Destruction of pentachlorophenol using glow discharge plasma process

Article abstract of DOI:10.1021/es981001i

Pentachlorophenol (PCP), found in wood preservatives and pesticides, is an a cutely toxic recalcitrant organochlorine carcinogenic compound. A point- to-plane glow discharge plasma (GDP) process was used to study the destruction of 30-50 ppm (120-188 μM) sodium salt of pentachlorophenol (PCP) in an aqueous solution. PCP was converted to less than 10% of its initial concentration in 1-3 h, at room temperature and low pressure (50 Torr). Effects of varying the headspace gas chemistry, stirring rate, pH, and current upon rate of PCP conversion were investigated. Organic acids, including formate, acetate, butyrate, and oxalate, were formed as byproducts after 1-3 h of GDP treatment of PCP. The chloride recovery suggests 50-70% dechlorination. The PCP removal rate exhibited mixed order kinetics. A reaction model developed to verify mixed order kinetics compares well with the experimental data. The increase in order may be due to the production and subsequent destruction of several reaction products throughout the process. An increase n current and stirring rate increased the rate of PCP removal in the GDP reactor. However, the rate of PCP removal decreased when the initial pH of the solution is raised to 11.4. Rapid removal of PCP was observed when the headspace gas was argon, air, or oxygen. Bench scale data was used to compare the power efficiency of the GDP process with atmospheric pressure corona discharge. Results suggest that the cost of power for PCP conversion by glow discharge was less than that of atmospheric pressure corona discharge. Additionally, the operating cost for PCP destruction in aqueous solution using UV based advanced oxidation technologies was found to be comparable with power costs for PCP conversion using GDP. Pentachlorophenol (PCP), found in wood preservatives and pesticides, is an acutely toxic recalcitrant organochlorine carcinogenic compound. A point-to-plane glow discharge plasma (GDP) process was used to study the destruction of 30-50 ppm (120-188 μM) sodium salt of pentachlorophenol (PCP) in an aqueous solution. PCP was converted to less than 10% of its initial concentration in 1-3 h, at room temperature and low pressure (50 Torr). Effects of varying the headspace gas chemistry, stirring rate, pH, and current upon rate of PCP conversion were investigated. Organic acids, including formate, acetate, butyrate, and oxalate, were formed as byproducts after 1-3 h of GDP treatment of PCP. The chloride recovery suggests 50-70% dechlorination. The PCP removal rate exhibited mixed order kinetics. A reaction model developed to verify mixed order kinetics compares well with the experimental data. The increase in order may be due to the production and subsequent destruction of several reaction products throughout the process. An increase in current and stirring rate increased the rate of PCP removal in the GDP reactor. However, the rate of PCP removal decreased when the initial pH of the solution is raised to 11.4. Rapid removal of PCP was observed when the headspace gas was argon, air, or oxygen. Bench scale data was used to compare the power efficiency of the GDP process with atmospheric pressure corona discharge. Results suggest that the cost of power for PCP conversion by glow discharge was less than that of atmospheric pressure corona discharge. Additionally, the operating cost for PCP destruction in aqueous solution using UV based advanced oxidation technologies was found to be comparable with power costs for PCP conversion using GDP.

Full text of DOI:10.1021/es981001i

Environ. Sci. Technol. 2000, 34, 2267-2272
discharge has been applied for the removal of organic
Destruction of Pentachlorophenol
Using Glow Discharge Plasma
Process
contaminants such as dyes (7, 8), phenol (9, 10), benzene
(11), and carbon tetrachloride (12) from aqueous solutions
at the laboratory scale. Further, formation of radicals in
distilled water by nonuniform pulsed discharge has been
investigated by Sun et al. (13, 14). Effect of chemical species
produced by pulsed high-voltage discharge upon microor-
ganisms in water was studied by Sato et al. (15). Corona
discharge is an ionic and electronic emission from a high
voltage source, characterized by the formation and flow of
negative ions, positive ions, free radicals, and electrons in an
electric field between two or more electrodes (8, 16, 17).
