Chemical transformation of 3-bromo-2,2-bis(bromomethyl)-propanol under basic conditions

Article abstract of DOI:10.1021/es0495157

The mechanism of the spontaneous decomposition of 3-bromo-2,2- bis(bromomethyl)propanol (TBNPA) and the kinetics of the reaction of the parent compound and two subsequent products were determined in aqueous solution at temperatures from 30 to 70 °C and pH from 7.0 to 9.5. TBNPA is decomposed by a sequence of reactions that form 3,3-bis(bromomethyl)oxetane (BBMO), 3-bromomethyl-3-hydroxymethyloxetane (BMHMO), and 2,6-dioxaspiro[3.3]-heptane (DOH), releasing one bromide ion at each stage. The pseudo-first-order rate constant of the decomposition of TBNPA increases linearly with the pH. The apparent activation energy of this transformation (98 ± 2 KJ/mol) was calculated from the change of the effective second-order rate constant with temperature. The pseudoactivation energies of BBMO and BMHMO were estimated to be 109 and 151 KJ/mol, respectively. Good agreement was found between the rate coefficients derived from changes in the organic molecules concentrations and those determined from the changes in the Br^- concentrations. TBNPA is the most abundant semivolatile organic pollutant in the aquitard studied, and together with its byproducts they posess an environmental hazard. TBNPA half-life is estimated to be about 100 years. This implies that high concentrations of TBNPA will persist in the aquifer long after the elimination of all its sources.

Full text of DOI:10.1021/es0495157

Environ. Sci. Technol. 2005, 39, 505-512
The present study focuses on a specific organic halide,
Chemical Transformation of
3-bromo-2,2-bis(bromomethyl)propanol, or, as it is known
commercially, tribromoneopentyl-alcohol (TBNPA). TBNPA
is used as a reactive intermediate for high molecular weight
flame retardants, e.g. in and as a reactive flame retardant
polyurethanes, and is abundant in bromine chemical in-
dustries (4). The TBNPA is the most abundant semivolatile
organic pollutant in the vadose zone and local chalk aquitard
under industrial complex in the northern Negev Desert, Israel.
It forms a major threat to the downstream coastal aquifer.
Although little has been published about TBNPA, its potential
to be an environmental hazardness is inferred from its close
similarity to other compounds such as 2,2-bis(bromomethyl)-
propan-1,3-diol or as more widely used dibromoneopentyl-
glycol (DBNPG), which has been defined as carcinogenic (5).
Moreover, one of the TBNPA decomposition products found
in this study is 3,3-bis(bromomethyl)oxetane (BBMO). This
compound is similar to its chlorinated equivalent 3,3-bis-
(chloromethyl)oxetane (BCMO), which was included in the
USA Environmental Protection Agency’s (EPA) list of Ex-
tremely Hazardous Substances (EHS) (6). BBMO and BCMO
are known mainly as polymerization agents. The two halogen
groups in these molecules are replaced by nucleophilic groups
from the targeted compounds, causing polymerization or
cross-linking of polymer chains (7, 8). In the same way, BBMO
can react with amino acid residues in proteins, connect
different parts of the proteins, change their structure, and
damage their functioning. As bromide is an even better
leaving group than chloride (3, 9), BBMO may be even more
toxic than BCMO.
The high degree of substitution at C2 in TBNPA confers
resistance to biodegradation (10). Together with its high
solubility in water (2 g/L at 25 °C) (11), TBNPA is considered
among the most rapidly transported and minimally retarded
contaminants. This is further corroborated by the widespread
distribution of TBNPA in the heavily polluted aquitard located
beneath a large industrial complex in the Negev area of Israel
(12). The wide distribution of TBNPA in the study area and
the potential hazard posed by its degradation byproduct make
the study of its fate in the vadose zone and in groundwater
highly important.
The studied subsurface environment is composed of
fractured chalk aquitard. The groundwater is brackish (ionic
strength 0.15), magnesium chloride type, with a relatively
high Ca2+ and SO42- content, as is typical for arid areas. Water
movement in this aquitard is confined mainly to the fractures
due to the low permeability of the chalk matrix (12). Flow
through a fracture reduces water-rock contact, and the chalk
walls act as a buffer that maintains basic conditions with pH
ranging between 7 and 8.
3-Bromo-2,2-bis(bromomethyl)-
propanol under Basic Conditions
S H A I E Z R A , * , † S H I M O N F E I N S T E I N , †
I T Z H A K B I L K I S , ‡ E I L O N A D A R , † , § A N D
J I W C H A R G A N O R †
Department of Geological and Environmental Sciences,
Ben-Gurion University of the Negev, 84105, Israel,
Institute of Biochemistry, Food Science and Nutrition,
Faculty of Agricultural, Food and Environmental Quality
Science, The Hebrew University of Jerusalem, POB 12,
Rehovot 76100, Israel, and Department of Environmental
Hydrology and Microbiology, Zuckerberg Institute for Water
Research, Blaustein Institutes for Desert Research, Ben-Gurion
University of the Negev, Sede Boqer Campus 84990, Israel
The mechanism of the spontaneous decomposition of
3-bromo-2,2-bis(bromomethyl)propanol (TBNPA) and the
kinetics of the reaction of the parent compound and two
subsequent products were determined in aqueous solution
at temperatures from 30 to 70 °C and pH from 7.0 to 9.5.
TBNPA is decomposed by a sequence of reactions that form
3,3-bis(bromomethyl)oxetane (BBMO), 3-bromomethyl-3-
hydroxymethyloxetane (BMHMO), and 2,6-dioxaspiro[3.3]-
heptane (DOH), releasing one bromide ion at each stage. The
pseudo-first-order rate constant of the decomposition of
TBNPA increases linearly with the pH. The apparent activation
energy of this transformation (98 ( 2 KJ/mol) was
calculated from the change of the effective second-order
rate constant with temperature. The pseudoactivation
energies of BBMO and BMHMO were estimated to be 109
and 151 KJ/mol, respectively. Good agreement was
found between the rate coefficients derived from changes
in the organic molecules concentrations and those
determined from the changes in the Br- concentrations.
