Degradation of fluorescent whitening agents in sunlit natural waters

Article abstract of DOI:10.1021/es950711a

Stilbene-type fluorescent whitening agents (FWAs), such as the distyryl biphenyl (DSBP) and the diaminostilbene types (DAS 1 and DAS 2), are commonly used in detergents and papers. They are not readily biodegradable, but due to their ability to absorb part of the terrestrial sunlight, they can be photochemically degraded in natural surface waters. Following a fast preceding photoisomerization, the three compounds are degraded by direct photochemical processes yielding mainly aldehydes and alcohols. Their degradation quantum yields are similar, about 10"SUP -4" . Nevertheless, in samples of a eutrophic Swiss lake water, DSBP is photochemically degraded three times faster (t"SUB 1/2" =87 min) in summer noon terrestrial sunlight at 25 degrees C than DAS 1 and DAS 2 (t"SUB 1/2" =278 and 313 min) because of a higher rate of sunlight absorption by the DSBP isomer mixture. All FWAs are degraded faster if the oxygen concentration is increased. Dissolved natural organic material partly inhibits the degradation of DSBP in the reaction with molecular oxygen. The behavior of these compounds illustrates the influence of a preceding isomer equilibrium on degradation rate coefficients. (Authors)

Full text of DOI:10.1021/es950711a

Environ. Sci. Technol. 1996, 30, 2227-2234
tant properties of FWAs are high affinity to cellulose, water
solubility, and stability with regard to other detergent
ingredients (1, 2).
Degradation of Fluorescent
Whitening Agents in Sunlit
Natural Waters
J O H A N N E S B . K R A M E R ,
S I L V I O C A N O N I C A , * A N D
J U¨ R G H O I G N EÄ
FWAs have been used since the 1940s. Thousands of
them have been described, but only a few are used world
wide (1, 3). Two of the stilbene-type FWAs investigated
here, DSBP and DAS 1 (see Figure 1), are among the most
commonly used whiteners in household detergents. Their
estimated world wide production in 1992 was 3000 and
14 000 t/ year, respectively (4). FWAs contained in deter-
gents at an average concentration of about 0.15% by weight
Swiss Federal Institute for Environmental Science and
Technology (EAWAG) and Swiss Federal Institute of
Technology (ETH), CH-8600 D u¨ bendorf, Switzerland
(
5) serve to replace the textile FWAs, which are photo-
chemically degraded during wearing (6) or washed out
during the washing process (7). Depending on the type of
FWA, washing temperature, type of fabric, and composition
of the detergent, about 20-95% of the FWAs will exhaust
onto the fabric during the washing process (5, 8). The
remaining amount is discharged with the wash liquor,
yielding sewage treatment plant raw influx concentrations
of typically 10-20 µg/ L FWAs (9). FWAs are partially
retained in municipal wastewater treatment plants due to
adsorption on the activated sludge. Biological degradation
by activated sludge is too slow compared to the residence
times in the treatment plants (9, 11). The concentration
of DAS 1 in secondary effluents of four Swiss sewage
treatment plants was found to range from 2.6 to 4.5 µg/ L,
which corresponds to average adsorption rates of 89% after
activated sludge treatment. DSBP was on average only 53%
adsorbed by the sludge, resulting in secondary effluent
concentrations of 3.3-8.9 µg/ L (8). Average concentrations
of DAS 1 in 35 U.S. rivers have been determined to range
from 0.06 µg/ L above to 0.7 µg/ L below sewage outfalls.
One additional diaminostilbene-type FWA was found in
comparable amounts (10). More recent studies in five Swiss
rivers showed concentrations ranging from 0.04 to 0.6 µg/ L
for DSBP and from 0.04 to 0.4 µg/ L for DAS 1 (9).
J U¨ R G E N K A S C H I G
Chemicals Division, Ciba-Geigy Ltd., Basel,
D-79630 Grenzach-Wyhlen, Germany
Stilbene-type fluorescent whitening agents (FWAs),
such as the distyryl biphenyl (DSBP) and the
diaminostilbene types (DAS 1 and DAS 2), are commonly
used in detergents and papers. They are not readily
biodegradable, but due to their ability to absorb part
of the terrestial sunlight, they can be photochemically
degraded in natural surface waters. Following a fast
preceding photoisomerization, the three compounds
are degraded by direct photochemical processes
yielding mainly aldehydes and alcohols. Their degrada-
-4
tion quantum yields are similar, about 10 . Never-
theless, in samples of a eutrophic Swiss lake water,
DSBPis photochemicallydegradedthree times faster
(t1/2 ) 87 min) in summer noon terrestrial sunlight at
2
5 °C than DAS 1 and DAS 2 (t1/2 ) 278 and 313 min)
because of a higher rate of sunlight absorption by the
DSBP isomer mixture. All FWAs are degraded faster
if the oxygen concentration is increased. Dissolved
natural organic material partly inhibits the degrada-
tion of DSBP in the reaction with molecular oxygen.
The behavior of these compounds illustrates the
influence of a preceding isomer equilibrium on
degradation rate coefficients.
The ecotoxicology of detergent-derived FWAs has been
studied extensively, especially during the mid 1970s,
showing low toxicity and no ready biodegradability (2, 5,
9
, 10, 12, 13). Studies examining the photofading of FWAs
in aqueous solution showed that the loss of fluorescence
is primarily due to isomerization from the E-isomer to the
Z-isomer, yielding a steady state between the two isomers.
