Dechlorination of lindane, dieldrin, tetrachloroethane, trichloroethene, and PVC in subcritical water

Article abstract of DOI:10.1021/es011186k

Pure water has been used to dechlorinate aliphatic organics without the need for catalysts or other additives. Dehydrohalogenation (loss of HCl with the formation of a double bond) occurred at temperatures as low as 105-200 °C for 1,1,2,2-tetrachloroethane, lindane (1,2,3,4,5,6-hexachlorocyclohexane, γ-isomer), and dieldrin (1,2,3,4,10,10-hexachloro-6,7-epoxy-1,4,4a,5,6,7,8,8a-octahydro-endo, exo-1,4:5,8-dimethanonaphthalene). Complete loss of the parent compounds was achieved in less than 1 h at 150, 200, and 300 °C for 1,1,2,2-tetrachloroethane, lindane, and dieldrin, respectively. The initial dechlorination of lindane had an activation energy of 84 kJ mol^-1 with an Arrhenius pre-exponential factor of 1.5 x 10⁶ s^-1. Dehydrohalogenation of lindane formed trichlorobenzenes, followed by subsequent hydrolysis and hydride/chloride exchange to form chlorophenols, lower chlorobenzenes, and phenol as the major final product. Reaction of poly(vinyl chloride) at 300 °C for 1 h formed aromatic hydrocarbons ranging from benzene to anthracene and a char residue with a ca. 1:1 carbon-to-hydrogen ratio (mol/mol). The residue contained <1 wt % of chlorine compared to 57 wt % chlorine in the original polymer. All compounds tested yielded chloride ion as the major product (at higher temperatures), indicating that complete dechlorination of some aliphatic organochlorines may be feasible.

Full text of DOI:10.1021/es011186k

Environ. Sci. Technol. 2002, 36, 1337-1343
maintain the liquid state). In most of those studies, water
Dechlorination of Lindane, Dieldrin,
Tetrachloroethane, Trichloroethene,
and PVC in Subcritical Water
was employed only as an intermediate media, and the
degradation was related to reagents such as zerovalent iron
or other metals (1-6, 8, 9), a strong base (7), or oxidants (10)
which were added to the water. The majority of investigations
using pure water to degrade chlorinated organics have been
performed using supercritical water (11, 12). For example,
the decomposition of poly(vinyl chloride) at temperatures
ranging from 400 to 600 °C was reported (13). However, the
use of subcritical water may have advantages over super-
critical water (in addition to requiring lower temperatures
and pressures) in that dechlorination reactions may be
enhanced. For example, the dechlorination of methylene
chloride was much faster using subcritical water than when
using supercritical water (14).
A L E N A K U B AÄ T O V AÄ ,
A R N A U D J . M . L A G A D E C , A N D
S T E V E N B . H A W T H O R N E *
Energy and Environmental Research Center,
University of North Dakota, Campus Box 9018,
Grand Forks, North Dakota 58202
Recently, pure water at so-called subcritical conditions
has been used to extract and degrade high explosives and
pesticides from highly contaminated soils (15, 16), and a
preliminary study has demonstrated the dechlorination of
polychlorinated dibenzodioxins on incinerator fly ash (17).
These results are apparently based on the ability of subcritical
water to enhance the solubility of nonpolar organics by as
much as 5 orders-of-magnitude (18, 19), as well as promoting
some organic reactions (20).
The present study investigates the use of subcritical water
to dehalogenate aliphatic organochlorine compounds, in-
cluding 1,1,2,2-tetrachloroethane and its degradation product
1,1,2-trichloroethene, 1,9-dichlorononane, lindane (1,2,3,4,5,6-
hexachlorocyclohexane, γ-isomer), dieldrin (1,2,3,4,10,10-
hexachloro-6,7-epoxy-1,4,4a,5,6,7,8,8a-octahydro-endo, exo-
1,4:5,8-dimethanonaphthalene), and PVC.
Pure water has been used to dechlorinate aliphatic
organics without the need for catalysts or other additives.
Dehydrohalogenation (loss of HCl with the formation of a
double bond) occurred at temperatures as low as 105-200
°C for 1,1,2,2-tetrachloroethane, lindane (1,2,3,4,5,6-
hexachlorocyclohexane, γ-isomer), and dieldrin
(1,2,3,4,10,10-hexachloro-6,7-epoxy-1,4,4a,5,6,7,8,8a-octahy-
dro-endo, exo-1,4:5,8-dimethanonaphthalene). Complete
loss of the parent compounds was achieved in less than
1 h at 150, 200, and 300 °C for 1,1,2,2-tetrachloroethane, lindane,
and dieldrin, respectively. The initial dechlorination of
lindane had an activation energy of 84 kJ mol-1 with an
Arrhenius pre-exponential factor of 1.5 × 106 s-1
.