TBNPA is the most abundant semivolatile organic pollutant
in the aquitard studied, and together with its byproducts
they posess an environmental hazard. TBNPA half-life is
estimated to be about 100 years. This implies that high
concentrations of TBNPA will persist in the aquifer long
after the elimination of all its sources.
In this study we try to estimate the half-life of TBNPA in
the polluted aquitard. Extensive efforts have been devoted
in the last 20-30 years in an attempt to determine the
pathways and rates of decomposition of halogenated alkanes
and alkenes in aqueous solutions (1-3, 9, 13-19). Nucleo-
philic substitution and 1,2-elimination are usually considered
as the major pathways of degradation of these compounds.
Hydrolysis half-lives of a large number of organic halides
under environmentally pertinent conditions have been
summarized in various publications (3, 20-23).
Neopentyl halides cannot undergo 1,2-elimination due
to the absence of the â-hydrogen atoms. Their hydrolysis in
aqueous solutions is attributed to the SN1 mechanism and
was found to be very slow (24, 25). TBNPA formally is a
neopentyl halide, but its chemical behavior, as it for other
3-halo-2,2-disubstituted propanols (26-28), is completely
Introduction
The enormous variety of industrial chemicals produced today
possess a growing threat to the environment and challenges
our ability to protect it. Organic halides are among the most
important chemical pollutant groups. This group is rapidly
transported, resistant to degradation, and in many cases toxic,
carcinogenic, and mutagenic, and, therefore, even low
concentrations of organic halides can preclude the use of an
aquifer (1-3).
* Corresponding author phone: +972-8-6477515; fax: +972-8-
6472997; e-mail: shaie@bgumail.bgu.ac.il.
† Department of Geological and Environmental Sciences, Ben-
Gurion University of the Negev.
‡ The Hebrew University of Jerusalem.
§ Blaustein Institutes for Desert Research, Ben-Gurion University
of the Negev.
9
10.1021/es0495157 CCC: $30.25
Published on Web 12/08/2004
2005 American Chemical Society
VOL. 39, NO. 2, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 505

SCHEME 1
TABLE 1. Conditions and the Kinetic Coefficients Calculated from the Experimentsf
temp/°C
pH
buffer concn/M
k1obs/s
R2
k2obs/s-
k3obs/s
R2
(Br)
(1obs)
9.25
0.025c
0.025e
0.051c
3.3 × 10-5 a
0.9562
0.9905
0.9912
0.9740
0.9688
0.9523
0.9872
0.9592
0.9636
0.9654
0.9841
0.9784
0.9913
0.9688
0.9849
0.9904
0.9884
0.9806
0.9927
0.9804
0.9909
8.84
8.50
7.88
7.06
9.51
8.98
1.3 × 10-5 a
6.3 × 10-6 a
1.5 × 10-6 a
2.9 × 10-7 a
1.7 × 10-6 a
6.3 × 10-7 a
70
0.036c
0.030d
0.025c
0.050d
0.018c
0.056d
0.025c
0.025e
0.051c
1.7 × 10-6-2.7 × 10-6
3 ( 1 × 10-6
2 ( 1 × 10-6
50
30
1.6 × 10-7-2.6 × 10-7
1-1.3 × 10-7
6-9 × 10-8
8.59
8.05
9.73
0.036c
0.030d
0.025c
0.050d
0.025c
0.025e
0.051c
0.025c
0.050d
2.6 × 10-7 a
9.3 × 10-8 a,b
9.8 × 10-8 a
0.9684
0.9595
0.9831
0.9969
0.9922
0.9872
0.9940
9.17
8.29
3.2 × 10-8 b
0.9989
0.9673
1.3 × 10-8
4.3 × 10-9 b
f
a Calculated directly from the TBNPA concentrations. b Calculated from the Br- concentrations. c Borax. d KH2PO4. e NaOH. k2obs is presented
as a range; to that range we present the fitted k3obs and the correlation of this set to the bromide results (R2(Br)).
different. The transformation of 3-halo-2,2-disubstituted
propanols in basic solutions involves the intramolecular
nucleophilic substitution (Scheme 1a) and the so-called 1,4-
elimination (Scheme 1b).
by the manufacturer laboratory. In most of the experiments,
the initial concentration of TBNPA was 1000 mg/L (3.08
mmol/L). Determination of the kinetic constant for slow
transformation rates (e.g. low temperature and/or low pH)
was based on four batch experiments with different initial
TBNPA concentrations of 1000, 750, 500, and 250 mg/L.
In an attempt to isolate 3-(bromomethyl)-3-(hydroxy-
methyl)oxetane (BMHMO), the second daughter product of
TBNPA, DBNPG (Aldrich, 98% purity), was decomposed in
pH 11 (NaOH solution) and 30 ( 0.2 °C.
Analytical Methods. Extraction of Organic Compounds.
The main extraction method used was Solid-Phase Extraction
(SPE). Five milliliters of the experimental solution was passed
by vacuum through Envi-18 Tube (Supelco) that was previ-
ously washed with dichloromethane (DCM) and conditioned
with 2 mL of methanol followed by 2 mL of double distilled
water (DDW). The tubes were dried for 18 min under vacuum.
The dried tubes were washed with 2 mL of DCM, and internal
standards of biphenyl (BP) and diphenyl ether (DPO) were
added to the extract. The extract solution was concentrated
under a weak flow of dry nitrogen. Several samples were
extracted via Liquid-Liquid Extraction (LLE), for analysis of
the more volatile compounds, using DCM or chloroform-d1.
The BBMO and BMHMO for Nuclear Magnetic Resonance
(NMR) analysis were initially extracted using the SPE method.
After evaporating all the DCM, the extracts were dissolved
again by chloroform-d1. For the DOH, a more volatile
compound, the LLE method was conducted initially with
chloroform-d1.
