Photochemical transformations of tetrabromobisphenol A and related phenols in water

Article abstract of DOI:10.1016/S0045-6535(03)00704-5

A method was developed for studies of the phototransformation at UV irradiation of aqueous solutions of tetrabromobisphenol A (TBBPA), tribromobisphenol A (TriBBPA), tetrachlorobisphenol A (TCBPA), 2,4-dichlorophenol at various pHs as well as 2-chlorophenol, 2-bromophenol, 3,4-dichlorophenol and bisphenol A at pH 11. The absorbance spectra of the compounds and the emission spectra of the light-source were determined and used to calculate disappearance quantum yields of the photochemical reactions that were taking place. No major differences between the disappearance quantum yields of TBBPA and TCBPA were observed at pH 10, while the disappearance quantum yield of TriBBPA was approximately two times higher. The rate of decomposition of TBBPA was six times higher at pH 8 than at pH 6. Identification of the degradation products of TBBPA and TriBBPA, by GC-MS analysis and by comparison to synthesised reference compounds, indicated that TBBPA and TriBBPA decompose via different mechanisms. Three isopropylphenol derivatives; 4-isopropyl-2,6-dibromophenol, 4-isopropylene-2,6-dibromophenol and 4-(2-hydroxyisopropyl)-2,6-dibromophenol, were identified as major degradation products of TBBPA while the major degradation product of TriBBPA was tentatively identified as 2-(2,4-cyclopentadienyl)-2-(3,5-dibromo-4-hydroxyphenyl)propane.

Full text of DOI:10.1016/S0045-6535(03)00704-5

Chemosphere 54 (2004) 117–126
www.elsevier.com/locate/chemosphere
Photochemical transformations of tetrabromobisphenol A
and related phenols in water
*
ꢀ
Johan Eriksson , Sara Rahm, Nicholas Green, Ake Bergman, Eva Jakobsson
Departments of Environmental Chemistry and Structural Chemistry, Stockholm University, SE-106 91 Stockholm, Sweden
Received 17 September 2002; received in revised form 4 July 2003; accepted 4 July 2003
Abstract
A method was developed for studies of the phototransformation at UV irradiation of aqueous solutions of tetra-
bromobisphenol A (TBBPA), tribromobisphenol A (TriBBPA), tetrachlorobisphenol A (TCBPA), 2,4-dichlorophenol
at various pHs as well as 2-chlorophenol, 2-bromophenol, 3,4-dichlorophenol and bisphenol A at pH 11. The absor-
bance spectra of the compounds and the emission spectra of the light-source were determined and used to calculate
disappearance quantum yields of the photochemical reactions that were taking place. No major differences between the
disappearance quantum yields of TBBPA and TCBPA were observed at pH 10, while the disappearance quantum yield
of TriBBPA was approximately two times higher. The rate of decomposition of TBBPA was six times higher at pH 8
than at pH 6. Identification of the degradation products of TBBPA and TriBBPA, by GC–MS analysis and by
comparison to synthesised reference compounds, indicated that TBBPA and TriBBPA decompose via different
mechanisms. Three isopropylphenol derivatives; 4-isopropyl-2,6-dibromophenol, 4-isopropylene-2,6-dibromophenol
and 4-(2-hydroxyisopropyl)-2,6-dibromophenol, were identified as major degradation products of TBBPA while the
major degradation product of TriBBPA was tentatively identified as 2-(2,4-cyclopentadienyl)-2-(3,5-dibromo-4-
hydroxyphenyl)propane.
ꢀ
2003 Elsevier Ltd. All rights reserved.
Keywords: Tetrabromobisphenol A; Brominated flame retardants; Reactivity; Photodegradation
1
. Introduction
market demand for TBBPA was estimated at 120 000
tonnes per year in 2001 (BSEF, 2003).
Tetrabromobisphenol A (TBBPA) is primarily used
TBBPA and its dimethylated derivative have been
detected in sewage sludge and sediments in Sweden near
a plastic industry using TBBPA (Sellstr o€ m and Jansson,
1995) and also in sediments and fish in Japan (WHO,
1995). Bioconcentration factors ranging from 20 to 3200
have been indicated in accumulation studies on aquatic
invertebrates and vertebrates. The half-life in human has
been reported to be less than 2 days (Hagmar et al.,
as a reactive flame retardant, i.e. it is covalently bound
into mainly epoxy and polycarbonate resins used in the
manufacture of printed circuit boards for information
technology, communication and other electronic equip-
ment (WHO, 1995). It is also used as an additive flame
retardant to a minor extent and is then mixed with the
preproduced polymer (WHO, 1995). The world wide
2
1
000), <1 day in fish and <5 days in oysters (WHO,
995). Irritation, abnormal swimming behaviour and
reduced survival and growth have been observed in fish
exposed to TBBPA (WHO, 1995). TBBPA has also been
reported to be toxic to quail embryos, causing 80%
embryo mortality at a dose of 45 lg/g egg (Berg, 2000).
*
Corresponding author. Tel.: +46-8-163686; fax: +46-8-
63979.
