Application of Fenton-based reactions for treating dye wastewaters: Stability of sulfonated azo dyes in the presence of iron(III)

Article abstract of DOI:10.1002/1522-2675(20010919)84:9<2589::AID-HLCA2589>3.0.CO;2-C

Fenton- and photo-assisted Fenton advanced oxidation processes generate reactive oxygen species from hydrogen peroxide and are candidates for the remediation of dye wastewaters. The purpose of this study was to investigate interactions of iron (III) with

Full text of DOI:10.1002/1522-2675(20010919)84:9<2589::AID-HLCA2589>3.0.CO;2-C

Helvetica Chimica Acta ± Vol. 84 (2001)
2589
Application of Fenton-Based Reactions for Treating Dye Wastewaters:
Stability of Sulfonated Azo Dyes in the Presence of Iron(III)
by Allison A. MacKay1) and Joseph J. Pignatello*
The Connecticut Agricultural Experiment Station, 123 Huntington Street, Box 1106, New Haven, CT 06504, USA
Â
Dedicated to Professor Andre M. Braun on the occasion of his 60th birthday
Fenton- and photo-assisted Fenton advanced oxidation processes generate reactive oxygen species from
hydrogen peroxide and are candidates for the remediation of dye wastewaters. The purpose of this study was to
investigate interactions of iron (III) with hydroxyazo dyes. The o-hydroxyazo dyes Acid Orange 7 (AO7; 4-[(2-
hydroxynaphthalen-1-yl)azo]benzenesulfonic acid sodium salt) and Acid Orange 10 (AO10; 7-hydroxy-8-
(phenylazo)naphthalene-1,3-disulfonic acid disodium salt) represent dyes allegedly able to chelate FeIII through
the chromophore. The p-hydroxyazo dye Acid Orange 20 (AO20; 4-[(4-hydroxynaphthalen-1-yl)azo]benzene-
sulfonic acid sodium salt) represents an analogous structure that is unable to chelate FeIII due to the position of
the OH group. Reactions were carried out at pH 2 ± 3 in perchlorate or chloride media in the absence of
peroxide. No evidence was found by UV/VIS spectroscopy for complexation of FeIII by the o-hydroxyazo
chromophore. Instead, FeIII apparently coordinated or formed an ion pair with the sulfonate group, and, when
only one sulfonate group was present (i.e., AO7), the dye formed a co-precipitate with iron(III) hydrous oxides
and perchlorate ion. Dye precipitation was seeded by colloidal iron hydrolysis product nuclei. By contrast, the p-
hydroxyazo dye (AO20) was rapidly oxidized by iron(III). The net Fe2‡/oxidized AO20 ratio was 2 :1, and a
minor yield of 1,4-naphthoquinone was obtained. The major initial oxidation product, which was not identified,
formed a reversible complex with Fe2‡. Results of this study indicate that the effectiveness of Fenton-based
methods for treating certain azo dyes that form insoluble ferric salts may be compromised by removal of the
catalyst from solution. However, the degradation of certain other azo dyes might be assisted by direct thermal
oxidation by iron(III).
Introduction. ± Fenton-based advanced oxidation processes (AOPs) seem to be
well-suited for treating wastes from dye manufacture and textile- and paper-dyeing
operations. An important class of dyes for these applications are acid dyes [1]. Acid
dyes comprise a large group of relatively low molecular mass, mainly monoazo
compounds that bear up to three sulfonic acid groups [2]. Wastewater is produced from
dying operations at a rate of 100 to 130 liters per kilogram of dyestuff applied [3].
Conventional wastewater treatment methods are ineffective for removing color from
these wastestreams. For example, highly soluble acid dyes cannot be destabilized by
coagulation, and are not efficiently removed by carbon adsorption [4]. And, since
dyestuffs are chosen in part for their resistance to microbial attack, biological treatment
processes are often ineffective, unless wastewaters are pre-treated with strong oxidants
or reductants [5][6]. Fenton and photo-assisted AOPs involving iron and hydrogen
1
)
Present address: University of Connecticut, Department of Civil and Environmental Engineering, 261
Glenbrook Rd., Unit 2037, Storrs, CT 06269-2037 (phone: 860-486-2450; fax: 860-486-2298; e-mail:
mackaya@engr.uconn.edu). (phone: (203)974-8518; fax: (203)974-8502; e-mail: joseph.pignatello@po.-
state.ct.us)

