Oxidation kinetics of phenolic and indolic compounds by ozone: applications to synthetic and real swine manure slurry

Water Research 36 (2002) 1513–1526

Oxidation kinetics of phenolic and indolic compounds byozone: applications to synthetic and real swine manure slurry

Jerry J. Wu1, Susan J. Masten*

Department of Civil and Environmental Engineering, Michigan State University, East Lansing, MI 48824, USA

Received 1 July 2000; received in revised form 1 April 2001; accepted 1 April 2001

Abstract

In this study, an oxidation model combining the mass transfer of ozone and ozonation kinetics was developed topredict the degradation of several phenolic and indolic compounds in a semi-batch reactor. The mass transfer andpartition coefficients were calculated at various physical and chemical conditions. In addition, the reaction rate

constants of ozone with phenolic and indolic compounds were also estimated independently using the method ofcompetition kinetics and relative reaction-rate constants. Incorporating mass transfer and chemical reaction concepts,an oxidation model that considers side reactions between ozone and byproducts has been established using non-linear

simultaneous differential equations. Thus, numerical computation is capable of simulating the degradation of phenolicand indolic compounds both in synthetic and real manure. r 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Ozone; Swine manure slurry; Malodorous metabolites; Mass transfer; Oxidation kinetics

1. Introduction

Phenolic (phenol, p-cresol, and p-ethylphenol) andindolic (indole and skatole) compounds are usuallyregarded as amongst the most odorous malodorous

compounds found in swine manure slurry. Althoughmany researchers have investigated the kinetics of thereaction of ozone with phenol and p-cresol [1–6], no

effort has been made to investigate the kinetics of thereaction of these compounds in a complex matrix suchas wastewater. While in full-scale contactors ozone-

containing gas is bubbled into the liquid being treated,few researchers have determined the mass transfer andkinetics of the reactions involving ozone in such

heterogeneous systems. In a few cases, empirical kineticmodels of ozonation have been proposed for hetero-geneous systems in order to determine an apparent rate

constant for the overall reaction as a function of both

flow and chemical parameters [7], but these constantsare limited to the specific reactor type used.The objectives of this study were threefold. Firstly, we

sought to describe the kinetics of the oxidation of

phenolic and indolic compounds by ozone in a semi-batch reactor. The effect of such system parameters,such as flowrate, temperature, pH, and solution

composition on the removal of these malodorouscompounds was determined. Lastly, a mathematicalmodel, which combines the mass transfer of ozone and

ozonation kinetics, to predict the degradation of thosecompounds in synthetic and real swine manure wasdeveloped.

2. Materials and methods

2.1. Determination of stoichiometric factors andreaction rate constants

The stoichiometric factors for the reactions wereobtained by mixing solutions of known concentrations

*Corresponding author. Tel.: +1-517-353-8539; fax: +1-

517-355-0250.

E-mail address: [email protected] (S.J. Masten).1Present address: Department of Environmental Engineering

and Science, Feng-Chia University, Taichung, Taiwan.

0043-1354/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved.

PII: S 0 0 4 3 - 1 3 5 4 ( 0 1 ) 0 0 3 5 2 - 9

of a single aqueous target compound (phenol, p-cresol,p-ethylphenol, indole, or skatole) and ozone. In order

to avoid, as much as possible, the interference ofother reactions (i.e., between ozone and byproducts),the initial ratios of concentrations of the targetcompounds [M]o to ozone [O3]o were chosen to

be between 5 and 10mol/mol, thus allowing ozoneto be consumed predominately by its reaction withthe parent compound(s). Experiments were conducted

at two different concentrations of the targetorganic chemical: 100 and 200mg/L. An appropriatevolume of the stock ozone solution (10.2mg/L for

pH=2.1 and 9.2mg/L for pH=6.7) was addedto the reaction vessel and allowed to react with thetarget compound. These experiments were conducted

in 30mL glass reaction vessels into which wereplaced mini-stir bars to keep the solution completelymixed. After the aqueous ozone was consumed, theresidual concentrations of the compounds were deter-

mined by high-performance liquid chromatography(HPLC). The stoichiometric factor was then computedas Z ¼ D[O3]/D[M]. All experiments were conducted in

triplicate.The rate constants for the reaction of the

target compounds and ozone were estimated using

the method of competition kinetics [8] usingrelative reaction rate constants [3]. A pair of organiccompounds (M1 and M2) in some aqueous solutionwill compete for ozone; the kinetics of these com-

petition reactions can be described mathematically as

follows:

d½M1�dt

¼ �k1;O3½M1�½O3�; ð1Þ

d½M2�dt

¼ �k2;O3½M2�½O3�: ð2Þ

Thus,

d½M1�d½M2�

¼k1;O3

½M1�k2;O3

½M2�; ð3Þ

k2;O3¼ k1;O3

lnð½M2�t=½M2�0Þlnð½M1�t=½M1�0Þ

: ð4Þ

p-Cresol was chosen as the reference compound since the

reaction rate constant of ozone with p-cresol ispublished in the peer-reviewed literature [6]. From thisreaction rate constant, the reaction rate constant of the

other target compounds can be calculated.