Several researchers have investigated various forms of corona
discharge processes, and many books, monographs, and
review articles are available on the subject (18-20). In the
present investigation, continuous corona discharges were
created across a high voltage metallic point and a grounded
aqueous plane using a dc power supply (0-30 kV). The
pressure in the reactor was maintained below atmospheric
conditions, at 50 Torr. This apparatus was used to convert
PCP to other products in a 50 mL aqueous solution. In a GDP
process, nonthermal plasma is created on the liquid surface
by a high voltage electrode, located above the plane of the
liquid, when the electrode is energized. The process is carried
out under reduced pressure. The plasma region originates
at the high voltage electrode and extends to the grounded
liquid surface. It is believed that the charged species in the
plasma is accelerated due to a steep potential gradient and
enters the solution with a distribution of energy covering the
range up to the accelerating potential (21). These energetic
species may be responsible for causing chemical change in
the liquid phase. The unique feature with this type of
discharge is that the aqueous solution itself is one of the
electrodes.
Corona discharge under low-pressure conditions has been
described as glow discharge electrolysis (22, 23). Glow
discharge electrolysis is analogous to radiolytic or electron
beam processes; therefore, concepts developed in radiation
chemistry can be directly applied to this type of plasma
discharge process (24). However, the operating conditions
for GDP require operating pressure (50-200 Torr) lower than
the atmospheric pressure (760 Torr), while radiolysis is usually
carried out under atmospheric pressure. Since GDP is a low
energy process, the equipment cost for GDP is less than that
of radiolytic or electron beam processes. The destruction of
PCP by corona discharge under atmospheric conditions was
investigated by Brisset (25). He reported the destruction of
PCP by corona discharge obeyed pseudo-first-order kinetics.
However, the rate constant was determined using a spec-
trophotometer, and potential interference from byproducts
of PCP degradation would not have been detected. It is
suggested that a HPLC separation of reaction products would
provide a better understanding of the underlying mechanism.
Brisset reported a linear dependence upon current intensity
with treatment time; varying the current from 20 to 59 µA
resulted in increased PCP removal with increasing current.
Fang et al. (26) reported PCP decomposition in aqueous
solution by pulse and gamma radiolysis. They observed that
PCP removal increased linearly with increasing dose rate of
γ radiolysis. Fang also reported that OH radical reacts with
PCP by both electron transfer (53%) and addition, followed
by rapid HCl elimination to form phenoxyl radical. Phenoxyl
radical is unstable and decays to form other products.
The specific goal of the present work was to investigate
the kinetics and the rate law for PCP conversion from an
aqueous solution by a glow discharge plasma (GDP) process.
A . K . S H A R M A , * G . B . J O S E P H S O N , *
D . M . C A M A I O N I , A N D S . C . G O H E E N
Pacific Northwest National Laboratory,
Richland, Washington 99352
Pentachlorophenol (PCP), found in wood preservatives
and pesticides, is an acutely toxic recalcitrant organochlorine
carcinogenic compound. A point-to-plane glow discharge
plasma (GDP) process was used to study the destruction
of 30-50 ppm (120-188 µM) sodium salt of pentachlorophenol
(PCP) in an aqueous solution. PCP was converted to
less than 10% of its initial concentration in 1-3 h, at room
temperature and low pressure (50 Torr). Effects of
varying the headspace gas chemistry, stirring rate, pH,
and current upon rate of PCP conversion were investigated.
Organic acids, including formate, acetate, butyrate, and
oxalate, were formed as byproducts after 1-3 h of GDP
treatment of PCP. The chloride recovery suggests 50-70%
dechlorination. The PCP removal rate exhibited mixed
order kinetics. A reaction model developed to verify mixed
order kinetics compares well with the experimental data.
The increase in order may be due to the production and
subsequent destruction of several reaction products
throughout the process. An increase in current and stirring
rate increased the rate of PCP removal in the GDP
reactor. However, the rate of PCP removal decreased
when the initial pH of the solution is raised to 11.4. Rapid
removal of PCP was observed when the headspace gas
was argon, air, or oxygen. Bench scale data was used to
compare the power efficiency of the GDP process with
atmospheric pressure corona discharge. Results suggest
that the cost of power for PCP conversion by glow discharge
was less than that of atmospheric pressure corona
discharge. Additionally, the operating cost for PCP destruction
in aqueous solution using UV based advanced oxidation
technologies was found to be comparable with power costs
for PCP conversion using GDP.
