Understanding and Improving Process Performance ...

Understanding and Improving Process Performance of Advanced Oxidation Processes (AOPs) The Croucher Foundation Advanced Study Institute (ASI) , June 23-27, 2008 The Hong Kong University of Science and Technology, HK, China.

John C. Crittenden, Ke Li, Daisuke Minakata, Paul Westerhoff, Hyunju Jeong Department of Civil and Environmental Engineering, Arizona State University

David Hokanson, and Rhodes Trussell Trussell Technologies, Inc.

1

Outline 1. RO retentate DOC removal –comparison of energy consumption among various Advanced Oxidation Processes (AOPs) 2. AdOx –advanced oxidation simulation software 3. Case study 1. UV/H2O2 process for methyl tert-butyl ether (MtBE) and tertiary butyl alcohol (tBA) removal from drinking water source: effect of pretreatment options and light source 4. Case study 2. Bromate mitigation during ozonation and MtBE and tBA removal during O3/H2O2 process 5. UV/H2O2 mechanism study 2

What are AOPs? 

AOPs are emerging and promising water treatment technology and attractive alternatives to traditional non-destructive water treatment technologies.



AOPs produce highly reactive oxidants, hydroxyl radicals (HO•) under conditions of ambient temperature and atmospheric pressure.

3

What are AOPs? 

According to Bolton and Carter (Bolton and Cater, 1994), the following general pattern of oxidation is observed for AOPs.

Organic  Aldehydes  Carboxylic  Carbon dioxide pollutant acids and mineral acids 

The most significant observed by-products are the carboxylic acids, due to the fact that the second order rate constants for these compounds are much lower than those for most organics. However, if adequate reaction time is provided, all by-products (>99% as measured by a TOC mass balance) are destroyed.

Ultimate goal of our AOP project 

Establish a computer-based kinetic model of reactions that are initiated by HO• in aqueous AOPs. 

Reaction Pathway Generator (Graph theory)



Reaction Rate Constant Calculator (QSARs, Quantum mechanical calculations)



Ordinary Differential Equations (ODEs) Solver Computer Reaction Generator Toxicity Screening

Reaction Rate Constants Calculator ODE Generator and Solver

Products

5

Outline 1. RO retentate DOC removal –comparison of energy consumption among various Advanced Oxidation Processes (AOPs) 2. Adox –advanced oxidation simulation software 3. Case study 1. UV/H2O2 process for methyl tert-butyl ether (MtBE) and tertiary butyl alcohol (tBA) removal from drinking water source: effect of pretreatment options and light source 4. Case study 2. Bromate mitigation during ozonation and MtBE and tBA removal during O3/H2O2 process 5. UV/H2O2 mechanism study 6

Reclamation Plant - System Diagram Screen

Primary Aeration Secondary Filtration Clarification Clarification

Microfiltration Reverse Stabilization Osmosis

Disinfection

Disinfection Aquifer Storage and Recovery

RO retentate DOC: 40 ~ 50 mg/L BDOC (biologically degradable DOC5): 4.5 mg/L pH = 7.5-7.8

Use

7

Parameters for evaluations EE/O (electrical efficiency per order of contaminant destruction)

, kWh-kgal/order P = lamp power output, kW Q = water flow rate, gal/h Ci = influent conc. of MtBE or tBA, µg/L Cf = effuluent conc. of MtBE or tBA, µg/L

8

Energy consumption v.s. DOC % reduction among AOPs • 10 kWh/lb O3, 4.9 kWh/lb H2O2, 13.6 kWh/lb TiO2 • Fenton’s was excluded due to the limited DOC removal (49%)

100

O3 /H2O2 (HiPOx) 730/380 mg/L 1100/550 mg/L

80 DOC reduction (%)

UV/TiO2/H2O2 (Bench) 960 J/cm2 2 g/L 10 mM pH = 5.0

UV/TiO2/H2O2 (Bench) 600 J/cm2 2 g/L 10 mM pH = 5.0

60 40

0 0

10

20

30

UV/TiO2 (Purifics) 15 W/m3 2 g/L pH = 5.0

UV/H2O2 960 J/cm2 10 mM pH = 4.0

UV/TiO2/H2O2 (Bench) 210 J/cm2 2 g/L 10 mM pH = 5.0

20

UV/TiO2 (Bench) 960 J/cm2 2 g/L pH = 4.0

40

50

60

Energy consumption (kWh/kgal) Source: Westerhoff. 2008

10

Removal of RO retentate DOC DOC removal 75% at about 70 kWh/m3 Lower power can be used when combined with downstream biodegradation (i.e., sand filter).

