Sonolytic destruction of methyl tert-butyl ether by ultrasonic irradiation: The role of O3, H2O2, frequency, and power density

Article abstract of DOI:10.1021/es9810383

The kinetics of degradation of methyl tert-butyl ether (MTBE) by ultrasonic irradiation in the presence of ozone as functions of applied frequencies and applied power are investigated. Experiments are performed over the frequency range of 205-1078 kHz. The higher overall reaction rates are observed at 358 and 618 kHz and then at 205 and 1078 kHz. The observed pseudo-first-order rate constant, k(O), for MTBE degradation increases with increasing power density up to 250 W L^-1. A linear dependence of the first- order rate constant, k(O3), for the simultaneous degradation of O₃ on power density is also observed. Naturally occurring organic matter (NOM) is shown to have a negligible effect on observed reaction rates. The kinetics of degradation of methyl tert-butyl ether (MTBE) by ultrasonic irradiation in the presence of ozone as functions of applied frequencies and applied power are investigated. Experiments are performed over the frequency range of 205-1078 kHz. The higher overall reaction rates are observed at 358 and 618 kHz and then at 205 and 1078 kHz. The observed pseudo-first-order rate constant, k₀, for MTBE degradation increases with increasing power density up to 250 W L^-1. A linear dependence of the first-order rate constant, k_O(3), for the simultaneous degradation of O₃ on power density is also observed. Naturally occurring organic matter (NOM) is shown to have a negligible effect on observed reaction rates.

Full text of DOI:10.1021/es9810383

Environ. Sci. Technol. 1999, 33, 3199-3205
In our preceding paper (14), the kinetics and mechanism
Sonolytic Destruction of Methyl
tert-Butyl Ether by Ultrasonic
Irradiation: The Role of O , HO ,
of sonolytic degradation of MTBE were investigated at a single
ultrasonic frequency of 205 kHz. The addition of ozone into
the aqueous sonolytic system resulted in an increased rate
of MTBE destruction. The generalized reactions can be
described by the following equations.
3
2 2
Frequency, and Power Density
)
)), H O
2
(
CH ) COCH
8 (CH ) COH + CH OH
(3)
3
3
3
3 3
3
J O O N - W U N K A N G , † H U I - M I N G H U N G ,
A N G E L A L I N , A N D
M I C H A E L R . H O F F M A N N *
)
)), O , H O
3 2
(
3 3
CH ) COCH3
8 (CH ) COH + HCO H (4)
3 3
2
)
)), H O
2
W. M. Keck Laboratories, California Institute of Technology,
Pasadena, California 91125
(
CH ) COH
8 (CH ) CO H, CH OH, HCHO (5)
3 3
3 3
2
3
)
)), O3
)))
(
CH ) COCH
8 (CH ) COOCH3
9
8 CH OH, HCO H
3
3
3
3 3
3
2
(
6)
The kinetics of degradation of methyl tert-butyl ether
(
MTBE) by ultrasonic irradiation in the presence of ozone
where the symbol ))) indicates ultrasonic irradiation. The
main intermediates observed during MTBE sonolysis and
ozonolyis are tert-butyl formate and tert-butyl alcohol. In
this paper, the combined effects of sonolysis and ozonolysis
on the kinetics of MTBE degradation are explored over a
broader frequency range.
as functions of applied frequencies and applied power
are investigated. Experiments are performed over the
frequency range of 205-1078 kHz. The higher overall reaction
rates are observed at 358 and 618 kHz and then at 205
and 1078 kHz. The observed pseudo-first-order rate constant,
k0, for MTBE degradation increases with increasing
-
1
Experimental Methods
power density up to 250 W L . A linear dependence of
the first-order rate constant, kO , for the simultaneous
MTBE (99.9%; EM science) and sodium bicarbonate (Reagent
Grade; EM Science) were used without further purification.
A humic acid reagent was prepared by dissolving solid humic
acid obtained from Fluka AG in a 0.1 N NaOH solution and
then filtering through 0.45 µm filter paper.
3
degradation of O3 on power density is also observed.
Naturally occurring organic matter (NOM) is shown to have
a negligible effect on observed reaction rates.
