalic acid)
8 20 detected in batch experiment, it seems likely that reaction pathway
4 on GIC may lead to the other organic acids discussed above.
0
0
0 2 4 6 8 10 12 14 16 18 20 22 4. Conclusions
Time / h
The results presented in this paper have demonstrated that
Fig. 13. The percentage of aromatics formed relative to the percentage of phenol intermediate oxidation products are formed during the electro-
removed during batch recycle treatment in the continuous electrochemical reactor,
chemical regeneration of the GIC adsorbent loaded with phenol in
based on the data shown in Figs. 10 and 11.
the batch and continuous processes. Studies in a simple batch elec-
trochemical cell has revealed that the main breakdown products
the phenol concentration in the outlet stream was reduced by 95%. include benzoquinone, hydroquinone, catechol, maleic and oxalic
Iniesta et al. [34] observed only aromatic intermediates including acids. However, these species were observed at low concentrations
hydroquinone, benzoquinone and catechol with low phenol con- compared to the initial concentration of the phenol.
version during bulk electrolysis of 20 mg L−1 phenol solution in It was found that hydroquinone is oxidized to benzoquinone in
acidic conditions using a synthetic boron-doped diamond electrode the presence of GIC, and consequently the concentrations of hydro-
(BDD) at 5 mA cm−2 . They found that the percentage of aromat- quinone detected were low. By contrast, it was found that catechol
ics formed was higher than 80%. However, during the continuous is not easily oxidized on the GIC, and significant concentrations of
process of adsorption with electrochemical regeneration using GIC catechol were detected. Previous studies of electrochemical oxida-
adsorbents only around 10% of the phenol was converted to aro- tion of phenol on other electrode materials have emphasized the
matics suggested that even by applying low current density and hydroquinone pathway [25,26,28], while for GIC it appears that the
high phenol concentration (500 mg L−1 ) the complete oxidation of catechol pathway also plays an important role.
phenol could be possible. The adsorption studies of phenol and its aromatic intermediates
In addition to the breakdown products for the continuous pro- including benzoquinone, catechol and hydroquinone showed that
cess, the observed decrease in TOC of the water both in feed tank all of these species have almost the same adsorption characteristics
and the outlet stream, as shown in Fig. 14, suggests that the organic for the GIC adsorbents and these are described by L-shaped adsorp-
contaminants and breakdown products have been removed and tion isotherms. On the other hand, carboxylic acids including maleic
destroyed. However, it may also be possible that the phenol ends and oxalic acids were observed to follow the S-shaped adsorption
up in another form such as a polymer which needs further inves- isotherms and thus they are poorly adsorbed onto GIC adsorbent
tigation. In addition, the identification and quantification of the at low concentrations. The oxidation of hydroquinone to benzo-
formation of possible chlorinated breakdown products will be the quinone at ambient conditions in the presence of GIC made accurate
subject of a separate paper. The results show that the removal of determination of the adsorption characteristics of hydroquinone
organic carbon was rapid at the beginning of the electrochemical difficult.
The continuous process of adsorption with electrochemical
400 regeneration of GIC adsorbent has similarities with the studies car-
ried out in the batch mode. Benzoquinone was found to be the main
350 intermediate product. The rate of formation of benzoquinone was
found to be higher than its rate of degradation, especially during the
300 first 12 h of treatment. Maleic acid was found in low concentration
Feed tank whereas the concentration of oxalic acid was found to increase with
250 Water outlet
-1

