Regeneration of dye-saturated quaternized cellulose by bisulfite- mediated borohydride reduction of dye azo groups: An improved process for decolorization of textile wastewaters

Article abstract of DOI:10.1021/es970395v

Cellulosics modified to contain quaternary ammonium groups have a strong affinity for anionic dyes. Therefore, ion exchangers based on quaternized cellulose or ligno-cellulose can be used to remove textile dyes from wastewater. However, restoration of exchanger binding capacity is poor using conventional, low-cost regenerants. Experiments were conducted with two monoazo dyes, Orange II (Acid Orange 7) and Remazol Red F3B (Reactive Red 180), to determine whether reductive cleavage of dye azo bonds improves exchanger regenerability. Treatment with the redox couple KBH₄/NaHSO₃ fully restored the binding capacity of Orange II-saturated quaternized cellulose. KBH₄/NaHSO₃ treatment of quaternized cellulose saturated with Remazol Red F3B (hydrolyzed, unreactive form) restored 74% of the exchanger binding capacity, which increased to 95% with a subsequent wash with NaOH or NaClO₄. High-performance liquid chromatography was used to confirm that KBH₄/NaHSO₃ reductively cleaved dye azo bonds. Bisulfite was found to form a stable adduct with Orange II but to not cleave the dye's azo bond. The efficiency of dye azo bond reduction was the same for dye in solution and exchanger-bound dye. These results indicate that reduction of monoazo dyes is an efficient method by which to regenerate the dye binding capacity of quaternized cellulosics used to decolorize textile wastewater. Cellulosics modified to contain quaternary ammonium groups have a strong affinity for anionic dyes. Therefore, ion exchangers based on quaternized cellulose or ligno-cellulose can be used to remove textile dyes from wastewater. However, restoration of exchanger binding capacity is poor using conventional, low-cost regenerants. Experiments were conducted with two monoazo dyes, Orange II (Acid Orange 7) and Remazol Red F3B (Reactive Red 180), to determine whether reductive cleavage of dye azo bonds improves exchanger regenerability. Treatment with the redox couple KBH₄NaHSO₃ fully restored the binding capacity of Orange II-saturated quaternized cellulose. KBH₄NaHSO₃ treatment of quaternized cellulose saturated with Remazol Red F3B (hydrolyzed, unreactive form) restored 74% of the exchanger binding capacity, which increased to 95% with a subsequent wash with NaOH or NaClO₄. High-performance liquid chromatography was used to confirm that KBH₄/NaHSO₃ reductively cleaved dye azo bonds. Bisulfite was found to form a stable adduct with Orange II but to not cleave the dye's azo bond. The efficiency of dye azo bond reduction was the same for dye in solution and exchanger-bound dye. These results indicate that reduction of monoazo dyes is an efficient method by which to regenerate the dye binding capacity of quaternized cellulosics used to decolorize textile wastewater.

Full text of DOI:10.1021/es970395v

Environ. Sci. Technol. 1997, 31, 3647-3653
ties. Chemically modified biomass products can have very
Regeneration of Dye-Saturated
Quaternized Cellulose by
Bisulfite-Mediated Borohydride
Reduction of Dye Azo Groups: An
Improved Process for Decolorization
of Textile Wastewaters
high dye-binding capacities but incur a concomitant increase
in cost. Quaternized cellulose and quaternized lignocelluloses
have been shown to be particularly effective in removing
anionic dyes (3, 4). However, because of the high affinity of
dye for the quaternized substrates, regeneration of the
adsorbent’s anion exchange capacity is difficult using con-
ventional salt and base regenerants. The very high cost of
commercially available quaternized cellulose necessitates that
the adsorbent undergo regeneration. Regeneration of the
much less expensive quaternized lignocellulose may be less
necessary but could provide some extra cost savings.
The use of bisulfite-mediated borohydride reduction for
the decolorization of wastewaters containing azo dyes has
had some commercial success (5). Cook (5) describes this
process as consisting of, first, the reduction of bisulfite to
dithionite (hydrosulfite) by borohydride (eq 1).
J O S E P H A . L A S Z L O *
USDA-ARS, National Center for Agricultural Utilization
Research, Biomaterials Processing Research Unit, 1815 North
University Street, Peoria, Illinois 61604
BH4- + 8HSO3- + H+
8 4S2O42- + B(OH)3 + 5H2O
pH 5-8
Cellulosics modified to contain quaternary ammonium
groups have a strong affinity for anionic dyes. Therefore,
ion exchangers based on quaternized cellulose or ligno-
cellulose can be used to remove textile dyes from wastewater.
