Chemosphere 68 (2007) 18461853

www.elsevier.com/locate/chemosphere

Advanced oxidation of natural organic matter using hydrogen

peroxide and iron-coated pumice particles
M. Kitis *, S.S. Kaplan
Department of Environmental Engineering, Suleyman Demirel University, 32260 Isparta, Turkey

Received 15 August 2006; received in revised form 12 March 2007; accepted 13 March 2007
Available online 25 April 2007

Abstract

The oxidative removal of natural organic matter (NOM) from waters using hydrogen peroxide and iron-coated pumice particles as
heterogeneous catalysts was investigated. Two NOM sources were tested: humic acid solution and a natural source water. Iron coated
pumice removed about half of the dissolved organic carbon (DOC) concentration at a dose of 3000 mg l1 in 24 h by adsorption only.
Original pumice and peroxide dosed together provided UV absorbance reductions as high as 49%, mainly due to the presence of metal
oxides including Al2O3, Fe2O3 and TiO2 in the natural pumice, which are known to catalyze the decomposition of peroxide forming
strong oxidants. Coating the original pumice particles with iron oxides signicantly enhanced the removal of NOM with peroxide. A
strong linear correlation was found between iron contents of coated pumices and UV absorbance reductions. Peroxide consumption also
correlated with UV absorbance reduction. Control experiments proved the eective coating and the stability of iron oxide species bound
on pumice surfaces. Results overall indicated that in addition to adsorptive removal of NOM by metal oxides on pumice surfaces, surface
reactions between iron oxides and peroxide result in the formation of strong oxidants, probably like hydroxyl radicals, which further
oxidize both adsorbed NOM and remaining NOM in solution, similar to those in Fenton-like reactions.
2007 Elsevier Ltd. All rights reserved.

Keywords: Adsorption; Advanced oxidation; Coating; Hydrogen peroxide; Iron oxide; Natural organic matter; Pumice

1. Introduction tion, sonolysis, and others (Glaze et al., 1987; Glaze

Advanced oxidation processes (AOPs) are the hydroxyl
radical-mediated oxidations which utilize hydroxyl radicals
(OH) as their primary oxidizing species. Hydroxyl radicals
are extremely reactive and non-selective due to their
unpaired electron. They often have kinetic rate constants
much greater for many organic compounds than other oxi-
dants such as ozone. Treatment processes used to generate
hydroxyl radicals in aqueous systems include hydrogen
peroxide/ozone (H2O2/O3), H2O2/ultraviolet (UV), O3/
UV, TiO2/UV, O3/OH, Fe2+/H2O2 (Fenton reaction),
Fe2+/H2O2/UV (photo-Fenton reaction), gamma irradia-
*
Corresponding author. Tel.: +90 246 211 1855; fax: +90 246 237 0859.
E-mail address: mkitis@mmf.sdu.edu.tr (M. Kitis).
et al., 1992; Hoigne, 1997; Wang et al., 2001; Neyens and
Baeyens, 2003; Gogate and Pandit, 2004; Vogelpohl and
Kim, 2004; Pignatello et al., 2006). Although a range of
AOPs has been tested for oxidizing natural organic matter
(NOM) in raw waters or humic and/or fulvic acid isolates
dissolved in synthetic waters (Bekbolet et al., 1996; Eggins
et al., 1997; Gracia et al., 2000; Fukushima et al., 2001;
Wang et al., 2001; Kasprzyk et al., 2004; Murray and Par-
sons, 2004; Park et al., 2004), full-scale application of such
processes to remove NOM in drinking water treatment
plants is still very limited, mainly due to high cost, lack
of experience, requirement of high degree of pre-treatment,
and operational diculties.
Hydrogen peroxide is a moderately strong oxidant;
however, its use alone is not eective for certain refrac-
tory contaminants or NOM due to kinetic limitations at