Dehydrohalogenation of lindane formed trichlorobenzenes,
followed by subsequent hydrolysis and hydride/chloride
exchange to form chlorophenols, lower chlorobenzenes, and
phenol as the major final product. Reaction of poly(vinyl
chloride) at 300 °C for 1 h formed aromatic hydrocarbons
ranging from benzene to anthracene and a char residue
with a ca. 1:1 carbon-to-hydrogen ratio (mol/mol). The residue
contained <1 wt % of chlorine compared to 57 wt %
chlorine in the original polymer. All compounds tested
yielded chloride ion as the major product (at higher
temperatures), indicating that complete dechlorination of
some aliphatic organochlorines may be feasible.
Experimental Section
Degradation Experim ents. Unless otherwise noted, all
reactions were performed using a static (no flow) 4-mL cell
constructed from an Inconel 600 (21) threaded (npt) pipe
fitting with end caps (63.5 mm long, 6.3 mm i.d.; Parker
Hannifin Corporation, Columbus, OH). Inconel cells were
selected on the basis of their inertness in corrosion tests
with subcritical and supercritical water (21). For each
experiment, 3 mL of water (HPLC grade; Fisher Scientific,
Pittsburgh, PA), which had been purged with nitrogen for ∼2
h to remove dissolved oxygen, was placed in the cell leaving
about 1 mL of headspace when the vessel was capped. With
this procedure, the internal pressure was governed by the
steam/ water equilibrium (i.e., pressures ranged from 1.3 to
170 bar for reactions performed over the 105 to 350 °C
temperature range used in this study (22)). Thus, all pressures
were substantially below the 428 bar rating of the Inconel
reaction cells. [Safety note: It is imperative to have sufficient
headspace in the vessel to ensure that interior pressure is
maintained by the steam/ water equilibrium to avoid excessive
pressures which can occur with a completely full cell (23).
At temperatures above 275 °C, the density of water drops so
that the cell is full of liquid water, which can cause pressures
above the gas/ liquid equilibrium (22). Therefore, the pro-
portion of water added to the cell should be <75% at higher
temperatures to ensure that adequate headspace remains).
Initial reactions were performed in water spiked with 10
µL of an acetone solution of tetrachloroethane (final con-
centration in water of 1 mM) and lindane (final concentration
of 0.57 mM). To aid in determining reaction products and
chlorine mass balance, experiments were also performed
with higher concentrations (i.e., near their saturation con-
centrations at ambient conditions). Neat compounds were
spiked in the water to obtain final concentrations of 16 mM
Introduction
Organochlorine compounds represent one of the largest
groups of anthropogenic compounds found in the environ-
ment. In addition to the concern for the environmental effects
of the parent compound (e.g., toxic effects of chlorinated
solvents and pesticides in groundwater), the presence of
organochlorines can inhibit the treatment and recycling of
waste streams. For example, poly(vinyl chloride) (PVC) is a
major source of chlorine for chlorinated dioxin formation
during the incineration of municipal wastes. A number of
studies have investigated degradation of chlorinated pol-
lutants via reductive dehalogenation to transform chlorinated
organics into their non-chlorinated analogues (1-10). Recent
degradation experiments involving chlorinated aliphatic (7)
and aromatic compounds (8-10) have been performed in
supercritical water (i.e., at temperatures and pressures higher
than the critical point of 374 °C and 221 bar) and subcritical
water (defined as hot water with sufficient pressure to
* Corresponding author phone: (701)777-5256; fax: (701)777-5181;
e-mail: shawthorne@undeerc.org.
9
10.1021/es011186k CCC: $22.00
Published on Web 02/07/2002
2002 Am erican Chem ical Society
VOL. 36, NO. 6, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 1 3 3 7

for tetrachloroethane, 8.9 mM for lindane, and 19 mM for
trichloroethene. Degradation experiments were also per-
formed on 1,9-dichlorononane and dieldrin, with both at
concentrations of 9 mM. PVC studies used 100 mg of powder
(number average molecular weight, Mn, ca. 35 000; weight
average molecular weight, Mw, ca. 62 000; supplied by Aldrich,
Milwaukee, WI) in 3 mL of water for each test.
Reaction cells were heated in a GC oven (Hewlett-Packard
model 5890). Because the heating rate of the interior solution
was not initially known, a cell was modified by replacing the
min-1 ramp to 320 °C. Quantitations were based on the
relative response factors of individual standards to the
internal standard. Similarly, the identification and quanti-
fication of degradation products were based on comparisons
of mass spectra and retention data of pure standards with
those of the sample species, unless otherwise noted.
Solid-phase microextraction (SPME) with 7 µm poly-
(dimethylsiloxane) (PDMS) fibers was used (24) in an attempt
to identify organics which were too volatile to be determined
in the methylene chloride extracts. The analytes were directly
recovered with SPME from the product water and the
headspace above the water followed by direct thermal
desorption in the injection port of the GC/ MS.