However, restoration of exchanger binding capacity is
poor using conventional, low-cost regenerants. Experiments
were conducted with two monoazo dyes, Orange II (Acid
Orange 7) and Remazol Red F3B (Reactive Red 180), to
determine whether reductive cleavage of dye azo bonds
improves exchanger regenerability. Treatment with the redox
couple KBH /NaHSO fully restored the binding capacity
(1)
Dithionite spontaneously dissociates to SO2 anion radicals
(eq 2)
S2O42- f 2SO2
(2)
•-
which in turn reduce the dye azo bond (eq 3).
4SO2•- + R1R2ArNdNArR3R4 + 4H2O + 2H+
f
4HSO3- + R1R2 ArNH3+ + R3R4ArNH3 (3)
+
4
3
of Orange II-saturated quaternized cellulose. KBH /NaHSO
4
3
treatment of quaternized cellulose saturated with Remazol
Red F3B (hydrolyzed, unreactive form) restored 74% of the
exchanger binding capacity, which increased to 95% with
Bisulfite is regenerated in the reduction step (eq 4) and so is
not consumed in the overall reaction scheme.
-
HSO3
BH4- + 2R1R2ArNdNArR3R4 + 3H2O + 5H+
8
a subsequent wash with NaOH or NaClO . High-performance
4
liquid chromatography was used to confirm that KBH /
4
B(OH)3 + 2R1R2 ArNH3+ + 2R3R4ArNH3 (4)
+
NaHSO reductively cleaved dye azo bonds. Bisulfite was
3
found to form a stable adduct with Orange II but to not
cleave the dye’s azo bond. The efficiency of dye azo bond
reduction was the same for dye in solution and exchanger-
bound dye. These results indicate that reduction of
monoazo dyes is an efficient method by which to regenerate
the dye binding capacity of quaternized cellulosics used
to decolorize textile wastewater.
However, the actual products generated from dyes treated
with the borohydride/ bisulfite redox couple have not been
reported, nor have the reduction stoichiometry and reaction
kinetics been detailed. Voyksner and co-workers found that
reduction by dithionite or tin(II) chloride produced the
anticipated aromatic amines from simple azo dyes, but azo
dyes containing reduction-sensitive substituents (e.g., -NO2,
-CO2CH3, and -CN) generated many products (6).
A potential drawback of using borohydride and bisulfite
is that the treatment of dilute dye solutions results in the
generation of large volumes of decolorized wastewater
contaminated by boron and potentially toxic aromatic amines.
In addition, side reactions of borohydride with water (eq 5),
and dithionite with oxygen (eq 6) results in a wasting of
reducing equivalents.
Introduction
The decolorization of textile wastewaters is a worldwide
problem to which many diverse technologies have been
applied. Methods for dye removal include physical adsorption
or flocculation, chemical destruction by oxidation or reduc-
tion, and metabolic conversion to uncolored products by
microorganisms (1). To gain wide acceptance, a treatment
method should decolorize or remove a broad range of dye
types at very low cost in an environmentally benign manner.
One or more of these features is lacking in commercially
available technologies.
Dye removal by adsorption to biomass, such as waste
agricultural byproducts, chitin, etc., has been the subject of
a considerable amount of research (2). Although this ap-
proach has the potential to provide a low-cost solution to the
problem, unmodified biomass has poor dye-binding proper-
BH4- + 3H2O + H+ f B(OH)3 + 4H2v
(5)
(6)
S2O42- + 1/ 2O2 f S2O5
2-
Therefore, this approach is more efficient when treating small
volumes of concentrated dyes.
In the present work, the hypothesis is tested that bisulfite-
mediated borohydride reduction of azo dyes can be efficiently
used to reductively cleave dyes bound to quaternized cellulose
and, by inference, quaternized lignocellulose. Furthermore,
it is demonstrated that the resulting dye fragments are more
readily removed from the exchanger than the intact dye, thus
* Phone: 309-681-6322; fax: 309-681-6686; e-mail: laszloja@
mail.ncaur.usda.gov.
S0013-936X(97)00395-7 This article not subject to U.S. Copyright.
Published 1997 by the Am erican Chem ical Society.
9
VOL. 31, NO. 12, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 6 4 7

allowing regeneration of the dye adsorbent using conventional
regenerants.
TABLE 1. Effect of Regeneration Treatments on Dye-Binding
Capacity
Experimental Section
Orange II capacity
(mequiv/g)
F3B capacity
(mequiv/g)
Materials. Remazol Brilliant Red F3B (C. I. Reactive Red 180)
was obtained from Hoechst Celanese Corp. Orange II (C. I.
Acid Orange 7) was purchased from Aldrich Chemical Co.