time indicating that the oxidation of maleic acid occurred more

TOC / mg L

rapidly than that oxidation of benzoquinone. Only 64% TOC removal

200
was possible in 21 h of treatment of 500 mg L−1 phenol solution in
150 a batch-recirculation mode using the continuous electrochemical
regeneration cell. It can be anticipated that lower phenol concen-
100 trations and operation with a higher current density could improve
the percentage TOC removal.
50 This study was a preliminary investigation about the forma-
tion of breakdown products during adsorption and electrochemical
0 regeneration using GIC adsorbents for water treatment. However,
0 4 8 12 16 20 a further investigation about the formation of intermediate prod-
Time / h ucts should be carried out under a range of operating conditions.
The optimum conditions for the minimal formation of the break-
Fig. 14. Total organic carbon during batch recycle treatment of 25 L of water contain- down products should be investigated. In addition, a detailed study
ing an initial concentration of 500 mg L−1 of phenol solution circulated through the
continuous adsorption and electrochemical regeneration reactor (shown in Fig. 3)
would be useful in understanding the mechanism of electrochem-
operating at a current density of 7 mA cm−2 . ical regeneration.
30 S.N. Hussain et al. / Electrochimica Acta 92 (2013) 20–30

Acknowledgements [17] N.W. Brown, E.P.L. Roberts, A. Chasiotis, T. Cherdron, N. Sanghrajka, Atrazine
removal using adsorption and electrochemical regeneration, Water Research
38 (2004) 3067.
The authors would like to acknowledge the financial and mate- [18] K.T. Eccleston, N.W. Brown, E.P.L. Roberts, Adsorbents for treating contami-
rial support received from Arvia Technology Ltd. UK and from the nated liquids, US Patent 0321361 A1, 2009.
Institute of Chemical Engineering and Technology, University of the [19] N.W. Brown, E.P.L. Roberts, Electrochemical pre-treatment of effluents con-
taining chlorinated compounds using an adsorbent, Journal of Applied
Punjab, Lahore, Pakistan. Electrochemistry 37 (2007) 1329.
[20] A. Vlyssides, E.M. Barampouti, S. Mai, Electrochemical oxidation of noncyanide
strippers wastes, Environmental Science and Technology 38 (2004) 6125.
References [21] F.M. Mohammad, E.P.L. Roberts, A. Hill, A.K. Campen, N.W. Brown, Continu-
ous water treatment by adsorption and electrochemical regeneration, Water
[1] J. Levec, A. Pintar, Catalytic oxidation of aqueous solutions of organics—an effec- Research 45 (2011) 3065.
tive method for removal of toxic pollutants from waste waters, Catalysis Today [22] S.N. Hussain, H.M.A. Asghar, A.K. Campen, N.W. Brown, E.P.L. Roberts, Break-
24 (1995) 51. down products formed due to oxidation of adsorbed phenol by electrochemical
[2] C.A. Martinez-Huitle, S. Ferro, Electrochemical oxidation of organic pollutants regeneration of a graphite adsorbent, Electrochimica Acta, under review.
for the wastewater treatment: direct and indirect processes, Chemical Society [23] C. Wu, X. Liu, D. Wei, J. Fan, L. Wang, Photosonochemical degradation of phenol
Reviews 35 (2006) 1324. in water, Water Research 35 (2001) 3927.
[3] J. Bratby, Coagulation and Flocculation in Water and Wastewater Treatment, [24] P. Canizares, J. Lobato, J.G. Gomez, M.A. Rodrigo, Electrochemical oxidation of
2nd ed., IWA Publishing, London, 2006. alcohols and carboxylic acids with diamond anodes: a comparison with other
[4] H. Strathmann, Membrane separation processes in advanced wastewater treat- advanced oxidation processes, Journal of Applied Electrochemistry 34 (2004)
ment, Pure and Applied Chemistry 46 (1976) 213. 111.
[5] T. Kuritz, C.P. Wolk, Use of filamentous cyanobacteria for biodegradation of [25] C. Comninellis, C. Pulgarin, Anodic oxidation of phenol for wastewater treat-