1
1
1
cap with a
/
in. npt × / 8 in. tubing fitting to allow a
/
8
2
diameter thermocouple to be placed directly into the water
inside the cell during the heating step. This modified cell
was used to determine the optimal heating procedure for
the subsequent reaction cell studies.
Chloride Ion Determ ination. The determination of
chloride ion content in the product waters from the reactions
using the higher concentrations of test compounds was
performed with a chloride ion selective electrode (Cole-
Parmer, Vernon Hills, IL). All determinations were performed
in triplicate or quadruplicate.
When the GC oven was heated from room temperature
after placing the cell in the oven, the heating rate of the
water inside the cell was quite slow (e.g., 15 min was required
for the water temperature to reach 240 °C with an oven
setpoint of 250 °C). Because more rapid heating of the reaction
water was desired, the procedure was modified so that the
oven was first preheated to 50 °C above the desired
temperature, and then the cell was quickly inserted into the
oven through the oven door. The GC oven temperature was
maintained at 50 °C above the desired temperature until the
cell interior temperature reached ca. 10 °C below the desired
temperature, and then the oven temperature was lowered to
the desired reaction temperature. With this procedure, the
interior cell temperature monitored by the thermocouple
reached within 10 °C of the desired temperature in 3-6 min
(for 150 and 300 °C reaction temperatures, respectively) and
never exceeded the desired temperature by more than 2 °C.
For example, when the GC oven was preheated to 250 °C (for
a desired reaction temperature of 200 °C), the interior cell
temperature reached 190 °C in 4 min. At this time, the oven
temperature was lowered to 200 °C (and was designated time
zero for the reaction studies). After the desired heating times,
the hot cells were immediately removed from the oven and
immersed in tap water to quench the reaction.
After cooling, each cell was carefully opened and its
contents transferred to a 7-mL vial. Organics were then
extracted with three 1-mL aliquots of methylene chloride.
1-Chloronaphthalene was added as an internal standard at
a concentration similar to the starting material, and the
extracts were analyzed by gas chromatography with mass
spectrometric detection (GC/ MS). Because the methylene
chloride extracts of the product water from the lindane
reactions showed phenolic compounds, the water from the
lindane reactions was acidified with 0.5 M HCl (to pH ∼ 2)
before methylene chloride extraction. To determine the
quantitative efficiency of this method, the identical procedure
(i.e., every step from spiking the water with the test compound
in the cell to the methylene chloride extraction and GC/ MS
analysis) was applied for each test compound except that
the cell was not heated. On the basis of a minimum of
triplicate cells, the recoveries of unreacted test compounds
were typically 95 ( 10%.
Results and Discussion
Lindane. Lindane was selected as a test compound for two
reasons. First, it is a widespread toxic pollutant (25, 26).
Second, our results obtained in pure subcritical water could
be compared with the data previously obtained in hot water
under alkaline conditions (7).
As shown in Table 1, the rate of lindane removal rapidly
increased with the increase in temperature from 125 to 200
°C. Using the data presented in Table 1, linear regressions
were obtained for the logarithm of lindane concentration
(mean values) versus reaction time (as shown in Figure 1,
top). Linear regression correlation coefficients (R2) were
0.6716, 0.7585, 0.8582, and 0.8384 for the 125, 150, 175, and
200 °C reactions, respectively. On the basis of the rate
constants for the initial dechlorination of lindane shown in
Table 1, the half-lives (t1/ 2) of lindane at different temperatures
decrease from ca. 1000 min at 125 °C to 12 min at 200 °C.
At 250 °C, the removal of lindane was completed in 15 min;
therefore, it was not possible to obtain its rate constant and
half-life.
On the basis of the data from Table 1 and Figure 1 (top),
activation energies were calculated assuming a pseudo-first-
order reaction (excess of water) and assuming a second-
order reaction where the concentration of hydroxide ion is
based on the dissociation constant of water at the different
temperatures (27). For the pseudo-first-order assumption,
the activation energy for the initial dechlorination of lindane
in pure water was 84 ( 24 kJ mol-1 (95% confidence interval
based on linear regression analysis), with an Arrhenius pre-
exponential factor A of 1.50 × 106 s-1. The second-order
assumption yielded an activation energy of 69 ( 24 kJ mol-1
and a pre-exponential factor of 1.40 × 1010 L mol-1 s-1. These
values correspond to the second-order activation energies
of 86 kJ mol-1 reported by Cristol et al. and 85 kJ mol-1
reported by Ngabe et al. for the alkaline dehydrohalogenation
of lindane at 30 °C (28, 29).
Degradation Experim ents in Glass. To verify the inertness
of Inconel cells, selected experiments were performed in 3-mL
reaction glass V-vials with Teflon stoppers (Aldrich) at
temperatures of up to 150 °C (the maximum possible without
leaks).