Both dyes were 98% pure according to their suppliers.
Hydrolyzed Remazol Brilliant Red F3B, hereafter referred to
as F3B, was prepared by heating 0.5 mmol in 100 mL of 100
mM Na2CO3 (initially pH 11.6) or 25 mN NaOH, as described
previously (3). F3B hydrolyzed in NaOH was used for
experiments conducted at pH 5.6, in which case 2-(N-
morpholino)ethanesulfonic acid (MES) was added to the F3B
solution prior to diluting the dye to a final concentration of
1.0 mM (in 25 mM MES/ NaOH).
Quaternary ammonium cellulose (QAC) was obtained from
Whatman International (Maidstone, England) and treated as
described previously (3). Quaternized, epichlorohydrin-
cross-linked sugarcane bagasse (QEB) preparation is de-
scribed by Laszlo (4). Orange II- and F3B-saturated QAC
samples were prepared by stirring QAC (3.8 g) with 1.5 L of
3.0 mM Orange II (in 5.0 mN HCl) or with 2 L of 1 mM F3B
(in 20 mM Na2CO3, pH 11.0), for at least 2 h at room
temperature. Dye-saturated QAC was collected on cellulose
filter paper and rinsed briefly with water and then dried under
vacuum at 25-30 °C overnight.
Dye-Binding Capacity Determination. Dye-binding mea-
surements were conducted on duplicate 100-mg samples of
QAC (or dye-containing QAC) equilibrated for 2 h with 100-
mL solutions of 3.0 mM Orange II (in 5.0 mN HCl) or 1.0 mM
F3B (in 20 mM Na2CO3, pH 11). A portion of the equilibrated
solutions were filtered (25 mm diameter Anotop filter, 0.2 µm
pore size, Whatman International) and then diluted 50-fold
into 5 mN HCl (for Orange II) or into 100 mM Na2CO3, pH
11.6 (for F3B). Orange II and F3B absorbances were measured
at 484 and 540 nm, respectively. Using suitably prepared
standards, residual dye concentrations were determined, from
which the amount of dye bound was calculated (3). The extent
of dye binding is reported as milliequivalents per gram (dry
weight basis of adsorbent), which assumes that Orange II
behaves as a monovalent anion and F3B behaves as a trivalent
anion.
UV/Vis Analysis. Spectra were collected with a Hitachi
U-2000 UV/ Vis spectrophotometer equipped with an auto-
sipper. Dye solutions (1 mM dye, in 25 mM MES/ NaOH, pH
5.6) were diluted 25-fold (Orange II) or 50-fold (F3B) into
water for analysis. Spectra (200-600 nm) were collected using
a baseline established with water containing 1.0 or 0.5 mM
MES/ NaOH.
Orange II-
saturated
F3B-
saturated
treatmenta
5 m M HCl
2 m M KBH4
10 m M bisulfite
2 m M KBH4,
QAC
QAC
QAC
QAC
0.06 (5)b
0.08 (7)
0.52 (45)
1.12 (98)
1.17
1.20
1.16
1.14
0.00 (0)
0.01 (1)
0.00 (0)
1.07
1.07
1.05
0.75 (74) 1.01
10 m M bisulfite
10 m M dithionitec
1 M NaCl
1.12 (94)
0.33 (27)
1.13 (94)
1.19
1.21
1.20
0.73 (72) 1.01
0.26 (25) 1.04
0.89 (88) 1.01
KBH4/bisulfited,
then 1 M NaCl
1 M NaClO4
0.52 (46)
1.13 (101)
1.13
1.12
0.65 (64) 1.01
0.94 (95) 0.99
KBH4/bisulfited,
then 1 M NaClO4
1 N NaOH
1.09 (90)
1.16 (95)
1.21
1.22
0.77 (75) 1.02
0.96 (94) 1.02
KBH4/bisulfited,
then 1 N NaOH
a
QAC (chloride form ) or dye-saturated QAC (0.3) equilibrated for 1
h in 200 m L of the indicated solution, then filtered, and rinsed briefly
with water before drying. Sam ples receiving
(indicated by “then”) were treated with 200 m L of salt or base solution
for 1 h prior to drying. Borohydride and/or bisulfite treatm ents were in
25 MES/NaOH, pH 5.6. Dye-binding capacity expressed as a percentage
of the capacity of QAC receiving the sam e treatm ent. Sodium
hydrosulfite, added as a solid, to 200 m L of 25 m M MES/NaOH, pH 6.2.
2.0 m M borohydride and 10 m M bisulfite.
a second treatm ent
b
c
d
in the negative ion mode. The sample injection volume was
10 µL. Column elution conditions were the same as described
above.