2.2. Ozonation of synthetic and real swine manure in asemi-batch reactor

In order to better understand the reaction kineticsbetween ozone and the target malodorous components,a formulation of synthetic swine manure was prepared

referring to the concentrations of major malodorouscomposition reported by Yasuhara [9]. The synthesisswine manure contains phenol (28.1mg/L), p-cresol

(210mg/L), p-ethylphenol (3.5mg/L), indole (5.1mg/L), and skatole (12.8mg/L) [10]. In these experiments,

Nomenclature

Cing gaseous concentration of inlet ozone, mol/L

Coutg gaseous concentration of outlet ozone, mol/L

E enhancement factor by chemical reactionKphenol equilibrium constant between phenol and

pheolate ion in solution, mol/LkLa overall mass transfer coefficient of ozone in

solution in the absence of ozone-reactive

compounds, 1/skd self-decomposition rate constant of dis-

solved ozone, 1/s for m equal to 1 and

L/mol s for m equal to 2ki reaction rate constant for malodorous

compound i with dissolved ozone, L/mol sks rate constant for the removal of malodor-

ous compound, i; by stripping or oxygena-tion, 1/s

kA reaction rate constant between phenol and

ozone, L/mol skB reaction rate constant between phenolate

ion and ozone, L/mol s

m self-decomposition order of dissolved ozonemi molar flux of ozone at the reactor inlet, mol/smo molar flux of ozone at the reactor outlet, mol/s

[M]i concentration of malodorous compound,mol/L

NO3actual ozone absorption rate defined by

equation, mol/L s[O3] dissolved ozone concentration, mol/L[O3

*] equilibrium dissolved ozone concentration

at the water–gas interface, mol/LR minimum residual for the observed and

predicted data of target compounds in thetested solution

V liquid volume of the reactor, LXj predicted concentration for the parent

compounds in the reactor, mol/L

XExperiment observed concentration during the experi-ment, mol/L

a partition coefficient of ozone in solution

Zi stoichiometric factor for malodorous com-pound, i; with ozone

J.J. Wu, S.J. Masten / Water Research 36 (2002) 1513–15261514

the temperature (141C, 221C, 301C) and flowrate (100,300, 500mL/min) were varied to investigate their effect

on the rate of ozonolysis of the target malodorouscompounds. The pH of the synthetic manure wascontrolled using a phosphate buffer system (pH=6.7).

The ionic strength of the synthetic manure was 0.1M. Asemi-batch stirred reactor (Fig. 1) made of pyrex glasswith a capacity of 1.5 L was utilized to investigate thedegradation of target compounds in the synthetic

solution.Ozone was generated from pure oxygen by corona

discharge using an ozone generator (Model T-408,

Polymetrics Inc., San Jose, CA). The oxygen streamwas dried using a molecular sieve trap (S/P Brand#G5301-21) prior to ozone generation. In this reactor, a

fritted glass diffuser that is able to generate fine bubbles(bubble diameter o1mm) was located near the bottomof the reactor and a magnetic stirrer bar was placed

under the diffuser to achieve complete mixing through-out the liquid. In addition, a water jacket around thereactor maintained the desired temperature. The con-centrations of gaseous ozone were monitored using a

UV spectrophotometer (UV 1201, Shimadzu, Japan) bypassing ozone gas, from both the inlet and outlet,through two 2-mm flow cells. The absorbance of the

ozone was monitored at the wavelength 258 nm. Anextinction coefficient of 3000M�1 cm�1 was used toconvert absorbances to concentration units [6]. A ozone

concentration of 64.8mg/L in the oxygen gas stream wasused in these experiments. The dissolved ozone concen-

tration was measured continuously by circulating waterfrom the reactor through the flow chamber of thespectrophotometer. Before initiating each experiment, a

three-way valve was used to pass the ozone gas into a2% potassium iodine solution (trap) until it wasobserved, spectrophotometrically, that the inlet concen-tration was constant. When the desired concentration of

inlet ozone was obtained, the experiment was com-menced by rotating this valve and admitting the ozone-enriched oxygen gas into the reactor. A control

experiment, in which the manure slurry was spargedwith oxygen using the same flowrate as that used in theozone experiments, was conducted to investigate the

effect of stripping and oxygenation. Phenolic and indoliccompounds in the synthetic and real manure wereextracted by methylene chloride (1 : 1). Gas chromato-

graphy then was employed to separate and quantifythese compounds [11].Swine manure was collected from the storage pits

under one of the swine houses at MSU Swine Teaching

and Research Facility. After collection, the manure wasplaced in a 4L glass bottle for two weeks to allowfermentation to occur. After storage, oxygenation and

ozonation experiments were conducted at a temperatureof 221C and a flowrate of 300mL/min. Samples weretaken at ozone applied dosages from 0 to 2 g/L at

Fig. 1. System setup for the semi-batch ozonation process.