GC/MS Analyses. Analyses were performed using a
Hewlett-Packard model 5890 GC with a Hewlett-Packard
model 5972 MS in the full-scan mode (50-400 m/ z), with
electron impact ionization. Chromatographic separations
were accomplished using a 30 m DB-5 column with a 0.25
mm i.d. and 0.25 µm film thickness (J&W Scientific, Rancho
Cordova, CA), with injections in splitless mode. The oven
temperature was held at 40 °C for 1 min followed by a 10 °C
The degradation mechanism of lindane is complicated
because of the high number of reaction products and
intermediates identified in the product water. Possible
degradation pathways are presented in Figure 2, and
quantitative data on the major reaction products are
presented in Figure 3. The first degradation intermediate,
pentachlorocyclohexene, was presumably formed by elimi-
nation (loss of HCl with the formation of a double bond, or
“dehydrohalogenation”) (30). Alternatively, the reaction may
occur by displacement of Cl- by OH- followed by rapid
dehydration (31). Hydrolysis of the pentachlorocyclohexene
yielded two isomers of hydroxytetrachlorocyclohexene at 150
°C, but both compounds accounted for less than 5% of the
9
1 3 3 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 6, 2002

TABLE 1. Effect of Time and Water Temperature on the Dehalogenation of Lindane (0.57 mM)
a
% degraded ( SD
time (min)
125 °C
150 °C
21 ( 7
175 °C
30 ( 8
200 °C
61 ( 7
250 °C
15
20
25
30
35
40
45
60
75
95
120
360
840
11 ( 3
ndb
c
54 ( 1
72 ( 3
90 ( 2
99 ( 1
34 ( 11
45 ( 8
45 ( 8
11 ( 1
37 ( 8
nd
nd
10 ( 3
45 ( 20
60 ( 2
68 ( 14
14 ( 5
55 ( 15
80 ( 10
50 ( 8
nd
nd
nd
38 ( 1
1080
61 ( 13
(1.4 ( 0.6) 10-5
1030
nd
nd
d
k (s-1
)
(8.2 ( 2.8) 10-5
(2.1 ( 0.4) 10-4
72
(8.6 ( 3.6) 10-4
t1/2 (m in)e
190
12
a
b
Standard deviations were calculated from triplicate degradation experim ents. nd: not detected; the detection lim it for lindane was 0.2% of
c
d
e
the original m ass. Not determ ined. Rate constants (k) at different tem peratures. Half-lives at different tem peratures.
FIGURE 1. (Top) Effect of temperature on degradation of lindane
FIGURE 2. Possible degradation pathways of lindane. The isomers
in boldface are typically found in ∼10-fold higher concentrations
than other isomers shown. All isomers for which a specific structure
is shown were verified by GC/MS retention times and mass spectra
of pure standards.
(0.57 mM) in subcritical water, where C is initial concentration of
0
lindane and C is actual concentration of lindane. (Bottom) Arrhenius
plot (pseudo-first-order) of the reaction rates for lindane degradation.
original lindane content. At higher temperatures, neither
isomer was observed.
detected was phenol, which accounted for 18 mol % of the
total lindane after 18 h at 350 °C. However, even phenol
continues to degrade, as demonstrated by ca. 50% loss after
1 h at 250 °C when phenol was used as the test compound.
On the basis of the isomer distribution of the products,
trichlorobenzenes with two adjacent chlorine atoms are
preferentially formed via dehydrohalogenation of lindane.
Regardless of reaction temperature, 1,2,4-trichlorobenzene
was the most abundant isomer, while 1,2,3-trichlorobenzene
was found at order-of-magnitude lower concentrations, and
1,3,5-trichlorobenzene was not observed. The same prefer-
The major reaction pathway for pentachlorocyclohexene
was further dehydrohalogenation to form trichlorobenzenes,
especially the 1,2,4-substituted isomer. Additional reactions
formed dichlorophenols, either via hydrolysis of trichlo-
robenzenes or dehydrohalogenation of hydroxytetrachlo-
rocyclohexenes. Dechlorination continued with the formation
of dichloro- and monochlorobenzenes, presumably by
H-/ Cl- exchange. Monochlorophenols were also produced,
by either the hydrolysis of dichlorobenzenes or from di-
chlorophenols by H-/ Cl- exchange. The final major product
9
VOL. 36, NO. 6, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 1 3 3 9

FIGURE 3. Degradation of 0.57 mM lindane and formation of dechlorination products in subcritical water at different temperatures. Standard
deviations presented are based on triplicate experiments.
ence was also reported previously in alkaline subcritical water
(7) and alkaline dehydrohalogenation at ambient temperature
(32). Subsequent dechlorination preferentially forms 1,4-
dichlorobenzene and 2,5-dichlorophenol, as opposed to their
other isomers (Figure 2).
with the hydroxide ion concentration raising to 2.2 × 10-6
M at 200 °C to a maximum of 2.5 × 10-6 M at 250 °C and then
dropping to 2.2 × 10-6 M at 300 °C.
