Modeling Aqueous Ozone/UV Process Using Oxalic

Acid as Probe Chemical

Temesgen Garoma1* and Mirat D. Gurol2

1 Formerly a doctoral student at UC - San Diego and San Diego State University, and

currently affiliated with Parsons Water and Infrastructure, Pasadena, CA 91124

2 Department of Civil and Environmental Engineering, San Diego State University, San

Diego, CA 92182

* Corresponding Author Current Address: 100 West Walnut Street, Pasadena, CA 92114,

Tel: 626-440-3020, Fax: 626-440-6337, Email: [email protected]

On the following pages the supporting information for the article entitled “Modeling Aqueous

Ozone/UV Process Using Oxalic Acid as Probe Chemical” is presented. First, based on proposed

reaction mechanisms for an Ozone/UV system in natural waters detailed derivation for the kinetic model

that describes the interaction of ozone with UV light is presented. The experimental results used in the

determination of the effective path length of UV light and volumetric mass transfer coefficient for the

experimental reactor is also presented. Finally, the effect varying bicarbonate concentration, pH, and

effective path length on removal efficiency of oxalic acid is presented. The effects of incident UV light

intensity and influent ozone gas concentration on the accumulation of hydrogen peroxide (H2O2) and

hydroxyl radical (OH•) in the Ozone/UV system is also presented.

S1
 APPENDIX

A. Derivation of Kinetic Model for an Ozone/UV Process

In this section derivation for the kinetic model that describes the interaction of ozone with UV

light is presented. Based on the proposed reaction mechanisms in Table 2 (ref. 10), the overall rate of

change in concentrations of aqueous ozone, hydrogen peroxide (H2O2), and hydroxyl radical (OH•) for

completely mixed batch system can be expressed as follows:

d [O3 ]
= nO3 + k 20 [O3−• ][CO3−• ] − k 1[O3 ][OH − ] − Φ P I a − k 2 [O3 ][ H 2 O2 ]
dt (A.1)
− k 3 [O3 ][ HO2− ] − k 4 [O3 ][O2−• ] − k 6 [O3 ][OH •] − ∑ k di [O3 ][M i ]

d [ H 2 O2 ]
= Φ P I a + k13 [OH •]2 + k17 [O2−• ][ HO2• ] + k18 [ HO2• ]2 − Φ 'P I a'
dt (A.2a)
− k 2 [O3 ][H 2 O2 ] − k 7 [ H 2 O2 ][OH •] − k11[ H 2 O2 ][CO3−• ]

d [ HO2− ]
= k1 [O3 ][OH − ] − k 3 [O3 ][ HO2− ] − k 8 [ HO2− ][OH •] − k12 [ HO2− ][CO3−• ] (A.2b)
dt

d [OH •]
= 2Φ 'P I a' + k 2 [O3 ][ H 2 O2 ] + k 3 [O3 ][ HO2− ] + k 5 [O3−• ][ H + ]
dt
− k 6 [O3 ][OH •] − k 7 [ H 2 O2 ][OH •] − k 8 [ HO2− ][OH •] − k 9 [OH •][ HCO3− ] (A.3)
2− −• •
− k10 [OH ][CO ] − k13 [OH •] − k14 [OH •][O ] − k15 [OH •][ HO ]
3
2
2 2

− k16 [OH •][O3−• ] − ∑ k bi [ M i ][OH •] − ∑ k j [ S j ][OH •]

Combining Equations (A.2a) and (A.2b) by using the equilibrium relationship between H 2 O2 and

HO2− , first Equation (A.2c) was obtained and then rearranging Equation (A.2c), Equation (A.2d) was

obtained:

 K a 3  d [ H 2 O2 ]
1 +  = Φ P I a + k1 [O3 ][OH − ] + k13 [OH •]2 + k17 [O2−• ][ HO2• ]
 [ H + ]  dt (A.2c)
+ k18 [ HO2• ] 2 − Φ 'P I a' − k 2* [O3 ][H 2 O2 ] − k 7* [ H 2 O2 ][OH •] − k11* [ H 2 O2 ][CO3−• ]

d [ H 2 O2 ]
= γ 1{Φ P I a + k1 [O3 ][OH − ] + k13 [OH •]2 + k17 [O2−• ][ HO2• ]
dt (A.2d)
+ k18 [ HO2• ] 2 − Φ 'P I a' − k 2* [O3 ][ H 2 O2 ] − k 7* [ H 2 O2 ][OH •] − k11* [ H 2 O2 ][CO3−• ]}