J.J. Wu, S.J. Masten / Water Research 36 (2002) 1513–1526 1515

increments of 0.25 g/L. At the ozone gas-phase concen-tration used, 19.3min was needed to obtain an incre-

mental increase of 0.25 g/L ozone.

2.3. Model development

A model combining mass transfer and chemicaloxidation kinetics was developed to predict the accu-

mulation of dissolved ozone concentration and thedecrease in the concentrations of malodorous com-pounds in the synthetic and real swine manure during

lab-scale ozonation. The parameters regarding thetransport of ozone in pure water had been determinedusing a mass transfer model of ozone developedpreviously [12]. Our study showed the transport of gas

through the reactor could be described as plug flow,supporting the work of Nishikawa et al. [13] and Zhouand Smith [14] describing the transport of gas in a

reactor. Incorporated into this model were the conceptsof enhancement factor, defined as the ratio between theactual and maximum physical absorption rates [15], and

chemical kinetics. The following basic assumptions weremade:

1. the liquid phase is completely mixed;2. the flow pattern of gaseous ozone through the system

is plug flow;

3. constant temperature and pressure are maintainedduring the experimental period;

4. ozone is consumed mainly by the malodorous parent

compounds (phenol, p-cresol, p-ethylphenol, indole,and skatole); and

5. the enhancement factor is constant during the

ozonation process.

Assuming plug flow conditions in the gas phase, theaverage driving force [16] can be described either as thearithmetic-mean driving force, [O3]

*=aðCin þ CoutÞ=2;or as the logarithmic-mean driving force, [O3]

*=

aðCin � CoutÞ=lnðCin=CoutÞ: Since the ozone concentra-tion of outlet gas ðCoutÞ is essentially zero duringreaction with our target compounds, the logarithmic-

mean driving force could not be applied because of themathematical error in logarithm that would haveoccurred. Model I describes mathematically the mass

transfer and the ozonation reactions by the followingequations:

d½O3�dt

¼EkLa aCin

g þ Coutg

2

!� ½O3�

!

�X

ZiðkiÞ½O3�½M�i � kd½O3�m; ð5Þ

�d½M�idt

¼ ks½M�i þ ðkiÞ½O3�½M�i; ð6Þ

where [O3] is the dissolved ozone concentration (mg/L),E the enhancement factor by chemical reaction (dimen-

sionless), kLa the mass transfer coefficient of ozone insolution in the absence of ozone-reactive compounds

(1/s), a the partition coefficient of ozone (dimensionless),Cin

g the gaseous concentration of inlet ozone (mol/L),Cout

g the gaseous concentration of outlet ozone (mol/L),

Zi the stoichiometric factor for malodorous compound i(mol/mol), ki the reaction rate constant for malodorouscompound i with dissolved ozone (L/mol s), [M]i theconcentration of malodorous compound (mol/L), kd the

self-decomposition rate constant of dissolved ozone (1/sfor m equal to 1; L/mol s for m equal to 2), m the self-decomposition order of dissolved ozone (dimensionless),

ks the coefficient for the removal of malodorouscompound i; by stripping or oxygenation (1/s).The enhancement factor can be expressed mathema-

tically as [17,15]

E ¼NO3

kLa½O*3 �; ð7Þ

NO3¼

mi � mo

V; ð8Þ

where NO3(mol/L s) is determined from the difference

between the molar flux of ozone at the reactor inlet andoutlet, mi and mo (mol/s), kLa is the liquid phase

volumetric mass transfer coefficient (1/s), [O3*] is the

equilibrium dissolved ozone concentration at the water–gas interface (mol/L), and V is the liquid volume in the

reactor (L). The enhancement factor was determined inthe solution containing all target compounds.Another important feature to be considered is the

occurrence of competition reactions involving ozone andthe intermediates formed during the ozonation. Ozona-tion of phenolic compounds yields dihydroxybenzenes

as the initial byproducts. These compounds are still asreactive with ozone as the parent phenols [1,5]. Thesecircumstances make kinetic studies more difficult tocarry out especially for an initial reaction system with

more than one parent compound. Therefore, assump-tion 4 in Model I should be discarded for this situation(Model II). To overcome the difficulty in using the

mathematical model, a first-order reaction rate constant,kf ; was used to describe the consumption of ozone by allintermediates or impurities. This reaction rate constant

ðkf Þ is incorporated into Model II, which can beexpressed mathematically as

d½O3�dt

¼EkLa aCin

g þ Coutg

2

!� ½O3�

!