organic pollutants, Applied and Environment Microbiology 61 (1995) 234. ment, Journal of Applied Electrochemistry 21 (1991) 703.
[6] R.C. Bansal, M. Goyal, Activated Carbon Adsorption, Taylor and Francis Group, [26] N.B. Tahar, A. Savall, Mechanistic aspects of phenol electrochemical degrada-
Boca Raton, 2005. tion by oxidation on a Ta/PbO2 anode, Journal of the Electrochemical Society
[7] K.A. Halhouli, N.A. Darwish, N.M. Al-Dhoon, Effects of pH and inorganic salts 145 (1998) 3427.
on the adsorption of phenol from aqueous systems on activated decolorizing [27] C. Comninellis, A. Nerini, Anodic oxidation of phenol in the presence of NaCl
charcoal, Separation Science and Technology 30 (1995) 3313. for wastewater treatment, Journal of Applied Electrochemistry 25 (1995) 23.
[8] G.S. Miguel, S.D. Lambert, N.J.D. Graham, The regeneration of field-spent gran- [28] X. Li, Y. Cui, Y. Feng, Z. Xie, J. Gu, Reaction pathways and mechanisms of the
ular activated carbons, Water Research 35 (2001) 2740. electrochemical degradation of phenol on different electrodes, Water Research
[9] T. Sutikno, K.J. Himmeistein, Desorption of phenol from activated carbon by 39 (2005) 1972.
solvent regeneration, Industrial and Engineering Chemistry Fundamentals 22 [29] C.H. Giles, T.H. Macewan, S.N. Nakhwa, D. Smith, Studies in adsorption. Xi. A
(1983) 420. system of classification of solution adsorption isotherms, and its use in diag-
[10] D.H. Hutchinson, C.W. Robinson, A microbial regeneration process for granular nosis of adsorption mechanisms and in measurement of specific surface areas
activated carbon — 1. Process modelling, Water Research 24 (1990) 1209. of solids, Journal of the Chemical Society (1960) 3973.
[11] R.V. Shende, V.V. Mahajani, Wet oxidative regeneration of activated carbon [30] J. Bertin, H.B. Kagan, J.L. Luche, R. Setton, Graphite electrolytic lamellar reagents
loaded with reactive dye, Waste Management 22 (2002) 73. in organic chemistry: esterifications in the presence of graphite bisulfate, Jour-
[12] J.L. Lim, M. Okada, Regeneration of granular activated carbon using ultrasound, nal of the American Chemical Society 96 (1974) 8113.
Ultrasonics Sonochemistry 12 (2005) 277. [31] J.S. Mattson, H.B. Mark, M.D. Malbin, W.J. Weber, J.C. Crittenden, Surface chem-
[13] Z. Minghua, D. Qizhou, L. Lecheng, W. Dahui, Activated carbon adsorption- istry of active carbon: specific adsorption of phenols, Journal of Colloid and
advanced electro-oxidative regeneration for the treatment of biorefractory Interface 31 (1969) 116.
organic pollutants, Chinese Science Bulletin 50 (2005) 489. [32] M.G. Conti-Ramsden, N.W. Brown, E.P.L. Roberts, The combination of adsorbent
[14] R.M. Narbaitz, J.Q. Cen, Electrochemical regeneration of granular activated car- slurry sorption with adsorbent electrochemical regeneration for VOC removal,
bon, Water Research 28 (1994) 1771. Chemical Engineering Journal 130 (2012) 198.
[15] R.M. Narbaitz, A. Karimi-Jashni, Electrochemical regeneration of granular acti- [33] M.G. Conti-Ramsden, N.W. Brown, E.P.L. Roberts, Towards an odour control
vated carbons loaded with phenol and natural organic matter, Environmental system combining slurry sorption and electrochemical regeneration, Chemical
Technology 30 (2009) 27. Engineering Science 79 (2012) 219.
[16] N.W. Brown, E.P.L. Roberts, A.A. Garforth, R.A.W. Dryfe, Electrochemical [34] J. Iniesta, M.A. Michaud, M. Panizza, G. Cerisola, A. Aldaz, C. Comninellis,
regeneration of a carbon-based adsorbent loaded with crystal violet dye, Elec- Electrochemical oxidation of phenol at boron-doped diamond electrode, Elec-
trochimica Acta 49 (2004) 3269. trochimica Acta 46 (2001) 3573.