Destruction of halogenated hydrocarbons with solvated electrons in the presence of water

Article abstract of DOI:10.1016/S0045-6535(99)00428-2

Model halogenated aromatic and aliphatic hydrocarbons and halogenated phenols were dehalogenated in seconds by solvated electrons generated from sodium in both anhydrous liquid ammonia and ammonia/water solutions. The minimum sodium required to completely dehalogenate these model compounds was determined by increasing the Na/substrate ratio until halogen loss was complete. Minimum sodium consumptions were determined in both anhydrous liquid ammonia and with a (5, 20, 50-fold molar excess of water per mole of halide). While more Na was consumed in the presence of water, these dehalogenations were still efficient when a 50-fold water excess was present. Dehalogenation is faster than competiting reactions with water. CCl₄ and CH₃CCl₃ in the presence of a stoichiometric deficiency of sodium produced only CH₄ and CH₃CH₃ and recovered CCl₄4 or CH₃CCl₃, respectively. No partially dechlorinated products were detected, indicating dechlorination was diffusion controlled. Na consumption per chlorine removed (as NaCl) was lower than that of Li, K or Ca and this advantage increased in the presence of water. Na consumption was lower using Na chunks instead of a thin Na mirror. Chloroaromatic compounds gave the parent aromatic hydrocarbon and aminated products in anhydrous ammonia but aminated products did not form when water was present. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0045-6535(99)00428-2

Chemosphere 41 (2000) 907±916
Destruction of halogenated hydrocarbons with solvated
electrons in the presence of water
*
Guang-Ri Sun, Jin-Bao He, Charles U. Pittman Jr.
Department of Chemistry, University/Industry Chemical Research Center, Mississippi State University, Mississippi State,
MS 39762-9573, USA
Received 25 June 1999; accepted 1 September 1999
Abstract
Model halogenated aromatic and aliphatic hydrocarbons and halogenated phenols were dehalogenated in seconds
by solvated electrons generated from sodium in both anhydrous liquid ammonia and ammonia/water solutions. The
minimum sodium required to completely dehalogenate these model compounds was determined by increasing the Na/
substrate ratio until halogen loss was complete. Minimum sodium consumptions were determined in both anhydrous
liquid ammonia and with a (5, 20, 50-fold molar excess of water per mole of halide). While more Na was consumed in
the presence of water, these dehalogenations were still ecient when a 50-fold water excess was present. Dehalogenation
is faster than competiting reactions with water. CCl4 and CH3CCl3 in the presence of a stoichiometric de®ciency of
sodium produced only CH4 and CH3CH3 and recovered CCl4 or CH3CCl3, respectively. No partially dechlorinated
products were detected, indicating dechlorination was di€usion controlled. Na consumption per chlorine removed (as
NaCl) was lower than that of Li, K or Ca and this advantage increased in the presence of water. Na consumption was
lower using Na chunks instead of a thin Na mirror. Chloroaromatic compounds gave the parent aromatic hydrocarbon
and aminated products in anhydrous ammonia but aminated products did not form when water was present. Ó 2000
Elsevier Science Ltd. All rights reserved.
Keywords: Solvated electron reductions; Dechlorination; Dehalogenation; Na/NH3; Chlorinated-aliphatic and aromatic hydrocarbons
1. Introduction
struction has been carried out by incineration, wet air
oxidation and catalytic dehydrochlorination reaction
with superoxide, photolysis and electrolytic reduction.
Metal-promoted borohydride or alkoxyborohydride re-
ductive dechlorination (Tabaei and Pittman, 1993a,b;
Tabaei et al., 1992), Ti-catalyzed borohydride dechlori-
nation (Liu et al., 1995), iron-promoted dechlorination
(Chung et al., 1995), borohydride reductive dechlorina-
tion in glymes (Yang and Pittman, 1997, 1998a,b) and
thermolysis over solid bases like CaO/Ca(OH)2 (Yang
and Pittman, 1996) were recently developed. However,
serious limitations exist and many methods cannot be
used economically with soils.
Halogenated organic compounds, such as solvents,
pesticides, herbicides, PCBs, CFCs, etc. have caused
serious environmental problems. PCBs, for example, are
distributed in soils, sludges, estuaries, etc. at over 400
sites and chlorinated aliphatic hydrocarbons (CAHs) are
serious contaminants at 358 major hazardous waste sites
in the United States. PCBs are regulated under TSCA
(Toxic Substances Control Act, 1976). Chemical de-
* Corresponding author. Tel.: +1-601-325-7616; fax: +1-601-
Alkali metal-based reductive dehalogenation is a well
known process. Knorre et al. (1980) added sodium to the
halogenated organic compounds without solvent, but
325-7611.
E-mail address: cpittman@ra.msstate.edu (C.U. Pittman
Jr.).
0045-6535/00/$ - see front matter Ó 2000 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 4 5 - 6 5 3 5 ( 9 9 ) 0 0 4 2 8 - 2

908
G.-R. Sun et al. / Chemosphere 41 (2000) 907±916
this heterogeneous reaction is slow and moisture sensi-
tive. Dissolving Li, Na, K or Ca in liquid NH3 generates
solvated electrons which dehalogenate dissolved organic
compounds at exceptionally fast rates (Watt, 1950; Birch
and Subba Rao, 1972; Aiired et al., 1974; Lahaniatis
et al., 1975; Jessup et al., 1977).
After drying over Ca, ammonia was distilled and con-
densed into a smaller stainless cylinder (120 ml).