E-mail address: johan.eriksson@mk.su.se (J. Eriksson).
1
0
045-6535/$ - see front matter ꢀ 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0045-6535(03)00704-5

118
J. Eriksson et al. / Chemosphere 54 (2004) 117–126
Due to TBBPA extensive use, bioconcentration poten-
tial in aquatic organisms and observed toxicity, it was of
interest to determine its photochemical decomposition in
water and to identify products formed.
with previously published data to study the robustness
of the suggested method.
A large number of organohalogen substances (OHS)
have been reported to undergo photochemical trans-
formations, both under experimental and natural con-
ditions as reviewed by M ꢁe allier (1999), Pagni and
Sigman (1999) and Richard and Grabner (1999). Ex-
tensive data exists on the identity of transformation
products of e.g. DDT (Mosier and Guenzi, 1969; Chu,
2. Experimental
2.1. Instruments and methods
1
H NMR spectra were recorded at room temperature
3
in CDCl , unless otherwise stated, with a JEOL EX270
1
instrument at 270 MHz. The H NMR for TriBBPA was
recorded at 400 MHz and with a digital resolution of
0.37 Hz.
1
1
999) and PCBs (Crosby and Moilanen, 1973; Waid,
986; Yao et al., 1997), and in a number of cases the
rates of transformation of selected OHS have been re-
ported (Kulovaara et al., 1995; Yao et al., 1997). Most
photolytic degradation in the environment occurs in the
atmosphere and in surface waters, however the low
volatility and low solubility of many OHS makes it
difficult to study their photolysis in these media in the
laboratory. To overcome these constraints a variety of
procedures for measuring the photolysis reactions of
OHS have been reported, although differences in the use
of solvents, apparatus and wavelengths can lead to dif-
ferent results being obtained for the same compound.
The requirements of the current study were for a simple
system for measuring the absorption spectrum, products
and rates of reaction under conditions that would be
relevant to the photolysis of these compounds in the
natural environment. Here we describe such a system
and we present some in-depth studies on TBBPA.
Melting points were determined with a B u€ chi 353
apparatus (Flawil, Switzerland).
The pH measurements were performed using a pH-
meter 718 Stat Titrino from Metrohm (Herisau, Swit-
a
zerland). Determinations of pK values were performed
under constant nitrogen flow according to the following
procedure. The compounds were dissolved in water. The
pH was adjusted to a value between 11 and 12 with
sodium hydroxide (2 M) and titrated with hydrogen
chloride (0.01 M) to pH 3.5. This was repeated five times
for each solution.
High-performance liquid chromatography (HPLC)
was performed on a Shimadzu (LC-4A) (Kyoto, Japan)
equipped with an UV detector (Shimadzu SPD-2AS)
and a C18 reversed-phase column (Supelcosil LC-18,
150 · 4.6 mm, 5 lm particles) from Supelco (Bellefonte,
USA). The detection wavelength was set at 230 nm and
the mobile phase was 0.1% trifluoroacetic acid in ace-
For phenolic compounds the rate of photodegrada-
tion is highly dependent on the pH (Richard and
Grabner, 1999) as the absorption of the associated and
dissociated forms can be very different. The present
method has been applied to aqueous solutions of bi-
sphenol A (2,2-bis(4-hydroxyphenyl)propane) (BPA),
tribromobisphenol A (TriBBPA), TBBPA, tetrachloro-
bisphenol A (TCBPA), 2-chlorophenol (2-CP), 2-bromo-
phenol (2-BP), 2,4-dichlorophenol (2,4-DCP), 3,
ꢀ1
tonitril:water (6:4). The flow rate was 2 ml min
.
Gas chromatography–mass spectrometry (GC–MS)
was performed on an ion trap GCQ instrument from
Finnigan Mat. The gas chromatograph was equipped
with a fused silica capillary column (DB-5, 30 m ꢁ 0.25
mm, 0.025 mm film thickness, J&W Scientific, Folsom,
USA). The temperature program was 80 ꢁC for 5 min,
ꢀ1
raised at 10 ꢁC min to 280 ꢁC which was held for 15
min. The injections were made in the splitless mode
using an injection temperature of 260 ꢁC. Helium was
used as carrier gas. Mass spectra were recorded in
electron ionisation mode at an ion source temperature of
140 ꢁC and an electron energy of 70 eV.
4
-dichlorophenol (3,4-DCP), 2,6-dibromophenol (2,6-
DBP), 2,4,6-tribromophenol (2,4,6-TBP) and 2,6-di-
bromo-4-isopropylphenol (2,6-DB-4-iPP) (c.f. Fig. 1).
Information on the photochemical properties of BPA,
TBBPA and TCBPA is limited although these com-
pounds are used in large quantities as monomers and
additives in plastic materials. On the other hand, pho-
tochemical reactions of more simple halophenols have
been quite extensively studied (Richard and Grabner,
The spectrophotometer was a double beam instru-
ment U 3000 from Hitachi (Tokyo, Japan). The scan-
ꢀ1
ning rate was 0.5 nm s with a slit width of 1 nm. UV
spectra were recorded in water for the determination of
disappearance quantum yields. The absorbance spectra
of TBBPA at different pH’s were recorded in a mixture
of ethanol:water (1:1).