The ratio between these two pathways depends on the
nature of the solvent and the structure of substituents in
position 2 of 3-halopropanols. The kinetics of these trans-
formations was not studied, and the available literature data
do not allow an estimation of TBNPA half-life under
environmental conditions. Nevertheless, TBNPA and its
byproducts are posing a major threat for a subsurface
saturated and unsaturated environment, including in the
study area, and the understanding of their fate under natural
conditions is crucial for sustainable environmental manage-
ment. In this study we investigate the kinetics and major
decomposition reactions of the TBNPA in moderate basic
solutions (pH ∼ 7-9) at different temperature (30-70 °C).
Methods
Experimental Setup. Batch experiments were carried out in
500 mL glass bottles, which were fully immersed in a water
bath held at constant temperatures of 30, 50, or 70 °C ( 0.2
°C and with a pH between 7.1 and 9.5 (Table 1). All the
experiments were stirred at 350-400 rpm using a submersed
stirring-plate. Solutions were prepared using KH2PO4/Borax
or NaOH/Borax buffer solution (reagent grade, BDH) (29)
(Table 1). To examine the effect of buffer concentration on
the reaction kinetics we conducted an additional set of
experiments at pH ∼ 9, 70 °C and three different buffer
concentrations, 0.025 M, 0.05 M, and 0.10 M. The experiments
were performed with TBNPA (<5% of impurities) provided
Analysis of Organic Compounds. A gas chromatograph
equipped with a flame ionization detector (GC-FID) was used
as the standard method for analysis of organic compounds.
9
506 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 2, 2005

One microliter of the extract sample was injected into a
Hewlett-Packard (HP) 5890 FID gas chromatograph with an
HP-5 fused silica column (30 m, 0.32 mm i.d., 0.25 mm film).
The injector and detector were held at 300 °C. The temper-
ature program was as follows: an initial stage 1 min at 70
°C followed by heating to 100 °C at 15 °C/min, followed by
isothermal hold for 1 min. The column was then heated to
126 °C at 1 °C/min and to 225 °C at 5 °C/min. The relative
response factor of TBNPA (0.20 area/area) and the recovery
(90%) were calculated based on preliminary calibration
measurements.
Several samples were also analyzed by gas chromatog-
raphy - mass spectrometry (GC-MS) for the identification
of daughter products. The extract sample was injected into
a gas chromatograph (HP 5890 GC) directly connected to a
mass spectrometer (HP 5971A MS ionization method - EI-
(70 eV)). The GC was equipped with a DB-5 fused silica
column (30 m, 0.25 mm i.d., 0.25 mm film). The temperature
program was usually similar to that of the GC-FID. To identify
FIGURE 1. An example of the change in solution composition with
time. The experiment was held at pH 8.5 and 70 °C. 0, 4, O, and
] represent the measured concentrations of TBNPA, BBMO,
BMHMO, and Br-, respectively. The lines represent the model
predictions.
assignment: δ1(13C) ) 77.86 and two doublets at δ1(1H) )
4.48 and δ1′(1H) ) 4.46 (A,A′ system, J(H,H′) ) 6.05 Hz) belong
to C, H, and H′ hydrogen atoms of the -CH2O- group; δ3-
(13C) ) 34.70 and δ3(1H) ) 3.80 belong to the -CH2Br group,
and δ2(13C) ) 64.37 and δ2(1H) ) 4.00 belong to the -CH2OH
group (30). The intensities of δ2(1H) and δ3(1H) are similar,
approximately half that of δ1(1H), implying a 1:1:2 -CH2-
Br:-CH2OH:-CH2O- fragments ratio in the molecule. All
fragments are bonded to the same quaternary carbon atom,
δ3(13C) ) 44.91. The MS spectrum of the intermediate is
consistent with the presence of one bromine atom in its
molecule: the isotope cluster with m/z ) 181/183 (1:1 ratio)
may be ascribed to M+1 ion and the isotope cluster with
m/z ) 150/152 (1:1 ratio) to an ion produced by oxetane ring
fragmentation with neutral loss of mass 30 (CH2dO fragment)
(see Figure 2, Supporting Information). These spectral data
suggest that the second TBNPA and first DBNPG decom-
position product have the structure of 3-bromomethyl-3-
hydroxmethyloxetane (BMHMO). The formation of the latter
from DBNPG and a base in an organic medium was reported
earlier (27, 28).
1
the daughter products we used NMR analysis. H- and 13C
NMR spectra were obtained at 500.1 MHz using a Bruker
DMX-500 NMR spectrometer. Chloroform-d1 was used as a
solvent and as an internal standard.
Analysis of Inorganic Compounds. Bromide ion measure-
ments were performed on a Dionex-600 high performance
liquid chromatograph (HPLC) with AS4A-SC column and
using Na2CO3/NaHCO3 as eluent (Dionex). The uncertainty
in measuring bromide ion was less than 5%. pH measure-
ments were performed on an Orion 720A pH-meter at
experimental temperature on an unstirred aliquot of solution
using BDH buffer standards 7, 9, and 11.
Results and Discussion
Decomposition of TBNPA and the Reaction Products. The
experimental results show that TBNPA slowly decomposes
in basic aqueous solutions via the formation of a sequence
of three products (BBMO, BMHMO, and DOH) and a
corresponding release of bromide ion (Figure 1). The
stoichiometric relation indicates that one Br- ion is released
to solution for each molecule that is decomposed, and
therefore the last daughter product (DOH) is a bromine-free
compound.
NMR analysis of the first daughter product shows two
signals of about the same intensity in the 1H NMR spectrum
and three in the 13C NMR spectrum. The COSY and 13C-1H
correlation spectra allow the attribution of theδ1(13C) ) 77.82
and δ1(1H) ) 4.43 signals to the CH2 group, bonded to an
oxygen atom and to a quaternary carbon atom (30). δ2(13C)
) 36.88 and δ2(1H) ) 3.86 belong to the -CH2Br group, also
bonded to a quaternary carbon atom. The δ3(13C) ) 45.00
identifies the quaternary carbon atom. With the two bromine
atoms unequivocally indicated by the GC-MS spectrum (see
Figure 1, Supporting Information), we infer that the molecule
contains two equivalent -CH2Br groups. The approximately
equal intensities of the two 1H NMR signals indicate that the
daughter product also contains two equivalent CH2O frag-
ments. The two CH2Br and two CH2O fragments are bonded
to the same carbon atom, δ3(13C). Two molecular structures
can consist with this description, DBNPG or BBMO. However,
NMR and GC-MS characterization of DBNPG (Aldrich)
performed by us are inconsistent with the NMR chemical
shift and GC-MS retention time obtained for the first daughter
compound, and, therefore, we attribute it to BBMO.