2590
Helvetica Chimica Acta ± Vol. 84 (2001)
.
peroxide enable dye degradation by OH radicals in homogeneous solution and have
the potential to achieve deep oxidation of the dye to carbon dioxide and other nontoxic
by-products.
The efficacy of various Fenton-based AOPs has been demonstrated with simulated
and actual acid dye wastewaters. Kuo [7] first demonstrated that acid dye solutions
were decolorized by 99% with up to 83% removal of chemical oxidation demand by
reaction with 3.3 mm ferrous sulfate (FeSO4) and 26 mm hydrogen peroxide (H2O2) at
pH 3.5 for 5 h at 508. Tang and Chen[8] compared dye removal using Fentonꢀs reagent
(Fe2‡/H2O2) with H2O2/Fe powder. Greater dye-removal efficiencies were obtained
with the Fe-powder system, but the increased removal was attributed to dye adsorption
on the Fe powder, not faster transformation reactions. Decolorization of acid dye
solutions by Fentonꢀs reagent was increased with the addition of UV light [9]. Complete
decolorization of Acid Orange 7 (AO7; Fig. 1) was observed within 2 h, and up to 95%
mineralization occurred in less than 8 h [9]. Comparable dye transformation rates were
achieved when iron was immobilized in Nafion membranes or alginate beads in the
presence of H2O2 and light [11][12]. In general, mineralization of acid dyes by Fenton-
based AOPs is significantly slower than transformation of the primary dye chromo-
phore; however, integration of Fenton-based AOPs with other wastewater-treatment
methods improves the overall removal of chemical and biological oxidation demand
[13][14].
Fig. 1. Structures of the sulfonated azo dyes shown in the hydrazone form, the dominant tautomer in polar
solvents [10]
Few studies have investigated the mechanisms of dye degradation by Fenton-based
AOPs. Spadaro et al. [15] followed the appearance of degradation products from the
oxidation of substituted (phenylazo)benzene and (phenylazo)naphthalene dyes by
Fe3‡ and H2O2 at pH 2.8. They suggested that degradation of such monoazo dyes is
initiated by hydroxyl-radical attack on the azo group and results in cleavage of the azo
bond from one of the aromatic moieties and formation of an unstable aryldiazene
intermediate. The products depend upon the structure of the dye and upon which side
.
of the azo group the OH radical attacks. Similar mechanisms have been proposed for
the oxidation of monoazo dyes by non-Fenton-based AOPs [16][17]. Naphthoquinone,
arene sulfonic acid and carboxylic acid products have been identified from the
oxidation of azo dyes having naphthalenyl and benzyl moieties [15][16][17].
The mechanisms by which hydroxyazo dyes (e.g., Fig. 1) are degraded with Fenton-
based AOPs are not well characterized. Hydroxy groups are common substituents of
acid dyes [18]. It is conceivable that a dye compound with the OH group located ortho
to the azo group could chelate the metal ion at the chromophoric site, as in A or B.
Nadtochenko and Kiwi [19] attributed decreases in AO7 absorbance in solutions