�X

ZiðkiÞ½O3�½M�t � kf ½O3� � kd½O3�m; ð9Þ

�d½M�tdt

¼ ks½M�t þ ðkiÞ½O3�½M�i; ð10Þ

where kf is the first-order reaction rate constant for the

reaction of ozone with the byproducts or othercompounds (1/s).

J.J. Wu, S.J. Masten / Water Research 36 (2002) 1513–15261516

The overall reaction coefficient was estimated byfitting our experimental data to find a minimum residual

ðRÞ; which is defined as below:

R ¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiPNj¼1 ðXj � XExperimentÞ

2

N

s; ð11Þ

where Xj is the predicted concentration of the parent

compounds in reactor, XExperiment is the observedconcentration during the experiment, and N is thenumber of total observations.

3. Results

3.1. Stoichiometric factors and rate constants for

ozonation reactions

The stoichiometric factors and rate constants for the

reaction of ozone with phenolic and indolic compoundsare summarized in Table 1. Results show that thestoichiometric factors for phenol are 1.8170.16 at pH

2.1 and 1.8670.16 at pH 6.7. For p-cresol, thestoichiometric factors were found to be 1.0070.07 atpH 2.1 and 1.3870.11 at pH 6.7. The value reported byZheng et al. [6] for cresol isomers was 3; the value

documented by Beltran et al. [18] for o-cresol was 2.Using the stoichiometric factor for p-cresol determinedin this work, it appears that approximately 1mol of

ozone is consumed by 1mol of p-cresol. Due to the lackof references relating to the ozonation kinetics of p-ethylphenol, indole, and skatole, no comparison could

be made for the stoichiometric factors of thesecompounds. In our work, the stoichiometric factor ofp-ethylphenol was found to be approximately 1.2 over

the pH range used. For indolic compounds, however,

the stoichiometric factor at pH 6.7 was less than 1 (butclose to 1).

To evaluate the effects of stripping and oxy-genation on the removal of the phenolic and in-dolic compounds, oxygen was used to purge the

solution. The maximum flowrate, 500mL/min, wasemployed. Almost no change occurred in the concentra-tion of any of the target compounds during a period of80min.

3.2. Determination of the enhancement factor and

reaction rate constant ðkf Þ

As mentioned previously, a model combining masstransfer and oxidation kinetics was developed to predict

the concentration profiles of phenolic and indoliccompounds in water under different pH values andflowrates. The mass transfer coefficients, partition

coefficients (Table 2), stoichiometric factors, and reac-tion rate constants were determined previously [12]. Theenhancement factor was determined at a temperature of

221C and is presented in Table 3. According to Eqs. (7)and (8), the enhancement factor can be regarded as aconstant since the gaseous ozone concentration in theoutlet stream was not detectable until the phenolic or

indolic compounds were removed from the reactor.Phenolic mixtures (phenol, p-cresol, p-ethylphenol)

and indolic mixtures (indole and skatole) were ozonated,

respectively, under the following conditions: pH 2.1,temperature of 221C, flowrate of 300mL/min, and ionicstrength of 0.1M. The reaction coefficients that resulted

in minimum residuals obtained using Model II are 0.8and 1.0 for the phenolic compounds and indoliccompounds, respectively. At pH 6.7, the reaction rate

constants ðkf Þ are much larger than those found at pH

Table 1

Stoichiometric factors and rate constants for the reaction of ozone with phenolic and indolic compounds under different pH conditions

Stoichiometric factor Reaction rate constant (M�1 s�1) Conc. at t0 (mM) Final conc. (mM)

pH=2.1, I ¼ 0:1; T ¼ 221C, initial DO3¼ 0:213mM

Phenol 1.8170.16 460710 2.13 2.01

p-Cresol 1.0070.07 5200a 1.89 1.68

p-Ethylphenol 1.1370.05 16507280 1.69 1.52

Indole 0.9870.07 45,00072900 1.71 1.49

Skatole 0.9070.03 10,20072200 1.55 1.31

pH=6.7, I ¼ 0:1; T ¼ 221C, initial DO3¼ 0:192mM

Phenol 1.8670.16 750,000733,000 2.12 2.01

p-Cresol 1.3870.11 850,000a 1.89 1.75

p-Ethylphenol 1.2970.11 990,0007110,000 1.69 1.54

Indole 0.9870.07 5,000,00073,600,000 1.70 1.50

Skatole 0.8970.04 4,500,0007780,000 1.54 1.32

aValue was determined by Zheng et al. [6] and CL were not reported.