Oxidation-Reduction Potential Measurem ent. Solution
potentials were determined with a Corning redox combination
platinum-Ag/ AgCl electrode (Eref ) 199 mV vs SHE) connected
to an ATI Orion 720A potentiometer.
Borohydride/Bisulfite Reductions. Reductions of dye in
solution or bound to QAC were conducted on stirred, buffered
(25 mM MES/ NaOH, pH 5.6) solutions or suspensions in glass
containers exposed to air or under a bed of N2 (where noted).
All experiments were conducted at 25 ( 2 °C. Sodium
metabisulfite was added as a solid, to a final bisulfite
concentration of 10 mM, approximately 5 min before the
addition of borohydride. Borohydride additions were made
from stock solutions (0.1 M KBH4 in 0.1 N NaOH) prepared
the day of the experiment.
Results
HPLC and HPLC/MS Analysis. Dye products produced
by bisulfite treatment and borohydride/ bisulfite reduction
of Orange II and F3B were determined by high-performance
liquid chromatography (HPLC) using a Spectra-Physics
SP8800 liquid chromatograph system consisting of an au-
tosampler, gradient pump, variable-wavelength UV/ Vis de-
tector, and integrator. The column used was a 4.6 mm i.d.
× 15 cm Microsorb MV (Rainin) C-18 reverse phase column
with a 5-µm particle size. The sample injection volume was
20 µL. The column was developed at a flow rate of 0.5 mL
min-1 with an aqueous mobile phase containing 1.0 mM
ammonium acetate, pH 7.2, for 15 min, followed by a linear
gradient to 75:25 (vol:vol) methanol/ water containing 1.0 mM
ammonium acetate, pH 7.2, over a period of 30 min. The
column was eluted for an additional 10 min with 75:25
methanol/ water containing 1.0 mM ammonium acetate and
then reconditioned to its initial state with a reverse gradient
to 100% water.
QAC Regeneration. The quaternized ammonium cellulose
anion exchanger bound maximally 1.17 mequiv/ g (383 mg/
g) of Orange II or 1.07 mequiv/ g (272 mg/ g) of F3B. Table
1 summarizes the effects of various regeneration treatments
on dye-binding capacity for Orange II-saturated and F3B-
saturated QAC as well as chloride-form QAC (i.e., no dye
bound). Regeneration treatments of chloride-form QAC
generally caused a slight decrease in F3B capacity (e8%) as
compared to 5 mM HCl-treated QAC, while Orange II-binding
capacities were not significantly altered (Table 1). These
findings indicate that QAC degradation is minimal with the
various regeneration treatments employed.
Treatment of dye-saturated QAC with 5.0 mM HCl
regenerated dye-binding capacities to a negligible extent
(e5%), which reflects the ineffectiveness of low Cl- concen-
trations in displacing the bound dyes (Table 1). Similarly,
borohydride treatment without bisulfite present afforded very
limited regeneration (e7%). Bisulfite alone (no borohydride
present) partially regenerated Orange II-saturated QAC but
had no influence on F3B-saturated QAC. Bisulfite does not
For HPLC/ MS, the reverse-phase column was coupled to
a Finnigan-MAT LCQ mass spectrometer via the Finnigan
electrospray interface (Finnigan MAT, San Jose, CA) operating
9
3 6 4 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 12, 1997

FIGURE 1. Orange II UV/Vis spectra. Solid line: 1.0 mM Orange II
in 25 mM MES/NaOH, pH 5.6, diluted 25-fold into water. Dashed line:
1.0 mM Orange II in 25 mM MES/NaOH, pH 5.6, reacted with 10 mM
bisulfite for 2 h under N2 and then diluted 25-fold into water. Dotted
line: 1.0 mM Orange II in 25 mM MES/NaOH, pH 5.6, reacted with
FIGURE 3. Structure of Orange II and its anticipated reduction
products.
2.0 mM KBH and 10 mM bisulfite for 2 h under N2 and then diluted
4
25-fold into water.
FIGURE 2. F3B UV/Vis spectra. Solid line: 1.0 mM F3B in 25 mM
MES/NaOH, pH 5.6, diluted 50-fold into water. Dashed line: 1.0 mM
F3B in 25 mM MES/NaOH, pH 5.6, reacted with 10 mM bisulfite for
2 h under N2 and then diluted 50-fold into water. Dotted line: 1.0
mM F3B in 25 mM MES/NaOH, pH 5.6, reacted with 2.0 mM KBH
4
and 10 mM bisulfite for 2 h under N2 and then diluted 50-fold into
water.
FIGURE 4. Structure of F3B and its anticipated reduction products.