0045-6535/$ - see front matter 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.chemosphere.2007.03.027
 M. Kitis, S.S. Kaplan / Chemosphere 68 (2007) 18461853
reasonable peroxide concentrations. On the other hand,
aqueous hydrogen peroxide decomposes over heteroge-
neous catalysts including metals (i.e., Fe, Cu, Pt, Ti, and
Ni) and metal oxides immobilized on various support
materials such as sand, silica, zeolites, and alumina. As in
Fenton reactions, the decomposition of peroxide through
interactions with the surface sites of such catalysts results
in the formation of strong oxidants including hydroxyl
radicals, which are shown to eectively oxidize various syn-
thetic organic chemicals and NOM (Abbot and Brown,
1990; Al-Hayek and Dore, 1990; Tyre et al., 1991; Watts
et al., 1993; Ravikumar and Gurol, 1994; Gates and Sie-
grist, 1995; Miller and Valentine, 1996; Lin and Gurol,
1996, 1998; Gurol and Lin, 2001; Neyens and Baeyens,
2003). It was suggested that the hydroxyl radicals react
with solutes adsorbed on the iron oxide surfaces, causing
oxidation of the organic compounds (Gurol and Lin,
2001). It was demonstrated that an increase in the
iron oxide (goethite particles) concentration signicantly
increased the decomposition rate of peroxide showing the
eectiveness of the iron oxide surface as a catalyst. The
increased rate of peroxide decomposition resulted in faster
removal for BuCl, an indication of faster generation of OH
in the reaction system. The oxidation rate was relatively
insensitive within the pH range of 59 and no pH
adjustment of the medium was needed for eective oxida-
tion (Gurol and Lin, 2001). No sludge production and
the pH-independence were suggested to be the major
advantages of this heterogeneous process compared to
the Fenton oxidation, which produces signicant quanti-
ties of iron hydroxide sludge and is only eective at pH val-
ues of about 35 (Lin and Gurol, 1998; Gurol and Lin,
2001).
The main objective of this work was to investigate the
oxidative removal of NOM from waters using hydrogen
peroxide and iron-coated pumice particles as heteroge-
neous catalysts. Pumice is a light, highly porous volcanic
stone, available at very low cost. Pumice particles, resem-
bling a sponge, consist of a network of irregular or oval
shape internal voids/pores or vesicles, some of which are
interconnected and open to the external surface, while oth-
ers are isolated inside the particle. Original or surface-mod-
ied pumice particles were shown to be eective as granular
support media in ltration, in heterogeneous catalytic reac-
tions and for biolm growth, and as adsorbent in removing
target pollutants from waters (Akbal et al., 2000; Rachel
et al., 2002; Farizoglu et al., 2003; Njau et al., 2003; Rao
et al., 2003; Kitis et al., 2005; Kitis et al., 2007). As an alter-
native to other support materials such as sand, the advan-
tages of pumice particles are that they are highly porous
(pore volumes up to 85%) and have higher surface areas
(about 530 m2 g1), which may immobilize more amounts
of catalysts thus providing more reaction sites on a mass
basis for the oxidation of NOM. To our knowledge, the
use of original or surface-modied pumice particles for
the catalytic oxidation of NOM in waters has not been
investigated before.
1847
2. Materials and methods
2.1. Materials
Natural pumice from Isparta, Turkey was used in this
work. The raw sample was ground by primary crusher
(hammer type), and then sieved to four dierent particle
size fractions (<63, 63125, 125250, and 2501000 lm).
Each fraction was coated by iron using reagent grade
FeCl3 6H2O, employing the method reported by Lai
et al. (2000) and Lai and Chen (2001) with some modica-
tions. The details of the coating process were reported else-
where (Kitis et al., 2007). As demonstrated in preliminary
coating trials, iron oxide precipitates or colloids in aqueous
phase were easily detected by the fast and practical turbid-
ity measurements. For many coating trials, the nal solu-
tion turbidities achieved were in fact less than 0.01 NTU.
To further demonstrate the absence of any potential solu-
ble or colloidal iron species, in one of the initial coating tri-
als, the nal soak distilled and deionized (DDW) solution
(with a pH of about 5.56.0) was analyzed for total iron
concentration. It was found to be about 0.7 mg l1. This
value can be accepted as very low given the fact that there
was 200 g of pumice in 1 l DDW, a concentration much
higher than those typically dosed in the experiments (Kitis
et al., 2007). Therefore, these preliminary experiments dem-
onstrated that iron oxide precipitates/colloids are eec-
tively removed from cleaning solutions during coating
and that NOM oxidation is due to iron oxide coated pum-
ice not to colloidal iron in the solution.
Two dierent NOM sources were used in the experi-
ments: a surface water (Alibeykoy reservoir) supplying
some portion of City of Istanbuls drinking water demand;
and humic acid (HA) isolate purchased from Acros Organ-
ics. These two NOM sources were selected since they repre-
sented a low- (Alibeykoy) and a high-SUVA (specic UV
absorbance) (HA) water, enabling the evaluation of vari-
ous pumices in removing NOM with a wide range of chem-
ical characteristics. As a simple bulk water parameter,
SUVA provides a quantitative measure of the unsaturated
double bonds or pp electron interactions such as in aro-
matic moieties of NOM. The raw water was ltered
(0.45 lm) before water quality analysis and oxidation
experiments and kept at 4 C in dark before use. HA solu-
tions were prepared in DDW with a target dissolved
organic carbon (DOC) concentration of 6 mg l1. DOC
concentration of DDW was below the detection limit of
the TOC analyzer (50 lg l1). The DOC concentration of
the natural water was 3.9 mg l1.
2.2. Analytical techniques
The physicochemical characteristics of each pumice frac-
tion were determined by measuring specic surface area,
point of zero charge (pHPZC), iron content, total surface
acidic and basic groups, and by SEM-EDX (energy disper-
sive X-ray spectrometer) and XRF (X-ray uorescence)
1848 M. Kitis, S.S. Kaplan / Chemosphere 68 (2007) 18461853