4-Chlorophenol,
4-¯uorophenol,
4-bromophenol,
2-chloro-4-¯uorophenol, 2, 4-dichlorophenol, 2, 4, 6-
trichlorophenol, chlorobenzene, 1, 2-dichlorobenzene,
1, 2, 3-trichlorobenzene,
1, 2, 3, 4-terachlorobenzene,
Both Commodore Solution Technologies (Weinberg
and Abel, 1989, 1992) and our group (Pittman and
Tabaei, 1993, 1994; Pittman and Mohammed, 1996)
have demonstrated that neat PCBs and PCB-contami-
nated soils (as received clay, loam and sandy soils con-
taining up to 30% water) can be decontaminated in
liquid ammonia slurries by adding Na or Ca. PCB-de-
struction eciencies >99.9% were achieved within 30 s
at room temperature (Weinberg and Abel, 1989, 1992;
Pittman and Tabaei, 1993, 1994; Pittman and Mo-
hammed, 1996). These results indicate that Na/NH3 or
Ca/NH3 dechlorinations are signi®cantly faster than
reactions with water. Soils typically contain water, ox-
ygen and iron, all of which can interfere with solvated
electron chemistry (Eastham and Larkin, 1959; Smith,
1968; Rabideau et al., 1984). Both oxygen and iron
catalyze the reaction of Na and Ca with ammonia to
produce the metal amide salt and hydrogen. The half-
life (t1=2) of solvated electrons in puri®ed NH3 is ca. 300 h
(Crooks and Bard, 1987), but in pure water the solvated
electron has a t1=2 of ca. 100 ls. (Gould, 1965). De-
chlorination of PCB-contaminated wet soils is possible
due to the relatively long solvated electron lifetime in
20% H2O/80% NH3 (t1=2 ˆ ca. 100s (Schindewolf, 1970;
Crooks and Bard, 1987)) coupled with fast dechlo-
rination by Na/NH3ꢀL† (Weinberg and Abel, 1989,
1992; Pittman and Tabaei, 1993, 1994; Pittman and
Mohammed, 1996).
4-chlorotoluene, 3, 4-dichlorotoluene, 2-chloro-p-xylene,
2bromo-p-xylene, 3-¯uoro-o-xylene, 1-chlorooctane,
1, 1, 1-trichloroethane, carbontetrachloride and 1, 2, 3,
4, 5, 6-hexachlorocyclohexane, octane, decane, dode-
cane, diethylether, p-xylene, sodium, lithium, potassium
and calcium were purchased (Aldrich) and used as re-
ceived.
Reactions were run in glass pressure vessels (Andrews
Glass) All valves and conduits were stainless steel (90 ml,
maximum safe pressure 200 psi).
2.2. Dehalgenation reactions
All dehalogenation reactions were carried out by
Procedure A, unless otherwise stated. Procedures B, C,
D and E were used for comparison with Procedure A.
Two work-up protocols, Method 1 and Method 2, were
used as described below. All data in Tables 1±8 were
determined from at least two (and many cases more)
experiments.
Procedure A: Halocarbon (1 mmol) and metal (par-
ticle size, >8 mm3) were added, under nitrogen, to vessel
A equipped with a glass covered stirring bar. Ammonia
(10 ml) and water (or just ammonia) were added to
vessel B. The pressure in vessel B was raised to 160 psi
with nitrogen from a cylinder. Vessel B (upside down)
was connected from above to vessel A via a metal con-
duit (Fig. 1). Ammonia was discharged, all at once, from
B into A by the higher pressure in B. The contents of A
were vigorously stirred before and after addition for 30 s
at room temperature. Sodium dissolution was complete
before 30 s.
No systematic study has appeared of solvated elec-
tron dechlorinations in NH3ꢀL† with excess water pre-
sent. PCBs, CAHs, etc. intercalated deep in clay layers
require di€usion of solvated electrons into soil or ex-
traction of chlorinated organics into the bulk solution.
This could allow time for competitive reactions with
water. Competitive reactions may accelerate in the
presence of oxygen, transition metal ions, acidic soil
sites, etc. Therefore, we determined the consumption of
Na required to acheive complete dechlorination of both
chloroaromatic and chloroaliphatic compounds in the
presence of excess water. The reactivity of Li, Na, K and
Ca was also brie¯y compared.
2. Experimental
2.1. Material and equipment
Ammonia was purchased from Dupont, National
Ammonia Division. It was condensed from a steel tank
into a stainless cylinder (1 l) containing calcium metal.
Fig. 1.

G.-R. Sun et al. / Chemosphere 41 (2000) 907±916
909
Procedure B: Procedure B was similar to Procedure A
except that both substrate (1 mmol) and ammonia were
added into vessel B. The vessel B was shaken vigorously
to dissolve the substrate in NH3ꢀL† or NH3/H2O solu-
tion. Then, this solution was added all at once to vessel
A containing sodium with vigorous stirring.
Chloride analysis procedure: After dehalogenation
and work-up by Method 2, H2O/NH3 was removed in
vacuo. Sodium chloride residue remained. Distilled wa-
ter was added to this residue to give 150 ml of solution.
Then, concentrated nitric acid (1 ml) was added. This
solution was titrated at room temperature with AgNO3
(0.1 N) until no further silver chloride precipitate
formed. The solution was heated to re¯ux for 5±10 min,
cooled and allowed to stand for 2 h. AgCl precipitate
was collected, rinsed with HNO3 aq. (1.2% (w/w), 10 ml)
and distilled water (10 ml ´ 2), dried in vacuo at 70°C for
8±15 h and weighed.
Procedure C: Metal was added, under nitrogen, into
vessel A, equipped with a glass stir bar. Liquid ammonia
(10 ml) was then added and stirred at ambient temper-
ature until the metal was completely dissolved. Ammo-
nia was slowly evaporated at room temperature to
generate a thin silver sodium mirror in vessel A. Sub-
strate (1 mmol) was added to the sodium mirror. Then
NH3ꢀL† or NH3/H2O was rapidly added to A from B
while vigorously stirring. Stirring was continued for 30 s
at room temperature.