Helvetica Chimica Acta ± Vol. 84 (2001)
2591
containing only the dyestuff and iron(III) to the formation of a 1:1 dye/Fe3‡ complex
through the chromophoric group as in B. They proposed that the initiating degradation
step under photo-Fenton reaction conditions (UV light, Fe3‡, H2O2) is ligand-to-metal
electron transfer within the excited state.
Other evidence suggests that iron complexation of the chromophore may not be
important for o-hydroxyazo dyes. Metal-ion complexes of monohydroxyazo dyes are
known to have low stabilities in acidic dye applications [20]. Drew and Fairbairn [21]
found that when sulfonate groups are present in monohydroxyazo dyes, coordination or
ion pairing with the sulfonate group takes place in preference to complexation with the
chromophore. Stable iron(III) complexes have been reported only for o,o'-dihydrox-
yazo dyes [22]. For these compounds, distinct hypsochromic shifts and peak broadening
in the dye absorbance spectra were observed on addition of Fe3‡. Thus it appears that
iron salts may be formed when Fenton-based AOPs are applied to acid dye wastewaters.
Under certain conditions, dye oxidation proceeds [9]; however, if the solubility of the
dye salts is low, the iron catalyst may be removed from solution and oxidation slowed or
halted.
The purpose of this study was to determine the stability of hydroxyazo acid dyes in
the presence of ferric ions (Fe3‡). To test whether dyes from this class formed soluble
iron(III) complexes or insoluble iron(III) salts, we chose three monoazo dyes with
related structures. Their structures are given in Fig. 1 in the dominant hydrazone
tautomeric form.
Experimental. ± Chemicals. The sulfonated azo dyes used in this study were Acid Orange 7 (AO7; 4-[(2-
hydroxynaphthalen-1-)azo]benzenesulfonic acid sodium salt; >85% pure; Aldrich, Milwaukee, WI), Acid
Orange 20 (AO20; 4-[(4-hydroxynaphthalen-1-yl)azo]benzenesulfonic acid sodium salt; >85% pure; Fluka,
Switzerland), and Acid Orange 10 (AO10; 7-hydroxy-8-(phenylazo)naphthalene-1,3-disulfonic acid disodium
salt; >85% pure; National Aniline and Chemical Co.). Iron(III) perchlorate, sodium perchlorate, benzene-
sulfonic acid sodium salt, 4-hydroxybenzenesulfonic acid sodium salt dihydrate, and FerroZine ꢁ iron reagent
(4,4'-[3-(pyridin-2-yl)-1,2,4-triazine-5,6-diyl]bis[benzenesulfonic acid]monosodium salt hydrate) were obtained
from Aldrich (Milwaukee, WI). Iron(III) chloride hexahydrate, NaCl, NaOH, EtOH, AcOEt, MeCN, and
alumina (80 ± 200 mesh) were from Fisher Scientific (Fairlawn, NJ), phosphoric, hydrochloric, and perchloric
acids from J. T. Baker (Phillipsburg, NJ), and ethylenediaminetetraacetic acid disodium salt (Na2EDTA) and
potassium dihydrogen phosphate from Mallinckrodt (Paris, KY). All compounds were used as received, except
AO7 and AO20 which were purified as described below. All solns. were prepared with high-purity (18 MW´ cm)
distilled, deionized H2O (Barnstead NANOpure system). Glassware was soap-and-water washed followed by
rinsing with high-purity water.
Dye Purification. AO7 was purified by recrystallization from a minimal amount of H2O. The hot soln. was
filtered through #41 ashless filter paper (Whatman) in a heated glass funnel. The filtrate was cooled slightly and
extracted with AcOEt to remove a hydrophobic impurity. The aq. phase was cooled overnight at 48. The crystals
were collected on #41 filter paper, rinsed with H2O, and vacuum dried. HPLC: only 1 peak on detection at
254 nm (general arom. compounds) and 478 nm (AO7 maxima). Anal. (Galbraith Laboratories, Knoxville, TN)
calc. for C16H12N2NaO4S ´ 0.5 H2O: C 53.5, H 3.2, N 7.8; found: C 53.5, H 3.35, N 7.76.