The third daughter product was identified to be 2,6-
dioxaspiro[3.3]heptane (DOH). Its formation from BMHMO
in an organic medium is described in ref 28. The GC-MS data
reveal that the molecule does not contain bromine atoms.
Moreover, the maximum ion mass of m/z ) 100, observed
in its mass spectrum, may be attributed to the molecular ion
of DOH and the base peak with m/z ) 70 to an ion produced
by oxetane fragmentation with neutral loss of mass 30 (CH2d
O fragment) (see Figure 3, Supporting Information). The only
1
signal observed in the H NMR spectrum, δ1(1H) ) 4.74, is
ascribed to CH2O- fragments of two oxetane rings in the
DOH structure (as was shown earlier for BBMO). The C13
NMR spectrum reveals two signals: the first at δ1(13C) ) 80.57
belongs to -CH2O- fragments of oxetane rings, and the
second δ2(13C) ) 35.72 is attributed to the quaternary carbon
atom.
Mechanism of TBNPA Decomposition Series. The de-
composition of TBNPA in basic aqueous solutions involves
release of bromide ion with sequential formation of three
products. These transformations are presented in Schemes
2-4. The decrease in the concentration of TBNPA coincides
with formation of the first daughter product, BBMO, and
with an increase in Br- concentration in the solution (Figure
1). The 1,4-elimination (Scheme 2b) is thus negligible. The
reaction rate strongly increases with the increase of tem-
perature and pH and is independent of the buffer concen-
tration. The concentration of BBMO increases with time,
reaches a maximum, and subsequently decreases, indicating
its instability under reaction conditions (Figure 1). The second
daughter product produced from decomposition of BBMO
was identified as BMHMO. This transformation involves the
substitution of one of the bromine atoms with an OH-group
The TBNPA second daughter product is produced in
relatively small amounts which is insufficient for a reliable
NMR analysis. However, GC-MS retention time and the mass
spectra indicate that TBNPA second hydrolysis intermediate
is the same compound as the first intermediate of DBNPG
decomposition. The NMR COSY and 13C-1H correlation
spectra of this intermediate allow the following signals
9
VOL. 39, NO. 2, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 507

SCHEME 2
from the data on the acidity of water (pKa∼15.7) and TBNPA
(pKa∼14, estimated using an online SPARC (33) calculator).
Two limiting cases are usually considered when analyzing
Scheme 2.
SCHEME 3
First limiting case: if k-1[H2O] , k2, the first stage is the
rate determining step, and the conditions of general base
catalysis are realized
dC
TBNPA ) -
ki[Bi-] ‚[TBNPA]
(1)
∑
dt
(Scheme 3). Two alternative pathways may be attributed to
this reaction, (1) a unimolecular reaction (SN1) or (2) a
bimolecular reaction (SN2), where both OH- and H2O may
play the role of a nucleophile. A possible indication for an
SN1 mechanism in hydrolysis of neopentyl halides in aqueous
solutions is the formation of rearranged products (24, 25).
No rearranged products were identified in the nucleophilic
transformation of the BBMO. However, their formation could
be precluded in our specific case due to stabilization of the
primarily obtained carbocation by the formation of 1,3-
bridged bromonium ion (31). Hence, the available data do
not allow us to decide what is the preferred mechanism for
the BBMO transformation.
BMHMO is also unstable in the basic aqueous solution
and transforms through loss of a bromide ion into the final
product, DOH (Scheme 4). The DOH could be detected clearly
by the GC analysis only when reaction products were
extracted by LLE, and we were unable to estimate its exact
yield. Thus, the formation of the 1,4-elimination product
with an exo-double bond (Scheme 4b) cannot be excluded.
Reaction Kinetics. TBNPA decomposition and formation
of BBMO (Scheme 2) is a two stage reaction. The first stage
is an acid/base equilibrium, in which an alkoxide (TBNPA-)
is formed from the parent alcohol. In the second stage,
TBNPA- predominantly undergoes an intramolecular nu-
cleophilic substitution (pathway 2a).
where Bi- is the concentration and ki is the constant of the
basic compound. A set of experiments performed in TBNPA
solutions with the same pH and different buffer concentra-
tions shows that the TBNPA decomposition rate is inde-
pendent of buffer concentration and/or composition and
their effects can be ignored. Thus, if the overall rate of TBNPA
decomposition is determined by the first step, Σki[Bi-] in eq
1 may be reduced to k1[OH-] (specific base catalysis).
Second limiting case: if k-1[H2O] . k2, then a fast acid-
base equilibrium precedes the second rate determining stage.
Because k1 and k-1 are of the same order of magnitude, the
steady-state conditions are not fulfilled: the concentration
of TBNPA- changes with the progress of the reaction. The
fraction of TBNPA- out of the total TBNPA ([TBNPA T] )
[TBNPA]+[TBNPA-]) is determined by
K
eq1[OH-]
1 + Keq1[OH-]
[TBNPA-]
[TBNPAT]
)
(2)
where Keq1 ) k1/k-1, and the overall reaction rate may be
described by
Keq1[OH-]
dC
TBNPAT
) -k2
‚[TBNPAT]
(3)
1 + Keq1[OH-]
dt
The general solution for the kinetic equations corre-
sponding to Scheme 2 was discussed in ref 32. According to
this solution the overall reaction rate is a complicated function
of the concentrations of different bases involved and is not
additive. In practice, several limiting cases are generally
realized in accord with the relations between rate constants.
The TBNPA decomposition belongs to a specific case where
k1 and k-1 are of the same order of magnitude; it follows
The observable rate constant will be proportional to the
concentration of OH-(specific base catalysis), when 1 .