S2
Similarly the rate of change of radicals O2−• , O3−• , and CO3−• can be expressed as follows:

d [CO3−• ]
= k 9 [ HCO3− ][OH •] + k10 [CO32− ][OH •]
dt (A.4a)
− k11[ H 2 O2 ][CO3−• ] − k12 [ HO2− ][CO3−• ]

d [O2−• ]
= k 3 [ HO2− ][O3 ] + k12 [ HO2− ][CO3−• ] − k 4 [O3 ][O2−• ] (A.5a)
dt

d [ HO2−• ]
= k 2 [ H 2 O2 ][O3 ] + k 6 [O3 ][OH •] + k 7 [ H 2 O2 ][OH •]
dt (A.5b)
+ k 8 [ HO2− ][OH •] + k11[ H 2 O2 ][CO3−• ]

d [O3−• ]
= k 4 [O3 ][O2−• ] − k 5 [O3−• ][ H + ] (A.6a)
dt

Combining Equations (A.5a) and (A.5b) by using the equilibrium relationship between O2−• and HO2−• ,

first Equation (A.5c) was obtained and then rearranging Equation (A.5c), Equation (A.5d) was obtained.

−•
 [ H + ]  d [O2 ]
1 +  = k 2* [ H 2 O2 ][O3 ] + k 6 [O3 ][OH •] + k 7* [ H 2 O2 ][OH •]
 K a 4  dt
 (A.5c)
+ k11* [ H 2 O2 ][CO3−• ] − k 4 [O3 ][O2−• ]

d [O2−• ]
= β 1{k 2* [ H 2 O2 ][O3 ] + k 6 [O3 ][OH •] + k 7* [ H 2 O2 ][OH •]
dt (A.5d)
+ k11* [ H 2 O2 ][CO3−• ] − k 4 [O3 ][O2−• ]}

In Equations A.1, A.2d and A.3, the terms involving radical-radical reaction between OH • ,

O2−• , O3−• , and CO3−• , can be neglected since low concentrations of radical specious make the radical-

radical reactions less likely than the radical-solute reactions. Furthermore, the radicals O2−• , O3−• , and

CO3−• are highly reactive species, and therefore don’t accumulate in the system. Thus, applying steady

state approximation on Equations A.4a, A.5d, and A.6a, one may derive the following expressions for

CO3−• , O2−• , and O3−• , respectively.

[CO3−• ] SS = {k 9* [ HCO3− ][OH •]} /{k11* [ H 2 O2 ]} (A.4b)

S3
 [O2−• ] SS = {k 2* [ H 2 O2 ][O3 ] + k 6 [O3 ][OH •] + k 7* [ H 2 O2 ][OH •]
(A.5e)
+ k11* [ H 2 O2 ][CO3−• ] SS } /{k 4 [O3 ]}

[O3−• ] SS = {k 4 [O3 ][O2−• ] SS / k 5 [ H + ]} (A.6b)

By substituting the steady state approximations of the radical intermediates into Equations A.1,

A.2d and A.3, and then simplifying the following expressions were derived for the overall rate of

change in concentrations of aqueous ozone, hydrogen peroxide (H2O2) and hydroxyl radical (OH•),

respectively.

d [O3 ]
= nO3 − k 1[O3 ][OH − ] − Φ P I a − 2k 2* [O3 ][ H 2 O2 ] − 2k 6 [O3 ][OH •]
dt (A.7)
− k 7* [ H 2 O2 ][OH •] − k 9* [OH •][ HCO3− ] − ∑ k di [O3 ][M i ]

d [ H 2 O2 ]
= γ 1{k1 [O3 ][OH − ] + Φ P I a − Φ 'P I a' − k 2* [O3 ][ H 2 O2 ]
dt (A.8)
− k 7* [ H 2 O2 ][OH •] − k 9* [OH •][ HCO3− ]}

d [OH •]
= 2Φ 'P I a' + 2k 2* [O3 ][ H 2 O2 ] − ∑ k bi [ M i ][OH •] − ∑ k j [ S j ][OH •] (A.9)
dt