Table 1
Characteristics of original and iron-coated Isparta pumice fractions
Pumice Specic surface area Iron content pHPZC Total surface acidic groupsa Total surface basic groupsb
(m2 g1) (mg Fe g1)
(meq g1) (meq m2) (meq g1) (meq m2)
Original (untreated) fractions
<63 lm 14.2 2.1 9.0 1.33 0.09 0.80 0.06
63125 lm 9.6 0.7 9.2 0.40 0.04 0.60 0.06
125250 lm 8.3 0.9 8.8 0.35 0.04 0.73 0.09
2501000 lm 2.1 7.9 9.2 0.80 0.38 0.95 0.45
Iron-coated fractions
<63 lm ICc 9.2 16.2 6.2 1.58 0.17 0.75 0.08
63125 lm IC 6.8 5.4 6.1 1.35 0.20 1.13 0.17
125250 lm IC 6.9 6.6 7.7 0.70 0.10 0.70 0.10
2501000 lm IC 7.9 23.1 8.4 1.20 0.15 0.65 0.08
Values reported are the averages of triplicate measurements.
a
NaOH uptake.
b
HCl uptake.
c
IC: Iron-coated.

analysis (Table 1). The detailed physicochemical character-
istics of pumice fractions are presented elsewhere (Kitis
et al., 2007). Pumice surface chemistry was characterized
in terms of pHPZC and acid and base neutralization capac-
ities. Values of pHPZC for pumice samples were determined
through a pH equilibration method (Dastgheib et al., 2004).
Total surface acidic groups (NaOH uptake) and total sur-
face basic groups (HCl uptake) were measured employing
the Boehm method (alkalimetric titration) with some mod-
ications (Summers, 1986; Dastgheib et al., 2004). Iron con-
centrations in solutions were measured using atomic
absorption spectrophotometry (Perkin Elmer). Acid diges-
tion analysis was performed to quantify the iron content
of the pumice samples (Lo and Chen, 1997). Specic surface
area measurements were conducted using a Micromeritics
Flowsorb II-2300 analyzer employing the conventional
multi-point BET technique (N2 gas adsorption). All analy-
tical measurements were conducted in triplicates.
A UVvisible spectrophotometer (UV-1601, Shimadzu)
was used to measure the UV absorbances (in triplicates)
in water samples. DOC analysis was performed by the high
temperature combustion method using a TOC Analyzer
(TOC-VCPH, Shimadzu). The minimum quantication
limit of the TOC analyzer was 50 lg l1. A 30% hydrogen
peroxide aqueous solution (Merck) was used as the stock
solution for all experiments. Hydrogen peroxide concen-
tration was measured with a titrametric test kit (Hach
HYP-1). All chemicals used were reagent grade. DDW
was used for stock solution preparations and dilutions.
2.3. Experimental procedures
All experiments were conducted in completely mixed
batch reactors (CMBRs) at various pumice (03000 mg l1)
and peroxide (01000 mg l1) dosages. Pumice and perox-
ide were also dosed alone to determine the extent of
NOM removal by adsorption and peroxide oxidation only.
All oxidation and/or adsorption experiments were con-
ducted in parallel CMBRs. Statistical analysis of the data
was performed based on t-statistics and 95% condence
intervals were calculated from parallel tests and triplicate
measurements. CBMRs used were 60-ml glass amber
bottles (solution volume of 50 ml) sealed with PTFE
screw-caps. All experiments were conducted at constant
temperature of 20 1 C. After dosing original or iron-
coated pumice (doses of 30, 100, 1000, or 3000 mg l1)
and peroxide (doses of 50, 100, 150, 300, 600, and
1000 mg l1), CMBRs containing natural water or HA
solution were kept well-mixed (150 rpm) in oxic conditions
in a temperature-controlled orbital shaker. CMBRs were
covered with aluminum foil to prevent the introduction
of light. Furthermore, the top portion of the orbital shaker
was also covered to provide dark conditions. After the pre-
selected oxidation time terminated, the bottles were opened
and samples were taken to measure residual peroxide and
determine the amount of sodium sulte (Na2SO3) solution
required to quench the residual peroxide and stop the oxi-
dation reaction. Sulte was added in slight excess (i.e., 1.2
times) of the stoichiometric requirements to ensure com-
plete quenching. Bottles were then closed and mixed about
3 min at 100 rpm in the shaker for quenching reaction.
Solutions were ltered (1 lm glass ber lters) to remove
pumice particles and analyzed for pH, UV absorbance
and DOC concentration to quantify NOM removal. Filter
papers were pre-washed with 900 ml of DDW to prevent
potential leaching of materials from the lter matrix. UV
absorbances were measured at a wavelength of 280 nm
rather than 254 nm since sodium azide was added
(50 mg l1) to Alibeykoy water samples to prevent any
potential microbial degradation of NOM before experi-
ments. Sodium azide absorbs UV light at 254 nm but not
at 280 nm. Furthermore, sodium sulte was also found to
interfere with UV measurements at 254 nm.
While no buer was employed for the natural water
(Alibeykoy) due to its relatively high level of alkalinity
(125 mg l1 as CaCO3), HA solution prepared in DDW
 M. Kitis, S.S. Kaplan / Chemosphere 68 (2007) 18461853