Errors are reported at the 95% confidence interval.

J.J. Wu, S.J. Masten / Water Research 36 (2002) 1513–1526 1517

2.1. The fitted values for the reaction rate constants atpH=6.7 are summarized in Table 4.

3.3. Oxidation of synthetic manure and stored swinemanure by ozone

Synthetic manure and stored ‘‘real’’ swine manurewere ozonated to investigate the oxidation of phenolicand indolic compounds in a complex matrix. In thesynthetic manure, at ozone dosages greater than 0.1 g/L,

p-ethylphenol, indole, and skatole were removed tobelow detection limits, whereas ozone dosages of greaterthan 0.5 g/L were necessary to remove phenol and

p-cresol to below detection limits. Stripping andoxygenation were ineffective at removing any of thephenolic or indolic compounds at the same hydrody-

namic conditions as those used with ozone-enrichedoxygen gas [12]. As such, the losses of phenolic andindolic compounds from the synthetic manure appear to

be due entirely to ozonation reactions. The oxidationmodel was used to predict and compare the decreases inconcentrations of the target compounds. Simulationsusing Model I resulted in a much more rapid reduction

in concentrations than those observed experimentally.As such, it is necessary to incorporate the reaction rateconstant ðkf Þ into the model to account for competing

reactions. The reaction rate constants are summarized inTable 5. The fitted values of the reaction rate constantsare in the same range as those obtained when the

individual phenolic or indolic compounds were ozo-nated in the deionized water.

Swine manure slurry was ozonated using the lab-scalereactor at a flowrate of 300mL/min and temperature of

221C. During ozonation, all of the target malodorouscompounds were oxidized within 2 h, i.e., at an ozonedosage of 1.5 g/L. The pH of the manure slurryincreased from 7.5 to 8.2 after 2.5 h of ozonation.

As the initial pH of the slurry was 7.5, and Hoign!eand Bader [3,4] showed that the ozonation rateconstants of phenolic compounds increase by a

factor of 10 for each pH unit increase, we can assumethat the reaction rate constants were 10 times greaterthan those obtained for a pH of 6.7. Thus, the rate

constants used in the oxidation model were 107M�1 s�1

for the phenolic compounds and 5� 107M�1 s�1 forskatole. A reaction rate constant ðkf Þ of 40,000 s

�1 wasfound to yield the minimum residual in fitting the

experimental data.

4. Discussion

Hoign!e and Bader [3,4] indicated that for theozonation of phenol-like compounds, 2.5mol of ozoneis needed to consume 1mol of phenol; Li et al. [19] also

reported a stoichiometric factor of 2. However, Eisen-hauer [20], Gould and Weber [1], and Roth et al. [7]found that for the total destruction of phenol, between 4

and 6mol of ozone is required per mole of phenol. Thesevalues are higher than that found in this work becausethey did not distinguish between the amount of ozone

that reacts with phenol and that which reacts with theozonation byproducts. Since the reaction conditions

Table 2

Fitting parameters obtained from the mass transfer equations

Inlet flowrate

(mL/min)

pH=2.1 pH=6.7

a KLa

(min�1)

a KLa

(min�1)

100 0.219 0.1574 0.232 0.1479

300 0.222 0.2357 0.222 0.2719

500 0.238 0.3188 0.240 0.3866

Table 3

Enhancement factors calculated under different flowrates and

pH values

Inlet flowrate,

Q (mL/min)

pH=2.1 pH=6.7

100 1.9 2.0

300 3.8 3.2

500 4.4 3.8

Table 5

Reaction rate constant kf for synthetic manure determined at

different flowrates

Flowrate,

Q (mL/min)

Reaction rate constant kf

(s�1)

100 300

300 400

500 600

Table 4

Reaction rate constant kf for phenols and indoles under

different flowrates at pH 6.7

Flowrate, Q (mL/min) Phenols (s�1) Indoles (s�1)

100 350 450

300 400 400

500 350 400

J.J. Wu, S.J. Masten / Water Research 36 (2002) 1513–15261518

used in this study favored the reaction of ozone with the

target chemical over side reactions, the stoichiometricfactors found in this work are lower than that found byothers and appear to be reasonable.