The quantum yield for E/ Z isomerization at 340 nm of DSBP
has been determined as 0.019 (14). Photofading rates in
river water exposed to natural sunlight have been reported
as 7% for DSBP and 71% for DAS 1 after 60 min (13). Further
degradation is much slower and has mostly been studied
with stronger light sources than sunlight. The degradation
rate depends on the oxygen concentration, but degradation
takes place also in degassed solutions (15). To date there
has been no study describing photochemical degradation
rate coefficients or predicting lifetimes for FWAs in sunlit
natural waters. We have therefore determined degradation
rate coefficients and half-lives in a sunlit Swiss lake water
for two detergent-derived FWAs (DSBP and DAS 1) and a
paper-derived FWA (DAS 2) (see Figure 1). The latter has
also been found in Swiss river water (9). Other topics
covered are isomer concentration ratios and product
formation under natural conditions with particular atten-
Introduction
Fluorescent whitening agents (FWAs) are used in textiles,
detergents, paper, and plastics to make products whiter
and brighter by compensating for the yellowish shade of
materials. FWAs act as an additional light source by
transforming the absorbed UV light of terrestrial solar
radiation and artificial light into visible, blue fluorescence
light. Efficient FWAs therefore must have high absorption
coefficients in the wavelength range of 300-400 nm. This
also makes them good candidates for undergoing photo-
chemical transformations under sunlight. Further impor-
*
Corresponding author telephone: +411 823 5453; fax: +411 823
5
471; e-mail address: canonica@eawag.ch.
S0013-936X(95)00711-5 CCC: $12.00

1996 Am erican Chem ical Society
VOL. 30, NO. 7, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2 2 2 7

of dissolved organic carbon (DOC) in the lake water DNOM
is 3.3 mg/ L, the average pH is 8.3. Solutions used to study
the DNOM concentration dependence were made by
diluting a Fluka humic acid stock solution (DOC ) 330
mg/ L) prepared according to Haag et al. (19). FWAsolutions
were made diluting 250 µM stock solutions. Generally 1
µM FWA solutions were used for irradiation. Except where
noted, all photolysis solutions were equilibrated with air
prior to photolysis. Oxygen-saturated and deoxygenated
solutions were prepared by purging with oxygen or nitrogen
for 10 min. Tubes containing these latter two solutions
were wrapped with Parafilm M (American National Can)
to minimize gas exchange between tubes and their sur-
rounding. All FWA solutions were prepared under red light
to prevent isomerization due to laboratory light.
Materials. All FWAs and some of their possible deg-
radation products were supplied by Ciba-Geigy AG, Basel,
Switzerland. Other chemicals were analysis grade and used
as received from commercial suppliers.
Analyses. The FWAs and their photochemical degrada-
tion products were analyzed by reverse-phase high-
performance liquid chromatography (HPLC) using a Hewlett
Packard System 1050 equipped with a quaternary gradient
pump and a variable wavelength absorbance detector.
Compounds were separated at room temperature on a 124
FIGURE 1. Structures of the three discussed FWAs. DSBP: 4,4′-
bis(2-sulfostyryl)biphenyl; DAS 1: 4,4′-bis[(4-anilino-6-morpholino-
1,3,5-triazin-2-yl)amino]stilbene-2,2′-disulfonate; DAS 2: 4,4′-bis[(4-
(
4-sulfoanilino)-6-bis(2-hydroxyethyl)amino-1,3,5-triazin-2-
yl)amino]stilbene-2,2′-disulfonate.
×
4 mm Spherisorb ODS-2 (5 µm pore size) C18-column.
tion to the influence of dissolved natural organic material
A ternary mixture of water (solution A), methanol (solution
B), and 1.7 g/ L tetrabutylammonium hydrogen sulfate
buffered with 0.05 M phosphate at pH 7 (solution C) was
used as eluent. For DSBP and DAS 1, a 14.5-min linear
gradient from 45% A, 45% B, 10% C to 20% A, 70% B, and
10% C was used. The initial detection wavelength for DSBP
and its degradation products was 310 nm and was switched
after 12 min to 345 nm. DAS 1 and its degradation products
were analyzed at 250 nm. DAS 2 and its products were
separated with a 15-min linear gradient from 60% A, 30%
B, 10% C to 50% A, 40% B, and 10% C and detected at 270
nm. Extinction coefficients of all FWAs and their isomers
were calculated from UV/ Vis spectra of pure standards,
except for those of (Z)-DAS 2, which have been calculated
by subtracting the spectrum of (E)-DAS 2 from that of a
mixture containing known amounts of (E)- and (Z)-DAS 2,
obtained by short illumination of an (E)-DAS 2 solution. All
UV/ Vis spectra were recorded on a Kontron Instruments
Uvikon 930 spectrophotometer at room temperature.
Identification of the Degradation Products. Degrada-
tion products of which pure standards were available were
identified on a Hewlett Packard System 1090 equipped with
a diode array detector (DAD) by comparing the retention
times and DAD spectra. Identification of the alcohol 2
formed from DAS 1 or DAS 2 was performed after isolation:
120 mg of (Z)-DAS 1 or (E)-DAS 2 were dissolved in 1000
mLofwater. This solution was divided into 20-mLportions,
which were filled into quartz tubes. After 10-min purging
with nitrogen, the samples were irradiated by simulated
sunlight in the MGRR at 700 W for 5 h. The combined
portions were evaporated to about 3-4 mL, then 20 mL of
2-propanol was added. The precipitate was isolated by
filtration and dried under vacuum yielding the alcohol 2
(purity over 90%). Identification was performed by NMR
and mass spectroscopy at Ciba-Geigy AG, Basel, and EMPA,
D u¨ bendorf.
