Adaptation of UV Advanced
Oxidation for Inland Potable
Reuse Treatment
WateReuse in Texas
San Marcos, TX
July 15, 2016
Michael Watts, PhD, PE, Steven Jones, PhD, PE – Garver
David Sloan, PE, BCEE – Freese and Nichols
Erik Rosenfeldt, PhD, PE – Hazen and Sawyer
What roles can UV AOP fill in a potable reuse
treatment train?
The precedent for UV and UV AOP in Texas direct
potable reuse projects
Does UV AOP always need RO pretreatment?
A collaborative research project to assess potential for UV
AOP treatment of reclaimed waters of varying quality
• H2O2/UV
• HOCl/UV
Both Texas DPR projects included UV treatment
following RO
Big Spring, TX
Microfiltration
UV AOP
Reverse Osmosis
Produced
Water
Hydrogen Peroxide
Both Texas DPR projects included UV treatment
following RO
Wichita Falls, TX
Microfiltration
UV
Reverse Osmosis
Not AOP
Produced
Water
Reverse osmosis pretreatment improves the
efficiency of H2O2/UV advanced oxidation
RO pretreatment reduces the dissolved organic carbon
concentration, which can limit targeted contaminant oxidation
RO pretreatment minimizes photon scavenging by substances other
than H2O2
RO pretreatment greatly reduces the concentration of most trace
organic contaminants, thereby reducing the competition for available
·OH
Both TX DPR facilities had neighboring streams
with capacity to assimilate RO concentrate TDS
Wichita Falls
RO
2.5 MGD Reject
Big Wichita
River
Big Spring
5 MGD
Permeate
RO
0.9 MGD Reject
Beals Creek
2.5 MGD
Permeate
A new approach to UV AOP (HOCl/UV) could see
effective treatment of marginal reuse waters
Feasibility Curves: 0.5-log MIB
oxidation with ≤ 8 mg/L as Cl2
or H2O2 (initial dose)
UV Dose = 750 mJ/cm2
HOCl/UV
DOC,
mg/L
H2O2/UV
RO permeate (typical)
pH
In 2015, WateReuse Texas sponsored a pilot
study of UV AOP after varying levels of filtration
Media Filtration
UV Pilot
Secondary
Effluent
Secondary
Effluent
Secondary
Effluent
NaOCl or
H2O2
In 2015, WateReuse Texas sponsored a pilot
study of UV AOP after varying levels of filtration
Secondary
Effluent
Submerged MF
UV Pilot
Secondary
Effluent
Secondary
Effluent
NaOCl or
H2O2
In 2015, WateReuse Texas sponsored a pilot
study of UV AOP after varying levels of filtration
Secondary
Effluent
or
NaOCl
Secondary
Effluent
Submerged MF
Secondary
Effluent
RO
H2O2
UV Pilot
Objectives for the 2015 pilot study
Assess feasibility of novel indirect [·OH] measurement technique in
reclaimed waters of varying quality
Test Watts, Rosenfeldt, and Hofmann (2012) steady-state [·OH]
model for predicting AOP performance in reclaimed waters of
varying quality
Predict degree of UV AOP treatment needed to see micropollutant
oxidation in reclaimed waters of varying quality
Initially, each sample from Lawton and Wichita
Falls was surveyed for water quality and trace
organic pollutant profiles
Media Filtration
UVT = 60%
15
10
5
0
pH
Ammonia-N Nitrate-N
TOC
Initially, each sample from Lawton and Wichita
Falls was surveyed for water quality and trace
organic pollutant profiles
Media Filtration
UVT = 60%
UVT = 77%
20
15
Submerged MF
10
5
0
pH
Ammonia-N Nitrate-N
TOC
Initially, each sample from Lawton and Wichita
Falls was surveyed for water quality and trace
organic pollutant profiles
Media Filtration
UVT = 60%
UVT = 77%
UVT = 99%
20
15
10
Submerged MF
5
0
RO
pH
Ammonia-N Nitrate-N
TOC
rtO
ct
A
pa
ra
be
n
M
Tr
ic
lo
ca
rb
an
Su
cr
al
os
e
Pr
op
yl
ND
Io
pr
om
id
e
Io
he
xa
l
Bu
ta
lb
ita
l
Di
clo
fe
na
c
G
em
fib
ro
zi
l
lfa
m
eK
ol
ph
en
ol
ylp
he
n
Ac
es
u
4te
4no
ny
l
600
CEC
2,
4D
37000
Sucralose
ng/L
4-nonylphenol and sucralose were most
prevalent CECs in Lawton effluent samples
38000
36000
35000
900
300
0
rtO
ct
ny
l
ph
en
o
l
4D
Ib
lo
A
ra
lo
se
ca
rb
an
Su
c