Product A: 2-amino-7-(2-hydroxyethylsulfonyl)-1-naphthalenesulfon-
ic acid (fw 330.2). Product B: 3-amino-4-hydroxy-5-phenylcarboxa-
mido-2,7-naphthalenedisulfonic acid (fw 436.2).
cleave the azo bond of either dye under the given reaction
conditions, but does modify Orange II (see below). Treatment
with borohydride/ bisulfite or dithionite regenerated Orange
II-saturated QAC almost completely (g94%) and substantially
regenerated F3B-saturated QAC (approximately 73%). These
findings indicate that reductive cleavage of dye azo bonds
wavelengths; see Figures 1 and 2), presumably by destruction
of the dyes through reductive cleavage of their azo bonds
(Figures 3 and 4). Bisulfite treatment alone slowly shifted
the absorption maximum of Orange II to shorter wavelengths
(Figure 1) but did not affect the color of F3B (Figure 2). The
color change in Orange II from treatment with bisulfite has
been noted by others (7). The shift in the Orange II absorption
maximum to a shorter wavelength occurred over a period of
several hours (Figure 5). These observations suggest that
bisulfite reacts with Orange II but does not reductively cleave
the molecule and that bisulfite is unreactive toward F3B.
The amount of borohydride required to decolorize solu-
tions of Orange II or F3B depended on whether air was
excluded from the system (Figure 6). Complete decolorization
of either dye was obtained using less borohydride when air
was excluded. This observation indicates that dissolved
oxygen competes with the dye azo bonds for reducing
equivalents, as suggested by eq 6. Alternatively, oxygen may
re-oxidize the hydrazo intermediate (Ar-NH-NH-Ar′), which
will also increase the amount of reducing equivalents required
to achieve complete decolorization.
•-
enhances regeneration and that SO2 likely mediates the
reaction (eq 3).
The data in Table 1 demonstrate the ineffectiveness of
either Cl- or ClO4- in removing Orange II and F3B from QAC.
Borohydride/ bisulfite treatment prior to extraction with these
anions resulted in very large increases in regenerated capacity.
Similarly, while 1 N NaOH regenerated a substantial amount
of dye-binding capacity, treatment of either dye-saturated
QAC with borohydride/ bisulfite, prior to extraction with
NaOH, increased the extent of regeneration. Thus, QAC dye-
binding capacity is more fully restored using a reduction
treatment, followed by salt or base extraction for QAC
containing F3B, than that achieved with simple salt or base
treatments alone.
Dye Reduction in Solution. Treatment of aqueous
solutions of Orange II or F3B with borohydride/ bisulfite led
to a large decrease in solution color intensity (i.e., at visible
9
VOL. 31, NO. 12, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 6 4 9

FIGURE 7. Change in F3B solution absorbance and redox potential
over time during titration with borohydride. Solution containing 1.0
mM F3B, 25 mM MES/NaOH, pH 5.6, and 10 mM bisulfite under N2.
Arrows indicate the time of borohydride addition, with the boro-
hydride:F3B molar ratio given above the arrows in parentheses.
Absorbance: solid line, filled squares. Redox potential: dotted line,
open circles.
FIGURE 5. Effect of bisulfite on dye color over time. Absorbance (at
max) of Orange II or F3B solutions (1.0 mM) treated with 10 mM
bisulfite under N2.
λ
dition, at which point the solution had no absorbance at 540
nm (λmax of F3B) and the dye is presumably fully reduced, the
potential remained below -400 mV for approximately 50 min
and then sharply increased (Figure 7). Borohydride/ bisulfite
solutions held under the same conditions but in the absence
of dye remained below -400 mV for at least 2 h (data not
shown). These observations indicate that azo bond reduction
is rapid in solution. However, measuring the completeness
of the reaction by monitoring solution redox potential may
be unreliable, perhaps because solution components other
than dye azo groups may continue to consume reducing
equivalents.
HPLC analysis of borohydride/ bisulfite-treated Orange II
solutions confirmed the presence of the two expected
reduction products, 4-aminobenzenesulfonic acid (sulfanilic
acid) and 1-amino-2-naphthol (Figure 8). Peaks eluting at
2.4 and 43.7 min co-chromatographed with authentic sul-
fanilic acid and 1-amino-2-naphthol, respectively. Orange
II treated with bisulfite for 2 h produced a single peak at 32.4
min, which did not coincide with Orange II (47.4 min retension
time) or either of the reduction products (Figure 8). HPLC/
MS analysis using electrospray ionization of Orange II
produced a mass spectrum dominated by the [M - H]- ion
at m/z 327 and a lesser amount of an apparent dimer of Orange
II (i.e., [2M - H]- at m/z 654, Figure 9). The mass spectrum
of the 32.4-min HPLC peak (bisulfite-treated Orange II)
contained the signature [M - H]- ion (m/z 327) of Orange II
as well as major signals at m/z 204 and 409, which are
tentitively attributed to the [M - 2H]2- and [M - H]- ions,
FIGURE 6. Effect of borohydride:dye molar ratio on extent of dye
decolorization. Absorbance (at λmax) of Orange II or F3B solutions
(1.0 mM) containing 10 mM bisulfite with successive additions of
KBH at 20-min intervals.