1.0
0.9
0.8
Normalized UV280 Abs.
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0 5 10 15 20 25 30 35
Time (h)
Fig. 1. Oxidation kinetics of humic acid with peroxide and iron-coated
Isparta pumice (<63 lm pumice fraction, H2O2 dose = 300 mg l1, pumice
dose = 3000 mg l1, UV280 absorbance and DOC concentration of initial
HA solution are 0.33 cm1 and 6.1 mg l1, respectively). Error bars
indicate the 95% condence intervals.
was buered at pH of 7.0 0.1 using 0.001 M phosphate
buer to prevent variations in pH after dosing of peroxide
and pumice and during oxidation. The pH of raw Ali-
beykoy water was 8.0. Final pH values after oxidations
were between 7.98.1 and 6.87.2 for Alibeykoy water
and HA solution, respectively, suggesting that constant
pH conditions were obtained in oxidation reactions for
each NOM source. Control experiments were conducted
without peroxide and pumice dosing to check for the pos-
sible losses of DOC (i.e., via photo-degradation, adsorp-
tion onto bottles, etc.) during 24 h of mixing. Change in
DOC concentrations in control bottles were within stan-
dard error of DOC measurements (0.1 mg l1), indicating
the stability of NOM during mixing.
Kinetic experiments in CMBRs were performed with
constant pumice and peroxide dose at oxidation periods
of 1, 2, 4, 8, 12, 24, 36, and 48 h. As shown in Fig. 1, initial
kinetic experiments indicated that no more statistically sig-
nicant oxidation of NOM (as measured by UV absor-
bance and DOC) occurs after 24 h of reaction with
peroxide and iron-coated pumice; therefore 24 h of reac-
tion period was employed for all remaining batch experi-
ments with variable original or coated pumice and
peroxide dosages. After an initial rapid removal kinetics,
the rate of removal decreased by reaction time, exhibiting
a typical rst-order kinetics. Similarly, rst-order rate
kinetics was reported for hydrogen peroxide decomposition
on supported iron oxides (Abbot and Brown, 1990) and
goethite (Lin and Gurol, 1998).
3. Results and discussion
After the initial kinetic work, tests were conducted by
only dosing pumice to determine the adsorptive removal
of NOM from waters. While original pumice reduced UV
absorbance by about 10%, iron coated pumice adsorbed
1849
a signicant portion of HA (59% UV absorbance reduc-
tion) (Fig. 2). Correspondent DOC removals were 22%
and 49%. Similar results were found for the Alibeykoy
water at the same pumice dosages (8% and 41% UV absor-
bance reductions by original and iron coated pumice,
respectively). Our previous work similarly indicated that
iron oxide coated pumice may be a promising novel adsor-
bent in removing NOM from waters, based on high NOM
adsorption capacities obtained even at neutral pH values.
Furthermore, iron oxide surfaces preferentially removed
high UV-absorbing fractions of NOM, with UV absor-
bance reductions up to 90% (Kitis et al., 2007). Mainly sur-
face complexation-ligand exchange reactions are favored
40 45 50
by many researchers as the mechanism of NOM adsorption
on iron oxide or aluminum oxide surfaces. In aqueous sys-
tems, the hydroxylated iron oxide surface sites behave like
diprotic acids, with three potential species being BFeOH 2,
FeOH and FeO. The replacement of surface-coordi-
nated H2O or OH groups from FeOH 2 or FeOH by anio-
nic functional groups of NOM (i.e., carboxyl and hydroxyl
groups) results in the site-specic NOM adsorption onto
oxide surfaces (Tipping, 1981; Gu et al., 1994; Edwards
et al., 1996; Chang et al., 1997).
For both HA solution and Alibeykoy water, peroxide
dosed alone (501000 mg l1) provided minor DOC and
UV absorbance reductions, always less than 7% (Fig. 2).
This result is expected since it is known that peroxide is
not eective in oxidizing NOM or refractory synthetic
chemicals. On the other hand, interestingly, original
pumice (3000 mg l1) and peroxide (1000 mg l1) dosed
together resulted in UV absorbance reductions of 38%
and 49% in Alibeykoy water and HA solution, respectively.
Even at lower dosages of original pumice (1000 mg l1) and
peroxide (300 mg l1), such reductions were 20% and 23%,
which were also statistically higher than those obtained by
peroxide alone based on 95% condence intervals. Similar
100
H2O2 only
90 H2O2+Original Pumice
80 H2O2+IC Pumice
UV280 Abs Reduction (%)
Original Pumice-Adsorption only
70 IC Pumice-Adsorption only
60
50
40
30
20
10
0
0 50 150 300 1000
H2O2 Dose (mg/l)
Fig. 2. UV absorbance reductions in humic acid solution by original or
coated pumice (adsorption only), hydrogen peroxide alone, and original or
coated pumice and peroxide together (<63 lm pumice fraction, pumice
dose = 3000 mg l1, reaction time = 24 h, IC: iron-coated). Error bars
indicate the 95% condence intervals.
1850 M. Kitis, S.S. Kaplan / Chemosphere 68 (2007) 18461853