K
eq1[OH-], and independent of it when 1 , Keq1[OH-].
The decomposition of TBNPA under basic conditions falls
into the second type of limiting cases. TBNPA and H2O belong
to the group of “normal” O-H acids (34). For the thermo-
SCHEME 4
9
508 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 2, 2005

FIGURE 2. An example of the variations in TBNPA (ln(TBNPAt/
TBNPA0) as a function of time. The experiment was held at 70 °C
and pH 8.84. The slope of the linear curve equals k1obs
.
dynamically favorable proton-transfer reaction between
“normal” acid and “normal” base the rate constant ap-
proaches the diffusion control limit kd ∼ 1010 M-1 s-1. For a
thermodynamically unfavorable process it will be equal to
about 1010.10-∆pKa M-1 s-1. Because TBNPA and H2O do not
differ much in their acidities (∆pKa is small) the rates of the
forward (k1) and reverse (k-1) proton-transfer reactions have
to be very high. For the mechanism with the first limiting
step (eq 1) the estimated reaction rate (1010 [OH-][TBNPA]M
s-1) is much higher than the experimentally observed. Hence
the mechanism involving fast acid-base equilibrium with
subsequent rate limiting intramolecular nucleophilic sub-
stitution is more likely for the TBNPA decomposition reaction.
Since the experiments were performed in a pH buffered
solution and at constant temperature, the activity of OH-
and Keq1 for each experiment should be constant throughout
the duration of the experiment. Therefore we can assume
that the observed (obs) kinetic coefficient, k1obs, integrates
all the kinetics-affecting factors and is equal to
FIGURE 3. Estimation of k1obs using the dependence of initial release
rate of Br- on TBNPA concentration in experiments conducted
under conditions in which the rate is slow (30 °C and pH 9.17).
FIGURE 4. The pH dependence of k1obs. ], 4, and 0 represent the
rate coefficient at 70, 50, and 30 °C, respectively.
and logk1obs is equal to
Keq1[OH-]
1 + Keq1[OH-]
k1obs ) k2
(4)
logk1obs ) pH - pKw + logKeq1 + logk2
(6)
where pKw is the water constant at the experimental
temperature. The concentration of TBNPA at any time during
an experiment can be predicted using the equation
Substituting eq 4 into eq 3 and solving for the concentration
of TBNPA gives
1obs‚t
[TBNPAT]t ) [TBNPA]0 ‚e-k
(5)
w/aH+)t
[TBNPAT]t ) [TBNPA]0 ‚e(-k′ ‚K
(7)
1
where the pH independent coefficient, k′1, is defined by
where t is time (sec) and the subscripts 0 and t refer to the
initial concentration and concentration at timet, respectively.
For each experiment the coefficient k1obs was calculated from
a least squares estimate of the slope of ln([TBNPAT]t/
[TBNPA]0) versus time (Figure 2) (18, 21). The excellent
linearity with high regression coefficient (Table 1) indicates
that, as anticipated, k1obs demonstrates the behavior of a
pseudo-first-order rate coefficient.
k1obs‚aH
k
+
k′1 )
)
1obs ) Keq1‚k2
(8)
Kw
aOH
-
For each experiment, the coefficient k′1 was calculated
using eq 8. The average k′1 values at 30, 50, and 70 °C are as
follows: 0.13 ( 0.01, 0.013 ( 0.002, and 0.0014 ( 0.0001 s-1
,
In experiments performed under low temperature and/
or pH conditions the variation in TBNPA concentration within
the experimental duration was small and within the range
of the analytical error. To calculate k1obs under these
conditions we performed a series of experiments with the
same pH, buffer concentration, and temperature and four
different initial TBNPA concentrations. The k1obs was cal-
culated from a least squares approximation of the slope of
the initial release rate of Br- versus TBNPA concentration
(Figure 3 and Table 1). This estimation is based on the
assumption that in the first stage of the experiment the release
of bromide ion is dominated by TBNPA decomposition.
Figure 4 shows that log(k1obs), obtained by both methods,
increases linearly with a slope ∼+1 (0.96 ( 0.11) as a function
of pH at all temperatures, indicating a first-order dependence
on the hydroxide ion in a pH ranging between 7 and 9.5.
Therefore we can assume that in this range Keq1[OH-] , 1
respectively, i.e., an order of magnitude rate variation for ∆T
) 20 °C.
The change in BBMO concentration with time is controlled
by its formation due to the decomposition of TBNPA (Figures
1 and 5; Schemes 2 and 3) and by its transformation via a
pseudo-first-order reaction
dC
BBMO ) k1obs[TBNPAT] - k2obs[BBMO]
(9)
dt
where k2obs is the kinetic coefficient observed. Solving eq 9
for initial conditions in which at t ) 0, [TBNPAT] ) [TBNPAT]0
and [BBMO] ) 0 gives
k1obs
2obst
1obs
[BBMO]t ) [TBNPAT]0
(e-k t - e-k
) (10)
k2obs - k1obs
9
VOL. 39, NO. 2, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 509

FIGURE 5. An example of the variation of BBMO concentration
with time (open diamond) at 70 °C and pH 8.5. The dashed and solid
lines represent possible fitting of the data to the proposed model
(see text).
FIGURE 6. An example of the variation of BMHMO concentration
with time (open diamond) at 70 °C and pH 8.5. The solid line is a
fitted line (see text).
where k3obs is the kinetic coefficient observed. Solving eq 12
for initial conditions in which at t ) 0, [TBNPAT] ) [TBNPAT]0
and [BBMO] ) [BMHMO] ) 0 gives
The time when [BBMO] concentration reaches a maxi-
mum value is calculated by
k2obs
ln
(
)
[BMHMO]t ) [TBNPAT]0 ‚k1obsk2obs
‚
k1obs
1obst
e-k
tBBMO(max)
)
(11)
k2obs - k1obs
+
+
(
(k2obs - k1obs)‚(k3obs - k1obs
)
2obst
e-k
Since k1obs values are known, k2obs can be resolved from
eq 11. For example, Figure 5 shows that at pH 8.5 and 70 °C
the maximum BBMO concentration was achieved after 2.9
( 0.1 days. Substituting this value and k1obs (6.3 ( 0.1 × 10-6
(k2obs - k1obs)‚(k2obs - k3obs
)
3obst
e-k
(13)
)
s-1) into eq 11 results in k2obs ) 2.3 ( 0.2 × 10-6 s-1
.