3. Results and discussion
3.1. Product distributions
Procedure D: Na/NH3 was generated in vessel B from
a speci®c weight of sodium and liquid ammonia (10 ml).
This solution was added with vigorous stirring, at am-
bient temperature, to a preformed substrate (1 mmol)/
NH3 (5 ml) or NH3/H2O (5 ml) solution in vessel A at
di€erent H2O/substrate ratios and stirred at room tem-
perature for 30 s.
Eighteen aliphatic and aromatic halogenated model
compounds were completely dehalogenated in seconds
by solvated electrons. These reductions required only the
time needed to dissolve the metal. Room temperature
reductions of six chlorinated aromatic molecules in an-
hydrous NH3ꢀL† are summarized in Table 1. Products
consist of the parent hydrocarbon and an amminated
product in each case except for 4-chlorophenol which
produced phenol.
Procedure E: Substrate (1 mmol), water and liquid
ammonia (10 ml) were added to vessel A at room tem-
perature. After cooling to )55°C, sodium was added all at
once while stirring. Stirring continued at )55°C for 30 s.
Work-up: Method 1 was used with halogenated xyl-
enes and phenols. Method 2 was used with aliphatic
chlorides, chlorobenzenes and chlorotoluenes.
General mechanistic features for these dehalogen-
ations are summarized in Scheme 1. Rapid electron
transfer to aromatic halides generates an anion radical.
Then, loss of halide ion gave an aromatic radical. This
radical might abstract hydrogen, dimerize or acquire
another electron to form an aryl anion which then ab-
stracts a proton to give ArH. Dimerization products
were not detected. This scheme accounts for the hy-
drocarbon products (Table 1). The aminated products,
formed when the solvent was anhydrous NH3ꢀL†, could
result from a benzyne route (Eq. (1) where the base
could be either NH2 or Ar ) or from NH2 addition to
the intermediate aryl radical (as shown in Eq. (2) for
chlorobenzene) followed by rapid electron transfer.
Aminated product yields increased when both sub-
strate and sodium concentrations were increased, a result
consistent with either pathway. This is illustrated in Ta-
ble 1 upon comparing entries 1 with 1a and 3 with 3a.
Reduction of 4-chlorotoluene (Eq. (3)) gave mostly
toluene but approximately equal amounts of m- and p-
toluidine were formed, consistent with a benzyne inter-
mediate. No animated products were produced when
water was added. Water protonates NH2 removing the
benzyne route.
Method 1: After dehalogenation, ammonia was
slowly evaporated at room temperature. The resulting
residue was extracted with ether (10 ml). In
chlorophenol reactions, the residue was neutralized with
H2SO4 (15% (w/w) aq.) prior to extraction with ether.
Method 2: After dehalogenation, water (30 ml) was
added to dilute the NH3ꢀL†. Then the mixture was ex-
tracted with ether (15 ml) at )20°C to 0°C. In the de-
halogenations of CCl4 and CH3CCl3, p-xylene (10 ml)
was used for extraction. The organic phase was sepa-
rated after shaking vigorously for several minutes. In
dechlorinations of CCl4 or CH3CCl3, a gas sample was
taken with a pressure syringe to identify CH4 or
CH3CH3 before separation.
2.3. Analysis
Capillary GC (Varian model 3300) determined con-
versions and identi®ed and quanti®ed products using
both Supel-Q PLOT fused silica (30 m, 0.32 mm ID) and
DB-5 (30 m, 0.32 mm ID) columns with n-octane,
dodecane and hexadecane internal standards. Minimum
conversion data were >99%, unless otherwise stated.
GC±MS analysis was performed on a Finnigan 4500
GC±MS. Chloride analysis was carried out after reduc-
tions as described below.
ꢀ1†

910
G.-R. Sun et al. / Chemosphere 41 (2000) 907±916
Table 1
Products from Na/NH3 dechlorinations of model chlorinated aromatic substrates at 25°Ca
Entry
Starting materials
Hydrocarbon products
(mol%)
Aminated products
(mol%)
Other unknown
productsb
1
Chlorobenzenec
Benzene (88%)
Benzene (84%)
Benzene (73%)
Toluene (87%)
Aniline (10%)
Aniline (15%)
Aniline (25%)
(2%)
(1%)
(2%)
(3%)
1a
2
Chlorobenzeneb
1, 2-Dichlorobenzenec
4-Chlorotoluenec
3
m-Toluidine (5%)
p-Toluidine (5%)
m-Toluidine (7%)
p-Toluidine (7%)
m-Toluidine (18%)
p-Toluidine (32%)
3a
4
4-Chlorotoluenec
Toluene (85%)
Toluene (47%)
(1%)
(3%)
3, 4-Dichlorotoluenec
5
6
7
3, 4-Dichlorotoluenec
2-Chloroxylenee
4-Chlorophenole
Toluene (100%)d
p-Xylene(60%)
Phenol (98%)
2-Aminoxylene (37%)
(3%)
(2%)
a Conducted by Procedure A where substrate (1 mmol), NH3ꢀL† (10 ml) and Na (1.5±2.8 mmol) were employed. However, entries 1a and
3a employed substrate (2 mmol), NH3ꢀL† and Na (4 mmol). Reproducibility in these experiments was ‹3%.
b Traces of several kinds of unknown products.
c Work-up by Method 2.
d Reduction was carried out in the presence of water where substrate (1 mmol), NH3ꢀL† (10 ml), H2O (5 mmol) and Na (4.2 mmol) were
employed.
e Work-up by Method 1.
Scheme 1.
NH3. For comparison, 4-chlorotoluene only gave 10%
of p- and m-toluidines at these conditions. The inter-
mediate anion formed after loss of the ®rst chlorine
might eliminate a second Cl to produce a benzyne
(Eq. (4)).