Ultrasonic irradiations were performed with an Ultrasonic
Transducer USW 51 (AlliedSignal ELAC Nautik, Inc.) in a
glass and titanium reactor, which has a vibrational surface
area of 25 cm2 and can be operated at four different
frequencies: 205, 358, 618, and 1078 kHz. The reaction
volume, as shown in Figure 1, is contained within a 600 mL
cylindrical, double-walled (i.e., water-cooled) reaction vessel,
that has four sampling ports on the top which are used for
gas venting, for withdrawing aqueous samples, and for the
introduction of background gases.
Introduction
Methyl tert-butyl ether (MTBE) which is blended into
gasoline, is frequently found in groundwater, where it is slowly
degraded by aerobic and anaerobic processes (1).
The application of ultrasound for the treatment of
chemical contaminants in water has been explored in recent
years (2-8). Ultrasonic waves, which consist of compression
and rarefaction cycles, produce cavitation bubbles in liquid
solution. After several compression cycles, the cavitation
bubbles collapse violently and adiabatically with extremely
high temperatures up to 5000 K and pressures of 975 bar
Temperature was maintained constant at 23 ( 3 °C with
a 20 °C refrigerated water bath (Haake Co., model A80).
Aqueous solutions were made with water obtained from a
MilliQ UV + purification system. Ozone solutions were
prepared with an Orec Ozonator (model V10-0) by bubbling
(
9,10). Under these extreme conditions, volatile chemical
compounds are destroyed by direct pyrolysis reactions and
indirectly by reactions with H , OH, O , and H O . The
2 2
-
1
•
•
•
ozone into deionized water at a flow rate of 100 mL min
through a glass fritted diffuser until the desired ozone
concentration in the aqueous phase was obtained. Ozone
concentrations were obtained spectrophotometrically using
a HP 8452 diode array spectrophotometer; the molar extinc-
addition of ozone into sonolytic systems often enhances the
overall rate of destruction of a wide variety of compounds
(
6,11,12).
The thermal decomposition of ozone in cavitation bubbles
-
1
-1
3
tion coefficient for O in water at 260 nm is 3300 M cm .
•
leads to the formation of OH as shown in eqs 1 and 2 (11,13).
In addition, the direct reaction of ozone with substrates may
occur.
MTBE stock solutions (100 mM) were prepared and stored
at 4 °C. After the desired aqueous ozone concentration was
obtained, an appropriate volume of the MTBE stock solution
was spiked into 500 mL of ozonated solution to initiate the
sonochemical reaction. Additional mixing was achieved by
3
O f O + O( P)
(1)
(2)
3
2
2 3
bubbling O / O gas for a few more seconds at a low flow rate
O( P) + H O f 2•OH
3
2
in order to minimize any sparging effects. The first sample
was taken immediately after closing off the ozone supply.
After the t ) 0 sample was taken, ultrasonic irradiation was
initiated. Sample aliquots (i.e., 1.0 mL) were taken at
appropriate time intervals during each kinetic run and stored
in 20 mL Teflon-capped, aluminum-sealed vials. In the case
*
To whom correspondence should be addressed. Phone: (626)
3
95-4391; fax: (626) 395-3170; e-mail: mrh@cco.caltech.edu.
†
Current address: Department of Industrial Environment and
Health, Yonsei University Wonju Campus, 234 Maeji, Wonju, KOREA
220-710).
(
2 2 3
of dissolved ozone, 10 µL of 1 N Na S O was used to quench
1
0.1021/es9810383 CCC: $18.00

1999 Am erican Chem ical Society
VOL. 33, NO. 18, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3 1 9 9
Published on Web 08/04/1999

FIGURE 1. Schematic diagram of the sonochemical-reactor system.
the residual aqueous ozone to preserve samples prior to the
analysis. Since the total amount of samples withdrawn from
the reactor was kept to less than 15 mL, only a small error
(
<3%) was introduced due to the differential volume change.
During each kinetic run, the ultrasonic power inputs were
kept in the range of 50-240 W, depending on the particular
ultrasonic frequency that was used. For frequencies greater
than 600 kHz, the maximum applied power was 50 W.
Compounds extracted into the gas phase of the HP 7694
headspace sampler were autoinjected into a HP 5890 series
II GC-FID equipped with a HP-624 capillary column (30 m
×
0.32 mm × 1.8 um) with a 70 °C isothermal oven
temperature. Sample aliquots for hydrogen peroxide analysis
were taken at 10 min intervals and were measured by
spectrofluorometer (9). Samples taken for TOC (total organic
carbon) analysis were filtered with 0.45 µm Teflon syringe
filters (Gelman) before injection and were measured with a
Shimadzu 5000A total organic carbon analyzer.