Because dehydrohalogenation generates HCl, the hy-
droxide ion concentration actually drops much below that
of room temperature (pure) water during dechlorination.
For example, with the 9 mM solution of lindane at 200 °C,
only 1% dechlorination is required to form enough HCl to
reduce the hydroxide concentration to ca. 1 × 10-8 M (based
on the water dissociation constants given in ref. 27), which
is even lower than the hydroxide concentration in pure water
at room temperature. In addition, the pH of the water (after
cooling) from the 1 h, 200 °C degradation of lindane was 2.6,
as would be expected for the dehydrohalogenation mech-
anism. Clearly, increased hydroxide concentration in high-
temperature water is not responsible for the dramatic increase
in dehalogenation reactions.
Because the reaction cells have a small gas headspace, it
is possible that the reactions occur in the vapor phase rather
than in the liquid. To test this possibility, the lindane reactions
from Table 3 were repeated using a reaction cell with a floating
piston to isolate the liquid (containing the lindane) and gas
phases (no lindane). Degradation results with only liquid
water in contact with lindane were identical (within typical
reproducibilities) to those in Table 3, thus demonstrating
that the headspace present in the routine cell did not affect
the lindane reactions.
Effect of Tem perature on Water Basicity. Previous studies
on dechlorinating aliphatic organochlorine species in water
have relied on added base (7, 28, 29, 32). Because dehydro-
halogenation and subsequent substitution reactions we
observed for lindane are catalyzed by base and because the
dissociation constant of water increases with temperature
(27), it is initially tempting to ascribe the rapid dechlorination
observed at higher temperatures to increased hydroxide ion
concentrations. The dissociation constant of water increases
from 1.0 × 10-14 at ambient to 5.0 × 10-12 at the highest
temperature used in this study (300 °C) (27). Assuming pure
water, the concentration of hydroxide ion increases from 1.0
× 10-7 M at room temperature to 1.1 × 10-6 M at 125 °C.
However, further increases in temperature have little effect,
Tetrachloroethane and Trichloroethene. Dehydrohalo-
genation of tetrachloroethane was rapid, even at tempera-
tures as low as 105 °C (Figure 4). The half-lives obtained for
105 and 150 °C, were 32 and 11 min, respectively, and the
nearly 100% of the tetrachloroethane was removed after 15
min at 200 °C. At each temperature tested, ca. 100% of the
tetrachloroethane found was converted to trichloroethene,
and a slow loss of trichloroethene was observed at long
reaction times (Figure 4).
Higher temperatures are required to degrade trichloro-
ethene, which might be expected because dehydrohaloge-
nation would require the formation of a triple bond. Studies
9
1 3 4 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 6, 2002

TABLE 2. Comparison of Degradation Efficiencies Using
Inconel and Glass Cells (1 h)
cell
a
reactant
material temperature % degraded ( SD
tetrachloroethane Inconel
tetrachloroethane glass
105 °C
105 °C
150 °C
150 °C
71 ( 14
68 ( 12
58 ( 10
71 ( 8
lindane
lindane
Inconel
glass
a
Standard deviations were calculated from triplicate experim ents.
TABLE 3. Percent of Total Chlorine Found as the Test
Compound, as Organochlorine Products, and as Chloride Ion
a
% chlorine ( SD
tetrachloro-
ethane (16 mM)
lindane (9 mM)
tem perature
tim e (h)
test com pound
organochlorine productsc 3.6 ( 0.3 2.0 ( 0.2 22 ( 10
chloride ion
total chlorine recovered
200 °C
1
25 ( 4
200 °C
18
150 °C
1
73 ( 9
300 °C
1
nd
ndb
19 ( 4
37 ( 7
66 ( 7
70 ( 6
72 ( 5
6.1 ( 0.5 62 ( 10
101 ( 2 80 ( 10
a
Standard deviations were calculated from triplicate experim ents.
nd: not detected. The detection lim its for lindane and tetrachloroethane
b
were 0.2 and 0.05%, respectively. cOrganochlorine products were
determ ined by GC/MS, as shown in Figure 3.
to give ca. 60-70% degradation of tetrachloroethane and
lindane so that the relative rates obtained in glass and Inconel
could be observed. Table 2 shows that data obtained in glass
were within the range of error of the Inconel cell results.
Unfortunately, the glass vials could not be used over 150 °C,
so some enhancement of reaction rates by the Inconel cell
may still be possible at higher temperatures. To guard against
this possibility, all experiments done at temperatures above
150 °C were performed using replicate cells, some which had
been heavily used and some which were nearly new. The
results from new and old cells were within the quantitative
reproducibilities typically seen for these experiments, indi-
cating that Inconel cell age did not affect the reaction rates.