ꢀ2†
ꢀ4†
ꢀ3†
In contrast, dehalogenation of chlorophenol gave no
aminated products. Furthermore, aryl chloride dechlo-
rinations with Na/NH3 in the presence of added water
gave no aminated products because NH2 was proto-
nated by water or phenol, removing amide and pre-
venting amination via Eqs. (1) or (2).
The yield of aminated products increased as the
number of ring chlorines increased. 3, 4-Dichlorotoluene
produced a mixture of m-toluidine (18%) and p-toluidine
(32%) (50% of the total yield) when treated with Na/
NH3 at a substrate concentration of 1 mmol in 10 ml

G.-R. Sun et al. / Chemosphere 41 (2000) 907±916
911
3.2. Minimum Na consumption required for complete
dechlorination
et al., 1977). The data in Table 2 demonstrate this point.
A total of 1.5 equivalents of sodium are required to
completely reduce chlorobenzene to benzene. However,
adding a 50 mol excess of water only increased the so-
dium requirement to 2.5 equivalents (entry 1). In CCl4
dechlorinations only 4.6 equivalents of Na per CCl4 (1.15
equivalents Na per Cl) were needed to reduce CCl4 to
CH4 when no water was added. Adding a 50 mol excess
of water versus CCl4 increased the sodium requirement
to only 6.4 equivalents (1.6 equivalents Na per Cl, entry
14). Similar results were obtained for the other chlorides
(entries 2±7, 12±14) except for chlorooctane (entry 11)
and the chlorophenols (entries 8±10).
The minimum amount of sodium required to com-
pletely dechlorinate both aromatic and aliphatic com-
pounds was determined by the stepwise increase of the
Na/substrate ratio in several reactions until complete
chlorine loss occurred. This ratio was obtained both in
anhydrous liquid ammonia and in the presence of water
(5, 20, 50-fold molar excess of water per mole of halide).
Representative results are summarized in Table 2 for
monochloroaryl compounds (entries 1±3), di-, tri- and
tetrachloroaryl compounds (entries 4±7), chlorinated
phenols (entries 8±10) and chlorinated aliphatic com-
pounds (entries 11±14).
The chlorophenols apparently consumed one extra
mole of Na to generate the corresponding phenoxide
‡
All dechlorinations required more than one equiva-
lent of Na to complete in anhydrous NH3ꢀL†. This might
be due to:
anions (or react with NH4 ). Chlorinated phenoxide
anions should be dicult to reduce due to the delocal-
ized negative charge. The data in entries 8±10 (Table 2)
show a higher sodium consumption per chlorine re-
moved, supporting this point. For example, when no
water was added, 4-chlorophenol, 2, 4-dichlorophenol
and 2, 4, 6-trichlorophenol required 2, 1.7 and 1.53
equivalents of Na, respectively, per chlorine removed.
However, chlorobenzene, 1.2-dichlorobenzene and
1, 2, 3-trichlorobenzene only required 1.5, 1.4 and 1.33
equivalents of Na, respectively, per chlorine removed.
Increasing chlorine substitution leads to a decreased
consumption of Na per chlorine lost, both with and
without water present.
1. reaction of solvated electrons, with NH3 to generate
NH2 , catalyzed by traces of Fe3‡
,
2. formation of aryl carbanions (after loss of chlorine)
or
3. competiting reactions with water (last 3 columns,
Table 2).
Aryl carbanion formation consumes two electrons per
chloride lost. The most important ®nding is that adding
substantial amounts of water did not greatly increase the
amount of Na needed for complete dehalogenation.
Dehalogenation was faster than reaction of solvated
electrons with water (Watt, 1950; Birch and Subba Rao,
1972; Aiired et al., 1974; Lahaniatis et al., 1975; Jessup
The Na required for total dehalogenation of 1-chlo-
rooctane increased from 1.7 to 4 equivalents upon
Table 2
Minimum amount of Na required to completely dechlorinate chloroaromatics at 25°Ca
Na/substrate mole ratios and (Na/Cl removed) required for complete dechlorination at di€erent H2O/substrate
mole ratios
Entry
substrate
H2O/substrate
!
0/1b
5/1
20/1
50/1
1
2
Chlorobenzene
4-Chloro-toluene
1.5 (1.5)c
1.5 (1.5)c
1.5 (1.5)c
2.8 (1.4)c
2.5 (1.25)c
4.0 (1.33)c
5.0 (1.25)c
2.0 (2)c
3.4 (1.7)c
4.6 (1.53)c
1.7 (1.7)c
3.6 (1.2)c
4.6 (1.15)c
6.7 (1.11)c
2.3
2.3
2.3
4.0
4.2
5.5
6.5
2.9
5.0
7.4
2.3
4.2
5.1
6.7
2.4
2.4
2.4
4.5
4.6
5.8
7.0
4.0
5.6
8.0
2.5d
4.5
5.5
7.2
2.5 (2.5)c
2.5 (2.5)c
2.5 (2.5)c
5.0 (2.5)c
6.0 (3)c
7.0 (2.33)c
8.6 (2.15)c
4.5 (4.5)c
8.0 (4)c
10 (3.33)c
4.0d (4)c
5.1 (1.7)c
6.4 (1.6)c
7.4 (1.23)c
3
2-Chloro-p-xylene
1, 2-Dichlorobenzene
3, 4-Dichlorotoluene
1, 2, 3-Trichlorobenzene
4
5
6
7
1, 2, 3, 4-Tetrachlorobenzene
4-Chlorophenol
8
9
2, 4-Dichlorophenol
2, 4, 6-Trichlorophenol
1-Chlorooctane
10
11
12
13
14
1, 1, 1-Trichloroethane
Carbontetrachloride
Lindane
a Experiments were conducted by Procedure A unless, otherwise, stated.
b No water added.
c Moles of Na consumed per mole of chlorine removed, derived from Na/substrate weighings and divided by the number of chlorines in
the substrate. Estimated errors ‹4%.
d Experiments were performed using Procedure B.