The reaction rate constants for the phenolic andindolic compounds were found to increase signifi-cantly with increasing pH. This increase for phenolic

compounds is thought to be due to the greater reacti-vity of the phenolate ion as compared to that ofphenol. Augugliaro and Rizzuti [2] and Hoign!e andBader [3,4] both showed that the phenolate ion has

a much larger rate constant towards ozone than that

of phenol (109M�1 s�1 for phenolate ion and

500M�1 s�1 for phenol). If using an overall reactionrate constant to represent the reaction of phenoland ozone, the following mathematical expressions can

be introduced:

½C6H5OH�tot ¼ ½C6H5OH� þ ½C6H5O��; ð12Þ

Kphenol ¼½Hþ�½C6H5O

��½C6H5OH�

; ð13Þ

Fig. 2. Comparison of phenol, p-cresol, and p-ethylphenol concentrations in the mixture of phenolic compounds (phenol+p-

cresol+p-ethylphenol) between experimental data and model prediction at pH 2.1, flowrate 300mL/min, and temperature 221C.

J.J. Wu, S.J. Masten / Water Research 36 (2002) 1513–1526 1519

�d½C6H5OH�tot

dt¼ kA½O3�½C6H5OH�

þ kB½O3�½C6H5O�� ð14Þ

¼kA½Hþ� þ kBKphenol

Kphenol þ ½Hþ�

� �½O3�½C6H5OH�tot; ð15Þ

where [C6H5OH]tot is the total molar concentration ofphenol and phenolate ion, Kphenol is the equilibriumconstant (10�9.9M), kA is the reaction rate constant

between phenol and ozone (500M�1 s�1), and kB is thereaction rate constant between phenolate ion and ozone(109M�1 s�1). Therefore, the overall reaction rateconstant at pH=6.7 for phenol is determined to be

6.3� 105M�1 s�1, which is close to the one we found inour experiment (7.5� 105M�1 s�1). Hoign!e and Bader[4] also demonstrated that the reaction rate constants for

phenolic compounds increase over a wide range of pHvalues by a factor of 10 per pH unit, corresponding tothe incremental increase in the degree of dissociation.

Although another explanation is that hydroxyl radicalscause the increase of the reaction rate with pH, Beltran

et al. [21] negated this hypothesis when they found thatthe rate constants for the ozonation of o-cresol inaqueous solution with and without tert-butanol, a

radical scavenger, were nearly identical. In our experi-ment, we used the phosphate buffer system to controlthe pH at 6.7. Masten et al. [22] demonstrated that, at

this pH, hydrogen phosphate (HPO42�) and dihydrogen

phosphate (H2PO4�) are the predominant OH radical

scavengers in oxidizing trichlorobenzene by advanced

oxidation process. Thus, the direct reactions betweenozone and the target compounds predominate becauseany free radicals generated from the decomposition of

ozone are likely to be scavenged by the phosphatespecies present in the buffer system we used.The reaction rate constants determined at pH 2.1 were

ranked, indole>skatole>p-cresol>p-ethylphenol>

phenol. The electrophilic reaction is thought to be themajor mechanism by which aromatic compounds reactwith ozone. This is especially true for aromatic

compounds that have electron donating groups sub-stituted at the ortho and para positions because of thehigh electron densities that result from this substitution.

Fig. 3. Comparison of indole and skatole concentrations in the mixture of indolic compounds (indole+skatole) between experimental

data and model prediction at pH 2.1, flowrate 300mL/min, and temperature 221C.

J.J. Wu, S.J. Masten / Water Research 36 (2002) 1513–15261520

As a result of the high electron density, thesecompounds are very reactive with ozone [23]. As p-

cresol contains two electron donating groups, –OH and–CH3 [24], the reactivity of p-cresol by the initial attackof the ozone molecule is expected to be greater than thatof phenol (with only a –OH group). The ethyl group

(–CH2CH3) on p-ethylphenol is also an electron donat-

ing group [25]. The reactivity of p-ethylphenol is,therefore, expected to be greater than that of phenol.

p-Cresol has a greater reaction rate constant thanp-ethylphenol because the methyl group is a betterelectron donating group than is the ethyl group. Inaddition, it was reported by Geissman [24] that indole

and skatole are very reactive nucleophiles, and undergo

Fig. 4. Comparison of phenol, p-cresol, and p-ethylphenol concentrations in the mixture of phenolic compounds (phenol+

p-cresol+p-ethylphenol) between experimental data and model prediction at pH 6.7, flowrate 300mL/min, and temperature

221C.

J.J. Wu, S.J. Masten / Water Research 36 (2002) 1513–1526 1521

electrophilic substitution reactions with ease. Never-theless, no explanation can be offered at this point for

why the reaction rate constants of indolic compoundsare much faster than those of phenolic compoundsbased upon structural differences in the two groups of

compounds. At pH 6.7, the reaction rate constants forthe three phenolic compounds are almost of the samemagnitude (close to 1� 106M�1 s�1) and the rateconstants of indole and skatole are 5� 106M�1 s�1.