2592
Helvetica Chimica Acta ± Vol. 84 (2001)
AO20 was purified by column chromatography (alumina). The dye was dissolved in MeOH and coated on
alumina by slow evaporation of the solvent. The dye-coated alumina was seated on a column (49 cm  3 cm i.d.)
of fully-activated alumina. Elution at 1.5 ml/min with EtOH (200 ml), EtOH/H2O 8 :2 (v/v; 500 ml), EtOH/H2O
6 :4 (250 ml), and EtOH/H2O 1:1 (250 ml) gave AO20 fractions which were identified by HPLC, evaporated
and freeze-dried. Anal. (Perkin-Elmer 2400 CHN elemental analyzer) calc. for C16H12N2NaO4S ´ 4 H2O: C 45.4,
H 2.83, N 6.62; found: C 45.5, H 3.21, N 6.15.
Dye Stability. Stock solns. containing ferric perchlorate (or chloride) and perchloric (or hydrochloric) acid
were diluted to the target iron concentration and pH. Sodium perchlorate (or chloride) was added to adjust the
ionic strength. At time zero, a small aliquot of concentrated dye soln. was added to the iron soln. At each
sampling time, the soln. was centrifuged for 20 min at 340 g, and several milliliters were removed from the top.
Absorbance measurements (Hewlett-Packard 8452A diode-array spectrophotometer) were made in 1-cm quartz
cuvettes at the absorbance maximum for the dye (486 nm for AO7, 476 nm for AO20, and 478 nm for AO10).
Replicate measurements were made on identically prepared solutions, 1) containing dye only with no iron, and
2) containing iron only with no dye. The effect of colloidal polynuclear iron hydrolysis species on AO7 stability
was investigated by adjusting the soln. pH to 2.7 with NaOH instead of by dilution [23]. At the end of the
experiment, Na2EDTA was added to solns. to enable mass-balance determination of the dye compound. All
sample manipulations for these and following experiments were undertaken in a darkened room, and all sample
vessels were foil-covered to minimize possible photochemical reactions.
Acid Orange 20 Oxidation. In the dark, an aliquot of a concentrated AO20 soln. was added to a large-
volume iron(III) perchlorate soln. to achieve nominal concentrations of 0.071 mm AO20, 0.125 mm FeIII, and
23 mm NaClO4 at pH 2 (to prevent formation of polynuclear Fe species). Over time, aliquots were removed to
monitor dye and product concentrations. AO20 and ionic org. products were quantified by ion-pair
chromatography. Separations were achieved on a C18 column (ODS-5, 250 mm, 4.6 mm i.d., 5 mm packing;
Alltech) in an Agilent-1100 instrument equipped with a diode-array detector gradient elution (modified from
that of Alonso [24]): 100% of B for 5 min ! B/A 4 :6 within 13 min; B ˆ 10 mm phosphate and 5 mm (Bu4N)OH
in MeOH/H2O 36 :63 (v/v) at pH 6.5, A ˆ MeOH. Compounds were detected by monitoring at 254 nm (general
arom. compounds), 272 nm (substituted benzene sulfonates), and 476 nm (AO20) with bandwidths of 2 nm. A
reference wavelength of 700 nm (10-nm bandwidth) was used. To prevent column contamination and
chromatographic interference, dissolved iron was removed by addition of NaOH, filtration through 0.2-mm
Teflon filters (Gelman Acrodisk) and re-acidification to pH 2 with perchloric acid. The limit of detection for
AO20 was 0.001 mm by this method.
Non-ionic products were determined by CH2Cl2 extraction of aliquots of the reaction mixture. Extracts
were dried (Na2SO4) and analyzed by GLC/MS (Hewlett Packard 5890GC/5970MSD). Upon product
identification, subsequent quantifications were performed with flame-ionization detection. Both methods
utilized a PTE-5 capillary column (30 m, 0.25-mm i.d., 0.25 mm film thickness; Supelco, Bellefont, PA), and a
temp. program beginning at 608 and ramping at 58 per min to 2508, with a final isotherm at 2508 for 10 min.
Injector and detector temp. were held at 2508.
Fe2‡ Concentrations were quantified by complexation with FerroZine ꢁ reagent [25] at pH 2. Absorbance
measurements were made at 562 nm to minimize interference from unreacted dye; however, background
corrections were still applied by subtracting the absorbance (562 nm) of a duplicate sample volume to which an
equivalent volume of pH 2 perchloric acid had been added in place of the FerroZine ꢁ reagent. The detection
limit was 0.002 mm Fe2‡
.
Results and Discussion. ± Reaction of FeIII Ion with o-Hydroxyazo Dyes AO7 and
AO10. Chelation of Fe3‡ by the o-hydroxyazo moiety is expected to affect the VIS
absorption band of the dye. Consistent with a recent report [19], the VIS absorbance at
486 nm of AO7 solutions decreased in the presence of excess FeIII (Fig. 2). However,
the changes were atypical for a chelation reaction in homogeneous solution; they often
occurred over long time periods (hours), and application of the Benessi-Hildebrand
method [26] or the method of variations [27] to the absorbance data to obtain a stability
constant gave nonlinear and/or non-reproducible correlations, and so failed to provide
a reliable value. Instead, we believe the spectral changes are due to formation of an
insoluble iron salt of the dye, although the appearance of a precipitate was not always

Helvetica Chimica Acta ± Vol. 84 (2001)
2593
evident to the eye. Presumably the iron is outer- or inner-sphere associated with the
sulfonate group (Eqn. 1).
dye SO3 ‡ FeIII ! FeIII ´ dye SO3
(1)
(#)
Fig. 2. Stability of Acid Orange 7 absorbance at 486 nm in the presence of iron(III) with varying counter ion and
pH. The absorbances of iron-only controls (D) were less than 0.008 and indicated minimal polynuclear iron
hydrolysis species in solution.
Several experiments confirmed precipitation rather than chelation by the chromo-
phoric group. First, the spectral changes depended markedly on the anion of the
background electrolyte. The decrease in dye absorbance at 486 nm was faster and more
extensive in NaClO4 than in NaCl media (Fig. 2,a vs. b). Since ClO4 and Cl have little
tendency to coordinate Fe3‡ ± complexation of Fe3‡ by ClO4 is negligible, while that by
Cl is less than 10% of total iron at these concentrations (MINEQL ‡ ; Environmental
Research Software) ± they are expected to have little effect on the spectrum of the
putative FeIII ´ AO7 complex or on the position of equilibrium at identical pH and ionic
strength. We interpret the dependence of spectral changes on the anion to mean that
FeIII ´ AO7 salt is less soluble in ClO4 solutions. Interestingly, the solubility of AO7
monosodium salt is lower in 0.1m NaClO4 than in 0.1m NaCl.
Secondly, the spectral changes depended on pH and on the way the pH was
adjusted, again supporting precipitation over chelation. In the most likely chelate
structure (A), acid-base equilibria of the dye itself are not involved in chelation.
Therefore, formation of the FeIII ´ dye complex should be favored as the pH declines
‡
due to increasing abundance of the hexaaquo species (Fe3‡ ‡ H2O > Fe(OH)2‡ ‡ H ;
pKa ˆ 2.45 at ionic strength of 30 mm [28]), which should more readily coordinate the
dye than the monohydroxy or higher hydrolysis species. However, the opposite was
observed (Fig. 2,b vs. c); the AO7 absorbance at 486 nm decreased much more at
pH 2.6 ([Fe3‡]/[Fe(OH)2‡] ˆ 0.7) than at pH 1.7 ([Fe3‡]/[Fe(OH)2‡] ˆ 5.7). Further
experimentation with 0.06 mm AO7 and 1 mm Fe3‡ revealed that the decrease in AO7
absorbance was sensitive to the presence of colloidal polynuclear iron hydrolysis
species, suggesting that the solutions were oversaturated with respect to FeIII
dye SO3 salt and that precipitation was nucleated by the colloids.
´