1
999).
Our studies were focused on TBBPA, which is the
most commonly used organic brominated flame retar-
dant (OECD, 1994; WHO, 1995, 1997). TriBBPA was
included in the study as a potential degradation product
of TBBPA, and as a test compound slightly different
from TBBPA. 2-CP, 2-BP, 2,4-DCP and 3,4-DCP were
included mainly as reference compounds for comparison
2.2. Chemicals
0
0
0
0
3,3 ,5,5 -TBBPA, 3,3 ,5,5 -TCBPA, 2-BP and BPA
were purchased from Aldrich Chemical Company, Inc.

J. Eriksson et al. / Chemosphere 54 (2004) 117–126
119
H C
3
CH3
H C
3
CH3
Br
Br
Br
HO
OH
HO
OH
Br
Br
Br
Br
TBBPA
triBBPA
H3C
CH3
H3C
CH3
Cl
Cl
HO
OH
Br
HO
OH
Cl
Cl
TCBPA
BPA
OH
OH
OH
OH
Cl
Br
Br
Br
Cl
Br
Cl
Cl
2
,4-DCP
3,4-DCP
2,6-DBP
2,4,6-TBP
OH
OH
OH
Cl
Br
Br
Br
H C
3
CH3
2
-CP
2
,6-DB-4-iPP
2-BP
0
0
Fig. 1. Structures of the compounds tested for their rate of photochemical decomposition. The compounds are: 3,3 ,5,5 -TBBPA,
0
0
0
3
,3 ,5-TriBBPA, 3,3 ,5,5 -TCBPA, BPA, 2,4-DCP, 3,4-DCP, 2,6-DBP, 2,4,6-TBP, 2,6-DB-4-iPP, 2-CP and 2-BP.
0
(
Gillingham, UK). 2,4-DCP, 3,4-DCP and 4-isopropyl-
3,3 ,5-TriBBPA: BPA (1.0 g, 4.3 mmol) was dissolved
in acetic acid (400 ml) and bromine (0.7 ml, 13 mmol)
was added to the solution. The reaction mixture was
stirred at room temperature for 1.5 h. Water (200 ml)
was added, the mixture was neutralised with sodium
hydrogen carbonate and then extracted with dichlo-
romethane. The organic phase was dried over sodium
sulphate. The solvent was evaporated and the product
was purified on an open silica gel column (Merck, 0.040–
0.063 mm) with dichloromethane as the mobile phase.
The product was recrystallised in hexane. Yield 1.1 g
phenol were from Fluka Chemie AG (Buchs, Switzer-
land). 2-CP was from Riedel-De Ha €e n AG (Hannover,
Germany). Methanol, hexane, toluene, dichlorome-
thane, bromine, acetic acid, sodium sulphate, sodium
dihydrogen phosphate, sodium hydrogen carbonate and
hydrochloric acid (37%) were from Merck (Darmstadt,
Germany) and carbon tetrachloride from Prolabo
(
Fontenay, Canada). All solvents were of pro analysis
grade. The water was deionised and further purified
from lipophilic contaminants by extraction with hexane.
Open silica gel chromatography was performed on
Matrex silica (Millipore 35–70 lm particles) from
Molsheim (France) or Kieselgel 60 (0.040–0.063 mm)
from Merck.
(2.4 mmol, 56%); M
p
115–117 ꢁC, [lit 107–110 ꢁC]
(Hennis and Thomson, 1966); GC–MS: m=z (rel. ab., %
and fragment identity): 462 [4, M ], 447 [31, (M–
þ
þ
þ
þ
CH
3
) ], 368 [4, (M–CH
3
Br) ], 289 [11, (M–CH
3
Br
2
) ],

120
J. Eriksson et al. / Chemosphere 54 (2004) 117–126
þ
1
0
2
75, [5, (M–C
2
H
5
Br
2
) ]. H NMR d: 7.30 (H2 , d,
dissolved in water (100 ml, pH 9). The solution was
stirred at room temperature for 3 days and extracted
with dichloromethane. The product, 4-(2-hydroxyiso-
propyl)-2,6-dibromophenol was isolated in quantitative
0
J ¼ 2:2 Hz), 7.26 (H2, H6, s), 7.01 (H6 , dd, J ¼ 8:4, 2.2
0
Hz), 6.93 (H5 , d, J ¼ 8:4), 5.77 (OH, s), 5.41 (OH, s),
1
3
.59 (CH , s).