912
G.-R. Sun et al. / Chemosphere 41 (2000) 907±916
adding a 50 mol excess of water (entry 5). In contrast,
the Na consumption per chlorine for complete dehalo-
genation of lindane (1, 2, 3, 4, 5, 6-hexachlorocyclohex-
ane) increased from 1.11 with no water to 1.23
equivalents with a 50 mol excess of water (entry 15).
Lindane might undergo E-2 eliminations of up to 3 mol
of HCl in the presence of NH3 and/or NH2 . Dissolving
lindane (1 mmol) in ammonia (10 ml) at ambient tem-
perature for 5 min with no Na present gave a 99% total
conversion to 1, 2, 4-trichlorobenzene (87%), 1, 3, 5-tri-
chlorobenzene (5%) and 1, 2, 3-trichlorobenzene (8%)
(Eq. (5)).
rine, respectively, for complete dechlorination. Chlo-
roaromatic compounds required distinctly larger
amounts of Na, except for 2, 3-dichlorotoluene and
1, 2, 3, 4-tetrachlorobenzene which both required 1.25
equivalents of Na per chlorine. With a 50 mol excess of
water present, the monochloroaromatic compounds
(entries 1±3, Table 2) required 2.5 equivalents of Na
while the mono-, di-, and trichlorinated phenols con-
sumed 4.5, 4 and 3.3 equivalents of Na per chlorine
removed.
3.3. Dechlorination with de®cient amounts of sodium
Partially dechlorinated intermediates were not ob-
served when 1, 2-dichlorobenzene, 3, 4-dichlorotoluene,
CCl4 and CH3CCl3 were reduced in NH3ꢀL† with a Na
de®ciency (Table 3) either with or without H2O present.
For example, 3, 4-dichlorotoluene (Eq. (6)) gave only
toluene and recovered starting material when treated
with Na/NH3 at Na/substrate ratios of 1, 2 and 3 in the
presence of 5 equivalents of water. A sample GC display
is shown in Fig. 2. No monochlorotoluenes could be
detected.
ꢀ5†
E-2 elimination may compete with dissociative elec-
tron transfer during the reduction of lindane in Na/NH3.
Fast dissociative electron transfer would produce radi-
cals which, upon abstraction of hydrogen, would result
in chloroaliphatic intermediates. They might undergo
subsequent E-2 elimination with NH3, after Na had
been consumed, leading to lower Na consumption.
Transfer of the electron to RCl with simultaneous loss of
chloride occurs for chlorinated aliphatic compounds
(Holm, 1999) (Scheme 1) because p* orbitals are not
available to give radical anions. CAHs could add an-
other electron after Cl loss and further reactions would
ensue.
ꢀ6†
Dechlorinations of CCl4 with 1, 2 and 3.4 equivalents
of Na in anhydrous NH4ꢀL† gave only CH4, recovered
CCl4 and traces of unknown products. The material
balances were almost complete. No CHCl3, CH2Cl2 or
CH3Cl were detected by GC analysis. Similarly,
CCl4, CH3CCl3 and C6H6Cl6 (lindane) required less
sodium per chlorine lost than halogenated aromatic
compounds, especially in the presence of excess water.
In anhydrous ammonia, CCl4, CH3CCl3 and lindane
required 1.15, 1.2 and 1.1 equivalents of Na per chlo-
Table 3
Products obtained in sodium-de®cient reactions in liquid NH3 at room temperaturea
Substrate
Na/substrate (mole ratio)
H2O/substrate (mole ratio)
Products
3, 4-Dichloro-toluene
1
2
3
5
5
5
82% SMb, 18% toluene
60% SMb, 40% toluene
25% SMb, 75% toluene
1, 2-Dichloro-benzene
CCl4
3.5
5
7% SMb, 93% benzene
c
1
0
0
0
79% CCl4, 20% CH4
54% CCl4, 45% CHc4
33% CCl4, 64% CHc4
2
3.4
CH3CCl3
2
0
38% CH3CCl3, 62%
d
CH3CH3
a Reactions were carried out by Procedure A and work-ups were performed by Method 2. Estimated experimental error ‹3%.
b SM ± starting material.
c No traces of CHCl3, CH2Cl2 or CH3Cl were detected.
d No traces of CH3CHCl2, CH3CH2Cl or CH2CHCl were found.

G.-R. Sun et al. / Chemosphere 41 (2000) 907±916
913
ꢀ8†
3.4. E€ects of Na particle size, mode of addition and
temperature. Consumption of solvated electrons in de-
chlorinations vs reaction with water
Fig. 2. Graph A ± toluene, 4-chlorotoluene, 3, 4-dichlorotolu-
ene and hexadecane in Et2O. Graph B, graph C and graph D ±
1/1, 1/2 and 1/3 of Na/substrate (mole ratio) were used,
respectively.
The e€ect of sodium particle size on sodium con-
sumption in 2-chloro-p-xylene dechlorination was stud-
ied by the use of chunks (>8 mm3, Procedure A), thin
Na mirrors (Procedure C) and preformed solvated
electron solutions (Procedure D) in anhydrous NH3ꢀL†
and with a 50 mol excess of water (Table 4). Without
water, the Na/2-chloro-p-xylene ratio required for
complete dechlorination was the same for reactions
employing chunks and reactions employing mirrors (1.5
equivalents of Na required). A modestly larger Na re-
quirement (1.85 equivalent) was needed when anhy-
drous, preformed, solvated electron solutions were used.