Introduction
Chlorinated organics such as CCl4, perchloroethylene (PCE),
and trichloroethylene (TCE) and aromatics such as poly-
chlorinated benzene (PCB) and chlorinated phenols have
been a major environmental concern for the past several
years. Most of these chlorinated organics are highly toxic
and very stable and therefore are difficult to treat. Several
oxidation techniques such as peroxide assisted oxidation (1),
incineration (2), ultrasound (3), bioremediation (4), and
photocatalysis (5, 6) have been applied for the removal of
halogenated hydrocarbons and other organics from aqueous
solutions. Recently, a process called point-to-plane corona
* Corresponding authors. A.K.S.: phone: (509)373-4675; fax:
(509)376-2329;e-mail: Amit.Sharma@pnl.gov. G.B.J.: phone: (509)376-
4325; fax: (509)372-1861; e-mail: Gary.Josephson@pnl.gov.
9
10.1021/es981001i CCC: $19.00
Published on Web 04/14/2000
2000 Am erican Chem ical Society
VOL. 34, NO. 11, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 2 6 7

FIGURE 2. The effect of current on PCP destruction in an aqueous
solution was investigated using a GDP reactor. The initial PCP
concentration was 50 ppm (188 µM) for both experiments. All data
points are connected using trendlines.
plane was kept constant at 1.5 cm. There were five other
openings, two on each side and one in front of the reactor
(0. 5 cm internal diameter). These openings were used to
recirculate the liquid in the reactor, monitor the gas pressure,
and ground the liquid surface, as shown in Figure 1.
A reagent grade sodium salt of PCP was obtained from
Aldrich Chemical Company and used without further pu-
rification. Reagent grade sulfuric acid and acetonitrile were
obtained from Fisher Scientific. Reagent grade potassium
sulfate was obtained from Mallinckrodt and used without
any further purification. All analyses were carried out in
duplicate using HPLC. Sample PCP concentrations were
determined using a five point calibration curve. The analytical
method indicated good reproducibility with an average error
of ( 2%. Experiments were also repeated to establish
reproducibility. We estimated an average experimental error
of ( 11% to (15% for the reproducibility of the data points.
The data points shown on the figures are the average of two
trials.
The formation of acids in the GDP treated solution was
confirmed using ion chromatography. Analysis of the PCP
solution for organic acids, after treating with glow discharge
in air for 1-3 h, was carried out using a Dionex Ion
Chromatography unit equipped with an AS11 anion exchange
column and a conductivity detector.
Iodom etry. We used iodometry for measuring total
oxidant yield and number of oxidation events per electron.
Typically, 50-60-mL volumes of deionized water containing
2 g of KI and 0.1 g of sodium bicarbonate were place in the
liquid corona reactor and treated for approximately 60 min
at 100 µA, as described elsewhere (12). A 50 mL volume of
the treated solution was then titrated with a standardized
sodium thiosulfate solution. Control experiments showed
that no I2 was produced in the absence of the discharge.
The cost of power consumption necessary for PCP
destruction was determined from the bench scale data (50
mL) assuming an average cost of $0.06/ Kw-h.
FIGURE 1. Schematic of the glow discharge plasma (GDP) batch
reactor.
The results from the kinetic study of PCP conversion by a
GDP process are reported and discussed. This information
will assist in determining the economic feasibility of applying
GDP to waste treatment.
Experimental Methods
All experiments were performed in a batch mode. A schematic
identifying our experimental conditions is shown in Figure
1.
A cylindrical Pyrex glass reactor, with 5 cm outer diameter
and 16.5 cm outer height, was used to treat 33-50 ppm
aqueous solutions of the sodium salt of PCP for 1-3 h. A
magnetic stirrer (not shown in Figure 1) was used for stirring.
A FMI Lab Pump (model RHV) recirculated the liquid solution
for sample withdrawal. Samples were withdrawn after a fixed
time interval from a three way valve. The dead volume within
the recirculating tubing was approximately 5.5 mL and
therefore does not affect the treatment process, as the reactor
volume was approximately 10 times larger (50 mL). The
removal of PCP from the aqueous solution was monitored
using high-pressure liquid chromatography (HPLC). About
1 mL of sample was withdrawn from the reactor, and a 100
µL aliquot was injected to the HPLC system. A mobile phase
consisting of 52% acetonitrile and 48% water buffered with
0.1 M H2SO4/ 0.025 M K2SO4 was used for PCP elution. The
flow rate of the mobile phase was kept constant at 1 mL/
min. A reversed phase C-8 column was used to separate PCP.
A variable wavelength UV detector, set at 284 nm, was used
to detect and quantify PCP.