1
2
,6-DB-4-iPP: 4-Isopropylphenol (0.98 g, 7.2 mmol)
yield. H NMR spectra were recorded at room temper-
was dissolved in acetic acid (55 ml) and bromine (0.75
ml, 14 mmol) was added to the solution. The reaction
mixture was stirred at room temperature for 5 h. Water
was added and the solution was neutralised with sodium
hydrogen carbonate. The product was extracted with
dichloromethane. The organic layer was dried over so-
dium sulphate. The solvent was evaporated and the
product was purified on an open silica gel column with
dichloromethane:hexane (2:1) as the mobile phase. Yield
ature in CD
ab., % and fragment identity): 308 (5, M ), 293 [55, (M–
3
OD. GC–MS of the compound: m=z (rel.
þ
þ
1
3 3
CH
2.3. UV irradiation experiments
) ]. H-NMR d: 7.55 (Ar–H, s), 1.50 (CH , s).
Studies of the degradation under UV irradiation
conditions were performed in a cylindrical glass vessel
with a height of 480 mm, an outer diameter of 85 mm and
a volume of 1.6 l (Fig. 2). A fluorescent tube (Philips TL
20W/09N, Holland) was placed in the middle of the
cylinder, isolated from the reaction medium by a glass
wall. The spectral irradiance of the fluorescent tube
(shown in Fig. 3) was measured at the Swedish National
and Research Institute (Bor ꢀa s, Sweden). The rates of
decomposition were determined at various pH (5.5–11)
using the following general procedure: The compound (5
lmol) and sodium dihydrogen phosphate (0.052 mol)
were dissolved in purified, degassed (with nitrogen) water
(5 l). The pH was adjusted with sodium hydroxide (2 M).
An aliquot (1.6 l) of the solution was transferred to the
reaction vessel and the illumination was started. Samples
(10 ml) were collected after suitable times depending on
the half-lives, which were determined in a preexperiment.
The samples were analysed by HPLC and the amount of
unreacted compound was determined based on peak area
quantification. Studies of decomposition products were
performed with higher concentrations of TBBPA (77
lM) and TriBBPA (74 lM). The pH in these experiments
was 9 and the illumination time was 50 min. The pH was
then adjusted to pH 2 with hydrogen chloride (2 M). The
solution was extracted with toluene and the products
were analysed by GC–MS.
1
.8 g (6.3 mmol, 87%); GC–MS: m=z (rel. ab., % and
þ
þ
fragment identity): 292 (23, M ), 277 [52, (M–CH
3
) ],
Br) ]. H NMR d: 7.31 (H3, H5, s),
.73 (OH, s), 2.82 (CH, m), 1.22 (CH
, d), J ¼ 6:9 Hz.
-Isopropylene-2,6-dibromophenol: 2,6-DB-4-iPP (1.4
þ
1
1
98 [23, (M–CH
3
5
3
4
g, 4.6 mmol) was dissolved in carbon tetrachloride (50
ml). The solution was illuminated with UV-light. After 5
min, bromine (2.5 ml, 4.9 mmol) dissolved in carbon
tetrachloride (10 ml) was added in small portions. After
about 4 min, the UV-lamp was switched off. The reac-
tion mixture was light yellow colour, indicating that all
bromine had reacted. The solvent was distilled off. The
product (140 mg) was heated in an ampoule for 4 min at
2
70 ꢁC. The ampoule was opened and the residue was
purified on an open silica gel column using hexane:
dichloromethane (1:2) as the mobile phase. The major
product, 4-isopropylene-2,6-dibromophenol, was iso-
lated and characterised as follows, GC–MS: m=z (rel.
þ
ab., % and fragment identity): 290 (50, M ), 275 [15,
þ
þ
þ
(
M–CH
3
) ], 211 [6, (M–Br) ], 196 [17, (M–CH
H NMR d: 7.55 (Ar–H, s), 7.54 (Ar–H, s), 6.29 (OH, s)
.28 (CH, s), 5.06 (CH, s) 2.08 (CH, d).
-(2-Hydroxyisopropyl)-2,6-dibromophenol:
3
Br) ].
1
5
4
4-Iso-
propylene-2,6-dibromophenol (30 mg, 0.10 mmol) was
Fluorescent tube
Inner mantel
77 mm
4
5 mm
3
8 mm
Outer mantel
Fluorescent tube
Fig. 2. Apparatus used in the photochemical experiments. The inner diameter of the outer cylinder is 77 mm.

J. Eriksson et al. / Chemosphere 54 (2004) 117–126
121
Fig. 3. Spectral irradiance of the fluorescent tube used in the photodecomposition experiments measured at 1 cm distance. Two
wavelength intervals are shown, 250–400 and 285–325 nm, respectively.
kt
2
.4. Calculation of disappearance quantum yield
C
I
0
ð1 ꢀ eꢀ
Þ
U ¼ P
ð6Þ
max
k¼min
0;ktð1 ꢀ eꢀ
2:303eklC
Þ
The disappearance quantum yields were calculated as
described below with the assumption that the optical
density in the medium is low and the decrease in light
intensity thereby negligible. The quantum yield (U) is
defined as the ratio between the number of reacted
molecules per unit time and unit volume and the total
number of photons absorbed per unit time and unit
volume:
In these experiments k ¼ 260 and k
¼ 400 in 1 nm
min
max
intervals and C is regarded to be the same as C . The I
0
0
values have been corrected for the absorption by the
glass wall between the fluorescent tube and the reaction
liquid (Fig. 2). Since the reaction vessel was cylindrical
the light intensity (I ) decreases with distance from the
0
fluorescent tube. We calculated the radius (r ¼ 2:94 cm,
i.e. 1.04 cm from the surface of the fluorescent tube)
at which the amount of light absorbed is equal on
both sides of this radius (i.e. 2:25 < r < 2:94 cm and
Number of reacted molecules
U ¼
ð1Þ
Number of absorbed photons
2
:94 < r < 3:85 cm), and the quantum yield was conse-
If the reaction is assumed to be of first order:
quently calculated based on the I0;k values at this point.