However, as a matter of caution, we typically limit the working
life of the Inconel cells at higher temperatures to ca. 100 h
with lower concentrations of reactant (e.g., 1 mM) and ca.
40 h with higher concentrations of reactant (e.g., 20 mM).
Chlorine/Chloride Mass Balance. Previous investigations
have generally not attempted to determine mass balance for
the products of dechlorination, particularly for the organic
components (2, 6), although Bunte et al. (7) did recover ca.
90% of the chloride ion from the complete degradation of
lindane at 350 and 400 °C from reactions with water and
added NaOH. In the present study, we attempted to account
for all chlorine atoms (whether organochlorines or chloride
ion) at conditions where the reaction was in its intermediate
stages and conditions where nearly complete dechlorination
occurred.
To aid in reproducibly detecting chloride ions, all experi-
ments were performed with higher concentrations (9-20
mM) of test compounds. Also, to ensure that the Inconel
cells did not remove chloride ions from the water, a recovery
test on chloride was performed by placing 3 mL of a NaCl
solution (2.85 mM in chloride) into triplicate Inconel cells
and heating them for 1 h at 300 °C. Recoveries of chloride
were 100 ( 3%, demonstrating that Inconel does not adsorb
chloride ion.
The results of attempts to achieve chlorine mass balance
for lindane and tetrachloroethane are presented in Table 3.
For lindane, the total mass balance for chlorine atoms was
ca. 70%, whether the removal of lindane was complete (200
FIGURE 4. Degradation of 1 mM tetrachloroethane and formation
of trichloroethene in subcritical water at different temperatures.
Standard deviations presented are based on triplicate experiments.
using trichloroethene as the reactant showed only slow losses
at 200 and 250 °C, and 300 °C was required to obtain a half-
life of 45 min based on percent degradations of 55 ( 5 (after
10 min), 64 ( 7 (30 min), 77 ( 6 (60 min), and 87 ( 11 (90
min). In comparison, Gillham et al. reported a half-life of 14
h with the zerovalent iron in water at room temperature (2),
while electrochemical reduction of trichloroethene with
palladized iron electrodes has been used to reduce half-lives
to a few minutes (6). It should be noted that these previous
studies used low trichloroethene concentrations (11 and 300
µM, respectively), while we used higher concentrations (19
mM) to aid in determining the reaction products, which may
result in reduced reaction rates (discussed later in the text).
No degradation products from the trichloroethene could
be identified in the methylene chloride extracts of the product
water. Because any such organic products might be very
volatile (and thus be lost in the solvent peak during GC
analysis), attempts were made to identify such volatiles using
a gastight syringe to collect a 1-mL headspace sample after
the reaction and SPME sorption both of the headspace and
the product water. Both techniques allow solventless injection
into the GC/ MS so that MS monitoring can begin immediately
upon injection. Unfortunately, no product species were
identified using either technique.
Reproducibility and Inertness of the Experim ental
System . As the number of experiments for a cell increased,
sporadic faster degradation rates were obtained. Close
observation showed slight corrosion of the cells. Therefore,
a brass brush was used in the cell during the washing
procedure, which eliminated this artifact and resulted in
better reproducibility among the replicate experiments.
Additional tests into the inertness of the Inconel cells were
performed by comparing reaction rates when the experiments
were conducted in glass. Reaction conditions were chosen
9
VOL. 36, NO. 6, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 1 3 4 1

FIGURE 5. GC/MS chromatogram of degradation products from 100 mg PVC after 1 h at 300 °C in subcritical water.
Dieldrin was investigated as a second example of a highly
TABLE 4. Degradation of 1,9-Dichlorononane and Dieldrin after
1 h
chlorinated pesticide which, like lindane, has six chlorine
atoms bound to aliphatic ring carbons. As shown in Table
4, removal of dieldrin at 200 °C (1 h) was similar to that
achieved for lindane. However, no significant intermediate
organic products could be found in the water extract from
the dieldrin degradation, in contrast to the large number of
reaction products found for lindane (Figure 2). Because only
5% of the chlorine could be found as chloride ion (Table 4),
we expect that the intermediate products from dieldrin were
too polar to be extracted by methylene chloride and
determined by GC/ MS and speculate that hydrolysis of
dieldrin’s epoxide group leads to the formation of carboxylic
acids, which are too polar to be determined by our methods.
PVC Degradation. The final investigations were performed
on poly(vinyl chloride). Because the weight average molecular
weight was 62 000 amu, the use of PVC allowed us to study
the dehalogenation of compounds which should have no
significant solubility in water, even at elevated temperatures.
To allow for gravimetric measurements of the mass loss of
the PVC, 100 mg of PVC was used per 3 mL of water for each
test. After the reaction was completed, the PVC residue was
air-dried and weighed to determine the mass lost. All
reactions were performed for 1 h.