40.0

DOC (mg/L)

35.0 30.0 25.0

UV/TiO2 UV/TiO2 + BDOC

20.0 15.0 10.0 5.0 0.0 0

2

4

6

illumination time (hrs)

8

10

Essentially incorporating bio can reduce power requirements by 20% to 50% depending upon what final DOC concentration you are targeting

Source: Westerhoff. 2008. based on the bench scale experiments.

11

Summary  UV/TiO2 was most energy efficient RO retentate DOC removal as compared to O3/H2O2, UV/H2O2 and UV/TiO2/H2O2.  UV/TiO2 required no addition of chemicals and hence is easy to operate.  UV/TiO2 treatment followed by biological treatment achieved more than 90% of RO retentate DOC removal. 12

Outline 1. RO retentate DOC removal –comparison of energy consumption among various Advanced Oxidation Processes (AOPs) 2. AdOx –advanced oxidation simulation software 3. Case study 1. UV/H2O2 process for methyl tert-butyl ether (MtBE) and tertiary butyl alcohol (tBA) removal from drinking water source: effect of pretreatment options and light source 4. Case study 2. Bromate mitigation during ozonation and MtBE and tBA removal during O3/H2O2 process 5. UV/H2O2 mechanism study 13

“All models are wrong but some are useful.” -- George E.P. Box

“Let’s develop some useful ones and make good use of them.” -- an optimistic modeler

AdOxTM, AOPs Simulation Software • AdOxTM is a part of the Environmental Technologies Design Option Tools (ETDOTTM). It contains mechanism-based models that can be used to evaluate and design advanced oxidation processes. • The final version of AdOx will include the following models: – H2O2/UV model (available in version 1.0) – H2O2/O3 model – UV/O3 model • A comprehensive database is also included in AdOxTM to provide powerful support to the above models as well as ease-of-use 15 features for the user.

Objectives of AdOxTM  Understand the chemistry of AOPs  Assess the preliminary design and feasibility of using AOPs  Plan pilot plant studies and interpret the results  Predict the effect of operational parameters and provide key parameters for process design  Trace the destruction of contaminants and the formation of byproducts, and provide valuable information for mechanic studies of AOPs 16

Features of Current Version AdOxTM 1.0 - overall • A kinetic model (AdOx) based on: – Chemistry -- reaction rate constants – Photochemistry -- photochemical parameters • Comprehensive Mechanism Based Model – Important elementary photochemical reactions – Reactor Options Includes, Plug Flow, Batch and Tanks in Series – Dye study evaluation for nonideal mixing – Process Operational Variables and Water Quality Parameters – Rate Constants Database 17

Features of Current Version AdOxTM 1.0 - Elementary Reactions H2O2 + hν → 2HO•

H2O2 = quantum yield of H2O2 (=0.5)

I0 = incident light intensity, einstein cm-2 sec-1 A=2.303b(εH2O2CH2O2+εRCR + εSCS + εHO2-CHO2-) b=pathlength, cm fH2O2= 2.303 b (εH2O2CH2O2 + εHO2-CHO2-)/A H2O2/HO2- + HO  H2O/OH- + HO2 H2O2 + HO2/O2-  HO + H2O/OH- + O2