reductions were found in DOC concentrations. It should be the sum of the individual processes, further proving that
noted that original Isparta pumice contains 16.6% Al2O3, the heterogeneous system is catalytically active.
4.8% Fe2O3 and 0.6% TiO2 by mass. All these metal oxides It was found that an increase in coated pumice and
are known to catalyze the decomposition of peroxide peroxide dosages signicantly enhanced NOM removals
resulting in the formation of strong oxidants. Therefore, (Fig. 4). As the peroxide dose increased at constant
even the natural pumice itself appears to be an eective cat- pumice dose, UV absorbance reductions increased in HA
alyst for the oxidative destruction of NOM. Similarly, solution. Furthermore, as the dose ratio of H2O2/coated
Ravikumar and Gurol (1994) showed that pentachlorophe- pumice (mg mg1) increased, NOM removals as mea-
nol and trichloroethylene adsorbed on natural sand parti- sured by DOC also increased (Fig. 5). Such increases in
cles can be oxidized by hydrogen peroxide even without DOC removals were statistically signicant. The studied
addition of an iron salt. Given the fact that natural pumice dose ratio ranged between 0.02 and 25. DOC concentra-
was found to be moderately eective even for the refractory tion of 1.7 mg l1 was achieved at 0.33 H2O2/coated pum-
NOM in this work, it can also be eective in the oxidation ice dose ratio (1000 mg l1 peroxide and 3000 mg l1
of synthetic organic chemicals (SOCs) when used in combi- pumice). Even at much lower peroxide (300 mg l1) and
nation with oxidants such as ozone or peroxide. pumice (100 mg l1) dosages (dose ratio of 3.0), DOC
Coating the original pumice fractions with iron oxides
further enhanced the removal of NOM with peroxide both
in the natural water and HA solution (Fig. 2). Such 1.0
enhancements were statistically signicant based on t-sta- 0.9
tistics. UV absorbance reductions of 70% and 49% were
0.8
achieved after dosing 150 mg l1 peroxide and 3000 mg l1 Normalized UV280 Abs.
0.7
coated pumice in HA solution and Alibeykoy water,
respectively. Maximum UV reductions obtained were 0.6
86% and 67% for HA solution and Alibeykoy water, 0.5
respectively. Similarly, correspondent maximum DOC 0.4
reductions were 73% and 35%. These results indicate that 0.3
P=0
in addition to metal oxides already present on natural pum-
0.2 P=30
ice surfaces coated iron oxides are also very eective in the P=100
generation of strong oxidants, thus in the heterogeneous 0.1 P=1000
P=3000
oxidation of NOM. This observation was further sup- 0.0
0 100 200 300 400 500 600 700 800 900 1000
ported by the strong linear correlation (R2 = 0.95) found
H2O2 Dose (mg/l)
between iron contents of coated pumices and UV absor-
bance reductions obtained at constant pumice and perox- Fig. 4. Impact of iron coated pumice and peroxide dosages on UV
ide dosages (Fig. 3). In addition, as can be noted from absorbance reductions in humic acid solution (<63 lm pumice fraction,
Fig. 2, UV absorbance reductions by combined adsorption pumice dosages are in mg l1). Error bars indicate the 95% condence
(coated pumice) and oxidation are statistically higher than intervals.

7
100
90 6
UV280 Abs Reduction (%)

80
5
70
DOC (mg/l)

60 4
y = 6.06x - 23.27
50 R2 = 0.95
3
40
2 HA Initial DOC
30
P=30
20 P=100
1
P=1000
10
P=3000
0 0
0 5 10 15 20 0.01 0.1 1 10 100

Fe Content (mg Fe/g pumice) H2O2 / Pumice dose (mg/mg)