When the reactions between intermediates and ozonewere ignored, the predicted concentration profiles of thetarget compounds decreased much faster than those

determined experimentally (Figs. 2 and 5). In Figs. 2–5,the experimental data and the predicted values deter-mined using Models I and II are shown. It should be

noted that competing reactions between the severaltarget parent compounds and side reactions between thereactive intermediates (generated) have the most sig-

nificant effect on the compound having the lowestreaction rate constant in our system, i.e., phenol. On thecontrary, the reaction of the compound with the greaterreaction rate constant, p-cresol in our system, can be

predicted using Model I, without considering sidereactions. The importance of these side reactions is also

substantiated when considering the results obtainedupon the ozonation of the real and synthetic manure.When ozonating the synthetic manure (Fig. 6.), an

ozone dosage of 0.5 g/L resulted in complete removalof all of the malodorous parent components. One wouldexpect that in real manure (COD=30,000mg/L) theextent of ozone depletion by side reactions would be

significantly greater than that required in the syntheticmanure (COD=3200mg/L), and this is exactly what isobserved.

If Model II is used to fit our experimental data for theozonation of the real swine manure, a reaction rateconstant of 40,000 s�1 is obtained (Fig. 7). This value,

which is much greater than that obtained for syntheticmanure, is likely to be larger because of greater numbersand concentrations of impurities in real swine manure

that can react with ozone. Any model used to predict theremoval of odorous compounds in swine manure wouldbe more effective if the reaction coefficient could bedetermined from a ‘‘gross parameter’’ such as COD. As

Fig. 5. Comparison of indole and skatole concentrations in the mixture of indolic compounds (indole+skatole) between experimental

data and model prediction at pH 6.7, flowrate 300mL/min, and temperature 221C.

J.J. Wu, S.J. Masten / Water Research 36 (2002) 1513–15261522

such, the relationship between kf and COD wasconsidered. Assuming that a mathematical relationshipbetween COD and kf exists, kf can be expressed as

kf ¼ k0f (COD). The intrinsic reaction rate constant, k0

f ;is calculated to be 1.33 for the real manure(COD=30,000mg/L) we tested. Similarly, k0

f is calcu-lated as 0.125 for the synthetic manure, in which the

COD is 3200mg/L. It should be noted that the pH ofreal manure is about a unit larger than that of syntheticmanure, possibly explaining why the intrinsic reaction

rate constant ðk0f Þ for the real manure is approximately

10 times greater than that obtained for the syntheticmanure. At the higher pH value the hydroxide ionconcentration would be approximately 10 times that in

Fig. 6. Comparison of phenolic and indolic concentrations in the synthetic manure between experimental data and model prediction at

pH 6.7, flowrate 300mL/min, and temperature 221C.

J.J. Wu, S.J. Masten / Water Research 36 (2002) 1513–1526 1523

the synthetic manure and the reaction rates of ozone

with byproducts or impurities would also increase by afactor of approximately 10 [4]. Thus, the future work

should be carried out to establish such a mathematical

expression between COD and reaction rate constantðkf Þ:

Fig. 7. Comparison of phenolic and indolic concentrations in the real manure between experimental data and model prediction at pH

7.5, flowrate 300mL/min, and temperature 221C.

J.J. Wu, S.J. Masten / Water Research 36 (2002) 1513–15261524

5. Conclusions

The combination of a mass transfer and kineticstudy has led to the development of a model tosuccessfully predict the degradation of malodorous,

phenolic and indolic compounds in a semi-batchozonation reactor. In this study, several importantconclusions can be drawn:

1. The reaction rate constants of the phenolicand indolic compounds increase with increasingpH.

2. No dissolved ozone and outlet gaseous ozone can

be observed prior to the complete removal ofthe phenolic and indolic compounds. Almost100% utilization of ozone was obtained under

our experimental conditions. Therefore, ozonedosage becomes a very important and reliableparameter for the design of an ozone reactor in

removing malodorous substances from the liquidmanure slurry.

3. The effect of pH on the oxidation of phenolic andindolic compounds is determined predominately by

their reaction rate constants and competition kineticswith ozone. For the solution containing phenoliccompounds at lower pH values, the degradation of

phenol is the slowest of all target compounds as itsreaction rate is the smallest of the target compounds;the degradation of p-cresol is the fastest as it has the

greatest reaction rate. As these reaction rates ofphenolic compounds are identical at a pH of 6.7, thedegradation of these compounds follows similar

patterns.4. A model combining mass transfer and oxidation

kinetics of ozone should include the effect of sidereactions involving ozone and byproducts. Using

Model I, which ignores the effect of side reactions,the predicted rates of the oxidation of thesemalodorous compounds are much greater than those

observed experimentally. The reaction coefficients forartificial and real swine manure were estimated as 400(s�1) and 40,000 (s�1), respectively. It is apparent that

the presence of ozonated reactive impurities results ina significant increase in the ozone dosage required tooxidize the target compounds.