(k3obs - k1obs)‚(k3obs - k2obs
)
Unfortunately, in many of the experiments, the time of
maximum BBMO concentration could not be determined
accurately due to the limitations of the sampling intervals
and analytical complexities. Since BBMO could not be
isolated or purchased, we could not determine to a sufficient
accuracy the relative response factor (RF) between the
normalized measured GC peak area ([BBMO]PA) and the
BBMO concentration. To overcome this uncertainty and to
utilize the information in the “tail” and “front” of the curve,
the k2obs was estimated based on the overall variations in
[BBMO]PA with time (Figure 5). Fitting the ([BBMO]PA) data
obtained at each experiment according to eq 10 using a
mutual assumed RF value gives the concordant k2obs (Table
1). The resulting k2obs shows linear dependence on the RF
value used. The best fit for the experiments at 70 °C was
obtained for RF 1.2 and a corresponding average k2obs ) 1.5
( 0.2 × 10-6 s-1. However, some of the graphs show
considerable data scatter, and other combinations of RF
(between 0.9 and 1.9) and k2obs adequately describe the
experimental data. The range of RF values that adequately
describes most of the data at 50 °C is between 1.2 and 1.7.
As RF should not depend on the temperature of the
experiments we conclude that the RF is in the range of 1.2
to 1.7 (Figure 5). Using these RF values and the BBMO
measurements the values of k2obs at 50 and 70 °C are bracketed
to the ranges of 1.6 × 10-7-2.6 × 10-7 and 1.7 × 10-6-2.7
× 10-6 s-1, respectively. Regardless of the RF used, the k2obs
calculated in all these experiments confirms pH indepen-
dence.
In principle, the method used for the determination of
BMHMO RF and k3obs is similar to that used for BBMO above.
For each experiment at 70 °C, RF(BMHMO) and k3obs were
estimated initially by substituting the value of k1obs and the
lower and higher estimations of k2obs into eq 13 and fitting
the experimental data to the equation using nonlinear
regression (Figure 6). The resulting RF(BMHMO) and k3obs using
the lower k2obs (1.7 × 10-6 s-1) are 0.42 ( 0.06 and 3 ( 1 ×
10-6 s-1, respectively. The higher k2obs value (2.7 × 10-6 s-1
)
yields RF(BMHMO) of 0.27 ( 0.03 and k3obs 2 ( 1 × 10-6 s-1. Due
to lower BMHMO concentrations in the experiments carried
out at 50 °C, the uncertainty of their measurements is much
larger than that at 70 °C, and therefore less useful for
constraining the RF values. Since RF is independent of the
experimental temperature, we used the RF(BMHMO) obtained
at 70 °C to fit the BMHMO data at 50 °C (eq 13). Nevertheless,
despite the utilization of the RF70°C, the values of k3obs at 50
°C could only be bracketed to a range of 8 × 10-8-5 × 10-7
s-1 for the two k2obs values.
The precision of Br- measurement ((5%) is significantly
better than that obtained for the organic derivative ((20%).
Since most of the experiments were conducted until all the
bromide was released, and with high sampling frequency
(Figure 1), it is possible to use the change in bromide ion
concentration with time to validate the rate coefficients and
to narrow down the range of the estimated k3obs at 50 °C. The
bromide concentration at any time ([Br-]t) is given by the
following mass balance equation
The change in BMHMO concentration with time is
controlled by its formation due to the decomposition of
BBMO (Scheme 3) and its subsequent transformation via a
pseudo-first-order reaction (Scheme 4)
[Br-]t ) 3[TBNPAT]0 - 3[TBNPA]t - 2[BBMO]t -
dC
BMHMO ) k2obs[BBMO] - k3obs[BMHMO] )
[BMHMO]t (14)
dt
k1obs
2obst
1obs
k2obs[TBNPA]0
(e-k t - e-k ) -
k2obs - k1obs
Substituting eqs 7, 8, 10, and 13 into eq 14 and rearranging
gives
k3obs[BMHMO] (12)
9
510 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 2, 2005

[Br]t ) [TBNPA]0 ×
k′1‚Kw
+
+
w/aH
w/aH
2obst
3 - 3e-(k ′ ‚K
)t - 2
(e-(k ′ ‚K
)t - e-k
)
1
1
k′1m‚Kw
aH+‚ k2obs
-
(
)
aH
+
+
1m‚Kw/aH )t
2mt
3mt
e-(k ′
e-k
k′1‚Kw
e-k
k′1‚Kw
k′1‚Kw
-
k2obs
‚
+
+
aH
k′1‚Kw
k′1‚Kw
+
[
]
k
-
‚ k3obs
-
k2obs
-
‚(k2obs - k3obs
)
k3obs
-
‚(k3obs - k2obs)
(
)
2obs
(
) (
) (
)
(
)
aH
aH
aH
aH
+
+
+
+
(15)
For each experiment at 50 and 70 °C, TBNPA initial
obtained for k3obs is 151 kJ/mol. However, this is a rough
approximation as the estimation of k3obs at 50 and 70 °C is
associated with a considerable error range.
concentration, k′1 values (obtained from the TBNPA mea-
surements), the higher or lower k2obs values (derived from
the BBMO measurements), and different k3obs values (ob-
tained from the BMHMO measurements) were substituted
into eq 15. A good fit (R2 > 0.959) between the bromide
(Br)
ion concentration measured and that predicted by eq 15 was
obtained for all the 70 °C experiments (Table 1). For data
obtained at 50 °C we obtained an adequate fitting only by
narrowing the k3obs range between 6 and 9*10-8 s-1 and 1.0-
1.3 × 10-7 s-1 for the higher (2.6 × 10-7 s-1) and lower (1.6
× 10-7 s-1) estimated k2obs values, respectively (Table 1). A
good agreement was also obtained with the tBBMO,(max) that
was calculated with the obtained k2obs using eq 11 and the
BBMO experimental data (Figure 5).