2594
Helvetica Chimica Acta ± Vol. 84 (2001)
To test whether this was true, we compared FeIII solutions at the same pH (2.7 ± 2.8)
adjusted by two different methods: a) diluting acidified concentrated stock solutions
(10 mm FeIII, pH 1.7) to the target pH with purified H2O, and b) neutralizing solutions
(1 mm FeIII, pH 1.7) to the target pH dropwise with NaOH. Addition of strong base
causes locally high pH at the interface between the drop and the bulk solution, forming
iron hydrous oxide colloids that do not redissolve readily [23]. The spectra of dye-free,
NaOH-adjusted FeIII solutions tailed off much further into the VIS region due to
scattering by colloids than did spectra of the same that were pH-adjusted by dilution
(Fig. 3), confirming that some iron hydrolysis species were formed by base
neutralization. When a small volume of concentrated AO7 stock solution was added
to the NaOH-adjusted FeIII solution, an immediate decrease in dye absorbance
occurred; whereas dye added to Fe3‡ solutions prepared by dilution exhibited no initial
decrease in dye absorbance relative to an Fe-free control (Fig. 3). Elemental analysis of
the orange-colored precipitate formed in a preparative-scale reaction (1 mm FeIII, 1 mm
AO7, pH 2.8, 23 mm total perchlorate) gave, after rinsing with electrolyte and vacuum
drying, a compound of composition C16H20Cl0.8Fe1.4N2 when normalized to the N
Fig. 3. Effect of iron colloids on the absorbance of Acid Orange 7. Iron was added as FeCl3 and the background
electrolyte was NaCl ([Cl]tot ˆ 24 mm). FeIII solutions without dye and pH-adjusted by NaOH addition
(± ´ ´ ±); FeIII solutions without dye and pH-adjusted by dilution ( ± ´ ±); AO7 added to FeIII solutions pH-
adjusted by NaOH (± ±); AO7 added to FeIII solutions pH-adjusted by dilution ( ´´´ ); dye only (Ð).

Helvetica Chimica Acta ± Vol. 84 (2001)
2595
content (38.4% C, 3.96% H, 5.44% N, 14.9% Fe, and 2.5% Cl). The Fe/dye ratio of 1.4
obviously indicates a co-precipitation of iron hydrous oxides with iron dye species as
opposed to adsorption of dye molecules on an iron oxide surface. H-Content in excess
of the stoichiometric amount expected for AO7 likely reflects the incorporation of
waters of hydration and hydroxide ions for charge balance.
Experiments conducted on precipitated reaction mixtures verified that AO7 had
not been transformed. The concentration of dissolved Fe2‡ was less than the detection
limit of 0.002 mm (<0.2% yield). Addition of EDTA with mixing for one to several
days resulted in re-solubilization of the precipitate and regeneration of more than 98%
of the initial AO7, as determined by HPLC.
Lastly, evidence against FeIII complexation of the o-hydroxyazo chromophore was
obtained in experiments with AO10, which is identical to AO7, except that it has two
sulfonate groups. The VIS spectrum at wavelengths above 425 nm of 0.06 mm AO10 in
a perchlorate solution at pH 2.8 was unchanged in the presence of 1 mm FeIII (Fig. 4).
This indicates no complexation of iron with the chromophore. The enhanced solubility
of the FeIII ´ AO10 sulfonate salt over the FeIII ´ AO7 salt is due possibly to the extra
sulfonate group in the AO10 molecule.
Nadtochenko and Kiwi [19] postulated that AO7 forms a 1:1 complex with FeIII
ions in solution. We prepared solutions to replicate the experimental conditions of
these authorsꢀ study. An aliquot of concentrated AO7 solution was added to a pH 2.8
Fig. 4. Absorbance spectrum of Acid Orange 10 in the absence of presence of iron(III) ions. Solutions were
prepared at pH 2.8 with a background electrolyte of NaClO4 (23 mm). Iron was added as the iron(III)
perchlorate salt.