(
DNOM). The photolysis of these FWAs is an excellent
example of how degradation kinetics are influenced by a
preceding isomer equilibrium.
Experimental Section
Irradiations in the Laboratory. For investigating the
degradation kinetics, irradiations were carried out in glass-
stoppered quartz photolysis tubes (1.5 cm i.d.) using a water-
cooled merry-go-round reactor (MGRR), equipped with a
7
00-W medium pressure mercury lamp (Type TQ 718,
Hanau). Appropriate combinations of cooling jacket and
filter solutions are used, either to roughly approximate
sunlight reaching the surface of natural waters or to create
monochromatic light. Sunlight conditions (λ > 320 nm)
are simulated with the 4-mm Duran 50 borosilicate glass
cooling jacket and a 0.15 M NaNO3 filter solution. Mono-
chromatic light at 366 nm was obtained using a cooling
jacket consisting of 2 mm of quartz, 2 mm of Duran 50
borosilicate glass, and a UVW-55 glass filter in combination
with 325 g/ L NaBr with 1.5 g/ L Pb(NO3)2 filter solution.
Filter solutions were thermostated at 25.0 ( 0.2 °C to keep
irradiated FWA solutions at constant temperature. Chemi-
cal actinometry for monochromatic experiments was based
on the method for weakly absorbing systems described by
Zepp (16) and performed as described by Canonica et al.
(
17).
Irradiations in Natural Sunlight. These were performed
in the same quartz tubes held on a rack at a 60° angle from
the horizontal in a water-filled flat-bottom container. The
water inside the photolysis tubes was below the exterior
cooling water level and was kept at 25.0 ( 0.3 °C. For more
detailed information see ref 18.
Solutions. DNOM-free solutions ofFWAs were prepared
with bidistilled water buffered at pH 8.0 with 0.05 M
phosphate. Water from Lake Greifensee was taken from
the summer epilimnion and was filtered through 0.45-µm
cellulose nitrate membrane filters (Sartorius AG) and stored
in the dark at 4 °C prior to use. The average concentration
1
Product 2 (DAS 1): (C40H40O9N12S2) H NMR (δ ppm,
CD3S(O)CD3, 360 MHz): 3.4 and 3.5 (2 multiplets, 2 H)
HOCCH2; 3.7 and 3.8 (2 broad singlets, 16H), 2 NCH2CH2-
2
2 2 8
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 7, 1996

be described as
kR,deg(λ)
φ
)
(2)
R,deg
2
.303ꢀ (λ)I(λ)
R
where kR,deg(λ) is the pseudo-first-order degradation rate
coefficient at wavelength λ, the factor 2.303 accounts for
the log 10 basis of ꢀ, I(λ) is the photon fluence rate at
-
2
-1
wavelength λ in Einstein m
s , and ꢀR(λ) is the apparent
extinction coefficient of the isomer mixture REZ(λ), calcu-
lated from the weighed contributions of both isomers at
photostationary state:
ꢀ
R(λ){[E]ps + [Z] } ) ꢀ (λ)[E] + ꢀ (λ)[Z]
ps
(3)
ps
E
ps
Z
FIGURE 2. Photolysis pathways of FWAs: OE,iso and OZ,iso are the
isomerization quantum yields, and Or,deg is the apparent degradation
quantum yield of the isomer mixture.
which by substituting [Z]ps ) REZ(λ)[E]ps and solving for
R(λ) leads to
ꢀ
OCH2CH2; 5.7 (multiplet, 1 H) HOCH; 6.0 (doublet, J ) 6
Hz, 1 H) COH; 6.9, 7.3, 7.5, 7.8, 8.1 (16 aromatic H); 9.1 and
ꢀ
(λ) + REZ(λ)ꢀ
+ R (λ)
(λ)
Z
E
ꢀR(λ) )
(4)
9
.2 (2 singlets, 4 H) 4 NH. In solvent mixture CD3S(O)-
1
EZ
CD3/ D2O no signals at 6.0, 9.1, and 9.2 (H atoms exchanged
by D atoms). Mass spectroscopy (sodium salt C40H39O9-
N12S2Na3) m/z: 964.3 (calculated 964.9).
For a batch-type reactor, kR,deg can be determined experi-
mentally from
1
Product 2 (DAS 2): (C40H42O17N12S4) H NMR (δ ppm,
CD3S(O)CD3, 400 MHz): 3.1 and 3.5 (2 multiplets, 2H)
OHCCH2;3.7 (multiplet, 16 H) 2 N (CH2CH2OH)2;4.8 (singlet
[FWA]t
ln
) -kR,deg
t
(5)
[FWA]0
4
6
4
H) 4 H2COH; 5.7 (singlet, 1 H) HCOH; 5.9 (doublet, J )
Hz, 1 H) CHOH; 7.4 and 7.7 (14 aromatic H) 9.1 (singlet,
H) 4 NH. In solvent mixture CD3S(O)CD3/ D2O no signals
where [FWA]t is the concentration of FWA at time t and
FWA]0 is the concentration of FWA at time t ) 0. It can
[
at 4.8, 5.9, and 9.1 (H atoms exchanged by D atoms).
be shown (20) that φR,deg is a function only of the photo-
isomerization quantum yields and the photodegradation
quantum yield of both isomers and can thus be assumed
as λ-independent. For polychromatic light, the numerator
and the denominator in eq 1 have to be integrated over the
wavelength range to give the isomer ratio REZ.