2.7×107, 7.5×109 3.0, 0.13

HO + HO  H2O2 HO + HO2/O2-  H2O/OH- + O2 HO2 + HO2/O2-  H2O2/HO2- + O2

5.5×109 6.6×109, 7.0×109 8.3×105, 9.7×107

R + hv  Products R + HO  Products

kMtBE=1.6×109, ktBA=6.0×108

HO  + CO32-/HCO3-  CO3-  + OH-/H2O HO + NOM  Products NOM + hv  Products

3.9×108, 8.5×106 2.4×104 (mgC/L)-1s-1 18

Features of Current Version AdOxTM 1.0 • Reactor Options – Completely mixed reactor (CMBR) – Completely flow type reactor (CMFR) w/o TIS – Plug flow reactor – Non-ideal Mixing • Main Process Operational Variables : – Initial hydrogen peroxide concentration – Incident UV-light intensity • Water quality parameters: – TOC concentration – pH – Concentrations of target compounds – Byproducts can be included if they are known 19

Features of Current Version AdOxTM 1.0  Rate Constant Database

A database was integrated into Software to Estimate Physical Properties (StEPP) to facilitate the use of AdOxTM. StEPP provides physical and chemical properties for over 600 compounds, many of which are covered on U.S. EPA’s list of priority pollutants. The database include all of necessary physico-chemical properties for AdOxTM. 20

References that were used for model validation 





Glaze, W. H.; Lay, Y.; Kang J. W. Advanced oxidation processes. A kinetic model for the oxidation of 1,2dibromo-3 chloropropane in water by the combination of hydrogen peroxide and UV radiation. Ind. Eng. Chem. Res. 1995, 34, 2314-2323. Li, Ke.; Stefan, M.I.; Crittenden, J.C. UV photolysis of trichloroethylene: Product study and kinetic modeling. Environ. Sci. & Technol. 2004, 38, 6685-6693. Li, Ke.; Stefan, M.I.; Crittenden, J.C. Trichloroethylene degradation by UV/H2O2 advanced oxidation process: Product study and kinetic modeling. Environ. Sci. & Technol. 2007, 41(5), 1696-1703.

21

Outline 1. RO retentate DOC removal –comparison of energy consumption among various Advanced Oxidation Processes (AOPs) 2. AdOx –advanced oxidation simulation software 3. Case study 1. UV/H2O2 process for methyl tert-butyl ether (MtBE) and tertiary butyl alcohol (tBA) removal from drinking water source: effect of pretreatment options and light source 4. Case study 2. Bromate mitigation during ozonation and MtBE and tBA removal during O3/H2O2 process 5. UV/H2O2 mechanism study 22

Background • MtBE was used as gasoline additive to enhance octane number. • Despite the ban in 1992, MtBE is still found nationwide as ground water contaminants. • Difficult to remove using adsorption or air stripping • Exposure to large dose causes significant non-cancerrelated-health risk (WHO). • Ruin taste of water at 5-15 µg/L • Established treatment target at MtBE ≤ 2.5 µg/L and tBA ≤ 6 µg/L, respectively, by California Department of Public Health (CDPH) 23

Initial Conditions + Major Raw Water Quality • 7000 gpm (=10.1MGD) • MtBE: 300 µg/L • tBA: 30 ug/L • TDS: 940 mg/L • Alkalinity: 318 mg/L as CaCO3 • Chloride: 138 mg/L • Nitrate: 0.9 mg/L Site of water source: • Iron: 0.44 mg/L Charnock Wellfield, • pH: 7.6 City of Santa Monica, CA, • TOC: 1.4 mg/L 1995

Objective MtBE ≤ 2.5 µg/L and tBA ≤ 6 µg/L

24

Pretreatment options -4 alternatives + dealkalization• NAIX: ion exchange softening with seawater • Pellet Softening: softening with Pellet reactor • WAIX: weak acid Ion exchange • RO: lime softening + reverse osmosis

Dealkalization (acid addition to lower pH and remove alkalinity)

Estimated major water quality after pretreatment Alternative Raw water NalX NalX + Dealkalization Pellet Softening Pellet Softening + Dealkalization WAIX WAIX + Dealkalization RO RO + Dealkalization