2596
Helvetica Chimica Acta ± Vol. 84 (2001)
FeCl3 solution to give nominal solution concentrations of 2.9 mm AO7 and 0.9 mm FeIII.
A flocculent precipitate formed immediately upon mixing. Nadtochenko and Kiwi did
observe precipitate formation in their solutions, but only after increasing the FeCl3
concentration to 4 mm.
Reaction of FeIII Ion with p-Hydroxyazo Dye AO20. Chelation of FeIII by the
chromophore of AO20 is not possible since the OH group is located in the para
position; hence, no change in the VIS spectrum of AO20 was expected in the presence
of FeIII ions. However, when AO20 was added to solutions containing an excess of FeIII
at pH 2 or 3, a rapid color change from orange to pink occurred, the intensity of the dye
band in the VIS region decreased, and the lmax shifted from 476 to 486 nm (Fig. 5). The
decrease in AO20 concentration was accompanied by the appearance of reduced iron
in solution (Fig. 6). EDTA failed to restore the spectrum corresponding to the lost
AO20, confirming that the dyestuff was degraded and not complexed. Furthermore,
AO20 degradation was inhibited by prior complexation of FeIII with EDTA. Together,
these results indicate that AO20 is easily oxidized by FeIII. Interestingly, p-hydroxyazo
dyes are more readily oxidized by oxo(porphyrinato)iron [29] and horseradish
peroxidase [30] than are o-hydroxyazo dyes, possibly due to intramolecular H-bonding
in the latter that may inhibit removal of an electron. The only by-product identified in
our experiments by liquid chromatography or by gas chromatography and MS or FID
detection was 1,4-naphthoquinone, produced in only 4% yield based on initial AO20.
The reaction stoichiometry was determined by following Fe2‡ spectrophotometri-
cally in aliquots of the reaction mixture after complexation with ferrozine. The typical
Fig. 5. Absorbance spectrum of Acid Orange 20 in the absence and presence of iron(III) ions. Solutions were
prepared at pH 2.8 with a background electrolyte of NaClO4 (23 mm). Iron was added as the iron(III)
perchlorate salt.

Helvetica Chimica Acta ± Vol. 84 (2001)
2597
Fig. 6. Oxidation of Acid Orange 20 with iron(III) perchlorate. Initial concentrations were 0.071 mm dye and
0.125 mm FeIII, at pH 2 and [ClO4]tot ˆ 23 mm. Fe2‡ Measurements were made 60 s after ferrozine addition to
sample aliquots and reflect an apparent 1:1 reaction stoichiometry (see text for discussion of reaction
stoichiometry).
development time for the ferrozine complex is 1 min, which also proved adequate for
control samples of Fe2‡ in the same solution matrix. However, in the samples, but not in
the controls, we found that after an immediate color change within 1 min, the color of
the Fe2‡ ´ ferrozine complex continued to develop slowly for an additional 30 min
(Fig. 7). A 1 min development time gave an apparent Fe2‡/oxidized dye ratio of 1.0
(e.g., Fig. 6); however, extending the development time to 30 min gave an Fe2‡/oxidized
dye ratio of 2.3. In several experiments similar to that in Fig. 6, the 30 min development
Fig. 7. Formation of Fe2‡ ´ ferrozine complex in N2-purged solutions. The reaction was initiated with 0.095 mm
iron(III) perchlorate and 0.0122 mm AO20 at pH 2. Solution aliquots were removed after reaction times of 3, 5,
and 30 min and mixed with ferrozine reagent. Complex formation is reported as reacted Fe2‡/dye stoichiometry
calculated for 100% dye oxidation.