No detectable loss in PVC mass occurred at 200 °C,
although a slight yellowing of the polymer was observed.
However, when the reaction temperature was raised to 250
°C, 50% of the mass was lost (Table 5). At both 300 and 370
°C, ca. 44 wt % of the initial PVC mass remained as residue.
Because PVC is 56.6% chlorine by weight, the ca. 56% weight
loss found for the 300 and 370 °C reactions corresponds
almost perfectly with the weight loss expected if all chlorine
was removed from the polymer.
Attempts to identify organic products by GC/ MS showed
no organochlorine molecules but did show a range of
aromatic hydrocarbons from benzene to anthracene (Figure
5). However, the total quantity of these products accounted
for only ca. 1% of the total PVC mass. Carbon and hydrogen
content of the residue and chloride ion concentrations in
the product waters were also performed in an effort to achieve
mass balance. Elemental analysis of the PVC residues showed
a 1:1 carbon/ hydrogen molar ratio, as would be expected if
% of test
compound
degraded
( SD
% of
chlorine
as chloride ( SD
temp
(°C)
a
1,9-dichlorononane
dieldrin
200
200
300
100
35 ( 5
100
98 ( 10
5 ( 1
74 ( 3
a
Standard deviations were calculated from triplicate experim ents.
°C for 18 h) or only partial (200 °C for 1 h). Note that the mass
balance attempted to account for the millimoles of chlorine
both as chloride ions and by quantitating all of the inter-
mediate organics formed from lindane, which are shown in
Figure 2. Similarly, the total mass balance of chlorine from
tetrachloroethane was 100 and 80% for partial removal (150
°C for 1 h) or complete removal (300 °C for 1 h) of the
tetrachloroethane.
It should be noted that somewhat slower reactions
occurred when using these higher concentrations of lindane
and tetrachloroethane. For example, the removal of lindane
at 200 °C after 1 h dropped from >99% (at 0.6 mM) to ca.
75% (with 9 mM). Similarly, the removal of tetrachloroethane
at 150 °C (1 h) dropped from 97% (with 1 mM) to 25% (with
16 mM).
Dieldrin and 1,9-Dichlorononane. Limited tests were
performed with 1,9-dichlorononane to determine whether
aliphatic chlorines which are isolated yield rapid dehydro-
halogenation (in contrast to the tetrachloroethane where all
four chlorines are found on two adjacent carbon atoms).
While reaction for 1 h at 150 °C resulted in ca. 50% loss, 1
h at 200 °C resulted in 100% removal of the 1,9-dichlo-
rononane (Table 4). Also, the ca. 100% recovery of chloride
ions dem onstrated that com plete dechlorination was
achieved. Only one organic product was found by GC/ MS,
which had a mass spectrum consistent with the expected
product, 1,8-nonadiene (however, standards were not avail-
able to allow the identity to be confirmed).
9
1 3 4 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 6, 2002

(5) McNab, W. W., Jr.; Ruiz, R.; Reinhard, M. Environ. Sci. Technol.
2000, 34, 149-153.
(6) Li, T.; Farrell, J. Environ. Sci. Technol. 2000, 34, 173-179.
(7) Bunte, G.; Eisenreich, N.; Hirth, Th.; Krause, H. Proceedings of
3rd International Symposium on Supercritical Fluids, Strasbourg,
France, 1994; pp 201-206.
TABLE 5. Mass Balance for PVC after 1 h Degradation in
Subcritical Water
product mass from 100 mg PVC (mg ( SD)a
residue
(8) Yak, H. K.; Wenclawiak, B. W.; Cheng, I. F.; Doyle, J. G.; Wai, Ch.
M. Environ. Sci. Technol. 1999, 33, 1307-1310.
(9) Chuang F.-W.; Larson, R. A.; Wessman, M. S. Environ. Sci.
Technol. 1995, 29, 2460-2463.
(10) Anitescu, G.; Tavlarides, L. L. Ind. Eng. Chem. Res. 2000, 39,
583-591.
(11) Katritzky, A. R.; Nichols, D. A.; Siskin, M.; Murugan, R.;
Balasubramanian, M. Chem. Rev. 2001, 101, 837-892.
(12) Savage, P. E. Chem. Rev. 1999, 99, 603-621.
(13) Sato, Y.; Kato, K.; Takeshita, Y.; Nishi, S. Jpn. J. Appl. Phys. 1998,
37, 6270-6271.
b
b
temp (°C)
total
as C
as H
chloride in water
250
300
370
49.5 ( 0.3 41 ( 2 3.6 ( 0.2
44 ( 3
46 ( 5
47 ( 8
43 ( 1
44 ( 1
40 ( 5 3.3 ( 0.5
39 ( 1 3.5 ( 0.1
a
Standard deviations were calculated from triplicate experim ents.