Due to the similarity between the decomposition of
TBNPA (Scheme 2) and BMHMO (Scheme 4) the dependence
of k3obs on the OH- concentration should be similar to that
of k1obs (eq 16)
FIGURE 7. The temperature dependence of the rate coefficient,
k′1m, plotted as ln(k) vs the temperature reciprocal (in degrees Kelvin)
(see text).
Keq4[OH-]
1 + Keq4[OH-]
k3obs ) k5
(16)
The measured change in the concentration of TBNPA,
BBMO, and BMHMO can be successfully fitted to the
prediction of our proposed mechanism. By using the rate
coefficients that were estimated using the measurements of
the organic molecules, it is possible to predict the change in
Br- concentration with time that was independently mea-
sured during the experiments. This good agreement (Table
1 and Figure 1) provides additional support both for the
proposed mechanism and for the estimations of the rate
coefficients.
Field Observation. The area underlying the industrial
complex located in the northern Negev Desert is contami-
nated by an enormous variety of inorganic and organic
pollutants, many of which are yet to be identified. The chalk
comprising this aquitard dictates pH’s ranging between 7
and 8. Our experimental results predict that under these
conditions, TBNPA will decompose to give BBMO, which
will subsequently decompose to BMHMO and DOH. Fol-
lowing this prediction, GC-MS data from the monitoring wells
were examined, and it was found that one of the thus far
unidentified organic compounds in the polluted groundwater
is indeed BBMO. BBMO is not known among the chemicals
produced and/or used in this industrial complex. We propose
that the BBMO found in the studied aquitard is generated
as a byproduct of spontaneous decomposition of TBNPA.
Taking into account an average temperature of 18 °C and
an average pH range of 7 to 8, the TBNPA decomposition
rate k1obs should be on the order of 2 ( 3 × 10-11-2 ( 3 ×
10-10 s-1 and the TBNPA half-life about 100 years. This implies
that high concentrations of TBNPA will persist in the aquitard
long after the elimination of all sources.
where Keq4 ) k4/k-4
.
According to eq 16, if 1 . Keq4[OH-] k3obs will depend
linearly on OH- concentration, and if 1 , Keq4[OH-] k3obs will
be independent of [OH-]. Due to the relatively large
uncertainty accompanying the estimation of k3obs, we could
not distinguish between these two behaviors. The limited
range for possible variation in k3obs as a function of change
in OH- concentration compared to the up to 102 variation
observed for k1obs (Table 1) corroborates that 1 , Keq4[OH-].
Temperature Dependence. The temperature dependence
of the pH independent coefficient k′1 can be described using
the Arrhenius law
k′1 ) Keq1 ‚k2 ) A1obse-E′
(17)
a1obs/RT
where A1obs is the pre-exponential factor, E′a1obs is the apparent
activation energy, R is the gas constant, and T is the absolute
temperature. By plotting ln(k′1) vs 1/T, a pre-exponent value
of 2.3 × 1014(2 mmol/s and an apparent activation energy
of 98 ( 2 kJ/mol were calculated using eq 17 from a least-
squares estimation of the intercept and the slope of the plot,
respectively (Figure 7). The apparent activation energy
includes a contribution from the temperature dependence
of the acid/base equilibrium between TBNPA and OH- (Keq1),
in addition to the activation energy of the oxetane ring
formation (k2).
The transformation rate coefficient of BBMO (k2obs) does
not depend on pH. Hence, its temperature dependence can
be described by eq 17 simply by replacing k′1 with k2obs. The
pre-exponent and the apparent activation energy estimated
for k2obs are 8.6 × 1010 s-1 and 109 ( 7 kJ/mol, respectively.
Since these coefficients were derived from a range and not
from an explicit value, they should be considered as rough
estimates. An estimate of the apparent activation energy
Acknowledgments
We thank Prof. Ronit Nativ and her working group for
providing the monitoring field database and Dr. Zeev Ronen,
9
VOL. 39, NO. 2, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 511

(16) Barbash J. E.; Reinhard M. Abiotic dehalogenation of 1,2-
dichloroethane 1,2-dibromoethane in aqueous solutions con-
taining hydrogen sulfide. Environ. Sci. Technol. 1989, 23(11),
1349-1358.
(17) Cooper W. J.; Mehran M.; Riusech, D. J.; Joens, J. A. Abiotic
transformations of halogenated organics. 1. Elimination reaction
of 1,1,2,2-tetrachloroethane and formation of 1,1,2-trichloro-
ethene. Environ. Sci. Technol. 1987, 21(11), 1112-1114.
(18) Vogel, T. M., Reinhard, M. Reaction products and rates of
disappearance of simple bromoalkanes, 1,2-dibromopropane,
and 1,2-dibromoethane in water. Environ. Sci. Technol. 1986,
20(10), 992-997.
Prof. Zeev Aizenshtat, and Prof. Reginald Shagam for their
useful comments.
Supporting Information Available
GC-MS mass spectrum of the TBNPA the first daughter
product the BBMO (Figure 1), TIC GC-MS mass spectrum of
the TBNPA the second daughter product the BMHMO (Figure
2), and TIC GC-MS mass spectrum of the TBNPA the third
daughter product the DOH (Figure 3). This material is
available free of charge via the Internet at http://pubs.acs.org.
(19) Burlinson, N. E.; Lee, L. A.; Rosenblatt, D. H. Kinetics and
products of hydrolysis of 1,2-dibromo-3-chloropropane. En-
viron. Sci. Technol. 1982, 16(9), 627-632.
(20) Mabey, W.; Mill, T. Critical review of hydrolysis of organic
compounds in water under environmental conditions. J. Phys.
Chem. Ref. Data 1979, 7(2), 383-416.