In these Equations ∑k di [O3 ][M i ] represents the reaction of organic pollutants with aqueous

ozone and is negligible for the target chemical in this research, ∑k bi [ M i ][OH •] represents the reaction

of target organic compounds with OH • and ∑ k [S j j ] represents the scavengers of OH • other than the

target organic compounds and bicarbonates, and yet is negligible for solutions prepared in purified

water. Furthermore, the following rate constants in these equations vary with pH, as shown below:

k 3 K a3 k8 K a3 k10 K a 2
k 2* = k 2 + k 7* = k 7 + k 9* = k 9 +
[H + ] [H + ] [H + ]

k K K a4 [H + ]
k = k11 + 12 +a 3
*
β1 = γ1 =
K a4 + [H + ] [ H + ] + K a3
11
[H ]

In Equations A.1 and A.7, nO3 represents the rate of ozone absorption per unit volume of water

sample and is given by the following expression:

S4
 N O3
nO3 = (A.10)
V

In which N O3 is the rate of ozone absorption and can be described by using a model developed

by Gurol (20) for ozone absorption in bubble reactors as follows.

 (O ) 
N O3 = Q g H [1 − exp(−θ O3 )] 3 inf − [O3 ] (A.11)
 H 

 k L a 
θO =   (A.12)
3

 H (Qg / V ) 


where k L a is the volumetric mass transfer coefficient for ozone, H is the dimensionless

Henry’s constant for ozone, which is 2.86 at T = 20oC, Q g is ozone gas flow rate, (O3 ) inf is influent

ozone gas concentration at steady state and V volume of water sample.

B. Effective Path Length and Volumetric Mass Transfer Coefficient

Effective Path Length ( b ): The effective path length for UV light in the experimental reactor

was determined by measuring the rate of photolysis of hydrogen peroxide. To make sure that Equations

12 and 13 would be applicable, low concentration of hydrogen peroxide (0.2 mM) was used. Allyl

alcohol at 0.1 M was added to the solution as hydroxyl radical (OH•) scavenger in order to quench

secondary reactions (33). In Figure B.1, the experimental results are presented and the results follow

first-order decomposition of H2O2, as expected. The first order reaction rate constant ( k obs ) calculated

for each I o was plotted as a function of corresponding I o in Figure B.2. For Φ 'P = 0.5 and ε ' = 19.6 M-

cm-1, b was estimated from the slope of the straight line in Figure B.2 as 7.5 ± 0.9 cm.
1

S5
 0 10 20 30 40
0

-1

Ln([H2O2]/[H2O2]o )

-2

-3 [H2O2]O = 0.20 mM
2.9x10-6 E/L-s
5.8x10-6 E/L-s Allyl Alcohol = 0.1 M
pH = 7.0
9.0x10-6 E/L-s

-4
Time (minutes)

Figure B.1 Photolysis of Hydrogen Peroxide by UV Light

2.0E-03

Slope = 2.3Φp'a'beff= 169 ± 20 L/Ein

b = 7.5 ± 0.9 cm
1.5E-03 R2 = 0.9889
kobs (s-1)

1.0E-03

5.0E-04

0.0E+00
0.0E+00 2.0E-06 4.0E-06 6.0E-06 8.0E-06 1.0E-05
Io (E/L-s)

Figure B.2 Hydrogen Peroxide Photolysis Rate Constant vs. Incident UV Light Intensity

S6
 Volumetric Mass Transfer Coefficient ( k L a ): The volumetric mass transfer coefficient ( k L a )

was determined by sparging ozone into a pure aqueous solution at pH of 2 since the rate of

decomposition of ozone under such acidic condition is known to be very slow. The solution was mixed

with a magnetic stirrer to provide homogeneity. Figure B.3 shows the accumulation of ozone in solution

for a gas flow rate of 1.5 L/min. The steady-state concentration of 15.4 mg/L was taken as saturation

concentration, [O3 ]* , the equilibrium concentration under the experimental conditions. The

experimental data are then plotted as shown in Figure B.4, and from the slope of the straight line, k L a

was estimated as 8.8x10-3 s-1.