FIGURE 2. Effect of power density on the pseudo-first-order rate
constant, k , for the destruction of MTBE by ultrasonic irradiation
at 205 and 358 kHz: [MTBE] ) 1 mM at 205 kHz; [MTBE] ) 0.05
mM at 358 kHz.
0
0
0
sonochemical reaction rate. However, the cavitation bubble
size, the bubble collapse time, the transient temperature,
and the internal pressure in the cavitation bubble during
collapse are all dependent on the power intensity. For
example, the power intensity, which is proportional to the
applied power density, is a function of the acoustic amplitude,
For some runs, the aqueous ozone concentrations were
continuously measured spetrophotometrically with circula-
tion of the aqueous solution from the reactor through the
spectrophotometer and back into the reactor.
Results and Discussion
Effects of Applied Power Input. In our earlier study, which
was carried out at 205 kHz, the sonolytic degradation of MTBE
was shown to follow pseudo-first-order kinetics. Plots of the
A
P as follows:
2
I ) PA / 2Fc
(7)
0
pseudo-first-order rate constants, k , versus the input power
density are shown in Figure 2 for sonolytic irradiation at 205
and 358 kHz with initial MTBE concentrations of 1 and 0.05
mM, respectively. The apparent rate constant, k0,358, at 358
kHz is higher than k0,205 by a factor of 5. From our previous
study (14), the MTBE sonication rate constant at 205 kHz
where I is the sound intensity (amount of energy flowing per
unit area per unit time), F is the density of the medium, and
c is the velocity of sound in the medium.
For a cavitation bubble, the maximum bubble size is
dependent on the density of the liquid, the applied frequency,
the hydrostatic pressure, and the acoustic pressure as follows
was found to be faster with [MTBE]
MTBE] ) 1 mM by a factor of 2 (i.e., the rate decreases with
increase initial substrate concentration). Furthermore, as
shown in Figure 2, k increases rapidly with an increase in
0
) 0.05 mM than for
[
0
(
9):
0
1
/ 2
1/ 3
-
1
4
ωa
2
2
3Ph
power density up to 240 W L . In general, an increase in
power density results in an increase in the observed
Rmax
)
(P - P )
1 +
(P - P )
(8)
A
h
A
h
(FPA
)
[
]
3
3
2 0 0
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 18, 1999

a h
where ω is the applied acoustic frequency and P is the
external (hydrostatic) pressure, which is 1 atm under our
experimental conditions. In addition, the bubble collapse
time, τ, is proportional to the maximum bubble size, R
m
(9):
1
/ 2
τ ) 0.915R (F/ P ) (1 + P / P )
(9)
m
m
vg
m
where P
m
is the pressure in the liquid (i.e., P
m
) P
h
a
+ P ) and
P
vg is the vapor pressure in the bubble. Therefore, at high
A
acoustic intensity (i.e., large P values) the cavitation bubbles
are able to grow larger in size during a rarefaction cycle such
that insufficient time is available for complete collapse during
a single compression cycle. As the above equations predict
and the experimental results show, there is an optimum
power density which can be applied during sonochemical
irradiation in order to obtain maximum reaction rates.
2 2
Effects of Variable Frequencies on H O Production. In
an attempt to determine the optimum frequency to obtain
the most beneficial chemical effects, hydrogen peroxide
production was used as a direct indicator of free radical
production. Hydrogen peroxide is produced during the
sonolysis of water (5,9,13). During sonolysis, hydrogen atoms
and hydroxyl radicals are formed due to the pyrolytic
decomposition of water.
•
•
H O f H + OH
(10)
2
Hydroxyl radicals can recombine to form hydrogen peroxide:
•
•
OH + OH f H O
(11)
2
2
FIGURE 3. (a) Effect of frequency on hydrogen peroxide production
rate at power density ) 240 W/L. (b) Effect of frequency on hydrogen
peroxide production rate at power density ) 100 W/L.