However, in the presence of a 50 mol excess of water, the
use of Na chunks led to a substantially lower Na con-
sumption than the use of the thin Na mirror (2.5 vs 4.5
equivalents). Using preformed solvated electron solu-
tions (Procedure D) with a 50 mol excess of water ±
required a far greater consumption of Na. Only 55% of
dechlorination of CH3CCl3 with 2 equivalents of Na
produced ethane (62%) and CH3CCl3 (38%) was re-
covered unchanged. The material balances con®rm that
NH3 or NH2 is not reacting with partially dechlorinated
intermediates to give aminated products or elimination.
Na=NH3
CCl43:4 equivalents of Na
!
CH4 ‡ CCl4
ꢀ7†
These results argue that the four dechlorinations in
Table 3 are di€usion controlled. Apparently, solvated
electrons react rapidly with partially dechlorinated spe-
cies before these chlorine-containing intermediates can
di€use away to other regions of the solution and be re-
placed by more substrate. Reaction of solvated electrons
with substrate goes to completion faster than substrate
escape from the vicinity of the dissolving Na particle,
despite rapid stirring. The supply of solvated electrons
(Na dissolution) must also be fast enough to completely
dechlorinate intermediates before they equilibrate with
more substrate in the vicinity of the dissolving particle.
Thus, dechlorination appears to be di€usion controlled.
Hence, solvated electrons are consumed before mixing
can equilibrate them with substrate.
Table 4
E€ects of Na particle size in halogenation of 2-chloro-p-xylene
at 25°C
Particle size
H2O/substrate
(mole ratio)
Na/substrate
required (mole
ratio) for complete
dehalogenationa
Chunk
(>8mm3)b
0/1
1.5
Chlorinated phenols do not react under di€usion
control, in contrast to CAHs and chloroaromatic hyd-
rocrbons. When 2, 4, 6-trichlorophenol was treated with
3 equivalents of Na by Procedure A, small amounts of
partially dehalogenated dichlorophenols (2%) and
monochlorophenols (8%) were detected (Eq. (8)) show-
ing this reaction is not completely di€usion controlled.
One would expect the most highly chlorinated products
to always lose chloride faster than partially dechlori-
nated intermediates. Thus, extremely rapid chlorine loss
occurs even though the 2, 4, 6-trichlorophenoxide anion
is the substrate present in NH3: However, chlorophen-
oxides dechlorinate slower than CCl4, CH3CCl3 or
chlorobenzenes.
50/1
2.5
Thin mirrorc
0/1
50/1
1.5
4.5
Preformed
solvated
Electron
solutiond
0/1
1.85
50/1
5 (55%)e
a S.D.s were ‹3%.
b Procedure A was employed.
c Procedure C was employed.
d Procedure D was employed.
e Only 55% of 2-chloro-p-xylene was dechlorinated after 5
equivalents of Na had been consumed.

914
G.-R. Sun et al. / Chemosphere 41 (2000) 907±916
Scheme 2.
Table 5
E€ect of di€erent modes of addition on sodium consumption
the 2-chloro-p-xylene was dechlorinated when 5 equiv-
alents of Na had been consumed.
Dechlorination can be discussed in two steps:
Na/substrate (mole ratio)
required for complete
dehalogenation in
Substrate
H2O/
substrate
(mole ratio)
1. dissolution of the metal (M) in ammonia to give sol-
vated electrons and
di€erent procedures
2. reaction of solvated electrons with the substrate
(Scheme 2).
Procedure
A
Procedure
B
Solvated electrons react at a di€usion controlled rate,
completely reducing polychlorinated species before in-
termediates di€use away, even when a stoichiometric Na
de®ciency exists. As they form, solvated electrons de-
stroy all the RCl in the volume of solution near the Na.
Therefore, water remaining in that volume can react
with solvated electrons also remaining in that region
once RCl is consumed. The highest dechlorination
selectivity (lowest Na consumption per chlorine), would
occur at solvated electron production rates which are
slow enough to permit RCl to di€use into this region
near the metal surface. This allows the solvated electrons
to be consumed rapidly by chloroorganic species rather
than by water. Solvated electron generation rates in-
crease with more surface area. Use of larger Na particles
will decrease the overall rate of generating solvated
electrons. Thus, a lower consumption of Na per mole of
chlorine removed should be observed as the Na particle
size increases.
2-Chloro-
p-xylene
0/1
1.5
1.6
5/1
2.3
2.4
2.5
2.4
2.5
2.6
20/1
50/1
n-Chloro-
octane
0/1
5/1
1.7
2.3
3.5 (97%)a
7 (80%)b
1.9
2.3
2.5
4.0
20/1
50/1
a Only 97% of n-chlorooctane was dechlorinated.
b Only 80% of n-chlorooctane was dechlorinated at a cost of 7
equivalents of sodium consumption.
Table 6
E€ects of temperature in Na/NH3ꢀL† dehalogenations
Na/substrate (mole
ratio) required
for complete
Substrate
H2O/
substrate
A comparison of addition protocols (Procedure A vs
Procedure B) found no di€erences in the amount of Na
required to totally dechlorinate 2-chloro-p-xylene
(Table 5) However, a signi®cant di€erence in the amount
of Na required was observed in chlorooctane dechlori-
nations upon comparing Procedures A and B in the
presence of a 50 mol excess of water. Four equivalents of
sodium completely reduced 1-chlorooctane in Procedure
B, whereas only 80% of the 1-chlorooctane was reduced
with 7 equivalents of sodium in Procedure A. In this
speci®c case, chlorooctane has a low solubility in liquid
ammonia solutions containing a high mole fraction of
water. Chlorooctane requires several seconds to dissolve
completely into solution (H2O 0.9 g/NH3 10 ml) with
vigorous shaking. Thus, substrate was delivered more
slowly to the solvated electrons being formed in Proce-
dure A, thereby allowing water to compete more
e€ectively.
dechlorinationa
25°C
)55°Cb
2-Chloro-p-
xylene
0/1
1.5
2.0
50/1
0/1
2.5
2.0
3.0
2.0
3-Fluoro-o-
xylene
50/1
4.0
4.5
4-Chlorophenol
0/1
50/1
2.0
4.5
2.1
2.5
n-Chlorooctane
0/1
50/1
1.7
4.0
2.5
5.0
(15%)c
a S.D. ‹3%.
b Procedure E used in these experiments.
c Only 15% of n-chlorooctane was dechlorinated.