18

15

[O3]* = 15.4 mg/L

12
[O3] (mg/L)

0
0 5 10 15 20 25 30
Time (minutes)

Figure B.3 Accumulation of Ozone in the Solution

S7
 0 2 4 6 8 10
0

kLa = 8.8x10-3 s -1
R2 = 0.9929

Ln{([O3]* -[O3])/[O3]*} -2

-4

-6
Time (minutes)

Figure B.4 Plot of Ln{([O3]*-[O3])/[O3]*} vs. Time

C. Sensitivity of the Kinetic Model

In the main manuscript of this article, the sensitivity of the kinetic model by varying the intensity

of UV light and influent ozone gas concentration was investigated. In this section, the effect varying

bicarbonate concentration, pH, and effective path length on removal efficiency of oxalic acid is

presented. The effects of incident UV light intensity and influent ozone gas concentration on the

accumulation of hydrogen peroxide (H2O2) and hydroxyl radical (OH•) in the Ozone/UV system was

also presented.

In Figure C.1, the effect of bicarbonate on the removal efficiency of the model compound is

presented. The result indicates that increasing bicarbonate concentration has significantly reduced the

removal of the model compound. This is expected because the reactivity of bicarbonate and carbonate

ions towards OH•, 8.5x106 and 3.9x108 M-1s-1 (34), respectively, is higher than the reactivity of oxalic

acid towards OH•, 5.3x106 M-1s-1 (14). The effect of pH on the removal of the model compound is

shown in Figure C.2. According to the result, at lower pH values the removal of the model compound is

S8
faster. This is because at lower pH values H 2 CO3* / HCO3− / CO32− equilibrium shifts towards H 2 CO3* ,

which does not scavenge OH•. On the other hand, as the pH increases the equilibrium shifts towards

CO32 − , which is an effective scavenger of OH• compared to HCO3− (34). From a practical point of

view, however, lower pH may not be applicable in the actual design of an Ozone/UV process because at

lower pH values corrosion can be a problem. In Figures C.3 the effective path of UV light ( b ) on the

removal of model compound is presented. According to the result, the removal of the oxalic acid

increased significantly as b increased up to 7.5 cm, however, remained insensitive to further increases

1.0

0.8
[OXA]/[OXA]o

0.6 10 mM
5 mM
2 mM
1 mM
0.4

[OXA]o=1.0 mM pH=7.0
0.2 Io=6x10-6 E/L-s kLa=0.0088 s -1
(O3)inf=48 mg/L b=7.5 cm
Qg= 1.5 L/min kOXA =5.3x106 M-1s -1

0.0
0 10 20 30 40 50 60
Time (minutes)

Figure C.1 Predicted Effect of Bicarbonate Concentration on Oxalic Acid Removal

S9
 1.0

0.8

9
7

[OXA]/[OXA]o
0.6 5
3

0.4

0.2 [OXA]o=1.0 mM Qg=1.5Ll/min

Io=6x10-6 E/L-s kLa=0.0088 s -1
(O3) inf=48 mg/L b=7.5 cm
[HCO3-]= 2 mM kOXA =5.3x106 M-1s -1
0.0
0 10 20 30 40 50 60
Time (minutes)

Figure C.2 Predicted Effect of pH on Oxalic Acid Removal

1.0

0.8

0.75 cm
7.50 cm
[OXA]/[OXA]o

0.6 10.0 cm
15.0 cm

0.4

[OXA]o=1.0 mM pH=7.0
Io=6x10-6 E/L-s kLa=0.0088 s -1
0.2 (O3) inf=48 mg/L Qg=1.5 L/min
[HCO3-]= 2 mM kOXA =5.3x106 M-1s -1

0.0
0 10 20 30 40 50 60
Time (minutes)

Figure C.3 Predicted Effect of Path Length on Oxalic Acid Removal

S10
 The effect of incident UV light intensity on the accumulation of hydrogen peroxide (H2O2) in an

Ozone/UV system is presented in Figure C.4. The result indicates that increasing incident UV light

intensity from 1x10-6 to 6x10-6 E/L-s increases the concentration of H2O2 in the system. However,

increasing the incident UV light intensity further resulted in a decrease in the concentration of H2O2 for

an influent ozone gas concentration of 48 mg/L.