This is also the radius at which the I values were mea-
0
sured.
ꢀ
ꢁ
C
t
ln
¼ ꢀkt
ð2Þ
C
0
ꢀ
1
where C
0
(mol l ) is the initial concentration, C
t
ꢀ1
ꢀ1
(mol l ) is the concentration after time t (s) and k (s ) is
a first order rate constant, the concentration of reacted
3
. Results
The uncertainty in the measurement of the irradiance
ꢀ
1
molecules C
¼ C ꢀ C
r
(mol l ) can be calculated as:
ð1 ꢀ eꢀkt
of the individual fluorescent tubes was ±15% for wave-
lengths < 320 nm and ±10% for wavelengths between
320 and 400 nm.
C
r
0
t
¼ C
0
Þ
ð3Þ
ð4Þ
The absorbance is given by
The emission from individual fluorescent tubes have
been generally observed to vary by not more than 10%.
The reproducibility of the photochemical degradation
method was tested over ten experiments, utilising dif-
ferent combinations of reaction vessel and fluorescent
tubes. In each experiment the first-order reaction con-
stant for degradation of TBBPA at pH 10 was mea-
sured. Values varied between 0.035 and 0.042 with mean
value of 0.038.
I
ln ¼ ꢀ2:303elC
I
0
ꢀ
1
1
ꢀ1
s ) is the intensity of the incoming
where I
0
(mol l
ꢀ
ꢀ1
light, I (mol l
s
) is the outgoing light intensity, e
ꢀ1
ꢀ1
(
l mol cm ) is the molar absorption coefficient, l (cm)
ꢀ1
is the cell path length and C (mol l ) is the concentra-
tion of the substance. The total amount of absorbed
photons Ia;k at wavelength k and for time t (s) is then:
The rates of photodecomposition, half-lives, absor-
bance spectra and disappearance quantum yields were
determined for TBBPA, TriBBPA, TCBPA, 2,4-DCP at
various pHs and are shown in Table 1. All experiments
on one compound at different pH were performed with
the same fluorescent tube and in the same reaction ves-
sel. No major differences between the disappearance
Þ ¼ tI0;kð1 ꢀ eꢀ2:303eklC
I
a;k ¼ tðI0;k ꢀ I
k
Þ
ð5Þ
The sum of this calculation for all wavelengths in the
emission spectra gives the total quantity of absorbed
photons. The final equation for the calculation of the
disappearance quantum yield is then obtained as

122
J. Eriksson et al. / Chemosphere 54 (2004) 117–126
Table 1
Physical properties of the studied compounds and experimental data for the UV irradiation experiments
Test
compounds
pKa1
pKa2
pH
Concen-
tration
Rate of
decomposition (min)
Half-life
Spectral absorption data Disappear-
ance
k
max
e
(cm
3
ꢀ1
(lM)
k ꢁ 10 (s
)
ꢀ1 ꢀ1
M )
quantum
yield (U)
(
nm)
a
a
TBBPA
7.6, 7.5
8.5
10
9
0.77
0.77
0.77
0.77
0.77
0.77
0.77
0.77
0.77
0.70
0.68
0.65
0.63
0.48
0.38
0.28
0.12
0.033
16
17
310
310
310
310
310
310
310
290
290
9170
9130
8600
8480
7310
6350
5310
2330
1040
0.045
0.042
0.043
0.042
0.044
0.039
0.033
0.031
0.018
8
7
7
7
6
6
5
.1
18
18
.9
.4
.1
.9
.1
.5
24
30
41
99
b
b
350
TriBBPA
7.8
7.5
11
0.74
0.74
0.74
0.74
0.74
0.74
0.78
0.77
0.75
0.50
0.33
0.16
15
15
15
23
35
69
1
0
9
8
7
6
306
7900
0.083
.5
.6
.6
.6
TCBPA
9.9
1.2
1.2
1.2
1.2
1.2
1.2
0.25
0.25
0.22
0.17
0.083
0.017
46
46
306
8760
0.040
8
7
7
6
5
.7
.9
.3
.4
.5
53
69
140
690
c
d
2,4-DCP
7.85, 7.85
11
9
1.0
1.0
1.0
1.0
1.0
1.0
0.32
0.28
0.23
0.17
0.10
0.017
36
41
305
3520
0.13, 0.10
.7
.9
.2
.5
.6
8
50
69
8
7
120
690
6
BPA
n.d.