The glow discharge was established between a high voltage
stainless steel needle and the aqueous plane, by using an
Acopian power supply (0-30 keV dc, negative polarity).
Current was monitored by measuring the voltage across the
ground electrode with a 1000 ohm precision resistor, by a
Fluke 77 multimeter. The pressure in the reactor chamber
was monitored by using a Heise vacuum gauge (0-1000
mmHg, absolute). A vacuum pump was used to keep the
pressure at 50 Torr and to evacuate air from the reactor so
as to change the headspace gas chemistry to argon or oxygen.
The headspace gas was purged three times with the desired
gas, and the reactor was operated at 50 Torr as a closed
system with no gas entering or exiting the reactor.
The high voltage needle cathode was inserted from the
top of the Pyrex glass reactor through an opening (1. 5 cm,
internal diameter). This opening was sealed using a Teflon
fitting. The distance between the needle tip and the liquid
Results
The effect of several parameters upon the rate of PCP
conversion by GDP was investigated under a variety of
conditions. The effect of varying the current input to the
glow discharge reactor is shown in Figure 2. The PCP removal
rate was found to increase with the increase in the current
input to the glow discharge reactor, from 100 to 500 µA in
an air atmosphere. A rapid decrease in the initial PCP
concentration was observed within the first 20 min, followed
by a slower rate of PCP removal at longer times.
The effect of stirring rate on the kinetics of PCP destruction
for both 100 and 500 µA current input is shown in Figure 3.
9
2 2 6 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 11, 2000

FIGURE 5. The effect of pH on PCP conversion in the GDP reactor
was studied by varying the initial pH. The experiments were carried
out in unbuffered solution, and the initial pH was adjusted using
a 1 M NaOH solution. The current was kept constant at 100 µA, and
the initial PCP concentration was approximately 30 ppm (120 µM).
The data points are connected using trend lines.
FIGURE 3. The effect of stirring on PCP conversion in the GDP
reactor was studied. “Slow stirring” was defined as the stirring
without vortex formation (∼2 rotations s-1) and “fast stirring” was
defined as stirring with vortex formation (∼5 rotations s-1). The
initial concentration of PCP was 50 ppm (188 µM) for all experiments.
The data points are connected using trend lines.
Discussion
The focus of the present investigation was to determine the
economic feasibility of the GDP process for treating an acutely
toxic nonvolatile compound such as PCP. We have inves-
tigated the effect of several parameters upon the rate of PCP
removal in the glow discharge reactor. This information will
assist in process optimization and scale-up.
Effect of Current. The kinetics of PCP removal in the
glow discharge reactor is shown in Figure 2. We observed a
rapid initial decline in PCP concentration, followed by a
slower PCP removal rate, suggesting that the rate law for
PCP removal in the glow discharge reactor may follow mixed
order kinetics. This result is similar to that observed by Mills
(6) where a TiO2 assisted photocatalytic destruction of 12.5
ppm (47 µM) PCP solution also followed similar mixed order
kinetics. We developed a simple model to verify the observed
mixed order kinetics. This model is based upon a simplified
mechanism. To explain this mechanism, we define “reaction
zone” as the area on the liquid surface where active species
are produced via direct plasma-liquid interactions. It is
assumed that the “reaction zone” PCP concentration and
bulk PCP concentration are approximately equal for the low
current (100 µA) case. However, this assumption may not be
valid for the higher current case (500 µA) as the bulk PCP
concentration and “reaction zone” PCP concentration are
not equal at elevated current. Also the measurement of the
mass transfer coefficient as well as a relationship between
the PCP reaction zone concentration and the bulk PCP
concentration is needed. In the initial step, it is further
assumed that the active species (Ox) are formed, due to glow
discharge process, as shown in eq i. The rate of formation
of these species depends on the current input to the reactor.
This rate was determined by measuring the total oxidation
of potassium iodide solution and was found to be 9.3 × 10-6
M/ L-min for 100 µA.
FIGURE 4. The effect of headspace gas chemistry (air, argon, and
oxygen) on PCP destruction was studied in the GDP reactor. The
current was kept constant at 500 µA. The initial PCP concentration
for these experiments was approximately 30 ppm (120 µM). The
data points are connected using trendlines.