2
If O is present in the vapor phase of the bubble, peroxide
can also be produced via the production of the hydroperoxyl
radical as follows:
•
•
H + O f HO
(12)
(13)
2
2
HO2• + HO2 f H O + O
.
2
2
2
•
•
Since the self-reaction rate constants of HO and OH at 298
2
5
9
-1 -1
K are 8.3 × 10 and 5.5 × 10 M
s
in solution (15) and 3
-
12
-11
3
-1 -1
×
10
(16) and 1.5 × 10
cm molecule
s
(17) in the
•
gas phase, respectively, the OH recombination reaction
should be the major route for the hydrogen peroxide
2 2 2
formation. The sonolytic production rates of H O in O -
saturated solutions relative to Ar-saturated solutions were
compared at 205 kHz with a power density of 240 W/ L. The
apparent H
2 2
O production rates for these two conditions
-
4
-
were found to be identical; i.e., d[H
2
O
2
]/ dt ) 3.5 × 10 mM
FIGURE 4. I
3
production rate at 205, 358, 618, and 1078 kHz with
-
1
-1
[KI] ) 0.1 M, temperature 23 ( 3 °C, and power density 84 W L .
min . This result supports the argument that the self-reaction
0
(
eq 11) of OH is the principal pathway for H
2
O
2
production.
2 2
Plots of hydrogen peroxide (H O ) production versus time
•
Recombination of OH to form H
2
O
2
most likely takes
with a power input of 240 W/ L are shown in Figure 3a for
four different sonolytic frequencies: 205, 358, 618, and 1078
place in the bubble phase instead of the liquid phase (18).
It should also noted that hydrogen peroxide is the most likely
-
1
kHz. At a power input of 100 W L , a similar trend with a
variation in frequency is observed in Figure 3b. However, in
•
7
-1 -1
OH scavenger (kOH,H O2 ) 2.7 × 10 M
s
) in the absence
2
•
of other OH scavengers (19).
this case the rates of H
lower applied power density (i.e., lower P
2
O
2
production are lower due to the
). The peroxide
2 2
production rate, d[H O ]/ dt, at 240 W L is approximately
A
•
•
OH + H O f HO + H O
(14)
-1
2
2
2
2
-
1
twice that at 100 W L and is clearly highest at 358 kHz and
lowest at 1078 kHz. From the data shown in Figure 3a, it is
worthwhile to note that peroxide production at 358 kHz is
nonlinear, whereas for the other frequency. This may be due
to the fact that the self-scavenging reaction between OH
radical and H O (eq 14) becomes significant as peroxide
The rate of hydrogen peroxide production during the
sonolysis of water is given by eq 15. The actual production
rates of OH and H
the bubble collapse temperature, the collapse pressure, and
the bubble lifetime.
•
2
O
2
during sonolysis will be affected by
2
2
accumulates in the liquid phase. Hua and Hoffmann (5)
reported that d[H ]/ dt increased with increasing frequency
for 20, 40, 80, and 500 kHz. However, it is noted that the
peroxide production rate is the lowest at the highest applied
d[H O ]/ dt ) kOH,OH[OH][OH] + k
[HO ][HO ] -
2 2
O
2
2
HO ,HO2
2
2
2
k
[OH][H O ] - k [H O ] (15)
2 2 pyr 2 2
OH,H O2
2
VOL. 33, NO. 18, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3 2 0 1

TABLE 1. Enhancement Effect of Ozone on the
Pseudo-First-Order Rate Constants of MTBE Destruction
k
0
, 10- s-1 at 205 kHz
4
k
0
, 10-4 s-1 at 358 kHz
Rc
[
MTBE],
O
3
,
3
O ,
mM
mM USa
O
3
-US Eb mM US
3 3
O O -US
-US Eb US
0
0
0
1
.01 0.30 8.5 33.2 3.9 0.20 16.5 88.3 5.4 1.9 2.7
.25 0.31 6.9 14.9 2.2 0.23 15.3 32.2 2.1 2.2 2.1
.50 0.34 5.4 12.2 2.3 0.23 10.4 18.6 1.8 1.9 1.5
.00 0.26 4.1
6.3 1.5 0.23 6.8 12.3 1.8 1.7 2.0
a
US ) ultrasound. b Enhancem ent factor ) k
0
3
(O -ultrasound)/
c
k
0
(ultrasound). Ratio ) k
0
(358kHz)/k
0
(205kHz).