2598
Helvetica Chimica Acta ± Vol. 84 (2001)
time gave an average 2 :1 stoichiometry for all dye-plus-FeIII reaction times of 1 min or
longer.
Complete disappearance of AO20 was observed only when FeIII was present
initially in at least two-fold molar excess over the dye. A dye oxidation experiment with
initial concentrations of 0.125 mm FeIII and 0.043 mm AO20 gave 0.085 mm Fe2‡
(30 min development time), and no dye compound was detectable by HPLC. Whereas,
experiments with a stoichiometric deficiency of FeIII had unreacted dye remaining in
solution. For example, in the experiment of Fig. 6, in which 1.48 mol of FeIII were added
per mol of dye, 97 ± 98% of FeIII was ultimately reduced, and 27% of the initial AO20
remained (expected: 14% based on 2 :1 stoichiometry). The unreacted AO20 in
aliquots treated with ferrozine reagent did not further react during the extended color
development; thus, the second equiv. of Fe2‡ produced is not coupled to further AO20
oxidation.
The proposed mechanism for AO20 oxidation by FeIII is shown in the Scheme. The
reaction is initiated by one-electron removal from the azo group yielding a dye radical
cation and the first equivalent of Fe2‡. A plausible minor pathway (Pathway 1) is
hydrolysis of the dye radical cation to give 1,4-naphthoquinone (4%) and an unstable
(sulfophenyl)hydraziniumyl radical. The hydraziniumyl radical is oxidized rapidly by
O2 or FeIII to give 4-hydroxybenzenesulfonic acid [31][32]. Due to chromatographic
interference, 4-hydroxybenzenesulfonic acid in yields comparable to 1,4-naphthoqui-
none could not be resolved by HPLC. If FeIII and not O2 is the oxidant in this step, the
Fe2‡ so generated would be very small relative to the total Fe2‡ produced in the
reaction.
Scheme. Postulated Mechanism for Oxidation of Acid Orange 20. FerroZineꢁ was added to solutions as a
colorimetric reagent for the Fe2‡ assay.

Helvetica Chimica Acta ± Vol. 84 (2001)
2599
The major route (Pathway 2) involves a second one-electron oxidation by FeIII to
produce a species that strongly complexes the reduced Fe2‡. This complex is formed
reversibly and exchanges slowly with ferrozine when exposed to excess ferrozine in the
assay. The exchange gives rise to the prolonged color development and accounts for the
observed second equivalent of Fe2‡. If an oxidation by-product were binding a portion
of Fe2‡, only the unbound species would be quantified in the 1 min ferrozine assay. We
have not identified the species responsible for complexation of Fe2‡ in Pathway 2.
An alternative explanation for the 2nd equiv. of Fe2‡ is reaction of the AO20 radical
cation in step 1 with O2 to produce H2O2, which then slowly reduces FeIII to Fe2‡ in the
presence of the ferrozine ligand (Eqn. 2). Reduction of FeIII by H2O2 in the presence of
the ferrozine reagent can be observed in control solutions. However, the pathway in
Eqn. 2 can be ruled out because FeIII reaction with AO20 in N2-purged compared to O2-
saturated solutions gave the same Fe2‡/oxidized dye stoichiometry and the same rate of
slow color development in the ferrozine assay.
.
.
‡
‡
AO20 ‡ O2 ‡ H ! AO202‡ ‡ HO2
(2a)
(2b)
(2c)
.
HO2 ! 1/2 H2O2 ‡ 1/2 O2
ferrozine ‡ FeIII ‡ 1/2 H2O2 ! Fe2‡ ´ ferrozine ‡ 1/2 O2 ‡ H
‡
Relevance to Dye Waste Treatment Technology. The results of this research point out
possible limitations or possible benefits of Fenton-based AOPs for treating acid dye
wastewaters depending on the dye. Typical wastestreams contain ca. 1 gram dyestuff
per liter of solution [3]. Such concentrations correspond to ca. 3 mm for the acid dyes
considered in our study. Fenton treatment would be initiated by adding H2O2 and
catalytic amounts (typically, 0.1 ± 1 mm) of iron. At a dye/FeIII ratio of 3 :1, we observed
immediate precipitation of AO7. Acid dyes that form poorly soluble FeIII salts could
remove the iron catalyst from solution and slow down or stop the oxidation process.
Furthermore, light scattering by particles could reduce the efficiency of photo-Fenton
treatments. Certainly, dye removal by Fenton-based AOPs has been achieved for a few
acid dyes tested [7][9]; however, no comprehensive study has been conducted to assess
whether azo-dye stability toward precipitation in the presence of FeIII affects reaction
rates. On the other hand, certain acid dyes may accelerate Fenton-degradation rates.
The key step in hydroxyl-radical generation is the reduction of H2O2 by Fe2‡ [33]. Azo
dyes such as AO20 are partially oxidized by FeIII and also provide reducing equivalents
for the Fenton reaction.
REFERENCES
[1] P. L. Short, Chem. Eng. News 2000, 78(2), 21.
[2] H. Zollinger, ꢂColor Chemistryꢀ, VCH Publishers, New York, 1987.
[3] USEPA, ꢂProfile of the Textile Industryꢀ, Washington, D.C., 1997, EPA/310-R-97-009.
[4] A. Reife, H. S. Freeman, in ꢂEnvironmental Chemistry of Dyes and Pigmentsꢀ, Eds. A. Reife and H. S.
Freeman, Wiley, New York, 1996.
[5] S. F. Dubrow, G. D. Boardman, D. L. Michelson, in ꢂEnvironmental Chemistry of Dyes and Pigmentsꢀ, Eds.
A. Reife and H. S. Freeman, Wiley, New York, 1996.