The stirring rate was varied (by magnetic stirrer) while keeping
the current constant. We defined “slow” stirring as the rate
at which the solution is stirred without forming a vortex at
a rate of approximately 2 rotations s-1. The “fast” stirring is
defined, as stirring with the formation of a vortex and it is
generally greater than 5 rotations s-1. Higher rates of PCP
removal were observed at faster stirring rates. Therefore, the
stirring was kept at a “fast” rate for subsequent experiments.
Figure 4 shows the effect of headspace gas composition
on PCP destruction rates in the GDP reactor. The rate of PCP
removal was initially slower in the oxygen atmosphere as
compared to air or argon; however, the reaction rapidly went
to completion in oxygen after approximately 30 min, as seen
in Figure 4.
e
- + H2O f Ox + e-
(i)
The effect of pH on PCP destruction rate in the GDP reactor
is illustrated in Figure 5. The variation in the normalized
PCP concentration with time has been plotted for pH 7.00
as well as pH 11.40. The rate of PCP removal was found to
be faster at pH 7.00 as compared to pH 11.40. Reactions at
a pH value below 7.00 were not investigated.
These active species react with PCP to form byproducts (B),
as shown in eq ii:
PCP + Ox f B
(ii)
In all experiments, conductivity of GDP treated PCP
solution was found to be significantly higher than that of
untreated PCP solution. For example, the initial conductivity
of a GDP treated PCP solution in an argon atmosphere was
0.28 mill-mho, and after 20 min treatment (at 2900 V, 1.7
mA), it was found to be 0.62 milli-mho.
Further, the byproduct B is also expected to react with the
active species Ox as shown in eq iii. It is assumed that the
product C formed in reaction iii is a relatively stable product
and does not consume a significant amount of active species:
Ox + B f C
(iii)
9
VOL. 34, NO. 11, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 2 6 9

The rate of reactions i, ii, and iii can be written as
r1 ) K1
(iv)
(v)
r2 ) K2[PCP][Ox]
r3 ) K3[B][Ox]
(vi)
Assuming pseudo steady state for the active species (Ox),
a mass balance on Ox leads to an algebraic equation:
K1
[Ox] )
(vii)
K2[PCP] + K3[B]
The rate of PCP consumption can be obtained by substituting
vii into v:
K1K2[PCP]
-d[PCP]
r2 )
)
(viii)
dt
K2[PCP] + K3[B]
Expressing [B] and [PCP] in terms of the extent of the first
reaction, (shown in eq i) “X” and integrating
K2(X + ln(1-X)) K1t
X -
(ix)
K2
[PCP]o
where t ) time and [PCP]o ) initial PCP concentration and
the extent of the first reaction “X” is defined as X ) ([PCP]o
- [PCP])/ [PCP]o ) (1 - [PCP]/ [PCP]o). From eq ii, B can be
expressed in terms of the extent of the first reaction, [B] )
[PCP]o x X, since all of B is formed via PCP and oxidant
reaction.
Using ix, values of X were determined as a function of
time, for a variety of K3/ K2 ratios. For first-order kinetics (K3/
K2 )1), the model was not in agreement with the experimental
data, as seen in Figure 6(a). However, we observed good
agreement between simulation and experimental data, for
100 µA and a K3/ K2 ratio of 1.5, indicating mixed order kinetics.
This is shown in Figure 6(b).
Initially, the products of the glow discharge process
(oxidants) are in gross excess of the PCP concentration.
Therefore, the reaction of PCP with the oxidants may yield
products (or intermediates) which may also react with the
oxidants at different rates. The decline in the reaction rate
during the later stages of the reaction may be due to the
increased consumption of the oxidants, due to their reaction
with the byproducts formed during the initial stages of the
PCP degradation.
Stirring Rate. The rate of PCP destruction was higher
when the stirring rate was fast (see Figure 4). This may have
been due to increased mass transfer as a result of efficient
mixing. The glow discharge process led to the formation of
volume filling plasma due to reduced gas density and
electrical resistance; therefore, under low pressure conditions,
the plasma filled the gap between the point electrode and
the liquid surface. This may lead to the production of most
oxidants, largely hydroxyl radicals, at the water surface, from
the water dissociation reaction:
FIGURE 6. (a and b): The kinetics for PCP degradation in the GDP
reactor at 100 µA current was compared with (a) a first-order kinetics
model and (b) the mixed order kinetics model developed using a
simple parallel reaction model. The first-order reaction model (K /
3
K ) 1) did not exhibit a strong fit with the experimental data (R 2
2
) 0.890). However the results from the mixed order, parallel reaction
model (K /K ) 1.5) and the experimental data fit more closely (R 2
3
2
) 0.991). The experimental data points are connected using trend
lines.
radicals. Our experiments with volatile organic compounds
(methanol and 1,1-dichloroethane) in the GDP reactor show
very high oxidation events/ electron (about 200-300 equiva-
lent/ Faraday) as opposed to the oxidation events measured
using a nonvolatile KI solution (about 9-10 equiv/ Faraday).