4
Equation 4 implies that the reaction stoichiometry of
borohydride to azo bond should be 0.5 (mole ratio). Titrations
with borohydride of F3B under N2 required a borohydride:
azo ratio of 0.7 to achieve complete decolorization of the
solution (Figure 6). Conversely, Orange II consumed less
than the expected amount of borohydride to become fully
decolorized (Figure 6) due to the concomitant reaction of
bisulfite with Orange II.
Reduction of F3B in solution by borohydride/ bisulfite was
rapid (Figure 7). Addition of a stoichiometric amount of
borohydride (i.e., a 0.5 borohydride to azo bond molar ratio)
to a F3B solution (under N2) decreased the absorbance of the
solution within 5 min to a value that was constant for at least
2 h. Subsequent additions of borohydride further reduced
solution color with similar speed. Solution potential also
dropped rapidly to about -500 mV with each addition of
borohydride. However, unlike the absorbance, solution
potentials increased steadily during the intervals between
borohydride additions. Following a final borohydride ad-
-
respectively, of a SO3 adduct of Orange II. This suggests
that bisulfite forms a stable adduct to Orange II under the
given reaction conditions but does not cleave the azo bond.
The nature of the product generated by bisulfite treatment
of Orange II was examined further. QAC bound 0.34 mmol/ g
of the product, which indicates that the product binds as
though it is a trivalent anion (i.e., the QAC capacity is 1.02
mequiv/ g; see Table 1 F3B binding data for comparison).
Examination of the product in D2O by nuclear magnetic
resonance (NMR) showed that bisulfite treatment of Orange
II results in the loss of proton signals at the naphthalene C3
(6.6 ppm) and C4 (5.6 ppm) positions (data not shown),
indicating that bisulfite substitutes at both positions. (NMR
data were obtained from a Bruker ARX 400 spectrometer
9
3 6 5 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 12, 1997

FIGURE 8. Orange II HPLC chromatograms. Upper chromatogram:
1.0 mM Orange II. Middle chromatogram: 1.0 mM Orange II treated
for 2 h with 10 mM bisulfite. Lower chromatogram: 1.0 mM Orange
II treated with 2.0 mM borohydride and 10 mM bisulfite for 20 min.
FIGURE 10. F3B HPLC chromatograms. Upper chromatogram: 1.0
mM F3B. Middle chromatogram: 1.0 mM F3B treated for 2 h with
10 mM bisulfite. Lower chromatogram: 1.0 mM F3B treated with 2.0
mM borohydride and 10 mM bisulfite for 2 h under N2.
to F3B. Bisulfite treatment of F3B caused the minor peak at
35.5 min to elute slightly ahead of the F3B major peak,
indicating that bisulfite reacts with this minor component.
The observation that bisulfite did not modify the mobility of
the major peak in the chromatogram is consistent with the
lack of effect of bisulfite on the F3B absorption spectrum
(Figure 2). Borohydride/ bisulfite treatment of F3B eliminated
all original peaks and generated a multitude of new peaks
(Figure 10), none of which had appreciable absorbance at
540 nm (data not shown). An HPLC/ MS analysis of F3B and
some of the products formed from its reduction are shown
in Figure 11. The mass spectrum of the major F3B peak (34
min retension time) showed characteristic ion peaks for the
[M - 3H]3- (m/z 254), [M - 2H]2- (m/z 382), and [M - H]-
(m/z 764) F3B species (Figure 11, upper spectrum). The 10-
min peak in the borohydride/ bisulfite-treated F3B chro-
matogram was identified as reduction product A (refer to
Figure 4) based on its strong signal at m/z 330 (Figure 11,
middle spectrum). The unresolved peaks eluting at around
30 min contained reduction product B (Figure 11, lower
spectrum) and related products. These results indicate that
borohydride/ bisulfite reductively cleaves F3B generally in the
manner implied by Figure 4 but generates other products as
well.