TOC (mg/L) 1.4 1.4 1.4 1.4

Alkalinity (mg/L as CaCO3) 318 318 0 203

7.6 7.6 4.65 9.1

Ferrous Iron (mg/L) 0.44 0 0 0

1.4

0

4.75

0

1.4 1.4 0.07 0.07

118 0 54.1 0

6 4.4 7 4.5

0 0 0 0

pH

25

Parameters for evaluations Impact to UV/H2O2 process by compound i

k HO•/Ri = second-order reaction rate constants of HO• with compound i Ci = concentration of compound i

Raw water H2O2 HO26.6% 0.2%Fe2+ 15.1% TOC 27.9% tBA 0.2% MtBE CO3-4.5% 2.9%

HCO342.6%

*H2O2 conc. is assumed 10 mg/L. ** Fe2+ is included only for raw water. 26

Pretreatment 1. -IX Softening with seawater (NAIX) w/o dealkalizationdownhole Cl2 Na IX

H2SO4 H2O2 Cl2 Cl2 Replacement GAC

mix

Seawater Brine

AOP Decarbonator

NaOH

contactor O2 Strip GAC* GAC Waste GAC

*Treatment scheme courtesy of R.Trussell

27

Pretreatment 2. – Softening with Pellet Reactor w/o dealkalization downhole Cl2 Waste Pellets

Pellet softener

wash water Cl2 Cl2 Replacement GAC

Filtration

mix AOP Decarbonator

NaOH HCl HCl H2O2

NaOH

contactor

HO2H2O2 4.5% 5.1% HCO3TOC 20.7% 21.5% CO3-44.6%

tBA MtBE 0.2% 3.5%

Pellet+Dealk H2O2 16.8%

O2 Strip GAC*

Pellet Soft

MtBE tBA 11.5% 0.5%

Waste GAC

GAC *Treatment scheme courtesy of R.Trussell

TOC 71.1% 28

Pretreatment 3. - Weak Acid IX w/o dealkalization downhole Cl2 Weak Acid IX H2O2

mix

Cl2

AOP Decarbonator

H2SO4

Brine

NaOH

WAIX HCO3- CO3-12.4% 0.0%

H2O2 14.7%

MtBE 10.1% tBA 0.5%

TOC 62.3%

contactor

O2 Strip GAC*

WAIX + Dealk

Replacement GAC

H2O2 16.8%

tBA MtBE 0.5% 11.5%

GAC Waste GAC *Treatment scheme courtesy of R.Trussell

TOC 71.1%

29

Pretreatment 4. - Lime softening + RO w/o dealkalization downhole Cl2

softening H2SO4 H2O2 Cl2

Brine

mix

GAC*

AOP

RO

Brine

LP RO

Decarbonator

Cl2

Seawater

GAC NaOH

H2O2 34.3%

TOC 7.3%tBA 1.0%

MtBE 23.5%

CO3-0.6%

RO + Dealk

contactor O2 Strip

Replaceme nt GAC

HCO333.1%

H2O2 51.8%

GAC* GAC

Waste GAC

*Treatment scheme courtesy of R.Trussell

MtBE 35.6%

TOC 11.0% tBA 1.6%

30

Elementary Reactions of UV/H2O2 H2O2 + hν → 2HO•

H2O2 = quantum yield of H2O2 (=0.5)

I0 = incident light intensity, einstein cm-2 sec-1 A=2.303b(εH2O2CH2O2+εRCR + εSCS + εHO2-CHO2-) b=pathlength, cm fH2O2= 2.303 b (εH2O2CH2O2 + εHO2-CHO2-)/A H2O2/HO2- + HO  H2O/OH- + HO2 H2O2 + HO2/O2-  HO + H2O/OH- + O2

2.7×107, 7.5×109 3.0, 0.13

HO + HO  H2O2 HO + HO2/O2-  H2O/OH- + O2 HO2 + HO2/O2-  H2O2/HO2- + O2

5.5×109 6.6×109, 7.0×109 8.3×105, 9.7×107

R + hv  Products R + HO  Products

kMtBE=1.6×109, ktBA=6.0×108

HO  + CO32-/HCO3-  CO3-  + OH-/H2O HO + NOM  Products NOM + hv  Products

3.9×108, 8.5×106 2.4×104 (mgC/L)-1s-1 31 31

Molar absorption coefficient & Absorption spectra

Absorbance (cm-1)