This result confirmed the formation of radicals in the gas
phase and that these gas-phase radicals may have minimal
impact upon liquid-phase reactions in the absence of good
mixing conditions. In addition, several researchers (24, 28-
30) have observed evidence of a variety of other oxidants in
glow discharge plasma. In the absence of good mixing, these
oxidants, which is a combination of several highly reactive
species including hydroxyl radical (23, 28), superoxide ion
(29), ozone (7, 8, 12, 30), and several other charged and
uncharged species (23, 28), may create a highly localized
concentration of reactive species at the gas-liquid interface.
This may render their recombination reactions significant
as compared to the reaction with a PCP molecule.
H2O glow discharge plasma8 OH• + H•
(x)
Thus, most oxidants may be present at the gas-liquid
interface and good mixing conditions would increase the
probability of the oxidants interacting with the pollutants
before they can recombine and become ineffective. At 50
Torr pressure and room temperature, approximately half of
the gas phase is composed of water vapor, since the vapor
pressure of water is 23.8 Torr at 25 °C (27); therefore, the
headspace gas may also contain an elevated level of hydroxyl
9
2 2 7 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 11, 2000

Headspace Gas Com position. To study the effect of
different gases upon the rate of PCP removal, the glow
discharge was carried out in air, argon, and oxygen by
evacuating the headspace of the GDP reactor and refilling it
with the desired gas. Since the reactor was sealed and isolated,
with only the liquid being recirculated, there was no exchange
of the headspace gas with the outside gas during the
experiment. We observed destruction of PCP in air, argon,
and oxygen gas. A rapid PCP degradation in argon, a noble
gas, suggests that reactive species were being produced via
direct interaction of the plasma with the water surface and
water vapor in the GDP reactor. Further, electron attachment
reactions are significant in glow discharges, and an argon
atmosphere will promote these reactions, since argon is a
monoatomic gas. We observed rapid PCP removal in the
presence of reactive gases such as air and oxygen. However,
it was found that the rate of PCP removal was initially slow
when the gas phase consisted of oxygen as compared to air
FIGURE 7. The cost of power utilized for PCP removal using corona
discharge was compared under a variety of conditions. The data
shown assumes power cost of $0.06/Kw-h. The data points are
connected using trend lines.
•
or argon. Higher initial yields of HO2 radicals, as opposed
to OH•, is expected in an oxygen atmosphere (23, 31). The
•
rate constant for reactions between HO2 radicals and PCP
of PCP in the GDP reactor was determined from the bench
scale data and extrapolated to the power cost for treating
1000 gallons of PCP. These data were compared with the
power cost for treating 1000 gallons of PCP by atmospheric
pressure corona discharge (see Figure 7) and two other
advanced oxidation technologies, a UV/ H2O2 based treatment
method, Perox-Pure (34), and a UV/ Ozone based process,
Ultrox (35).
is significantly smaller than that of hydroxyl radicals reacting
with PCP (32). This may explain lower initial rates of PCP
removal in an oxygen atmosphere. Rapid PCP removal, to
below detection limits (<0.01 ppm), was observed after
approximately 35 min of the experiment in the oxygen
atmosphere.
pH. All experiments were carried out in an unbuffered
neutral solution. The pH of the treated solution was found
to be 4.0 ((0.010) in most cases, indicating formation of
acids during the course of the experiment.