(21) Vogel, T. M.; Criddle, C. S.; McCarty, P. L. Transformations of
halogenated aliphatic compounds. Environ. Sci. Technol. 1987,
21(8), 722-736.
(22) Roberts, A. L.; Jeffers, P. M.; Wolfe, N. L.; Gschwend, P. M.
Structure-reactivity relationships in dehydrohalogenation re-
actions of polychlorinated and polybrominated alkanes. Crit.
Rev. Environ. Sci. Technol. 1993, 23(1), 1-39.
(23) Larson, R. A.; Weber, E. J. Reaction mechanisms in environmental
organic chemistry; Lewis Publishers: Boca Raton, Ann Arbor,
London, Tokyo, 1994.
(24) Hughes, E. D.; Ingold, C. K.; Martin, R. J. L.; Meigh, D. F. Walden
inversion in unimolecular reactions of secondary and tertiary
alkyl halides. Nature 1950, 166, 679-680.
Literature Cited
(1) Bouwer, E. J.; McCarty, P. L. Transformations of halogenated
organic compounds under denitrification conditions. Appl.
Environ. Microbiol. 1983, 45(4), 1295-1299.
(2) Schwarzenbach, R. P.; Giger, W.; Schaffner, C.; Wanner, O.
Groundwater contamination by volatile halogenated alkanes:
Abiotic formation of volatile sulfur compounds under anerobic
conditions. Environ. Sci. Technol. 1985, 19, 322-327.
(3) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M.
Environmental organic chemistry; John Wiley & Sons, Inc.: New
York, 1993.
(4) Risk reduction monograph no. 3: Selected brominated flame
retardants. Background and national experience with reducing
risk. Environment directorate. OCDE/GE(94)96, Paris 1994.
(5) Dunnick, J. K.; Heath, J. E.; Farnell, D. R.; Prejean, J. D.; Haseman,
J. K.; Elwell, M. R. Carcinogenic activity of the flame retatrdant,
2,2-bis(bromomethyl)-1,3-propanediol in rodents, and com-
parison with the carcinogenicity of other NTP brominated
chemicals. Toxicol. Logic Pathol. 1997, 25(6), 541-548.
(6) Extremely Hazardous Substances (EHS) Chemical Profiles and
Emergency First Aid Guides. http://www.epa.gov/swercepp/
ehs/ehslist.html.
(7) Krespan, C. G. Mecroheterocycles. The oxetane function spiro
to macrcrocyclic polyether rings. J. Org. Chem. 1974, 31(16),
2351-2355.
(8) Krespan, C. G. The R effect in R-Chlorofluoro ketones. J. Org.
Chem. 1978, 43(4), 637-640.
(9) Jeffers, P. M.; Wolfe, N. L. Homogenius hydrolysis rate constants
- Part II: additions, corrections and halogen effects. Environ.
Toxicol. Chem. 1996, 15(7), 1066-1070.
(10) Ashby, J.; Callander, R. D.; Gilman, D. Lack of mutagenicity to
S. typhimurium of neopentyl bromide and pentaerythrityl
tetrachloride: relation to chemical structure. Mutat. Res. 1984,
140(2-3), 71-74.
(25) Ingold, C. K. Structure and mechanism in organic chemistry;
Cornell University Press: Ithaca and London, 1969.
(26) Searles, S.; Gortatowsky, M. J. Cleavage of 3-bromo-2,2-dimethyl-
1-propanol by base. J. Am. Chem. Soc. 1953, 75(12), 3030-3031.
(27) Searles, S.; Nickerson, R. G.; Wietziepe, W. K. Oxetanes. IX.
Structural and solvent effects in the reaction g-bromo alcohols
with base. J. Org. Chem. 1960, 24(12), 1839-1844.
(28) Abdun-Nur, A.; Issidorides, C. H. P. Entaerythritol derivatives.
V.1 Preparation of diethers of Pentaerythritol by reduction of
acetals and ketals. J. Org. Chem. 1962, 27, 67-70.
(29) Dean, J. A. Lange’s handbook of chemistry, 13th ed.; McGraw-
Hill: New York, 1985.
(30) Lozac’h, R.; Braillon, B. Analysis of nuclear magnetic resonance
spectra of thietane and oxetane. Chim. Phys. 1970, 67(2).
(11) Material safety data sheet MSDS; Bromine Compounds Ltd.:
1999.
(12) Nativ, R.; Adar, E. M.; Becker, A. Designing a monitoring network
for contaminated groundwater in fractured chalk. Ground Water
1999, 37(1), 38-47.
(13) Pagan, M.; Cooper, W. J.; Jones, J. A. Kinetic studies of
homogenious abiotic reactions of several chlorinated aliphatic
compounds in aqueous solutions. Appl. Geochem. 1998, 13(6),
779-785.
(14) Jeffers, P. M.; Brenner, C.; Wolfe, N. L. Hydrolysis of carbon
tetrachloride. Environ. Toxicol. Chem. 1996, 15(7), 1064-1065.
(15) Jeffers, P. M.; Ward, L. M.; Woytowich, L. M.; Wolfe, N. L.
Homogenius hydrolysis rate constants for selected chlorinated
methanes, ethanes, ethenes, and propanes. Environ. Sci. Tech-
nol. 1989, 23(8), 965-969.
(31) Reineke C. M.; McCarthy, J. R., Jr. 1,3-bromine migration. The
deamination of 3-bromo-2,2-bis(bromomethyl)propylamine. J.
Am. Chem. Soc. 1970, 92(21), 6376-6377.
(32) Bell, R. P. Proton in chemistry; Chapmann and Hall: London,
1973.
(33) SPARC on-line calculator. http://ibmlc2.chem.uga.edu/sparc.
(34) Eigen, M. Proton Transfer, Acid-Base Catalysis, and Enzymatic
Hydrolysis Part I: Elementary Processes. Angew. Chem., Int.
Ed. 1964, 3(1), 1-19.
Received for review March 30, 2004. Revised manuscript
received September 15, 2004. Accepted September 29, 2004.
ES0495157
9
512 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 2, 2005