Calculated f (H2O2) = 0.0044 * 200-300 nm, 10 mg/L of H2O2 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0

Molar absorption coefficient -1 -1 (mgC/L)-1cm-1 M cm H2O2 Wavelength (nm) NOM Fe(II) Nitrate 0.764 179 642 9624 200 0.425 144 650 7943 210 0.180 100 551 3744 220 0.051 64 494 788 230 0.021 38 584 93 240 0.017 22 518 9 250 0.017 18 444 3 254 0.016 12 354 1 260 0.015 7 313 1 270 0.014 3 288 3 280 0.014 2 255 5 290 0.013 1 255 7 300

Raw water NO3-

Table 4. Absorbance of raw water of the Charnock wells, iron(II), ni NOM in the range from 200 to 300 nm of wavelength

H2O2 10mg/L Fe(II)

200

220 240 260 wavelength (nm)

280

300

** H2O2 absorbs photons efficiently in the range of 200-300 nm 32

Property of Low Pressure UV system and Medium Pressure UV system LPUV MPUV Diameter, D (inches) 30 48 Length, L (inches) 148 180 Number of lamps 144 18 Nominal power per lamps (kW) 0.25 15 UV efficeincy of lamps (%) 35-40 10-15 Wavelength of emit (nm) 254 200-300* * wavelength for hydrogen peroxide to absorb photons efficiently

Relative lamp output

Relative lamp output

LPUV

MPUV

1.00 0.75 0.50 0.25 0.00 200

Medium Pressure UV 220

240

260

280

300

220

240

260

280

300

1 0.75

0.5 0.25 0 200

Wavelength, nm

• LPUV lamps generate light more efficiently than MPUV lamp. Table 6. • Above 254 nm,Compari physical MPUV lamps generate less HO• due to the lower extinction coefficient of H2O2. 33

Low Pressure UV System (LPUV)

Lamp Configuration LPUV System

34

Medium Pressure UV System (MPUV)

MPUV System Lamp Configuration 35

Approach Tool utilized for simulations • Solved with AdoxTM (AOP Simulation Software) • no pseudo-steady-state assumption •pH is not constant • Gear algorithms utilized to solve stiff ODEs • completely mixed flow reactors with tank-in-series Assumptions • MtBE and tBA are the only target compounds. • Direct photolysis of MtBE and tBA is negligible. • 15 mole% of tBA is formed from MtBE oxidation. • NO3-, Fe2+ and NOM are the only interference with UV. • Fe2+, NOM and HCO3-/CO32- are the only HO• scavenger. • Decrease of UV irradiation due to scaling and bulb aging is 70%. • All UV irradiation is absorbed by the water matrix. 36

Approach (cont’d) Configuration of reactors and Modeling approach (LPUV/H2O2)

• Consist of four parallel train (each Q= 1750 gpm). Each train includes the required number of LPUV reactors in series. • Dye study data suggests 8 tanks-in-series (TIS) described reactor mixing conditions. • If both MtBE and tBA do not meet treatment objectives, the model run up to 9 maximum reactors (72 TIS). • If proved impossible to meet the criteria above, H2O2 dose is increased. Each Q = 1750 gpm Q = 7000 gpm

Target effluent conc. MtBE ≤ 2.5 µg/L tBA ≤ 6 µg/L

37

Approach (cont’d) Configuration of reactors and Modeling approach (MPUV/H2O2)

• No more than two reactors in series is allowed in design. • 4 TIS was chosen to described the mixing condition. • The number of trains would be increased until 9 parallel trains (18 reactors) to achieve the treatment target. • If proves impossible to meet the objectives, H2O2 dose would be increased. Target effluent conc. MtBE ≤ 2.5 µg/L tBA ≤ 6 µg/L Q = 7000 gpm