Theoretical Background
The isomer ratios of stilbene-type FWAs play an important
role in determining the photodegradation quantum yields,
although only the E,E-isomer of DSBP and the E-isomers
of DAS 1 and DAS 2 are the active whitening agents (see
Figure 1). The photochemical kinetics applicable to these
FWAs can be summarized according to the scheme in Figure
Results and Discussion
To estimate maximum degradation rate coefficients under
natural light conditions, degradation experiments were
carried out on clear summer days in D u¨ bendorf, Switzerland
2
. The quantum yields for the reversible photoisomeriza-
tion process, φE,iso and φZ,iso, can be assumed to be much
larger than the photodegradation quantum yield, φR,deg,
which results in a photostationary equilibrium between
the isomers on the time scale of photodegradation. In the
case of monochromatic irradiation at wavelength λ, this
equilibrium is described by the isomer concentration ratio,
REZ(λ) given by
(
latitude 47.4° N), around noon. All FWAs were dissolved
in water from Lake Greifensee, having a dissolved organic
carbon content of 3.3 mg/ L, and in DNOM-free water to
examine the influence of DNOM on the degradation rates.
The results given in Table 1 show that degradation of
DSBP in terrestrial sunlight is 3.2 times faster than that of
DAS 1 and 3.6 times faster than that of DAS 2 in Lake
Greifensee water, although their quantum yields show only
minor differences. The difference in rate coefficients can
be explained in terms of the different light absorption of
the isomer mixture, REZ, resulting from fast photoisomer-
ization. The extinction coefficient ofthe isomer distribution
ꢀR(λ) can be calculated using eq 4, by inserting the
experimentally determined REZ instead of REZ(λ). Figure 3
shows how this preceding isomer equilibrium influences
light absorption. The absorption spectra of the E,E-isomer
of DSBP and E-isomers of the DAS-type FWAs are all very
similar, with a maximum at ∼350 nm and an extinction
[
Z]ps
kE,iso(λ)
ꢀE(λ)φ
E,iso
R (λ) )
(λ) )
)
(1)
EZ
[E]ps
kZ,iso(λ) ꢀZ(λ)φZ,iso
where [X]ps(λ) is the photostationary state molar concen-
tration, ꢀX(λ) is the extinction coefficient of the X-isomer (X
)
E or Z), and kX,iso(λ) is the rate constant for photoi-
somerization from the X-isomer to the other isomer. The
photoisomerization quantum yields are assumed to be
λ-independent within the wavelength region of interest (λ
>
295 nm). [Note: To use this and the following equations
for DSBP, substitute E,E for E and E,Z for Z. The Z,Z-isomer
of DSBP could not be detected during our irradiation
experiments; it is apparently formed in only very small
ratios. Other authors also were unable to detect this isomer
after photo-irradiation (4). For these reasons, its contribu-
tion is neglected in our study.] For the degradation, we
only have to consider the isomer mixture REZ(λ), whose
apparent degradation quantum yield at wavelength λ can
-
1
-1
coefficient at the maximum of 60 000-70 000 M cm
.
However, the spectrum of the E,Z-isomer of DSBP is
remarkably different from that of the Z-isomers of DAS 1
and DAS 2. In the wavelength region that is relevant for
sunlight-induced photodegradation (λ > 295 nm), (E,Z)-
DSBP still exhibits a strong absorption band that is only
shifted by 12 nm toward shorter wavelengths with respect
VOL. 30, NO. 7, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2 2 2 9

TABLE 1
Degradation Experiments in Natural Sunlight and under Monochromatic Light Conditions at 25 °Ca
Lake Greifensee water
DNOM-free water
FWA
rEZ ()[Z]ps/[E]ps)b,c t1/2 (min) (sunlight)c Or,deg (MGRR, 366 nm) t1/2 (min) (calculated)d t1/2 (min) (sunlight)c Or,deg (MGRR, 366 nm)
(1.2 ( 0.1) × 10-
4
4
63
245
302
54 ( 4
324 ( 76
236 ( 20
(2.4 ( 0.2) × 10
(1.1 ( 0.1) × 10
(0.9 ( 0.1) × 10
-4
-4
-4
DSBP
DAS 1
DAS 2
20/80
85/15
87/13
87 ( 5
-
278 ( 30
313 ( 18
(1.0 ( 0.1) × 10
(0.74 ( 0.07) × 10-
4
a
Errors represent 95% confidence intervals and were calculated according to the error propagation law of Gauss taking the photolysis rates of
FWAs and actinom eters as affected by errors. Relative isom ers contents, see eq 1. c Determ ined in sum m er noon sunlight at D u¨ bendorf (latitude
b
4
7.4° N). d Minim um half lives (sum m er noon) for the isom er m ixtures REZ at the water surface calculated with the GCSOLAR program (21) for the
sam e days when the experim ents were carried out.
More details about the photoisomerization kinetics will be
described in a study using picosecond laser spectroscopy
(
20).