As seen in Figure 7 the power cost for treating over 90%
PCP by the GDP reactor is between $1-$2 per 1000 gallons,
which is significantly less than the cost of PCP treatment
using atmospheric pressure corona discharge ($7-8/ 1000
gallons). The notable difference in the two costs are due to
significantly higher power required to establish corona dis-
charge at atmospheric pressure as compared to low-pressure
conditions for the same current. In a corona discharge
process, the product of gas density “n” and the gap width
“h” (30) can influence the average density of the electrons
generated. As the product “n‚h” decreases, the average elec-
tron energy increases. For constant h, n would be lower at
low-pressure resulting in corona onset at lower power. Under
atmospheric pressure conditions, n is high for constant h,
requiring much higher power for the corona onset. However,
the information obtained by extrapolating bench scale data
from a batch reactor operating with 100 µA current input for
1-3 h should be interpreted with caution. In practice a con-
tinuous operation is desirable with very high flow rates (∼500
gallons/ h), and more current will have to be delivered. Fur-
thermore, a shorter residence time than 1-3 h is desired for
continuous operation. Nevertheless, it appears from the per-
formance enhancement obtained from the parametric studies
that the GDP process can compete with other waste treatment
technologies. Finally, GDP could become a viable alternative
for treating large quantities of hazardous organic waste.
To slow the recombination of HO2 radicals, we raised the
pH of the PCP solution to 11.40 using 1 M NaOH. Contrary
to our expectations, the rate of PCP destruction by GDP was
slower at pH 11.40 as compared to neutral pH. Hickling (22)
did not detect any hydrogen peroxide in a GDP process using
a strong alkaline solution (pH 13.00-12.00). He suggested
-
that this may be due to the decomposition of HO2 ion by
the discharge. Excess base may have scavenged the hydroxyl
radicals.
Product Analysis. The GDP converted at least 50% of the
chlorine in PCP to Cl- ion. In addition, organic acids were
formed from PCP conversion. Ion chromatography was used
to confirm the formation of the various organic acids. In air,
oxalate, formate, acetate, and butyrate were major degrada-
tion products. Both formate and acetate constituted ap-
proximately 60-70% of the organic acids, while the remaining
30-40% was found to be oxalate and butyrate. Recovery of
50-70% of the original chlorine in PCP as chloride ion was
also measured by ion chromatography. Complete mineral-
ization of 1 PCP molecule should lead to the formation of 5
Cl- ions:
C6Cl5O-Na+ + 11H2O glow discharge8
6CO2 + 5Cl- + 22H+ + 18e- + Na+ (xii)
Acknowledgments
Pacific Northwest National Laboratory is operated by Battelle
Memorial Institute for the U.S. Department of Energy under
contract DE-AC06-76RLO 1830. The authors would like to
extend thanks to Georgia Ruebsaman for text processing and
editing. Assistance provided by Wendy Shaw and Scott Clauss
is gratefully acknowledged. Peer review suggestions made
by Dr. Karen Wahl and Ms. Therese R. W. Clauss is greatly
appreciated. Part of the results from this investigation were
presented at the ACS symposia, “Emerging Technologies in
Hazardous Waste Management VII” in Atlanta, GA in
September 1995 and at the “International Symposium on
Environmental Applications of Advanced Oxidation Tech-
nologies” organized by EPRI and DOE in San Francisco, CA
in February 1996.
It has been reported that aqueous phenolics can undergo
oxidative coupling induced by ozone and oxygen (33).
However, we did not find “dimers” of PCP when a methylene
chloride extract of the PCP solution (treated by glow discharge
in air) was analyzed by GC-MS. We found tetrachlorophenol
(TCP) after 20 min of glow discharge exposure as measured
by GC-MS. Thus, TCP may be an early breakdown product
of oxidant attack on PCP.
Cost Com parison. As seen in Figure 2, the rate of
destruction of PCP was proportional to the current input to
the GDP reactor. In other words, the cost of power used
during the process of PCP destruction governs the operating
cost. Therefore, the cost of power for over 90% destruction
9
VOL. 34, NO. 11, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 2 7 1

(19) Loeb, L. B. In Electrical Coronas, Their Basic Physical Mecha-
nisms; University of California Press: Berkeley, CA, 1965.
(20) Gallo, C. F. IEEE Trans. IAS 1977, 13, 6, 550.
(21) Tezuka, M.; Iwasaki, M. Asian Pacific Conference on Plasma
Science and Technology III; 1996; Vol. 2, p 423.
(22) Davies, R. A.; Hickling, A. J. Chem. Soc. 1952, 3595.
(23) Kanzaki, Y.; Nishimura, N.; Matsumoto, O. J. Electroanal. Chem.
1984, 167, 297.
(24) Hickling, A. In Modern Aspects of Electrochemistry; Bockris, J.