38

Parameters for evaluations EE/O (electrical efficiency per order of contaminant destruction)

, kWh-kgal/order P = lamp power output, kW Q = water flow rate, gal/h Ci = influent conc. of MtBE or tBA, µg/L Cf = effuluent conc. of MtBE or tBA, µg/L

39

Simulation results (LPUV/H2O2) -overallH2 O 2 (mg/L)

Pretreatment process None (raw) NalX NalX + Dealk Pellet Pellet + Dealk WAIX WAIX + Dealk RO RO + Dealk

in 25 25 10 70 10 10 10 7.0 4.0

out 13 13 6.9 35 6.9 6.3 6.9 4.9 2.9

EE/O Effluent concentration Number of (kwh-kgal/order) (μg/L) reactors per train MtBE tBA MtBE tBA 1.1 1.0 0.77 1.4 0.77 0.83 0.77 0.15 0.11

4.4 4.0 3.0 5.3 3.0 3.1 3.0 0.49 0.29

0.5 0.6 0.6 0.4 0.6 0.3 0.6 1.3 0.2

5.7 6.0 5.8 5.4 5.8 4.7 5.8 5.7 1.8

9 8 6 11 6 7 6 1 1

Number of trains 4 4 4 4 4 4 4 4 4

Table 7. • Pretreatment (i.e. Dealkalization) significantly improves the Summary of results for using the LPUV reactor with all pretreatment alternatives. treatment efficiency and decrease the EE/O and the # of reactors. • NaIX + Dealk would be preferred because less residues of H2O2 and less complexity. 40

Significance of number of reactors (LPUV/H2O2) - NaIX with/without dealkalization(1) NaIX with Dealkalization (6 reactors required) Acid CO2 Caustic addition Stripping addition

(2) NaIX without Dealkalization (8 reactors required)

41

Simulation Results (LPUV/H2O2) – NAIX + Dealkalization MtBE

tBA

16.0

4.00

14.0

3.50

12.0

3.00

10.0

2.50

8.0

2.00

6.0

1.50

4.0

1.00

2.0

0.50

0.0

0.00 5

10

15

EEO (kWh/kgal-order)

H2 O2 residual conc. (mg/L)

Residual of H2O2

20

H2 O2 Dosage (mg/L)

• 10 mg/L of H2O2 dosage would be the better choice although the optimum dosage of H2O2 is over 20 mg/L. This is because the cost required for over 20 mg/L of H2O2 dose overweigh the energy cost. • At 10 mg/H2O2 dosage, UV dose was calculated as 2,100 mJ/cm422.

Simulation results (MPUV/H2O2) -overallPretreatment process None (raw) NalX NalX + Dealk pellet Pellet + Dealk WAIX WAIX + Dealk RO RO + Dealk

H2 O 2 (mg/L) in 30 30 10 50 10 10 10 4.0 4.3

out 15 12 6.6 19 6.6 5.9 6.6 1.9 2.2

EE/O (kwh-kgal/order) MtBE 7.3 6.2 4.6 8.3 4.6 5.2 4.6 0.99 0.48

tBA 23 18.0 15.0 27 15.0 16 15.0 2.8 1.3

Effluent concentration (μg/L) MtBE tBA 1.3 5.2 0.5 3.4 1.4 5.8 1.4 5.9 1.4 5.8 1.0 4.8 1.4 5.9 2.1 5.0 1.8 4.5

Number of reactors per train

Number of trains

2 2 2 2 2 2 2 1 1

8 8 5 9 5 6 5 2 1

• Pretreatment (i.e. Dealkalization) significantly improves the treatment efficiency and decrease the EE/O and the # of trains. • NaIX + Dealk would be desired as well as in the case of LPUV/H2O2 system. 43

Significance of number of trains (MPUV/H2O2) - NaIX with/without dealkalization(1)NaIX with Dealkalization (5 trains required) Acid CO2 Coustic addition Stripping addition

(2) NaIX without Dealkalization (8 trains required)