The degradation quantum yields at λ ) 366 nm in Table
1
are at least 2 orders of magnitude smaller for all FWAs
than the published isomerization quantum yields at 340
nm of 0.019 for DSBP and 0.185 for a FWA of the DAS-type
such as DAS 1 and DAS 2 (14). Using monochromatic light
at 366 nm, DSBP has a ratio of 80% E,E- to 20% E,Z-isomer
resulting in an apparent extinction coefficient of 46 000
-
1
-1
M
1
cm . The ratio of 10% E- and 90% Z-isomer for DAS
yields an apparent extinction coefficient at 366 nm of
-
1
-1
only 6600 M cm . This difference in light absorption
controls the difference in degradation rate coefficients
obtained using Lake Greifensee water and monochromatic
light: 0.034 min for DSBP and 0.0036 min for DAS 1.
Degradation rate coefficients are independent of FWA
concentration in the range of 1-10 µM, indicating that self-
sensitization is negligible in these experiments. As outlined
in the Introduction, FWA concentrations found in rivers
and lakes are even much below this range, and therefore
self-sensitization can be excluded.
-
1
-1
The extinction coefficients from Figure 3 and the
quantum yields from Table 1 are the only chemical
constants needed for calculating the half-lives of the FWAs
in sunlight with the computer program GCSOLAR (21).
These calculations yielded half-lives in good agreement
with the ones experimentally determined (see also Table
1
, column 5). The results of extensive calculations to
determine the degradability as a function of time of year,
latitude, and depth of the water body will be published
later.
In natural waters, light absorption by DNOM leads to
1
•
the formation ofphotooxidants (e.g., O2, OH, excited triplet
states, etc.) that are able to react with pollutants (17, 22).
Direct photodegradation, which we assume to be the only
possible process in DNOM-free water, might therefore be
accompanied by these indirect photochemical degradation
pathways. However, because all FWAs studied here show
degradation rate coefficients that are not faster in Lake
Greifensee water than in water which does not contain
DNOM (see Table 1), we conclude that direct photochemical
reactions are the main degradation routes for all FWAs,
and probably degradation due to photooxidants can be
neglected.
A further observation that can be derived from Table 1
is that for DSBP the degradation in DNOM-free water is
somewhat faster than in Greifensee water. Although some
screening by chromophoric water constituents associated
with the dissolved organic material takes place in this
system, its contribution to the decrease in degradation rate
is only small; for wavelengths >300 nm less than 8% at the
FIGURE 3. FWAs extinction coefficients for E-isomers (full line),
Z-isomers (dotted line), and isomer mixtures that are effective under
sunlight irradiation at the surface of natural waters (heavy full line).
The spectra of the isomer mixtures were calculated using eq 4 and
the experimental values of rEZ given in Table 1.
to the spectrum of (E,E)-DSBP due to the presence of one
of the stilbene moieties in its E-configuration. The molar
extinction of (Z)-DAS 1 and (Z)-DAS 2, however, is strongly
reduced, and the spectrum is shifted by 26 nm with respect
to the E-isomer. The spectra of the isomer mixture of DAS
1
and DAS 2 show correspondingly lower molar extinction
in the wavelength region of maximum overlap with sunlight
than that of DSBP, which results in a slower photodegra-
dation for DAS 1 and DAS 2 than for DSBP. The presence
of DNOM did not affect the isomer distribution of the FWAs.
2
2 3 0
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 7, 1996

FIGURE 4. Photochemical degradation rate coefficient of DSBP as
a function of dissolved organic carbon (DOC) concentration using
Fluka humic acid solutions. Irradiation in MGRR at 366 nm.
FIGURE 5. Photochemical degradation rate coefficient of DSBP as
a function of phenol and aniline concentration. Irradiation in MGRR
at 366 nm.
half-length of the irridiated tubes (in the important range
of 350 nm this shielding is only 3%). The values for the
quantum yields at 366 nm in Lake Greifensee water are
therfore corrected bya factor of0.98. Apossible explanation
for the decrease of the degradation rate coefficients in the
presence of DNOM could be due to an association of the
FWA with DNOM, which could make the FWA less available
for degradation. For example, Milligan and Holt have
already shown that isomerization is hindered if FWAs are
bound on fabrics (23). No such associations could be
observed by means of absorption and fluorescence spec-
troscopy of solutions containing different concentration
ratios of DSBP and DNOM at pH 8.0 (24).
DOC Dependence of the Photodegradation Kinetics of
DSBP. To further investigate the influence of DNOM on
the degradation of DSBP, irradiations have been carried
out in the presence of different concentrations of dissolved
organic carbon (DOC). Figure 4 shows the pseudo-first-
order degradation rate coefficients as a function of DOC
using Fluka humic acid solutions. The degradation rate
coefficient of DSBP decreases if the DOC content is
increased. However, this deceleration is at most by a factor
of 2, a further decrease is not observed. Therefore, these
results do not concur with a single quenching mechanism
according to the Stern-Volmer equation. If quenching of
the excited states of DSBP by DNOM was the reason for the
decrease in degradation rate coefficient, a linear relation
between the inverse of this and the DNOM concentration
would be expected, because a higher concentration of
quencher would lead to a shorter lifetime of the excited
state, up to a level where the lifetime is too short to undergo
reactions and all excited molecules are quenched.