Fig. 3. Impact of iron contents of coated pumices on UV absorbance Fig. 5. Impact of H2O2/pumice (iron coated) dose ratios on DOC
reductions in humic acid solution (H2O2 dose = 300 mg l1, pumice removals in humic acid solution (<63 lm pumice fraction, pumice dosages
dose = 3000 mg l1). Error bars indicate the 95% condence intervals. are in mg l1). Error bars indicate the 95% condence intervals.
 M. Kitis, S.S. Kaplan / Chemosphere 68 (2007) 18461853 1851

concentration of 2.8 mg l1 (54% reduction) was obtained.
Similar trends were also observed for the natural water.
The optimum dose ratio of H2O2/coated pumice could be
selected based on target DOC removal and cost consider-
ations in a specic application. It was found that, at con-
stant peroxide dose, as the iron coated pumice dose
increased, peroxide consumption increased (Fig. 6). Fur-
thermore, peroxide consumption positively correlated with
UV absorbance reductions at each pumice dose level
(Fig. 7), proving that metal oxide surfaces are responsible
for the decomposition of peroxide resulting in the genera-
tion of strong oxidants. No correlation was found between
peroxide consumption and UV absorbance reduction in the
case of peroxide dose only.
1000
H2O2 Demand (mg/l)
All these results overall indicate that possibly three
mechanisms simultaneously play role in the removal of
NOM from solutions: (1) a portion of NOM adsorbs onto
iron oxide on pumice surfaces depending on the coated
pumice dose, (2) remaining NOM fractions (i.e., more
hydrophilic and lower UV-absorbing, as demonstrated by
UV measurements) in solutions are further oxidized with
strong oxidants, probably hydroxyl radicals, which are
produced from the surface reactions among iron oxides
and hydrogen peroxide, similar to those in Fenton-like
reactions, and (3) adsorbed NOM fractions on iron oxide
surface sites are also oxidized by strong oxidants. Similarly,
it was concluded that the hydroxyl radicals, produced
through the interactions of peroxide with the surfaces of
iron oxide (goethite) particles, react with solutes adsorbed
on the iron oxide surfaces, causing oxidation of the organic
compounds (Lin and Gurol, 1996; Lin and Gurol, 1998;
Gurol and Lin, 2001). In addition, in the catalytic ozona-
tion of NOM on alumina, Kasprzyk et al. (2006) found
that adsorption plays a major role only during the rst

100 few cycles of ozonation after which the catalytic ozonation

system becomes in dynamic balance of continuous adsorp-
tion and oxidation of NOM at the solidliquid interface.
Control experiments performed in DDW indicated that less
10
HP=50
than 0.2 mg l1 of total iron was released to solution after
HP=100
24 h of mixing even with the largest pumice dose of
HP=300
3000 mg l1 and peroxide dose of 150 mg l1 (solution
HP=600 pH 58.3), suggesting the eective coating and the stability
HP=1000 of iron oxide species bound on pumice surfaces. This result
1
0 500 1000 1500 2000 2500 3000
conrmed that strong oxidants are generated through
Pumice dose (mg/l)
interactions among iron oxides bound on surfaces and per-
oxide, not through iron species in solutions as in conven-
Fig. 6. Impact of iron coated pumice and hydrogen peroxide dosages on tional Fenton reactions.
hydrogen peroxide demands in Alibeykoy water (<63 lm pumice fraction,
hydrogen peroxide (HP) dosages are in mg l1). Error bars indicate the
95% condence intervals. 4. Conclusions

A total of four dierent particle size fractions of natural

pumice was coated by iron oxides and used as heteroge-
100
neous catalysts in the oxidation of NOM by hydrogen per-
90 oxide. As an alternative to other support materials such as
sand, the advantages of pumice particles are that they are
UV280 Abs Reduction (%)