8.29
11
11
11
11
10
10
10
1.0
1.9
1.8
1.0
2
0.020
0.10
0.062
0.15
0.13
0.73
0.35
690
115
187
77
293
295
293
303
302
316
308
3540
4010
3840
2780
4770
4640
5440
0.064
e
f
2
2
3
2
2
2
-BP
-CP
0.33, 0.29
0.32, 0.28
0.12, 0.11
0.096
g
f
8.49
8.58
h
d
,4-DCP
,6-DBP
,4,6-TBP
,6-DB-4-iPP
n.d.
6.31
87
15
i
1.1
2.9
0.046
0.046
n.d.
33
n.d. ¼ not determined.
a
WHO (1995).
Measured at 310 nm.
b
c
Hennis and Thomson (1966), WHO (1995).
Boule et al. (1984).
Vandenbelt et al. (1954).
d
e
f
Richard and Grabner (1999).
Mock and Morsch (2001).
Migchels and Zeegers-Huyskens (1995).
g
h
i
Broderius et al. (1995).
quantum yields of TBBPA and TCBPA at pH 10 were
observed, while the disappearance quantum yield of
TriBBPA was two times higher (Table 1). As can be seen
in Fig. 4, the reaction rates decrease with decreasing pH
the studied compound. A similar trend was observed for
the quantum yield of TBBPA although the difference
between the quantum yield of the associated and disso-
ciated forms is small.
for pH lower than the pK
remains fairly constant for pH higher than the pK
a
of the substance, while it
of
The same data were also collected for the following
substances at a single pH: BPA, 2-CP, 2-BP and 3,4-
a

J. Eriksson et al. / Chemosphere 54 (2004) 117–126
123
0
.05
.045
.04
.035
.03
.025
.02
.015
.01
Section 2) or commercially available compounds. TriB-
BPA decomposed to 3, 4 and a third major product 6
which gave a mass spectrum indicating a compound
containing two bromine atoms and with a molecular ion
at m=z ¼ 356. This compound is tentatively identified as
0
0
0
0
0
0
TBBPA
TriBBPA
TCBPA
2.4-DCP
0
0
2
-(2,4-cyclopentadienyl)-2-(3,5-dibromo-4-hydroxyphe-
nyl)propane (6). Fig. 6 shows the MS with ascribed
fragmentation ions. This product was also found in very
low yield in the photodecomposition of TBBPA. Com-
pounds 3 and 4 can originate from the same substance
since there seems to be an equilibrium in the water phase
between these two. 2,6-DB-4-iPP (5) degraded to the
compounds 7, 8 and 9.
0
.005
0
5.5
6.5
7.5
8.5
9.5
10.5
pH
Fig. 4. The correlation between the pH and the decomposition
rate of TBBPA, TriBBPA, TCBPA and 2,4-DCP.
DCP at pH 11, 2,6-DBP, 2,4,6-TBP and 2,6-DB-4-iPP at
pH 10 as shown in Table 1.
4. Discussion
Some of the major decomposition products of
TBBPA were identified by comparison to authentic
reference standards using GC–MS. The transformation
products identified were; 4-hydroxy-2,6-dibromophenol,
Photochemical degradation of chemicals in general
has been an area of research for a long time and of
environmental contaminants ever since the discovery of
their environmental occurrence. Much of the focus for
environmental contaminants has been on determination
of degradation products and routes of degradation, e.g.
2 in Fig. 5, 2,6-DBP (7), 4-(2-hydroxyisopropyl)-2,6-
dibromophenol (3), 4-isopropylene-2,6-dibromophenol
(
4), 2,6-DB-4-iPP (5) by reference to synthesised (c.f.
HO
H3C CH3
Br
Br
HO
OH
Br
Br
Br
Br
OH
2
?
HO
H C
CH3
H3C
CH2
CH3
3
H C
3
H C
CH3
3
Br
Br
+
HO
OH
Br
Br
Br
Br
Br
Br
Br
OH
OH
OH
3
4
5
1
H3C
CH3
H C
CH3
3
H C
CH3
3
Br
+
+
HO
Br
Br
Br
Br
OH
Br
OH
OH
OH
6
7
8
9
Fig. 5. Proposed degradation path for TBBPA.

124
J. Eriksson et al. / Chemosphere 54 (2004) 117–126
Fig. 6. Mass spectrum of the major degradation product from TriBBPA, and the proposed structure: 2-(2,4-cyclopentadienyl)-2-(3,5-
dibromo-4-hydroxyphenyl)propane (structure 6 in Fig. 5).
of DDTs (Mosier and Guenzi, 1969; Chu, 1999) and
PCBs (Crosby and Moilanen, 1973; Waid, 1986; Yao
et al., 1997). In the present study we have focused on
determination of quantum yields and rate constants for
the disappearance of chemicals present, or potentially
present, in the environment. Our intention was to in-
vestigate the likely photodegradation that would be
caused by sunlight at the Earth’s surface. For this we
used a fluorescent tube designed for sunbeds to represent
the range of solar UV wavelengths which penetrate the
full atmosphere. This approach contains several benefits
and some caveats. The tubes are cheap, reliable and
produce a spectrum that has relevant similarities to solar
UV spectra. The variability in the irradiance of the tubes
has generally been found to be acceptable and resulted
in total apparatus variability of <15% for different tubes,
different containing vessels and over periods of time.