The effect of influent ozone gas concentration on the accumulation of H2O2 is presented in

Figure C.5. According to the result, early in the oxidation process increasing influent ozone gas

concentration up to 48 mg/L resulted in an increase in H2O2 concentration. Further increase in ozone gas

concentration beyond this level resulted in a peak and then decrease in the concentration of H2O2. This

is because the extra dissolved ozone, beyond what is consumed by photolysis, is expected to react with

H2O2/HO2− as described in Table 2 ref. 10, causing reduction in the concentration of H2O2 in the system.

However, later during the oxidation process, the highest concentration of H2O2 corresponds to the

smallest influent ozone gas concentration.

1.0E-04
[OXA]o=1.0 mM pH=7.0
Qg=1.5 L/min kLa=0.0088 s -1
8.0E-05 (O3)inf=48 mg/L b=7.5 cm
[HCO3-]= 2 mM kOXA =5.3x106 M-1s -1

9x10-6 E/L-s
6x10-6 E/L-s
6.0E-05
[H2O2] (M)

4.0E-05 3x10-6 E/L-s

2.0E-05 1x10 -6 E/L-s

0.0E+00
0 10 20 30 40 50 60
Time (minutes)

Figure C.4 Predicted Effect of Incident UV Light Intensity on H2O2 Accumulation

S11
 1.0E-04

8.0E-05
48 mg/L

6.0E-05
[H2O2] (M)
24 mg/L
4.0E-05 96 mg/L

[OXA]o=1.0 mM pH=7.0
Io=6x10-6 E/L-s kLa=0.0088 s -1
2.0E-05
Qg=1.5 L/min b=7.5 cm
[HCO3-]= 2 mM kOXA =5.3x106 M-1s -1 72 mg/L

0.0E+00
0 10 20 30 40 50 60
Time (minutes)

Figure C.5 Predicted Effect of Influent Ozone Gas Concentration on H2O2 Accumulation
In Figures C.6, the effect of incident UV light intensity on the accumulation of hydroxyl radical

(OH•) in an Ozone/UV system is presented. The result indicates that increasing incident UV light

intensity increases the concentration of OH• in the system. The increase in OH• concentration was more

significant as the incident UV light intensity increased from 1x10-6 to 3x10-6 E/L-s compared to 3x10-6

to 6x10-6 E/L-s or 6x10-6 to 9x10-6 E/L-s. On the other hand, increasing the influent ozone gas

concentration for given incident UV light intensity resulted in continuous increase in OH•

concentration, Figure C.7. This is because once the decomposition of ozone is initiated with UV light,

increasing the influent ozone gas concentration accelerates the decomposition of ozone by the reaction

of dissolved ozone with OH• and other radicals as indicated in Table 2 of ref. 10. This result also

supports the conclusion made previously regarding the effect of influent ozone gas on the rate of

removal of oxalic acid in an Ozone/UV system.

S12
 9.0E-11

6.0E-11
9x10-6 E/L-s

[HO.] (M)
6x10-6 E/L-s
3x10-6 E/L-s
1x10-6 E/L-s

3.0E-11

[OXA]o=1.0 mM pH=7.0
Qg=1.5/L kLa=0.0088 s -1
(O3) inf=48 mg/L b=7.5 cm
[HCO3-]= 2 mM kOXA =5.3x106 M-1s -1
0.0E+00
0 10 20 30 40 50 60
Time (minutes)

Figure C.6 Predicted Effect of Incident UV Light Intensity on OH• Accumulation

2.0E-10
[OXA]o=1.0 mM pH=7.0
Io=6x10-6 E/L-s kLa=0.0088 s -1
Qg=1.5 L/min b=7.5 cm
1.5E-10 [HCO3-]= 2 mM kOXA =5.3x106 M-1s -1
96 mg/L

72 mg/L
[HO.] (M)

1.0E-10
48 mg/L

5.0E-11 24 mg/L

0.0E+00
0 10 20 30 40 50 60
Time (minutes)

Figure C.7 Predicted Effect of Influent Ozone Gas Concentration on OH• Accumulation

S13