Tr
ic
M
e
NP
YR
ND
om
id
l
of
en
Io
he
xa
up
r
il
ita
l
ro
z
lb
em
fib
Io
pr
G
Bu
ta
ylp
he
no
l
Ac
es
ul
fa
m
eK
4te
no
2,
2000
CEC
4-
45000
Sucralose
ng/L
The concentrations of iohexol and sucralose
were greatest in MF filtrate from Wichita Falls
55000
50000
40000
3000
1000
0
4te
rtO
A
os
e
M
Tr
icl
os
an
al
ND
ylp
he
no
l
he
no
l
Su
cr
ct
4no
ny
lp
ng/L
Fewer micropollutants were detected following
RO
300
200
100
A sample of each water was also tested for Total
·OH-Scavenging (∑
)
,
∑
,
,
Real-time decay
data collected
for an ·OH-probe
under
UV AOP conditions
(H2O2 and UV)
As expected, Lawton effluent had the greatest
Total ·OH-Scavenging
1.20E+06
,
, 1/
1.00E+06
8.00E+05
6.00E+05
4.00E+05
2.00E+05
0.00E+00
Lawton
Rate of Total ·OH Scavenging
WF MF
WF RO
Total ·OH-Scavenging in RO permeate sample
equivalent to nitrite-less MBR effluent
1.00E+06
No Nitrite
, 1/
No Nitrite
,
1.00E+05
1.00E+04
MBR Effluent
MBR Effluent
Activated Sludge
Effluent
Activated Sludge
Effluent
WF RO
WF MF
Lawton
Source: Grant and Hofmann (2016) Water Science and
Technology May 2016, 73 (9) 2067-2073
Each sample was dosed with a chemical
oxidant and pumped at controlled rates through
the annular UV reactor
NaOCl (mg/L as Cl2)
H2O12.5
2 (mg/L)
Lawton:
Lawton: 25
WFWFMF:
MF:7.8
12.5 -17
RO:4.8
7–8
WFWFRO:
pH
Lawton: 8
WF MF: 7.8
WF RO: 6.5
UV Fluence (mJ/cm2)
Lawton: 282 – 364
WF MF: 458 – 493
WF RO: 711 – 735
Each sample was spiked with 100 ppb 1,4dioxane as a probe compound for monitoring
AOP treatment performance
! · #$
∑
,
,
%#$,&'()*+, -.
A prevalent artificial sweetener appeared
susceptible to multiple oxidation pathways
Acesulfame-K
Influence of NO3, ·Cl?
1000
ng/L
800
600
400
200
0
Lawton
Influent
WF MF
Effluent
Model
The most prominent micropollutant in RO
permeate samples, NDMA, was most efficiently
mitigated with UV alone
For equivalent UV fluence
40
20
V(
1)
Cl
/U
HO
H2
O
2/
UV
(1
)
0
UV
NDMA Removal, %
60
Regulated DBPs may be a concern for
HOCl/UV AOP
Initial Cl2: 7.8 mg/L
pH: 7.8
Wichita Falls MF Filtrate
30
20
µg/L
test
H2O2/UV(1)
HOCl/UV(1)
UV
10
Ac
i
ds
(H
AA
5)
ac
id
To
ta
lH
al
oa
ce
tic
hl
or
oa
ce
tic
Tr
ic
et
ic
hl
or
oa
c
Di
c
Di
br
om
oa
ce
tic
ac
ac
i
id
d
0
Short-term (<30 minutes)
HAA formation
However, RO removed this short-term HAA
formation potential
RO Permeate
UV
HOCl/UV
H2O2/UV
Dibromoacetic acid
ND
ND
ND
Dichloroacetic acid
ND
ND
ND
Monobromoacetic acid
ND
ND
ND
Monochloroacetic acid
ND
ND
ND
Total Haloacetic Acids (HAA5)
ND
ND
ND
Conclusions Drawn
Trace anthropogenic contaminants (targets for UV AOP oxidation) were
detected in all reclaimed water samples, including RO permeate.
In RO permeate, NDMA was detected at greater concentrations than
other micropollutants
• UV at AOP doses was most effective treatment for residual NDMA
Is UV AOP necessary following RO? The Wichita Falls DPR Model =
RO for organics removal + UV for NDMA destruction
Conclusions Drawn
Upstream treatment processes that can have a significant impact on
UV AOP efficiency:
• Denitrification
• Chloramination
• RO
Remaining Questions to Answer
Can disinfection by-product formation be mitigated for
HOCl/UV? Without RO?
Which lamp technology will be most effective for
HOCl/UV? MP or LP UV?
Expand kinetic model for HOCl/UV to assess potential
impact of
• Nitrate/Nitrite in Fully Nitrified Reuse Water
Acknowledgements
• The WateReuse Texas Association
• City of Wichita Falls: Mark Southard, Daniel
Nix, and Hunter Adams
• City of Lawton: Afsaneh Jabbar, and Lyna
Neal
• Trojan Technologies: Adam Festger
• Freese and Nichols: Chris Connolly
• Garver: Kyle Kruger
• Eurofins Analytical: Andy Eaton
Questions?
[email protected]
Garver
Frisco, TX
(972)377-7480