FIGURE 5. Effect of frequency on the first-order ozone degradation
-1
-
3
rate with power intensity ) 100WL ,[O
3
] ) 0.14-0.15 mM, [HCO ]
ultrasonic frequency of 1078 kHz. At these very high
frequencies, the rarefaction (and compression) cycles are
very short. The time required for the rarefaction cycle is too
short to permit a cavitation bubble to grow to a size sufficient
)
1 mM, and pH ) 8.25.
to cause optimal disruption of the liquid.
-
The I
2
(I
3
) production rate during sonolysis at four
different frequencies is shown in Figure 4. The iodine
•
formation rate can be related to the OH production rate by
-
a stoichiometric factor of 2. The sequence of I oxidation is
as follows:
•
-
-
•
OH + I f OH + I
(16)
(17)
•
-
•-
I + I f I
2
I2•- + I f I3
•
-
(18)
(19)
(20)
I2•- + OH f I + OH
•
-
FIGURE 6. Effect of power density on the first-order ozone decay
rate constant, k , for the sonolytic ozonolysis at 358 kHz: [O ] )
0.17-0.18 mM; [HCO
2
d
3
I2 + I h I3-
-
-
3
] ) 1 mM; pH ) 8.25.
-) production at 293
2 3
Again, the optimum frequency for I (I
ment in the sonochemical process, the role of ozone in the
sonolysis process was studied at the other target frequencies.
In this study, the sonolysis rate of ozone was investigated as
-
1
K and I
result is similar to that observed for H
the reaction volume for H can be with the vapor phase
a
) 84 W L is located between 358 and 618 kHz. This
2
O
2
production although
a function of frequency, power density, and [MTBE]
0
. Aqueous
2
O
2
ozone concentrations were determined by using continuous
of the collapsing bubble or within the interfacial region of
-
1
-
flow spectrophotometry at λ ) 260 nm, where ꢀ ) 3300 M
1
the bubble, whereas I
3
formation is restricted to the liquid
-
cm
.
phase (20).
In order to observe the relative effects of frequency on
Enhancem ent Effects of O
05 and 358 kHz. The combination of O
05 kHz has been shown to enhance the MTBE degradation
in
3
on MTBE Degradation at
the ozone decomposition rate during sonication, initial
aqueous ozone concentrations were adjusted to 0.13-0.15
mM before sonication and the aqueous ozone concentrations
were monitored at 10 s intervals. Experiments were performed
2
2
3
and ultrasound at
rate (14). In this study, the enhancement effect of O
3
sonolytic MTBE degradation at 358 kHz was investigated
and compared with the 205 kHz data (Table 1). The direct
reaction of MTBE with ozone has a pseudo-first-order
-
1
at a fixed power density of 100 W L for all frequencies in
-
a 1.0 mM bicarbonate [HCO
3
] solution. As shown in Figure
-
5
-1
5, the sonolytic degradation rate of ozone is found to be first
order at all frequencies. Thus, the apparent rate equation for
constant of 6 × 10
s
for [MTBE]
0
) 0.01 mM and 6.7 ×
-
6
-1
1
0
s
for [MTBE]
0
) 1 mM. These reaction rates are small
the sonolytic degradation of O
1, where k is the apparent first-order degradation rate
constant. On the basis of the data, k is lowest at the highest
3
by sonolysis is given by eq
enough to be ignored when compared with the effects of
sonication. From Table 1, it is clear that the combination of
2
d
d
O
3
and ultrasound increases k
greater at lower [MTBE] . The enhancement factors (k
ultrasound)/ k
similar at higher initial MTBE concentrations and increase
slowly as [MTBE] decreases. However, at lower concentra-
tions (i.e., 0.01 mM), the enhancement factor for 358 kHz
5.4) is higher than it is at 205 kHz (3.9). The k ratios at 358
and 205 kHz for the ultrasound system alone and the
combined O -ultrasound system are also given in Table 1.
For the ultrasound alone system, k at 358 kHz was greater
than at 205 kHz by a factor of 1.7-2.2. Similarly, in the case
of the O -ultrasound system, the k at 358 kHz was greater
at lower [MTBE]
0
. The enhancement in k
0
is
-
ultrasonic frequency, 1078 kHz, while the measured rates at
the other frequencies are nearly the same.