Since DNOM consists ofundefined compounds ofwhich
it is difficult to predict the chemistry, we used the well-
defined organic compounds 1-octanol, phenol, and aniline
to mimick DNOM. 1-Octanol, used to simulate the organic
carbon that is resistant to oxidation, showed no effect on
the degradation rate coefficients of DSBP. Phenol and
aniline were used to mimick the reductive properties of
DNOM. Figure 5 shows that phenol and aniline have an
analogous effect as DNOM on the degradation rate coef-
ficients of DSBP: with increasing phenol or aniline con-
centration the degradation rate is decreased, again to a
maximum factor of 2 as for DNOM itself. No significant
degradation of phenol and aniline could be observed. Since
we have already indicated that quenching may be excluded,
it seems plausible that a radical mechanism is involved in
the photochemical degradation of DSBP. Special attention
should be paid to the extreme low concentrations at which
phenol and aniline already have an influence: even at only
1
µM there is a noticeable interference in the reaction. Thus,
even if diffusion-controlled reaction with phenol or aniline
1
0
-1 -1
is assumed (rate constant of ∼10
M
s ), reactive
intermediates with lifetimes of at least 100 µs are involved
in the photodegradation of DSBP hindered by these
substances. These intermediates cannot be singlet or triplet
excited states of DSBP, which as those of stilbenes are much
shorter-lived (25), but may be radical intermediates.
Unidentified intermediates with lifetimes of over 100 µs
have been reported during laser-flash photolysis of DSBP
in air-saturated aqueous solution (26). In the same study,
one component was found of which the lifetime was
estimated to be several milliseconds.
Oxygen Concentration Dependence. Figure 6 shows
pseudo-first-order degradation rate coefficients as a func-
tion of the oxygen concentration. All FWAs are degraded
faster as the oxygen concentration is increased, the isomer
distributions are not affected. In Lake Greifensee water,
the degradation of DSBP in oxygen-saturated solution is
2
.9 and 5.8 times faster than in air- and nitrogen-saturated
solution, respectively. Noteworthy is that for DSBP only in
the absence of oxygen are the degradation rate coefficients
not influenced by the presence of DNOM. In oxygen- and
air-saturated solution the degradation proceeds slower in
DNOM-containing water. Obviously, DNOM inhibits only
reaction paths involving molecular oxygen. However, in
the presence of DNOM, the degradation rate under air or
oxygen atmosphere is not as slow as under a nitrogen
atmosphere. Thus, DNOM does not exclude all reactions
with oxygen but interferes only in some pathways. This
indicates that at least three reaction paths are involved in
the degradation of DSBP in air-saturared Lake Greifensee
water. It seems reasonable to assume that at least one is
included in the reactions which also proceeds in the absence
of oxygen and at least two other reactions involve oxygen.
At least one of the latter is inhibited by the presence of
DNOM.
DAS 1 is degraded 2.4 and 2.8 times faster if Lake
Greifensee water is saturated with oxygen as compared to
air- and nitrogen-satured solutions, respectively. None of
the three atmospheres showed a significant difference in
degradation rate coefficients between Lake Greifensee water
and DNOM-free water. The oxygen concentration depen-
VOL. 30, NO. 7, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2 2 3 1

FIGURE 6. Photochemical degradation rate coefficients of the FWAs as a function of the oxygen concentration in Lake Greifensee water
open) and DNOM-free water (closed). Irradiation in MGRR at 366 nm (DSBP) and with simulated sunlight (DAS 1 and DAS 2).
(
dence of the degradation rate coefficients of DAS 2 in Lake
Greifensee water was qualitatively similar to that of DAS 1,
although the rate coefficients in oxygen-saturated solution
were somewhat faster: 4.2 and 4.8 times faster than in air-
and nitrogen-saturated solutions, respectively. A possible
interpretation of the observed oxygen concentration de-
pendence would be that oxygen reacts with intermediates
generated upon excitation of the FWAs giving rise to stable
photodegradation products or precursors of them. These
intermediates might be excited states of the FWAs (singlet
or triplet) or radical ions derived from them. The present
results do not allow for a distiction between singlet and
triplet mechanisms. Both excited singlet and triplet states
of stilbene derivatives are enough short-lived (25) to expect
an approximately linear increase of the photodegradation
rate coefficient with increasing oxygen concentration.
Guglielmetti (32). With the aid of pure standards, the
corresponding carboxylic acids identified by Guglielmetti
during photolysis of DSBP could not be found. Although
we have not been able to identify the alcohol 2 so far, there
are indications that during phototransformation of DSBP
this water addition takes place. If DSBP is irradiated in the
absence of oxygen, an unidentified compound becomes
the main product. The same compound, which is not one
of the possible aldehydes, was also detected in the air-
saturated solution. Since photolysis of DAS 1 and DAS 2
in the absence of oxygen led to a water addition product,
it seems reasonable that for DSBP this addition would also
be possible. Because photochemical degradation of DSBP
had to be carried out at low concentrations (<10 µM) to
exclude the formation of dimers (cyclobutane compounds),
it was impossible to isolate the alcohol 2 in amounts
required for analytical identification. There are no indica-
tions that phenanthrene derivatives are formed during
photolysis of any of the FWAs. These compounds are often
found in photolysis of stilbenoid compounds by cyclization
of the Z-isomer followed by oxidation (dehydration) in the
presence of oxygen (33).
Photodegradation Products. Based on product analy-
sis, we propose for the three FWAs the photoreaction
scheme depicted in Figure 7. Photolysis of DAS 1 yields
mainly the water addition product 2, accompanied by the
aldehyde 1 and several unidentified minor products. This
result is in contrast with earlier studies which reported the
oxidative cleavage of the double bond yielding the aldehyde
Influence of DNOM on Product Form ation. Figure 8
shows the relative amount of the primary photoproducts
1b and 2 during irridation of DSBP in Lake Greifensee water
and DNOM-free water. In Lake Greifensee water, a greater
fraction of photoproduct 2 and a smaller fraction of the
aldehyde 1b are formed than in DNOM-free water. Thus,
the observed decrease in photodegradation rate coefficient
due to the presence of DNOM is accompanied by the
formation of smaller amounts of aldehyde and greater
amounts of the possible alcohol. These results support the
earlier suggestion that DNOM is hindering the reaction of
DSBP with oxygen, but not excluding it, and that at least
two reaction paths involving oxygen lead to aldehyde
formation.