Nevertheless, each tube needs to be checked before use,
and rechecked after long periods of use. Further, weak
transitions induced by higher wavelengths may intro-
duce error into estimations of the number of absorbed
photons. The weaker a compound absorbs in the 290–
the neutral form. There is a good correlation between
reaction rate and anionic concentration of mono-
hydroxylated aromatics. For 2,4-DCP the ratio between
the reaction rates at pH 11 and 6.6 is 19, while the ratio
between the anionic concentrations at these same pH’s is
a
calculated as 18.7 using a pK of 7.85. For aromatic
compounds substituted with two hydroxy-groups, such
as TBBPA, the situation may be more complicated. In
a
the case of TBBPA the difference between the two pK ’s
is small, indicating that the two rings may function, at
least in part, independently of each other. It is also quite
a
difficult to get accurate pK values when the experiments
are performed in water since the solubility of the proto-
nated form of this type of substance is very low.
Furthermore, although the two rings function indepen-
dently, the double phenolate compound is not necessary
twice as sensitive to light as the mono-phenolate form.
The half-lives given in Table 1 are calculated to in-
dicate the relative rate of decomposition of the com-
pounds tested and must not be used for anything but
relative comparisons within this study. However, the
relative rate of decomposition is most likely the same
even under different irradiation conditions but within
the same wavelength area leading e.g. to the conclusion
that TBBPA would be more rapidly transformed by
sunlight than TCBPA.
325 nm range, the more pronounced this effect may be
using our light source (and even for sunlight).
The disappearance quantum yields of 2-CP, 2-BP,
,4- and 3,4-DCP were comparable with previously re-
2
ported data (Table 1). As may be expected, the reaction
rate decreased with decreasing pH for pH lower than the
In general direct photolysis of halogenated aromatic
compounds in water proceeds via dehalogenation
(Richard and Grabner, 1999), although the subsequent
reactions/rearrangements depend on the position of
substitiuents. On the other hand BPA and related com-
pounds were found to decompose in a different way
pK
for pH higher than pK
a
of the substance, while it remains almost constant
(Fig. 4). This behaviour seems
a
reasonable since the absorption for the anionic form is
much higher and kmax is at a higher wavelength than for

J. Eriksson et al. / Chemosphere 54 (2004) 117–126
125
(
Fig. 5). For this type of compound the major primary
pounds. Ph.D. Thesis, Department of environmental toxi-
cology, Uppsala University, Sweden.
photochemical reaction involves cleavage between one
of the benzene rings and the isopropyl group as shown
both for TBBPA and TriBBPA. This mechanism of
decomposition has earlier been reported by (Ohko et al.,
Boule, P., Guyon, C., Lemaire, J., 1984. Photochemistry and
the environment. VIII. Photochemical behavior of dichloro-
phenols in a dilute aqueous solution. Chemosphere 13, 603–
6
12.
2
BPA on TiO
001) who investigated the photochemical behaviour of
.
Broderius, S.J., Kahl, M.D., Hoglund, M.D., 1995. Use of joint
toxic response to define the primary mode of toxic action for
diverse industrial organic chemicals. Environ. Toxicol.
Chem. 14, 1591–1606.
2
The TBBPA as commercial product is relatively pure,
but there is always small amounts of TriBBPA. The first
order reaction constant for the disappearance of TriB-
BPA in TBBPA is 0.034 which is somewhat lower than
from pure TriBBPA (0.046) and suggests that one of the
decomposition pathways of TBBPA could proceed via
TriBBPA. The uncertainty of the reaction constant of
TriBBPA in TBBPA is however very large due to the
low concentration. Even though a small part of TBBPA
might decompose via TriBBPA, the main decomposition
products found are those numbered 3, 4 and 5 in Fig. 5.
TriBBPA itself decomposes to 3 and 4 and also to a
compound that has tentatively been identified as 6.
Analogous cyclopentadiene derivatives have previously
been reported as degradation products of mono-ortho-
chlorinated phenols (Larson and Weber, 1994).
BSEF, 2003. Major brominated flame retardants volume
estimates.. Available from <http://www.bsef-site.com/docs/
BFR_vols_2001.doc>, Bromine Science and Environmental
Forum. Ref Type: Electronic Citation from April 10.
Chu, W., 1999. Photodechloronation mechanism of DDT in a
UV/surfactant system. Environ. Sci. Technol. 33, 421–
4
25.
Crosby, D.G., Moilanen, K.W., 1973. Photodecomposition of
chlorinated biphenyls and dibenzofurans. Bull. Environ.
Contam. Toxicol. 10 (6), 372–377.