0
0
(O
3
0
(ultrasound)) for 205 and 358 kHz are very
-
d[O ]/ dt ) k [O ]
(21)
3
d
3
0
The relative effect of applied power density on the ozone
decay rate was investigated at 358 kHz in 1.0 mM bicarbonate
spiked distilled water. As shown in Figure 6, a linear
dependence of the first-order rate constant for ozone with
a variation of power density can be quantified as follows:
(
0
3
0
3
0
k ) γ + γP
(22)
0
.
d
0
The Role of Ozone in Sonochem istry. The combination
of ozone and sonolysis was found to be significantly more
Where γ
0
-
is the ozone decay rate in the absence of ultrasound
1
effective in degrading MTBE than either ultrasound or O
alone (14). To illustrate the mechanism of ozone enhance-
3
(0.0011 s ), i.e., the rate due to the base-catalyzed reaction
-
-5
k
O ,OH-[OH ], γ is the slope of the k
d
vs P line (7.0 × 10
3
3
2 0 2
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 18, 1999

FIGURE 9. Comparison of the calculated and measured ozone
degradation rate by the O -ultrasound process at 358 kHz: power
3
-
1
-
3
density ) 100 W L ; [MTBE]
1 mM; pH ) 8.25.
0
) 0.5 mM; [O ] ) 0.16 mM; [HCO ]
3
0
FIGURE 7. Effect of [MTBE] on the first-order ozone degradation
)
rate for the O -ultrasound process at 358 kHz: power density )
3
-
1
-
3
1
3
00 W L ; [O ] ) 0.16-0.18 mM; [HCO ] ) 1 mM; pH ) 8.25.
where the first term represents the ozone decay rate in the
absence of MTBE and the second and third terms are the net
increased rates of ozone decay when MTBE is present.
Under these conditions, the sonolytic degradation of
MTBE in the presence of O
3
is given by
-
k0t
[
MTBE] ) [MTBE] e
(26)
t
0
∆
[H
system in the presence of MTBE compared to the MTBE-free
sonolysis. The ∆[H ] vs time profile obtained from
measuring the [H ] production rate at [MTBE] ) 0.5 mM
relative to [MTBE]
2 2 2 2
O ] is the net increase of [H O ] in the combined
2
O
2
2
O
2
0
0
)0 mM is shown in Figure 8. The apparent
increase in d[H
2
O
2
]/ dt can be attributed to the following
FIGURE 8. Net increased peroxide concentration, ∆[H
2
O
2
], versus
reactions (14):
time profile by the O -ultrasound process at 358 kHz: power density
3
-
1
-
)))∆
)
100 W L , [MTBE]
0
) 0.5 mM; [O
3
] ) 0.16 mM; [HCO
3
] ) 1 mM;
•
•
H O
8 H + OH
(27)
(28)
2
pH ) 8.25.
)
))∆
•
L s-1 W-1) and P is the power density in watts per liter. The
overall decomposition rate, k , is the sum of the rate due to
the base-catalyzed reaction pathway, the in situ thermal
pyrolysis) decomposition rate, and the ozone decay rate
O3
8 O + O
2
d
•
•
(
CH ) COCH + OH f (CH ) COCH + H O (29)
3 3
3
3 3
2
2
(
•
•
(
CH ) COCH2 + O f (CH ) COCH O
(30)
due to the conjugate base of H
intermediate.
2
O
2
produced as a sonolysis
3 3
2
3 3
2
2
•
•
(
CH ) COCH O + OH f (CH ) COCHO + H O (31)
3
3
2
2
3 3
2
2
-
+
H O T HO + H , pK ) 11.8
(23)
2
2
2
a
The [H
process. We suppose the ∆[H
polynomial equation as given by
2 2
O ] contributed from MTBE is a complicated
2
O
2
] vs time is a second-order
Thus, k
d
is given by
-
+ 10pH-pKak
3
kd ) k
O ,OH
-
[OH ] + k
O ,pyr
O ,HO2
-
[H O ] (24)
2
2
2
3
3
∆[H O ] ) A t + A t
(32)
2
2 t
2
1
-
8
The effects of [MTBE]
position rates are shown in Figure 7. In the absence of
sonication, the ozone decay rates were not affected by the
0
on the ozone sonolytic decom-
The fitted result is shown in Figure 8 with A ) 5.56 × 10
2
-
4
and A ) 2.0 × 10 . The influence to [H O ] from the first
1
2
2
term is smaller than the second one at t < 100 min.