FIGURE 9. Orange II mass spectra. Upper spectrum: Orange II (HPLC
47.4 min peak, Figure 8 upper chromatogram). Lower spectrum:
Orange II-bisulfite adduct (HPLC 32.4 min peak, Figure 8 middle
chromatogram).
operating at 400 MHz.) Thus, the NMR and QAC binding are
consistent with the formation of a disubstituted trivalent
anion, while the mass spectrum indicates that the product
is a monosubstituted divalent anion. Whether the product
is a tri- or divalent anion, there is no ready explanation for
the partial (45%) regeneration of Orange II-saturated QAC
observed upon bisulfite treatment (Table 1).
The HPLC chromatogram of F3B showed one major peak
at 34 min and minor peaks at 35.5 and 42.2 min (Figure 10).
The minor peaks absorbed strongly at 540 nm (as did the
major peak), which indicates that they are structurally related
Reduction of QAC-Bound Dye. The extent of F3B
reduction in solution as compared to that obtained with F3B
bound to QAC was examined to determine whether the ion
exchanger lowered the utilization efficiency of reducing
equivalents. In order to measure the extent of QAC-bound
F3B reduction, it was necessary to find conditions under which
F3B could be quantitatively extracted. Treatment with an
aqueous solution containing 8 M urea and 1 M NaOH was
found to release at least 96% of QAC-bound F3B (data not
shown). Table 2 summarizes the results obtained at two
different borohydride:F3B molar ratios, using the urea/ NaOH
9
VOL. 31, NO. 12, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 6 5 1

Discussion
This work has shown that the reduction of azo dyes bound
to quaternized cellulose permits facile regeneration of
exchanger dye-binding capacity. Dye reduction products are
more readily displaced from QAC than the intact dye because
the reduction in the number of anionic charges and hydrogen
bonding sites per molecule lowers the binding affinity of the
products. Some azo dyes with multiple sulfonic and/ or
carboxylic acid groups have all of their anionic sites attached
on just one side of the azo bond. In such cases, reduction
of the dye will not lead to a lower charge in all of the products;
even so, the reduction products likely will have less affinity
for QAC than the intact dye.
Bisulfite-mediated borohydride reduction of dye bound
to QAC occurs with the same efficiency as with dye in solution
(Table 2). Thus, the cellulose backbone of the exchanger
does not consume reducing equivalents, which might occur
if there were a large number of oxidized (i.e., carbonyl) groups
introduced during the preparation of the cellulose. The
consumption of bisulfite by reaction with cellulose chain
reducing ends is another potential unwanted side reaction
that was not manifest. Therefore, use of QAC (or quaternized
bagasse) to bind dye before treatment with the borohydride/
bisulfite couple does not impose a penalty from unproductive
consumption of reducing equivalents. Furthermore, since
reduction of exchanger-bound dye can be conducted in a
volume that is much smaller than the volume of wastewater
from which the dye was removed using QAC, fewer reducing
equivalents will be wasted on side reactions (see eqs 5 and
6). An additional efficiency may be found by repeated use
of the borohydride/ bisulfite solution for regenerating dye-
saturated QAC. This would require addition of only more
borohydride since there is no net consumption of bisulfite
in the reaction (see eqs 1 and 3).
Treating large volumes of dye-containing wastewater with
borohydride/ bisulfite may increase the toxicity of the waste-
water, necessitating additional remediation steps. The
reduction of azo dyes can produce carcinogenic aromatic
amines (8). Boron does not pose a human health risk (9) but
can be toxic to sensitive plant species at concentrations as
low as 0.5 mM (5.4 mg/ L; 10, 11). The examples provided by
Cook (5) on the treatment of various industrial dye manu-
facturing or textile effluents indicated borohydride dose levels
were 0.4-0.8 mM. Removal of the dye from the effluent
stream by QAC eliminates these problems or at least
minimizes the volume of wastewater requiring additional
treatment. The aromatic amines accumulated in the QAC
regeneration solution (and salt-wash solutions) may be readily
removed and perhaps put to productive use by enzymatic
polymerization (12).
FIGURE 11. F3B mass spectra. Upper spectrum: F3B (HPLC 34 min
peak, Figure 10 upper chromatogram). Middle spectrum: HPLC 10
min peak (see Figure 10) of borohydride/bisulfite treated F3B. Lower
spectrum: HPLC approximately 30-min peak of borohydride/bisulfite
treated F3B.