1
as the main product (27-29). Photochemical water
addition at olefinic moieties has been known for a long
time (30, 31) but has no significant role in preparative
organic chemistry because of the very long irridation times
needed. In air-saturated Lake Greifensee water, the yield
of the alcohol 2 was up to 65% and that of the aldehyde
only 9%. The alcohol 2 has been identified after isolation,
described in the Experimental Section; the aldehyde 1 has
been identified by HPLC/ DAD comparison with a pure
standard. Photolysis ofDAS2 showed the same degradation
pattern as DAS 1, the alcohol 2 yield being 72%. The
dihydrated product in which the ethene moiety of DAS-
type FWAs is reduced to an ethane moiety, by abstracting
two hydrogen atoms from the solvent water (28), could not
be detected by comparison with the pure standard in our
studies.
This reaction model is also applicable to the photore-
action kinetics of DAS 1 and DAS 2 to explain why no DNOM
effect was observed for these FWAs. Their main primary
stable phototransformation product is formed in a reaction
that does not involve oxygen, hence experiments in air-
and nitrogen-saturated solutions showed only small dif-
ferences in kinetics and product formation. We therefore
assume that DNOM only interferes in the oxidative cleavage
of the double bond, which leads to the formation of the
aldehydes. Since the contribution of this product is small
for DAS 1 and DAS 2, even in an oxygen-saturated solution,
Photochemical degradation of DSBP proceeds mainly
by an oxidative cleavage of the double bond. The following
aldehyde products were identified by HPLC/ DAD com-
parison with pure standards: 2-sulfonic acid benzaldehyde
(
1a), 4-aldehyde-4′-(2-sulfostyryl)biphenyl (4-benzalde-
hyde-2′-sulfonic acid-stilbene) (1b), and 4,4′-bisaldehyde
biphenyl (1c), a photodegradation product of 1b. The
products 1a and 1c have already been reported before by
2
2 3 2
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 7, 1996

FIGURE 8. Relative formation of aldehyde 1b (squares, detected at
45 nm) and the photoproduct 2 (circles, detected at 310 nm) of
3
DSBP in Lake Greifensee water (open) and DNOM-free water (filled)
versus time. Irradiation in MGRR with simulated sunlight.
waters. Effective half-lives in the environment are longer,
due to decreased intensity of the terrestrial sunlight (time
of day, season, clouds, shading by trees, etc.) and due to
screening of sunlight by natural waters in deeper layers.
Despite the decreased light intensity, the relative difference
in degradation rate between DSBP and the DAS-type FWAs
will not change. The degradation of DSBP will still be
approximately 3 times faster. Neither will the decreased
light intensity change the composition of the isomer
mixtures.
In natural waters, the presence of DNOM leads to a
somewhat slower degradation of DSBP and to a lower yield
of the aldehyde products than in pure water. The kind of
DNOM does not seem to be very important, since we
observed the same effect in Lake Greifensee-derived DNOM,
Fluka Humic acid, and model compounds such as phenol
and aniline. Degradation of DAS 1 and DAS 2 is not
hindered by the presence of DNOM because for these
compounds the aldehyde formation is only a minor process.
The degradation of the FWAs due to reaction with DNOM-
derived photooxidants should be a negligible process in
natural waters since the presence ofDNOM did not increase
the degradation rate in our experiments.
For this study, we used only water that was free of
particles (diameter > 0.45 µm), and we showed that there
is probably no association of FWAs with DNOM. Still, the
influence of particles has to be studied since FWAs are
known to readily exhaust onto fabrics during the washing
process and adsorb onto sludge during wastewater treat-
ment. Therefore, our current research focuses on the
importance of natural suspended solids for the photo-
chemical degradation of FWAs.
FIGURE 7. Product formation of DAS 1, DAS 2, and DSBP during
photo-irradiation (see text for relative amounts).
the influence of DNOM has no noticeable effect on the
their photodegradation rate coefficients. The observed
enhanced photodegradation rate coefficient in oxygen-
saturated solution is mainly due to the formation ofreaction
products other than the aldehyde. The similarities in
kinetics and mechanisms of DAS 1 and DAS 2 during
photochemical degradation lead us to suggest that other
FWAs of the DAS-type will show a comparable behavior as
long as photochemical reactions involving the stilbene
moiety of the molecules occur.
Environm ental Relevance. In this study we have shown
that the three discussed FWAs are photochemically de-
gradable within times relevant for environmental aquatic
systems. The half-lives for photochemical degradation
given in Table 1 have to be considered as minimum half-
lives for the latitude of 47.5° N at the surface of natural
Acknowledgments
The authors thank R. Hochberg, Ciba-Geigy AG, Basel, for
providing the FWAs and their isomers and some of the
degradation products and for his contributions during
helpful discussions. We thank Ms. C. Schneider for
performing many of the experiments with DAS 2 and T.
Poiger for critically reading the manuscript. We gratefully
acknowlegde financial support of this work by Ciba-Geigy
AG, Basel.