Substituting eqs 26 and 32 into eq 25 and rearranging in
an integrated form yield
[
MTBE]
of ozone with MTBE at ambient temperature. However, the
effect of [MTBE] on the ozone decomposition rate is
0
due to the negligibly low rate of the direct reaction
0
-
k0t
significant for sonolytic ozonolysis, as shown in Figure 7. In
this case, the ozone decomposed at a faster rate with
-
ln[O ]/ [O ] ) k + (1 - e )k
[MTBE] / k +
3
3 0
d
O ,MTBE
0
3
0
3
2
1
0pH-pKa
k
-
(A t / 2 + A t / 3) (33)
increasing [MTBE]
0
. Furthermore, as [MTBE]
0
increases, the
O ,HO2
1
2
3
ozone decomposition rates deviate from a strict first-order
relationship. This effect may be attributed to the direct
reaction of MTBE with ozone at the cavitation bubble
interface area or to the enhanced ozone decomposition rate
where kO ,HO2
-
3
is the second-order rate constant of O and
3
-
-
6
-1 -1
HO
2
(kO3,HO2 ) 5.6 × 10 M
s
) (21). The pK
a
-
for H
]. The rate
2
O
2
)
11.8 and the pH ) 8.25 in the 1.0 mM of [HCO
3
due to the increased H
MTBE (14). Thus, in the presence of MTBE, the sonolytic
decomposition rate of ozone, d[O ]/ dt, is given by
2
O
2
production in the presence of
constant, kO3,MTBE, is estimated by fitting eq 33 to the
experimental data. The fitted result is shown in Figure 9. In
this case, kO3,MTBE ) 2.5 M
-
1
-1
3
s . The third term of eq 33
accounts for the deviation from pseudo-first-order kinetics.
In addition, the apparent enhancement in kO3,MTBE may be
due to enhanced mass transfer between the bubble and liquid
in the presence of ultrasound.
-
d[O ]/ dt ) (k + k
[MTBE] +
O ,MTBE
3
d
3
1
0pH-pKa
k
-
[H O ])[O ] (25)
O ,HO2
2 2 3
3
VOL. 33, NO. 18, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3 2 0 3

of the sonolysis of MTBE has an optimal value at 358 kHz
with a power density of 100 W/ L. At the same power density,
Petrier and Francony (22,23) reported that the sonolysis of
4
phenol, a less volatile compound than CCl , has an optimal
rate constant at 200 kHz compared to 20, 500, and 800 kHz.
The resonance radius of the bubble as described by Mason
-
1
(
21) is proportional to frequency so the surface area-to-
volume ratio is larger at higher frequencies. The rectified
diffusion rate could also be enhanced at higher frequencies.
For this reason the observed rate constant appears to be
faster at 358 kHz than at 205 kHz.
Effect of Background Organics. Naturally occurring
organic matter (NOM) and certain inorganic species in natural
•
waters are potential competitive reactants for OH. Therefore,
these background components of natural waters may affect
the observed oxidation rates. In order to test the potential
impact of this type of competition for hydroxyl radical during
sonication, MTBE was degraded in the absence and presence
of NOM (i.e., made from Fluka AG) at concentrations of 2.1
and 4.2 mg/ L. As shown in Figure 11a, the effect of NOM on
the MTBE decomposition rate appears to be negligible. The
concentration of TOC as a function of time in the presence
of humic acid is shown in Figure 11b. This result confirms
previous observations that the major reaction site for MTBE
FIGURE 10. Effects of ozone and frequency on the pseudo-first-
order rate constants for MTBE destruction: power density ) 100
-
1
W L ; [MTBE] ) 0.05 mM.
0
•
with OH is in the vapor phase of the cavitation bubble and
not in the bulk aqueous phase. For example, highly volatile
compounds such as MTBE and CCl are more rapidly
4
decomposed during sonolysis than are semivolatile or
nonvolatile compounds.
The results of this study demonstrate that the combination
of ozonation and sonication can degrade MTBE more
effectively at an ultrasonic frequency of 358 kHz compared
to other available frequencies.