44

Simulation Results (MPUV/H2O2) – NAIX + Dealkalization Residual of H2O2

MtBE

tBA 30.0

14.0

25.0

12.0 20.0

10.0 8.0

15.0

6.0

10.0

4.0

EEO (kWh/kgal-order)

H2 O2 residual conc. (mg/L)

16.0

5.0

2.0 0.0

0.0 5

10 15 H2 O2 Dosage (mg/L)

20

• 10 mg/L of H2O2 dosage would be the best choice although the optimum dosage of H2O2 is over 20 mg/L. This is because the cost required for over 20 mg/L of H2O2 dose overweigh the energy cost. • At 10 mg/H2O2 dosage, UV dose was calculated as 10,700 45 mJ/cm2.

Cost comparison NaIX + Dealkalization pretreatment

LPUV/H2O2 MPUV/H2O2

# of reactors

Total power per day (kWh)

Cost of energy ($/day)

H2O2 dose (mg/L)

Total amount per day (lb)

Cost of chemical ($/day)

24 10

21,300 108,000

$2,130 $10,800

10 10

600 600

$899 $899

* Unit prices for H2O2 and electrical energy are $ 1.5/lb and $0.10 kWh, respectively. Table 11. Cost comparison for the desing comparing of the LPUV reactor and the MPUV reactor

• Cost of energy for MPUV system is 5 times higher than for LPUV system. • In the design process, a comparison of EE/O versus the H2O2 dosage provides valuable insight into the tradeoffs and support determination an appropriate H2O2 dosage.

46

Summary • Kinetic Models can be used to simulate the complex interaction of UV/H2O2 chemical physical reactions with water matrixes (e.g. TOC, carbonate species). • Removal of carbonate species using pretreatment processes can reduce the energy cost by much as 1/3 and reduce the number of reactors from 36 to 24 which would translate into a capital cost savings by 1/3. • For the LPUV system, the cost of chemicals can be as high as the energy cost. 47

Outline 1. RO retentate DOC removal –comparison of energy consumption among various Advanced Oxidation Processes (AOPs) 2. AdOx –advanced oxidation simulation software 3. Case study 1. UV/H2O2 process for methyl tert-butyl ether (MtBE) and tertiary butyl alcohol (tBA) removal from drinking water source: effect of pretreatment options and light source 4. Case study 2. Bromate mitigation during ozonation and MtBE and tBA removal during O3/H2O2 process 5. UV/H2O2 mechanism study 48

Application of ozone to water treatment • Direct reactions with O3

• Indirect reactions with HO• produced by O3 with NOM O3 + NOM → HO• + byproducts • HO• is quenched by the reaction with NOM HO• + NOM → byproducts

• O3/H2O2 Advanced Oxidation Process (AOP) H2O2 ↔ HO2- + H+ pKa = 11.6 O3 + HO2- → HO• + •O2k = 2.2 × 106 M-1s-1 • Overall Reaction for HO• formation 2O3 + H2O2 → 2HO• + 3O2

49

Concern about bromate • Formation of bromate (BrO3-) during ozonation in the presence of bromide ion (Br-) • A nationwide survey of Br- in drinking water sources: approximately 80 μg/L (Amy et al., 1994) • Br- in costal area is expected higher • 10 μg/L of BrO3- standard of MCL associated with cancer risk, Stage 1 of the Disinfectant/Disinfection By-Product (D/DBP) Rule (EPA, 1998) • When ozone is applied to disinfection, tradeoffs between inactivation of cryptosporidium and bromate formation should be evaluated. 50

Mechanisms of bromate formation Simplified reaction scheme for bromate formation during ozonation Disproportionation

Haag and Hoigné, 1983 Gunten and Hoigné, 1994

pKa = 8.8 Ozone involving pathway

k (M-1 s-1)

HO• involving pathway

O3 + Br - → OBr- + O2 OBr- + O3 → 2O2 + Br OBr- + O3 → O2 + BrO2 HOBr + O3 → O2 + BrO2 - + H+ BrO2 - + O3 → BrO3 - + O2 O3 + Br• → BrO• + O2

160 330 100

HO• + HOBr → BrO• + H2O