VOL. 30, NO. 7, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2 2 3 3

(
(
(
19) Haag, W. R.; Hoign e´ , J.; Gassmann, E.; Braun, A. M. Chemosphere
984, 13, 631.
20) Canonica, S.; Kramer, J. B.; Reiss, D.; Gygax, H. R. Environ. Sci.
Technol., submitted for publication
Literature Cited
1
(
1) Siegrist, A. E.; Eckhardt, C.; Kaschig, J.; Schmidt, E. Ullmann’s
Encyclopedia of Industrial Chemistry; VCH: Deerfield Beach,
FL, 1991; Vol. A18, p 153.
21) Zepp, R. G.; Cline, D. M. Environ. Sci. Technol. 1977, 12, 327.
(
2) Anliker, R.; M u¨ ller, G. Fluorescent Whitening Agents; Coulston,
F., Korte, F., Eds.; Environmental Quality and Safety Vol. IV;
Thieme Publishers: Stuttgart, 1975.
(22) Hoign e´ , J. In Aquatic Chemical Kinetics; Stumm, W., Ed.; John
Wiley & Sons Inc: New York, 1990; Chapter 2.
(
(
(
3) Siegrist, A. E.; Hefti, H.; Meyer, H. R.; Schmidt, E. Rev. Prog.
Color. Relat. Top. 1987, 17, 39.
4) Poiger, T.; Field, J. A.; Field, T. M.; Giger, W. Anal. Methods
Instrum. 1993, 1, 104.
5) Kramer, J. B. In The Handbook of Environmental Chemistry;
Hutzinger, O., Ed.; Vol. 3F; de Oude, N., Ed.; Springer Publish-
ers: New York, 1992; p 351.
6) (a) Leaver, I. H. Aust. J. Chem. 1977, 30, 87. (b) Leaver, I. H.
Photochem. Photobiol. 1978, 27, 451.
(23) Milligan, B.; Holt, L. A. Aust. J. Chem. 1974, 27, 195.
(
24) Kramer, J. B. Ph.D. Dissertation, Swiss Federal Institute of
Technology, Z u¨ rich, Switzerland, in preparation.
(
25) G o¨ rner, H.; Kuhn, H. J. In Advances in Photochemistry; Neckers,
D. C., Volman, D. H., von B u¨ nau, G., Eds.; John Wiley & Sons:
New York, 1995; Vol. 19, Chapter 1.
(
26) Smit, K. J.; Ghiggino, K. P. Dyes Pigm. 1990, 13, 45.
(
(27) Leaver, I. H.; Milligan, B. Dyes Pigm. 1984, 5, 109.
(28) Holt, L. A.; Milligan, B. Text. Res. J. 1974, 44, 181.
(
(
7) Bode, K.-D. Tenside Deterg. 1975, 12, 69.
8) Schuessler, U. Fluorescent whitening agents for detergents; World
conference on detergents; American Oil Chemistry Society:
Champaign, IL, 1986.
9) Poiger, T. Ph.D. Dissertation, Swiss Federal Institute of Technol-
ogy, Z u¨ rich, Switzerland, 1994.
(
(
(
29) Needles H. L.; Seiber, R. P. Text. Res. J. 1974, 44, 183.
30) Stoermer, R.; Stockmann, H. Berichte 1914, 47, 1786.
31) (a) Wang, S. Y.; Apicella, M.; Stone, B. R. J. Am. Chem. Soc. 1956,
(
7
8, 4180. (b) Stoll, A.; Schlientz, W. Helv. Chim. Acta 1955, 38,
5
85.
(
(
10) Burg, A. W.;Rohovsky, M. W.;Kensler, C. J. CRCCrit. Rev. Environ.
Control 1977, 7, 91.
11) Ganz, C. R.; Liebert, C.; Schulze, J.; Stensby, P. S. J. Water Pollut.
Control Fed. 1975, 47, 2834.
12) Katayama, M. Nippon Kasei Gakkaishi 1989, 40, 1025.
13) Komaki, M.; Kawasaki, M.; Yabe, A. Sen-I Gakkaishi 1981, 37, 95.
14) Smit, K. J.; Ghiggino, K. P. Dyes Pigm. 1987, 9, 83.
15) Ikuno, H.; Honda, M.; Komaki, M.; Yabe, A. Nippon Kaqaku
Kaishi 1985, 8, 1603.
16) Zepp, R. G. Environ. Sci. Technol. 1978, 12, 327.
17) Canonica, S.; Jans, U.; Stemmler, K.; Hoign e´ , J. Environ. Sci.
Technol. 1995, 29, 1822.
(32) Guglielmetti, L. In Fluorescent Whitening Agents; Anliker, R.,
M u¨ ller, G., Eds.; Environmental Quality and Safety, Coulston,
F., Korte, F., Eds.; Thieme Publishers: Stuttgart, 1975; Vol. IV,
Chapter VI/ 6.
(
(
(
(
(33) Meier, H. Angew. Chem. 1992, 104, 1425.
Received for review September 22, 1995. Revised manuscript
received March 11, 1996. Accepted March 20, 1996.
X
(
(
ES950711A
(
18) Wegelin, M.;Canonica, S.;Mechsner, K.;Fleischmann, T.;Pesaro,
F.; Metzler, A. J. Water SRT-Aqua 1994, 43, 154.
X
Abstract published in Advance ACS Abstracts, May 15, 1996.
2
2 3 4
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 7, 1996