Acknowledgments
Financial support from the U. S. Department of Energy (Grant
DOE 1 963472402) via the Argonne National Laboratory (Dr.
Robert Peters) is gratefully acknowledged.
Literature Cited
(
(
(
(
(
(
(
(
1) Suflita, J. M.; Mormile, M. R. Environ. Sci. Technol. 1993, 27,
76-978.
2) Serpone, N.; Terzian, R.; Hidaka, H.; Pelizzetti, E. J. Phys. Chem.
994, 98, 2634-2640.
3) Kotronarou, A.; Mills, G.; Hoffmann, M. R. Environ. Sci. Technol.
992, 26, 1460-1462.
4) Kotronarou, A.; Mills, G.; Hoffmann, M. R. J. Phys. Chem. 1991,
5, 3630-3638.
5) Hua, I.; Hoffmann, M. R. Environ. Sci. Technol. 1997, 31, 2237-
243.
6) Hua, I.; Hoffmann, M. R. Environ. Sci. Technol. 1996, 30, 864-
71.
7) Hua, I.; Hochemer, R. H.; Hoffmann, M. R. J. Phys. Chem. 1995,
9, 2335-2342.
9
1
1
9
2
8
9
8) Hua, I.; Hochemer, R. H.; Hoffmann, M. R. Environ. Sci. Technol.
1995, 29, 2790-2796.
FIGURE 11. (a) Effect of NOM on the MTBE degradation rate by
-
1
ultrasound at 358 kHz with a power density ) 100 W L . (b)
Concentration of TOC as a function of time in the presence of humic
acid.
(9) Mason, T. J.; Lorimer, J. P. Sonochemistry: Theory, Application
and Uses of Ultrasound in Chemistry; John Wiley & Sons: New
York, 1988.
(
(
10) Flint, E. B.; Suslick, K. S. Science 1991, 253, 1397-1399.
11) Hart, E. J.; Henglein, A. J. Phys. Chem. 1986, 90, 3061-3062.
Effects of Frequency and Ozone Concentration. The
effects of frequency and ozone dose on the net MTBE
degradation reaction rate were explored in order to obtain
the overall optimal conditions. The pseudo-first-order rate
(12) Olson, T. M.; Barbier, P. F. Water Res. 1994, 28, 1383-1391.
(
(
13) Hart, E. J.; Henglein, A. J. Phys. Chem. 1985, 89, 4342-4347.
14) Kang, J. W.; Hoffmann, M. R. Environ. Sci. Technol. 1997, 32,
3
194-3199.
(
15) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B. J.
Phys. Chem. Ref. Data 1988, 17, 513-886.
constant, k
shown in Figure 10. For each frequency, ozone doses were
varied from 0 to 0.19 mM at [MTBE] ) 0.5 mM at a power
0
, versus ozone concentration and frequency is
(16) Tsang, W.; Hampson, R. F. J. Phys. Chem. Ref. Data 1986, 15,
087-1279.
17) Zellner, R.; Ewig, F.; Paschke, R.; Wagner, G. J. Phys. Chem.
988, 92, 4184-4190.
1
0
(
(
-
1
density of 100 W L . From the 3-D plot, it can be seen that
1
ultrasonic frequencies of 358 and 618 kHz are the most
effective for MTBE degradation. Without O , the rate constant
3
18) Henglein, A.; Kormann, C. Int. J. Radiat. Biol. Relat. Stud. Phys.
Chem. Med. 1985, 48, 251-258.
3
2 0 4
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 18, 1999

(
(
19) Christensen, H.; Sehested, K.; Corfitzen, H. J. Phys. Chem. 1982,
6, 1588-1590.
20) Hung, H.-M.; Hoffmann, M. R. The Effect of Frequency,
Hydrostatic Pressure, and Sonication Power on Sonochemical
Reactions. Manuscript in preparation.
(22) Petrier, C.; Francony, A. Water Sci. Technol. 1997, 35, 175-180.
(23) Petrier, C.; Francony, A. Ultrason. Sonochem. 1997, 4, 295-300.
8
Received for review October 7, 1998. Revised manuscript
received March 29, 1999. Accepted June 16, 1999.
(
21) Staehelin, J.; Hoigne, J. Environ. Sci. Technol. 1982, 16, 676-
6
81.
ES9810383
VOL. 33, NO. 18, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3 2 0 5