Content of each elem ent in the residues was determ ined with a C,H,N
b
analyzer.
the polymer residue was formed of high molecular weight
conjugated alkenes (-CHdCH-CHdCH-)n, as would be
formed from dehydrohalogenation. In addition, mass balance
for the residue (the sum of carbon and hydrogen as compared
to residue mass) was ca. 98%, which demonstrates that no
significant chlorine remains in the PVC residue after water
treatment. Finally, ca. 80% of the total chorine was found as
chloride ion in the product water (Table 5).
(14) Marrone, P. A.; Arias, T. A.; Peters, W. A.; Tester, J. W. J. Phys.
Chem. A 1998, 102, 7013-7028.
(15) Hawthorne, S. B.; Lagadec, A. J. M.; Kalderis, D.; Lilke, A. V.;
Miller, D. J. Environ. Sci. Technol. 2000, 34, 3224-3228.
(16) Hawthorne, S. B.; Miller, D. J.; Lagadec, A. J. M. U.S. Patent
6,001,256, 1999.
(17) Windal, I.; Hawthorne, S. B.; De Pauw, E. Organohalogen Compd.
1999, 40, 591-594.
(18) Miller, D. J.; Hawthorne, S. B. J. Chem. Eng. Data 2000, 45,
78-81.
Acknowledgments
(19) Ro¨ ssling, G. L.; Franck, E. U. Ber. Bunsen-Ges. Phys. Chem. 1983,
87, 882-890.
The authors are grateful to Euguenii Kozliak (Department of
Chemistry, University of North Dakota) and Ed Olson (EERC,
University of North Dakota) for helpful discussions on
reaction mechanisms. The financial support of the U.S.
Environmental Protection Agency, Office of Research and
Development, NCERQA, is gratefully acknowledged. Although
the research described in this paper has been funded wholly
or in part by the United States Environmental Protection
Agency through Grant No. R825394-1 to the University of
North Dakota Energy & Environmental Research Center, it
has not been subjected to the Agency’s required peer and
policy review and, therefore, does not necessarily reflect views
of the Agency, and no official endorsement should be inferred.
This paper was also prepared with the support of the U.S.
Department of Energy, under Cooperative Agreement DE-
FC26-98FT40320. However, any opinions, findings, conclu-
sions, or recommendations expressed herein are those of
the author(s) and do not necessarily reflect the views of the
DOE.
(20) Kuhlmann, B.; Arnett, E. M.; Siskin, M. J. Org. Chem. 1994, 59,
3098-3101.
(21) Krammer, P.; Vogel, H. J. Supercrit. Fluids 2000, 16, 189-206.
(22) Haar, L.; Gallagher, J. S.; Kell, G. S. National Bureau of Standards/
National Research Council Steam Tables; Hemisphere Publishing
Corp.: Bristol, PA, 1984.
(23) Hawthorne, S. B.; Grabanski, C. B.; Hageman, K. J.; Miller, D.
J. J. Chromatogr. 1998, 814, 151-160.
(24) Zhang, Z.; Yang, M. J.; Pawliszyn, J. Anal. Chem. 1994, 66, 844A-
853A.
(25) Meharg, A. A.; Wright, J.; Leeks, G. J. L.; Wass, P. D.; Osborn, D.
Environ. Sci. Technol. 1999, 33, 2001-2006.
(26) Walker, K.; Vallero, D. A.; Lewis, R. G. Environ. Sci. Technol.
1999, 33, 4373-4377.
(27) Marshall, W. L.; Franck, E. U. J. Phys. Chem. Ref. Data 1981, 10,
295-304.
(28) Cristol, S. J.; Hause, N. L.; Meek, J. S. J. Am. Chem. Soc. 1951,
73, 674-679.
(29) Ngabe, B.; Bidleman, T. F.; Falconer, R. L. Environ. Sci. Technol.
1993, 27, 1930-1933.
(30) Solomons, T. W. G. Organic Chemistry, 2nd ed.; John Wiley &
Sons: New York, 1980.
(31) Narayan, R.; Antal, M. J., Jr. J. Am. Chem. Soc. 1990, 119, 1927-
1931.
Literature Cited
(1) Matheson, L. J.; Tratnyek, P. G. Environ. Sci. Technol. 1994, 28,
2045-2053.
(32) Van der Linden, T. Ber. 1912, 45, 231-247.
(2) Gillham, R. W.; O’Hannesin, S. Ground Water 1994, 32, 958-
967.
(3) Arnold, W. A.; Robert, A. L. Environ. Sci. Technol. 1998, 32, 3017-
Received for review August 2, 2001. Revised manuscript
received November 4, 2001. Accepted December 13, 2001.
3025.
(4) Orth, W. S.; Gillham, R. W. Environ. Sci. Technol. 1995, 30, 66-
71.
ES011186K
9
VOL. 36, NO. 6, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 1 3 4 3