Temperature did not seriously a€ect the Na con-
sumption required to complete dehalogenation (Table 6)
except for 1-chlorooctane and 4-chlorophenol. The Na/
substrate ratios required with 2-chloro-p-xylene and
3-¯uoro-o-xylene (Table 6) did not change very much
with or without added water at room temperature or at
)55°C. The high Na consumption required for chlo-

G.-R. Sun et al. / Chemosphere 41 (2000) 907±916
915
Table 7
Activity of di€erent halides in dehalogenation of aromatic halides in Na/NH3 at 25°C
Substrate
H2O/substrate
(mole ratio)
Na/substrate required for complete dehalogenation
F
Cl
Br
2-X-p-xylene (X ˆ Br and Cl) or 3-¯uoro-o-xylene
4-X-phenol (X ˆ F, Cl and Br)
0/1
50/1
2.0
4.0
1.5
2.5
1.5
2.5
0/1
50/1
2.0 (18%)a
2.0
4.5
2.2
6.0
a Only 18% of 4-¯uorophenol was de¯uorinated in 20 hour reaction time.
rooctane was related to its poor solubility in aqueous
ammonia, especially at )55°C.
Only 2.5 equivalents of Na were required for com-
plete dehalogenation of 4-chlorophenol at )55°C vs 4.5
equivalents to achieve the same result at room tem-
perature. 4-Chlorophenol was soluble (as the phenox-
ide) in liquid NH3 at )55°C. However, ice crystals were
observed at this low temperature so the water con-
centration in solution was lower than at room tem-
perature. Substantial amounts of water were frozen at
)55°C in all dehalogenations, so it is dicult to assess
reasons for the change in Na consumption going from
ꢀ9†
The amount of Li, Na, K and Ca required for com-
plete dechlorination of 2-chloro-p-xylene was compared
(Table 8). The electrode potentials for loss of a single
electron in NH3 at )55°C are: Li (2.99 V), Na (2.59 V),
K (2.73 V) and Ca (2.39 V) (Pleskov, 1937a,b). If de-
chlorination rates followed the order of the electrode
potentials, the expected order of rates would be
Li > K > Na > Ca. However, if solvated electrons are
transferred at di€usion-controlled rates, the rate of
metal dissolution and other factors could play a more
important role in processes competiting for those elec-
trons (see discussion on particle size). Sodium gave the
lowest stoichiometric consumption of the four metals to
complete 2-chloro-p-xylene dechlorination, with or
without water. In the absence of water, 1.5 equivalents
of sodium were required, versus 2, 1.8 and 2 equivalents
of lithium, potassium and calcium, respectively. With a
50 mol excess of water, only 2.5 equivalents of Na were
25°C to )55°C.
3.5. Activity of di€erent halogens and metals in dehalo-
genations
Since the strength of C±X bonds increases in the
order C±I < C±Br < C±Cl < C±F, the reactivity decreases
in going from C±I to C±F (Austin et al., 1990). Mac-
kenzie et al. (1986) reported that no obvious halogen-
based (X ˆ F, Cl, Br) selectivity was found during
reductions of a-halogenated naphthalenes. Reductions
of haloxylenes and 4-halophenols were carried out in
Na/NH3 (Table 7) to compare the e€ect of di€erent
halides. In each case, more sodium was required for
de¯uorination than for dechlorination or debromina-
tion. This was true with or without added water. This
di€erence was larger in 4-halophenol dehalogenations.
Chlorophenol and bromophenol were completely de-
stroyed by 2 and 2.2 equivalents of Na in 30 s at
ambient temperature, respectively, with no water pre-
sent. However, 4-¯uorophenol was only 18% de¯uori-
nated after 20 h using 2 equivalents of sodium. Adding
water increased the Na requirement for all three halides
(Table 7).
Table 8
Activity of di€erent metals in dehalogenations of 2-chloro-p-
xylene at 25°C
H2O/substrate
(mole ratio)
Metal/substrate (mole ratio) required
for complete dehalogenationa
Li
Na
K
Ca
0/1
50/1
2.0
6.0
(76%)b
1.5
2.5
1.8
5.7
1.0
2.5
Fluorophenoxide ions are extremly dicult to de-
¯uorinate. Dechlorination of 4-¯uoro-2-chlorophenol,
with 3 equivalents of sodium, was fast, but after 3 h only
3% de¯uorination to phenol occurred. 4-Fluorophenol
was obtained in 97% yield (Eq. (9)).
(23%)c
a S.D. ‹3.5%.
b Seventy-six % of the substrate was dechlorinated.
c Twenty-three % of the substrate was dechlorinated.

916
G.-R. Sun et al. / Chemosphere 41 (2000) 907±916
Chuang, F.-W., Larson, R.A., Wessman, M.S., 1995. Environ.
Sci. Technol. 29, 2460.
required, versus 5.7 equivalents potassium to complete
dechlorination at those conditions. Lithium and calcium
reactions were very inecient in the presence of water,
giving only 76% and 23% dechlorination with 6 and 5
equivalents of metal, respectively. Surface passivation by
insoluble hydroxides (LiOH or Ca(OH)2), generated by
reaction with water, prevented rapid metal dissolution.