TABLE 2. Efficiency of Borohydride/Bisulfite Reduction of F3B
Bound to QAC or QEB in Comparison to F3B in Solution
F3B reduction (%)
borohydride:F3B
b
molar ratio
in solutiona
on QAC
on QEB
0.35
0.70
49
98
46
92
52
98
a
A 50-m L solution of 1.0 m M F3B in 25 m M MES/NaOH, pH 5.6, was
stirred under N2 for 1 h, and then bisulfite was added to a concentration
of 10 m M. Following bisulfite addition, KBH4 was added to give the
indicated borohydride:F3B m olar ratio. After 1 h of treatm ent under N2,
the solution was brought to 8 M urea and 1 N NaOH (added as solids),
and stirred for an additional 1 h. The concentration of unreacted F3B
in solution was determ ined by diluting an aliquot 50-fold into 100 m M
Na2CO3, pH 11.6, and m easuring the absorbance at 540 nm . The F3B
concentration was corrected for the dilution caused by the addition of
The very high cost of commercially available quaternized
cellulose makes its use for wastewater decolorization too
expensive unless the QAC can be used repeatedly. Use of the
borohydride/ bisulfite couple, or dithionite alone, on QAC
saturated with azo dyes may provide a means by which to
recycle the exchanger economically. Thus, a coupling of two
treatment technologies, each separately having limited
potential, may provide an improved, economically feasible
process.
b
urea and NaOH. The sam e procedure was followed as described above
for the solution phase reduction of F3B, except that 0.15 g of QAC or
QEB (exchange capacity in excess of that needed to bind all F3B in
solution) was added 1 h before addition of bisulfite. Following the 1-h
borohydride/bisulfite treatm ent, the suspension was stirred 1 h open
to air to deplete residual reducing equivalents, if present, prior to
extraction of unreacted F3B with 8 M urea and 1 N NaOH. The suspension
was centrifuged (50g, 10 m in) to rem ove solids prior to sam pling for
the F3B concentration determ ination. Reported values are the m ean of
two separate experim ents.
Acknowledgments
extraction procedure to determine the amount of unreacted
F3B remaining. The extent of F3B reduction with the dye
bound to QAC was only slightly lower than that obtained
with F3B in solution. A similar experiment was conducted
using quaternized sugarcane bagasse (QEB). The extent of
F3B reduction was the same whether bound to QEB or in
solution (Table 2). This indicates that the cellulose or
lignocellulose backbone of the exchanger does not compete
with the dye azo bond for reducing equivalents.
The author wishes to thank Leslie J. Smith for her excellent
technical assistance, Ronald D. Plattner for making available
the mass spectrometer, David Weisleder for collecting and
interpreting the NMR data, and the manuscript reviewers for
their helpful comments. Names are necessary to report
factually on available data; however, the USDA neither
guarantees nor warrants the standard of the product, and the
use of the name by the USDA implies no approval of the
product to the exclusion of others that may also be suitable.
9
3 6 5 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 12, 1997

(9) Murry, F. J. Regul. Toxicol. Pharmacol. 1995, 22, 221-230.
(10) Choi, J.-M.; Pak, C.-H.; Lee, C. W. J. Plant Nutr. 1996, 19, 901-
915.
(11) Jackson, M. B.; Lee, C. W.; Schumacher, M. A.; Duysen, M. E.;
Self, J. R.; Smith, R. C. J. Plant Nutr. 1995, 18, 1337-1349.
(12) Klibanov, A. M.; Alberti, B. N.; Morris, E. D.; Felshin, L. M. J.
Appl. Biochem. 1980, 2, 414-421.
Literature Cited
(1) Reife, A. In Encyclopedia of Chemical Technology, 4th ed.; Howe-
Grant, M., Ed.; John Wiley & Sons, Inc.: New York, 1993; Vol.
8, pp 753-783.
(2) Laszlo, J. A. Am. Dyest. Rep. 1994, 83, 17, 18, 20, 21, and 48.
(3) Laszlo, J. A. Text. Chem. Color. 1995, 27 (4), 25-27.
(4) Laszlo, J. A. Text. Chem. Color. 1996, 28 (5), 13-17.
(5) Cook, M. M. In Environmental Chemistry of Dyes and Pigments;
Reife, A., Freeman, H. S., Eds.; John Wiley & Sons, Inc.: New
York, 1996; pp 33-41.
(6) Voyksner, R. D.; Straub, R.; Keever, J. T.; Freeman, H. S.; Hsu,
W.-N. Environ. Sci. Technol. 1993, 27, 1665-1672.
(7) Ufimtsev, V. N.; Levin, E. S. Chem. Abstr. 1944, 38, 6203.
(8) Pu¨ ntener, A.; Fennen, J.; Page, C. T. J. Soc. Leather Technol. Chem.
1996, 80, 1.
Received for review May 5, 1997. Revised manuscript received
September 2, 1997. Accepted September 10, 1997.X
ES970395V
X Abstract published in Advance ACS Abstracts, October 15, 1997.
9
VOL. 31, NO. 12, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 6 5 3