80
highly porous and have higher surface areas, which immo-
70
bilize more amounts of metal oxide catalysts, thus provid-
60
ing more reaction sites. For both HA solution and the
50 natural water, iron coated pumice with a dose of
40 3000 mg l1 removed about half of the DOC by adsorption
only. Iron oxides on pumice surfaces preferentially
30 Only H2O2
P=30
adsorbed high UV-absorbing fractions of NOM. Peroxide
20 dosed alone provided minor DOC and UV absorbance
P=1000
10
P=3000
reductions. Original pumice and peroxide dosed together
0 provided UV absorbance reductions as high as 49%,
0 50 100 150 200 250 300 350 400 450 500 mainly due to the presence of metal oxides including
H2O2 Demand (mg/l) Al2O3, Fe2O3 and TiO2 in the natural pumice. On the other
hand, coating the original pumice particles with iron oxides
Fig. 7. Relationships among hydrogen peroxide demands and UV
absorbance reductions in Alibeykoy water (<63 lm pumice fraction, enhanced the removal of NOM with peroxide both in the
coated pumice dosages are in mg l1). Error bars indicate the 95% natural water and HA solution. Such enhancements were
condence intervals. statistically signicant based on t-statistics (95% condence
1852 M. Kitis, S.S. Kaplan / Chemosphere 68 (2007) 18461853
intervals). Maximum DOC reductions obtained for HA
solution and the natural water were 73% and 35%, respec-
tively, indicating that in addition to metal oxides already
present on natural pumice surfaces, coated iron oxides
are also eective in the heterogeneous oxidation of
NOM. This observation was further supported by the
strong linear correlation found between iron contents of
coated pumices and UV absorbance reductions. An
increase in coated pumice and peroxide dosages signi-
cantly enhanced NOM removals. At constant peroxide
dose, as the iron coated pumice dose increased, peroxide
consumption increased, which also positively correlated
with UV absorbance reductions, proving that metal oxide
surfaces are responsible for the decomposition of peroxide
resulting in the generation of strong oxidants. Control
experiments proved the eective coating and the stability
of iron oxide species bound on pumice surfaces.
The use of hydrogen peroxide and iron-coated pumice
particles in removing NOM through both adsorption and
advanced oxidation process may be an eective tool during
water treatment operations. Studies are in progress to
determine the impacts of natural pumice source, particle
size fraction, coating process, solution pH and tempera-
ture, reactor congurations (completely mixed/suspended
and xed-bed reactor), and pumice regeneration on the
removal of NOM with this process.
Acknowledgements
This work was supported in part by a Research Grant
(Career Award Program, ICTAG 104I122) from the Scien-
tic and Technical Research Council of Turkey (TUBI-
TAK); however, it has not been subjected to TUBITAKs
peer and policy review and therefore does not necessarily
reect the views of TUBITAK and no ocial endorsement
should be inferred. The authors thank graduate research
assistants Nevzat O. Yigit and Emine Karakaya for their
technical assistance in the lab work.
References
Abbot, J., Brown, D.G., 1990. Kinetics of iron-catalysed decomposition of
hydrogen peroxide in alkaline solution. Int. J. Chem. Kinet. 22, 963974.
Akbal, F.O., Akdemir, N., Onar, A.N., 2000. FT-IR spectroscopic
detection of pesticide after sorption onto modied pumice. Talanta 53,
131135.
Al-Hayek, N., Dore, M., 1990. Oxidation of phenols in water by hydrogen
peroxide on alumina supported iron. Water Res. 24, 973982.
Bekbolet, M., C zkosemen, G., 1996. Photocatalytic oxidation
ecen, F., O
and subsequent adsorption characteristics of humic acids. Water Sci.
Technol. 34, 6572.
Chang, Y., Li, C.W., Benjamin, M.M., 1997. Iron oxide-coated media for
NOM sorption and particulate ltration. J. Am. Water Works Ass. 89,
100113.
Dastgheib, S.A., Karanl, T., Cheng, W., 2004. Tailoring activated
carbons for enhanced removal of natural organic matter from natural
waters. Carbon 42, 547557.
Edwards, M.E., Benjamin, M.M., Ryan, J.N., 1996. Role of organic
acidity in sorption of natural organic matter to oxide surfaces. Colloid
Surf. A 107, 297.
Eggins, B.R., Palmer, F.L., Byrne, J.A., 1997. Photo-catalytic treatment of
humic substances in drinking water. Water Res. 31, 12231226.
Farizoglu, B., Nuhoglu, A., Yildiz, E., Keskinler, B., 2003. The perfor-
mance of pumice as a lter bed material under rapid ltration
conditions. Filtr. Sep. 40, 4146.
Fukushima, M., Tatsumi, K., Nagao, S., 2001. Degradation characteris-
tics of humic acid during photo-fenton processes. Environ. Sci.
Technol. 35, 36833690.
Gates, D.D., Siegrist, R.L., 1995. In-situ chemical oxidation of trichloro-
ethylene using hydrogen peroxide. J. Environ. Eng.-ASCE 121, 639644.
Glaze, W.H., Kang, J.W., Chapin, D.H., 1987. The chemistry of water