Reaction times of about 2 min were required in these
cases. Thus, sodium has a substantial advantage over
calcium, potassium and lithium in the presence of water
and it might be the metal of choice for use in soil/NH3
slurry remediations.
Crooks, R.M., Bard, A.I., 1987. J. Phys. Chem. 91 (5), 1274.
Eastham, J.E., Larkin, D.R., 1959. J. Am. Chem. Soc. 81,
3652±3655.
Gould, R.F. (Ed.), 1965. Advances in Chemistry. American
Chemical Society, Washington, DC, pp. 50.
Holm, T., 1999. J. Am. Chem. Soc. 121, 515.
Jessup, D.W., Paschal, J.W., Rabideau, P.W., 1977. J. Org.
Chem. 42, 2620±2621.
Knorre, H., Langer, M., Pohl, G., 1980. Degussa AG Patent
DE 2813200 C2.
Lahaniatis, E.S., Parlar, H., Klein, W., Korte, F., 1975.
Chemosphere 2, 83±88.
Liu, Y., Schwartz, J., Cavalloro, C.L., 1995. Environ. Sci.
Technol. 29, 836.
4. Conclusions
Mackenzie, K., Kopinke, F.D., Remmler, M., 1986. Chemo-
sphere 33 (8), 1495±1513.
Solvated electrons react at di€usion-controlled rates
with CCl4, CH3CCl3 and chlorinated aromatic com-
pounds under the conditions applied. The addition of
substantial amounts of water only modestly increased the
amount of Na needed to completely dechlorinate organic
substrates. These di€usion-controlled reductions were
complete in about 30 s, which was the time needed for
complete metal dissolution in liquid ammonia. Di€usion
control was supported by the absence of partially de-
chlorinatedproductsinreductionsofmultiplychlorinated
molecules in sodium-de®cient partial reductions. Only the
chlorinated phenols exhibited minor amounts of partially
dechlorinated products inaccordwitha greaterreluctance
of the corresponding phenoxides to add an electron.
The blue color of solvated electrons in NH3 near the
surface of the Na disappeared immediately in the pres-
ence of chlorinated substrates even at )55°C. Ecient
stirring, larger particle size and control of the dissolution
rate into NH3ꢀL† lead to the lowest Na consumption in
the presence of water. The consumption of sodium was
less than that of calcium, potassium and lithium, espe-
cially in the presence of water, suggesting sodium is the
metal of choice for remediation.
Pittman Jr., C.U., Mohammed, M.K., 1996. Preprint. Extended
Abstracts of the Special ACS Symposium: Emerging Tech-
nologies in Hazardous Waste Management VII. Birming-
ham, AL, 9±11 September, pp. 557±560.
Pittman Jr., C.U., Tabaei, S.M.H., 1994. Unpublished results,
Mississippi State University. Presented at the 1994±1995
SAC/IAC Meeting of the Gulf Coast Hazardous Substance
Research Center. Beaumont, TX.
Pittman Jr., C.U., Tabaei, S.S.H., 1993. Preprint. In: Tedder,
D.W. (Ed.), Extended Abstracts of the Special ACS
Symposium: Emerging Technologies in Hazardous Waste
Management V, vol. 2. Atlanta, GA, 27±29 September,
1993, pp. 557±560.
Pleskov, V.A., 1937a. Zh. Fiz. Khim. 9, 12.
Pleskov, V.A., 1937b. Acta Physicochim. URSS 6, 1.
Rabideau, P.W., Wetzeland, D.M., Yong, D.M., 1984. J. Org.
Chem. 49, 1544±1549.
Schindewolf, U., 1970. In: Lagowski, J.J., Sienko, M.J. (Eds.),
Metal±Ammonia Solution. Butterworths, London, pp. 199±
218.
Smith, M., 1968. In: Augustine, R.L. (Ed.), Dissolving Metal
Reductions, in Reduction: Techniques and Application in
Organic Synthesis. Marcei Dekker, New York, pp. 99±170.
Tabaei, S.M.H., Pittman Jr., C.U., 1993a. Hazardous Waste
and Hazardous Mater. 10 (4), 431±440.
Tabaei, S.M.H., Pittman Jr., C.U., 1993b. Tetrahedron Lett. 34
(20), 3263±3266.
Acknowledgements
Tabaei, S.M.H., Pittman Jr., C.U., Mead, K.T., 1992. J. Org.
Chem. 57, 6669.
Toxic Substances Control Act, 1976. Public Low, pp. 94±469.
Watt, G.W., 1950. Chem. Rev. 317±379.
Support of this research was provided by the US
Environmental Protaction Agency, grant no. GAD
#R826180 and by the Department of Interior, US
Geological Survey, Grant number HQ 96GR02679-12.
Weinberg, N.D.J., Abel, A.E., 1989. US Patent 4 853 040, 1
August.
Weinberg, N.D.J., Abel, A.E., 1992. US Patent 5 110 364, 5
May.
Yang, C.-M., Pittman Jr., C.U., 1996. Hazardous Waste and
Hazardous Mater. 13 (4), 445±464.
References
Yang, C.-M., Pittman Jr., C.U., 1997. Tetrahedron Lett. 38
(37).
Aiired, E.L., Beck, B.R., Voorkees, K.J., 1974. J. Org. Chem.
39, 1426.
Yang, C.-M., Pittman Jr., C.U., 1998a. Synthetic Commun. 28
(3), 517±525.
Austin, E., Alonso, R.A., Rossi, R.A., 1990. J. Chem. Res.,
190±191.
Birch, A.J., Subba Rao, G., 1972. Adv. Org. Chem. 8, 33.
Yang, C.-M., Pittman Jr., C.U., 1998b. Synthetic Commun. 28
(11), 2027±2041.