Hagmar, L., Sj o€ din, A., H o€ glund, P., Thuresson, K., Rylander,
ꢀ
L., Bergman, A., 2000. Biological half-lives of polybromi-
nated diphenyl ethers and tetrabromobisphenol
exposed workers. Organohalogen Compd. 47, 198–201.
A in
Hennis, H.E., Thomson, L.R., 1966. Syntheses of some
0
brominated 4,4 -isopropylidenebisphenols. J. Chem. Eng.
Data 11, 135–136.
The present study represents a first step to develop a
more general concept on chemical reactivities including
also other types than photochemical reactions. Since
photochemically induced reactions do play a central role
for the fate of a large number of environmental con-
taminants we have prioritised this type of reactivity in
this study. Subsequent papers will deal with the photo-
chemical degradation of polybrominated diphenyl ether
congeners and other traditional environmental pollu-
tants.
Kulovaara, M., Backlund, P., Corin, N., 1995. Light-induced
degradation of DDT in humic water. Sci. Total Environ.
1
70, 185–191.
Larson, R.A., Weber, E.J., 1994. Reaction Mechanisms in En-
vironmental Organic Chemistry. CRC Press, Boca Raton,
FL.
M ꢁe allier, P., 1999. Phototransformation of pesticides in aque-
ous solution. In: Boule, P. (Ed.), Environmental Photo-
chemistry, vol. 2, part l. Springer Verlag, Berlin, Heidelberg,
Germany, pp. 241–261.
In conclusion, the described procedure provides ki-
netic and mechanistic data on the photodegradation of
an environmentally important contaminant, using an
inexpensive radiation source that emits a range of
wavelengths that are relevant to incident sunlight. The
data indicate that TBBPA is readily photodegraded in
aqueous solution, and highlight the sensitivity of de-
gradation rates to pH.
Migchels, P., Zeegers-Huyskens, Th., 1995. Infrared study of
the proton acceptor ability of metyrapone. J. Phys. Org.
Chem. 8, 77–83.
Mock, W.L., Morsch, L.A., 2001. Low barrier hydrogen bonds
within salicylate mono-ions. Tetrahedron 57, 2957–2964.
Mosier, A.R., Guenzi, W.D., 1969. Photochemical decomposi-
tion of DDT by a free-radical mechanism. Science 164,
1083–1085.
OECD, 1994. Selected brominated flame retardants. In: Risk
Reduction Monograph, vol. 3. OECD, Environment Direc-
torate, Paris, France.
Acknowledgements
Ohko, Y., Ando, I., Niwa, C., Tatsuma, T., Yamamura, T.,
Nakashima, T., Kubota, Y., Fujishima, A., 2001. Degra-
Dr. Peter Baeckstr o€ m are gratefully acknowledged
for valuable discussions and constructive comments on
the manuscript. The project was financially supported
by the Swedish Foundation for Strategic Environmental
Research (MISTRA) and by EU R&D program FIRE.
dation of bisphenol A in water by TiO
Environ. Sci. Technol. 35 (11), 2365–2368.
2
photocatalyst.
Pagni, R.M., Sigman, M.E., 1999. The photochemistry of
PAHs and PCBs in water and on solids. In: Boule, P. (Ed.),
Environmental Photochemistry, vol. 2, part l. Springer
Verlag, Berlin, Heidelberg, Germany, pp. 139–179.
Richard, C., Grabner, G., 1999. Mechanism of phototransfor-
mation of phenol and derivatives in aqueous solution. In:
Boule, P. (Ed.), Environmental Photochemistry, vol. 2, part
l. Springer Verlag, Berlin, Heidelberg, Germany, pp. 218–
240.
References
Berg, C., 2000. Environmental pollutants and the reproductive
system in birds: Developmental effects of estrogenic com-

126
J. Eriksson et al. / Chemosphere 54 (2004) 117–126
Sellstr o€ m, U., Jansson, B., 1995. Analysis of tetrabromobi-
sphenol in product and environmental samples.
Chemosphere 31 (4), 3085–3092.
Chemical Safety. World Health Organization, Geneva,
Switzerland.
A
a
WHO, 1997. Environmental Health Criteria 192. Flame retar-
dants: A general introduction. International Program on
Chemical Safety. World Health Organization, Geneva,
Switzerland.
Vandenbelt, J.M., Heinrich, C., Vanden Berg, S.G., 1954.
Comparison of pK
titration and ultraviolet absorption methods. Anal. Chem.
6 (4), 726–727.
a
values determined by electrometric
2
Yao, Y., Kakimoto, K., Ogawa, I., Kato, Y., Hanada, Y.,
Shinohara, R., Yoshino, E., 1997. Photodechlorination
pathways of non-ortho substituted PCBs by ultraviolet
irradiation in alkaline 2-propanol. Bull. Environ. Contam.
Toxicol. 59, 238–245.
Waid, S., 1986. PCBs and the Environment I–III. CRC Press,
Boca Raton, FL.
WHO, 1995. Environmental Health Criteria 172. Tetrabromo-
bisphenol A and derivatives International Program on