treatment processes involving ozone, hydrogen peroxide and ultravi-
olet radiation. Ozone-Sci. Eng. 9, 335352.
Glaze, W.H., Beltran, F., Tuhkanen, T., Kang, J.W., 1992. Chemical
models of advanced oxidation processes. Water Qual. Res. J. Can. 27,
2342.
Gogate, P.R., Pandit, A.B., 2004. A review of imperative technologies for
wastewater treatment I: oxidation technologies at ambient conditions.
Adv. Environ. Res. 8, 501551.
Gracia, R., Cortes, S., Sarasa, J., Ormad, P., Ovelleiro, J.L., 2000.
Heterogeneous catalytic ozonation with supported titanium dioxide in
model and natural waters. Ozone-Sci. Eng. 22, 461471.
Gu, B., Schmitt, J., Chen, Z., Liang, L., McCarthy, J.F., 1994. Adsorption
and desorption of natural organic matter on iron oxide: mechanisms
and models. Environ. Sci. Technol. 28, 3846.
Gurol, M.D., Lin, S.S., 2001. Hydrogen peroxide/iron oxide-induced
catalytic oxidation of organic compounds. Water Sci. Technol.: Water
Supply 1, 131138.
Hoigne, J., 1997. Inter-calibration of OH radical sources and water
quality parameters. Water Sci. Technol. 35, 18.
Kasprzyk, H.B., Dabrowska, A., Swietlik, J., Nawrocki, J., 2004. The
application of the peruorinated bonded alumina phase for natural
organic matter catalytic ozonation. J. Environ. Eng.-ASCE 3, 4150.
Kasprzyk, H.B., Raczyk, S.U., Swietlik, J., Nawrocki, J., 2006. Catalytic
ozonation of natural organic matter on alumina. Appl. Catal. B:
Environ. 62, 345358.
Kitis, M., Karakaya, E., Yigit, N.O., Civelekoglu, G., Akcil, A., 2005.
Heterogeneous catalytic degradation of cyanide using copper-impreg-
nated pumice and hydrogen peroxide. Water Res. 39, 16521662.
Kitis, M., Kaplan, S.S., Karakaya, E., Yigit, N.O., Civelekoglu, G., 2007.
Adsorption of natural organic matter from waters by iron-coated
pumice. Chemosphere 66, 130138.
Lai, C.H., Chen, C.Y., 2001. Removal of metal ions and humic acid from
water by iron-coated lter media. Chemosphere 44, 11771184.
Lai, C.H., Lo, S.L., Chiang, H.L., 2000. Adsorption/desorption properties
of copper ions on the surface of iron-coated sand using BET and
EDAX analyses. Chemosphere 41, 12491255.
Lin, S.S., Gurol, M.D., 1996. Heterogeneous catalytic oxidation of organic
compounds by hydrogen peroxide. Water Sci. Technol. 34, 5764.
Lin, S.S., Gurol, M.D., 1998. Catalytic decomposition of hydrogen
peroxide on iron oxide: kinetics, mechanism and implications. Envi-
ron. Sci. Technol. 32, 14171423.
Lo, S.L., Chen, T.Y., 1997. Adsorption of Se(IV) and Se(VI) on an iron-
coated sand from water. Chemosphere 35, 919930.
Miller, C.M., Valentine, R.L., 1996. Hydrogen peroxide decomposition
and quinoline degradation in the presence of aquifer material. Water
Res. 29, 23532359.
Murray, C.A., Parsons, S.A., 2004. Removal of NOM from drinking
water: Fentons and photo-Fentons processes. Chemosphere 54,
10171023.
Neyens, E., Baeyens, J., 2003. A review of classic Fentons peroxidation as
an advanced oxidation technique. J. Hazard. Mater. B 98, 3350.
Njau, K.N., Minja, R.J.A., Katima, J.H.Y., 2003. Pumice soil: a potential
wetland substrate for treatment of domestic wastewater. Water Sci.
Technol. 48, 8592.
Park, J.S., Choi, H., Ahn, K.H., Kang, J.W., 2004. Removal mechanism
of natural organic matter and organic acid by ozone in the presence of
goethite. Ozone-Sci. Eng. 26, 141151.
 M. Kitis, S.S. Kaplan / Chemosphere 68 (2007) 18461853
Pignatello, J.J., Oliveros, E., MacKay, A., 2006. Advanced oxidation
processes for organic contaminant destruction based on the Fenton
reaction and related chemistry. Crit. Rev. Environ. Sci. Technol. 36, 1
84.
Rachel, A., Lavedrine, B., Subrahmanyam, M., Boule, P., 2002. Use of
porous lavas as supports of photocatalysts. Catal. Commun. 3, 165171.
Rao, K.V.S., Rachel, A., Subrahmanyam, M., Boule, P., 2003. Immobi-
lization of TiO2 on pumice stone for the photocatalytic degradation of
dyes and dye industry pollutants. Appl. Catal. B: Environ. 46, 7785.
Ravikumar, J.X., Gurol, M.D., 1994. Chemical oxidation of chlorinated
organics by hydrogen peroxide in the presence of sand. Environ. Sci.
Technol. 28, 394400.
Summers, R.S., 1986. Activated Carbon Adsorption of Humic Substances:
Eect of Molecular Size and Heterodispersity. PhD thesis, Stanford
University, USA.
1853
Tipping, E., 1981. The adsorption of aquatic humic substances by iron
oxides. Geochim. Cosmochim. Acta 45, 191199.
Tyre, B.W., Watts, R.J., Miller, G.C., 1991. Treatment of four biorefrac-
tory contaminants in soils using catalyzed hydrogen peroxide. J.
Environ. Qual. 20, 832838.
Vogelpohl, A., Kim, S.M., 2004. Advanced oxidation processes (AOPs) in
wastewater treatment. J. Ind. Eng. Chem. 10, 3340.
Wang, G.S., Liao, C.H., Wu, F.J., 2001. Photodegradation of humic
acids in the presence of hydrogen peroxide. Chemosphere 42, 379
387.
Watts, R.J., Udell, M.D., Monsen, R.M., 1993. Use of iron minerals in
optimizing the peroxide treatment of contaminated soils. Water
Environ. Res. 65, 839844.