References

[1] Gould JP, Weber WJ. Oxidation of phenols by ozone.

J WPCF 1976;48:47–60.

[2] Augugliaro V, Rizzuti L. The pH dependence of the ozone

absorption kinetics in aqueous phenol solutions. Chem

Eng Sci 1978;33:1441–7.

[3] Hoign!e J, Bader H. Rate constants of reactions of ozone

with organic and inorganic compounds in waterFI (non-

dissociating organic compounds). Water Res 1983;17:

173–83.

[4] Hoign!e J, Bader H. Rate constants of reactions of ozone

with organic and inorganic compounds in waterFII

(dissociating organic compounds). Water Res

1983;17:185–94.

[5] Singer PC, Gurol MD. Dynamics of the ozonation of

phenolFI: experimental observations. Water Res

1983;17:1163–71.

[6] Zheng Y, Hill DO, Kuo CH. Rates of ozonation of cresol

isomers in aqueous solutions. Ozone Sci Eng 1993;15:

267–78.

[7] Roth JA, Moench WL, Debalak KA. Kinetic modeling

of the ozonation of phenol in water. J WPCF 1982;54:

135–9.

[8] Xiong F, Graham NJ. Rate constants for herbicide

degradation by ozone. Ozone Sci Eng 1992;14:283–301.

[9] Yasuhara A. Relation between odor and odorous

components in solid swine manure. Chemosphere

1980;9:587–92.

[10] Wu JJ, Park S, Hengemuehle SM, Yokoyama MT, Person

HL, Gerrish JB, Masten SJ. The use of ozone to reduce the

concentration of malodorous metabolites in swine manure

slurry. J Agric Eng Res 1999;72:317–27.

[11] Wu JJ, Park S, Hengemuehle SM, Yokoyama MT,

Person HL, Masten SJ. The effect of storage and ozonation

on the physical, chemical, and biological character-

istics of swine manure slurries. Ozone Sci Eng

1998;20:35–50.

[12] Wu JJ. The use of ozone for the removal of odor from the

swine manure. PhD thesis, Department of Environmental

Engineering, Michigan State University, E. Lansing, MI,

USA, 1998.

[13] Nishikawa M, Nakamura M, Yagi H, Hashimoto K. Gas

absorption in aerated mixing vessels. J Chem Eng Jpn

1981;14:219–26.

[14] Zhou H, Smith DW. Process parameter development for

ozonation of kraft pulp mill effluent. Water Sci Technol

1997;2–3:251–9.

[15] Charpentier JC. Mass transfer rates in gas–liquid absor-

bers and reactors. In: Advances in chemical engineering,

vol. 11. New York: Academic Press, 1981. p. 3–133.

[16] Beltran FJ, Garcia-Araya JF, Encinar JM. Henry and

mass transfer coefficients in the ozonation of wastewater.

Ozone Sci Eng 1997;19:281–96.

[17] Danckwerts PV. Gas–liquid reactions. New York:

McGraw-Hill Book Company, 1972. p. 96–151.

[18] Beltran FJ, Encinar JM, Garcia-Araya JF. Ozonation

of o-cresol in aqueous solutions. Water Res 1990;24:

1309–16.

[19] Li KY, Kuo CH, Weeks Jr JL. A kinetic study of ozone–

phenol reaction in aqueous solutions. AIChE

J 1979;25:583–91.

[20] Eisenhauer HR. The ozonation of phenolic wastes.

J WPCF 1968;40:1887–99.

[21] Beltran FJ, Encinar JM, Garcia-Araya JF. Absorption

kinetics of ozone in aqueous o-cresol solutions. Can

J Chem Eng 1992;70:141–7.

J.J. Wu, S.J. Masten / Water Research 36 (2002) 1513–1526 1525

[22] Masten SJ, Galbraith MJ, Davies SH. Oxidation of

trichlorobenzene using advanced oxidation processes.

Ozone Sci Eng 1996;18:535–48.

[23] Langlais B, Reckhow DA, Brink DR, editors. Ozone in

water treatment: application and engineering. Chelsea, MI:

Lewis Publishers, 1991. p. 12–3.

[24] Geissman TA. Principles of organic chemistry, 3rd ed.

San Francisco: W.H. Freeman and Company, 1968.

p. 547–8, 720–1.

[25] Hart H. Organic chemistryFa short course, 8th ed.

Boston: Houghton Mifflin Company, 1991. p. 107–34.

J.J. Wu, S.J. Masten / Water Research 36 (2002) 1513–15261526