2600
Helvetica Chimica Acta ± Vol. 84 (2001)
[6] M. Cooke, in ꢂEnvironmental Chemistry of Dyes and Pigmentsꢀ, Eds. A. Reife and H. S. Freeman, Wiley,
New York, 1996.
[7] W. G. Kuo, Water Res. 1992, 26, 881.
[8] W. Z. Tang, R. Z. Chen, Chemosphere 1996, 32, 947.
[9] J. Bandara, C. Morrison, J. Kiwi, C. Pulgarin, P. Peringer, J. Photochem. Photobiol., A 1996, 99, 57.
[10] C. V. Stead, in ꢂThe Chemistry of Synthetic Dyesꢀ, Ed. K. Venkataraman, Academic, New York, 1970.
[11] J. Fernandez, M. R. Dhananjeyan, J. Kiwi, Y. Senuma, J. Hilborn, J. Phys. Chem. B 2000, 104, 5298.
[12] J. Fernandez, J. Bandara, A. Lopez, P. Albers, J. Kiwi, Chem. Commun. 1998, 14, 1493.
[13] S. H. Lin, M. L. Chen, Desalination 1997, 109, 121.
[14] S.-F. Kang, H.-M. Chang, Water Sci. Technol. 1997, 36, 215.
[15] J. T. Spadaro, L. Isabelle, V. Renganathan, Environ. Sci. Technol. 1994, 28, 1389.
[16] C. Nasr, K. Vinodgopal, L. Fisher, S. Hotchandani, A. K. Chattopadhyat, P. V. Kamat, J. Phys. Chem. 1996,
100, 8436.
[17] K. Vinodgopal, D. E. Wynkoop, P. V. Kamat, Environ. Sci. Technol. 1996, 30, 1660.
[18] ꢂColour Indexꢀ, 3rd edn., The Society of Dyers and Colourists, Bradford, Yorkshire, England, 1971.
[19] V. Nadtochenko, J. Kiwi, J. Chem. Soc., Faraday Trans. 1997, 93, 2373.
[20] R. Price, in ꢂThe Chemistry of Synthetic Dyesꢀ, Ed. K. Venkataraman, Academic, New York, 1970, Vol. 3,
p. 303.
[21] H. D. K. Drew, R. E. Fairbairn, J. Chem. Soc. 1939, 141, 823.
[22] J. Sokolowska-Gajda, H. S. Freeman, A. Reife, Textile Res. J. 1994, 64, 388.
[23] R. J. Knight, R. N. Sylva, J. Inorg. Nucl. Chem. 1975, 37, 779.
[24] M. C. Alonso, M. Castillo, D. Barcelo, Anal. Chem. 1999, 71, 2586.
[25] L. L. Stookey, Anal. Chem. 1970, 42, 779.
[26] R. S. Mulliken, W. B. Person, ꢂMolecular Complexesꢀ, Wiley, New York, 1969.
[27] A. E. Martell, M. Calvin, ꢂChemistry of the Metal Chelate Compoundsꢀ, Prentice-Hall, Englewood Cliffs,
NJ, 1952.
[28] R. M. Milburn, W. C. Vosburgh, J. Am. Chem. Soc. 1955, 77, 1352.
[29] G. R. Hodges, J. R. Lindsay Smith, J. Oakes, J. Chem. Soc., Perkin Trans. 2 1998, 617.
[30] M. B. Pasti-Grigsby, A. Paszczynski, S. Goszczynski, D. L. Crawford, R. L. Crawford, Appl. Environ.
Microbiol. 1992, 58, 3605.
[31] P.-K. C. Huang, E. M. Kosower, J. Am. Chem. Soc. 1968, 90, 2367.
[32] J. Nicholson, S. G. Cohen, J. Am. Chem. Soc. 1966, 88, 2247.
[33] C. Walling, Acc. Chem. Res. 1975, 8, 125.
Received May 29, 2001