O’M., Conway, B. E., Eds.; Butterworth: London, 1971; Vol. 6,
pp 329-373.
(25) Brisset, J. L. J. Trace Microprobe Technol. 1998, 16(3), 363.
(26) Fang, X.; He, Y.; Liu J.; Jilan, W. Rad. Phys. Chem. 1998, 53, 411.
(27) CRC Handbook of chemistry and physics, 72nd ed.; Lide, D. R.,
Ed.; CRC Press: Boca Raton , FL, 1991; Section 6.
(28) Cserfalvi, T.; Mezei, P.; Apai, P. J. Phys. D: Appl. Phys. 1993, 26,
2184.
(29) Kellogg III, E. W.; Yost, M. G.; Barthaleur, N.; Krueger, A. P.
Nature 1974, 281, 400.
(30) Ellision, B.; Kogelschatz, U. IEEE Trans. Plasma Sci. 1991, 19,
309.
(31) Mazzocchin, G.; Bontempelli, G. J. Electroanal. Chem. 1976, 67,
191.
(32) Hoigne, J.; Bader, H. Water Res. 1983, 17, 185.
(33) Chrostowski, P. C.; Dietrich, A. M.; Suffet, I. H. Water Res. 1983,
17, 1627.
Literature Cited
(1) Franklin, T. C.; Greg, O.; Nnodimele, R.; Couch, K. J. Electrochem.
Soc. 1992, 139, 2192.
(2) Dickson, L. C.; Karasek, F. W. J. Chromatogr. 1987, 389, 127.
(3) Petrier, C.; Micolle, M.; Merlin, G.; Luche, J.; Reverdy, G. Environ.
Sci. Technol. 1992, 26, 1639.
(4) Jarvinen, K. T.; Melin, E. S.; Puhakka, J. A. Environ. Sci. Technol.
1994, 28, 2387.
(5) Davis, A. P.; Huang C. P. Water Res. 1990, 24, 543.
(6) Mills, G.; Hoffmann, M. R. Environ. Sci. Technol. 1993, 27, 1681.
(7) Goheen, S. C.; Durham, D. E.; McCulloch, M.; Heath, W. O.
Proceedings of the Second International Symposium on Chemical
Oxidation: Technologies for the Nineties; 1992; p 356.
(8) Clements, J. S.; Sato, M.; Davis, R. H. IEEE Trans. Industrial
Applications 1987, IA-23, 224.
(9) Sharma, A. K.; Locke, B. R.; Arce, P.; Finney W. C. Hazard. Waste
Hazard. Mater. 1993, 10, 209.
(10) Joshi, A. A.; Locke, B. R.; Arce, P.; Finney, W. C. J. Hazard. Mater.
1995, 41, 3.
(11) Warner, K.; Sharma, A.; Locke, B. R.; Arce, P.; Finney, W. C.
Presented at the Annual Meeting of the Florida Sections of
American Chemical Society; Orlando, FL, May 1993; Paper 113.
(12) Sharma, A. K.; Camaioni D. M.; Josephson, G. B.; Goheen, S. C.;
Mong, G. M. J. Adv. Oxid. Technol. 1997, 2(1), 239.
(13) Sun, B.; Sato M.; Clements, J. S. J. Electrostatics 1997, 39, 189.
(14) Sun, B.; Sato M.; Harano A.; Clements J. S. J. Electrostatics 1998,
43, 115.
(15) Sato, M.; Ohgiyama, T.; Clements, J. S. IEEE Trans. IAS 1996,
32, 106
(16) Joshi, A. A. M.S. Thesis, FAMU/ FSU College of Engineering,
1994.
(34) Perox-pure Chemical Oxidation Technology Peroxidation Sys-
tems; Applications Analysis Report, EPA/ 540/ AR-93/ 501; 1993.
(35) Ultrox International UV Radiation/ Oxidation Technology; Ap-
plications Analysis Report, EPA/ 540/ A5-89/ 12; 1990.
(17) Sharma, A. K. M.S. Thesis, FAMU/ FSU College of Engineering,
1993.
(18) Chang, J. S.; Lawless, P. A.; Yamamoto, J. IEEE Trans. Plasma
Sci. 1991, 6(19), 1152.
Received for review September 28, 1998. Revised manuscript
received August 2, 1999. Accepted March 1, 2000.
ES981001I
9
2 2 7 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 11, 2000