6.
References
APHA­
AWWA­
WEF.
1998.
Standard
methods
for
the
examination
of
water
and
wastewater,
20th
ed.
Washington
D.
C.

Aquafine
Corporation.
2001.
CSL
series:
Installation,
maintenance
and
operation
manual.
Part
no.
115­
1.

Bourgine,
F.
P.,
J.
I.
Chapman,
H.
Kerai,
J.
I.
Duval,
J.
G.
Green
and
D.
Hamilton.
1995.
The
degradation
of
atrazine
and
other
pesticides
by
photolysis.
The
chartered
institution
of
water
and
environmental
management
9:
417­
422.

Brandt,
C.
L.,
and
A.
C.
Giese.
1956.
Photoreversal
of
nuclear
and
cytoplasmic
effects
of
short
ultraviolet
radiation
on
paramecium
caudatum.
J.
Gen.
Physiol.
39,
no.
5:
735­
751.

Bukhari,
Z.,
T.
M
Hargy,
J.
R.
Bolton,
B.
Dussert,
and
J.
L.
Clancy.
1999.
Medium­
pressure
UV
for
oocyst
inactivation.
Journal
AWWA
91:
86­
94.

Cabaj,
A.,
R.
Sommer,
W.
Pribil,
and
T.
Haider.
2001.
What
means
"
dose"
in
UV­
disinfection
with
medium
pressure
lamps.
Ozone
Science
&
Engineering
23:
239­
244.

Carollo
Engineers.
2001.
Weber
basin
water
treatment
plant
no.
3
expansion.
Layton,
Utah.

Chang,
J.
C.
H.,
S.
F.
Osoff,
D.
C.
Lobe,
M.
H.
Dorfman,
C.
M.
Dumais,
R.
G.
Qualls,
and
J.
D.
Johnson.
1985.
UV
inactivation
of
pathogenic
and
indicator
microorganisms.
Applied
and
Environmental
Microbiology
49,
no.
6:
1361­
1365.

Cushing,
R.
S.,
E.
D.
Mackey,
J.
R.
Bolton,
and
M.
I.
Stefan.
2001.
Impact
of
common
water
treatment
chemicals
on
UV
disinfection.
Proceedings
of
the
AWWA
Annual
Conference
and
Exposition,
June
17­
21,
Washington,
D.
C.

Dinkloh,
L.
2001.
Interview
by
Ben
Hauck.
Telephone
conversation.
Tucson,
AZ.,
August
10.

Downes,
A.
and
T.
P.
Blunt.
1877.
Researches
on
the
effect
of
light
upon
bacteria
and
other
organisms.
Proceedings
Royal
Society
London.
Volume
26.

Dulbecco,
R.
1950.
Experiments
on
photoreactivation
of
bacteriophages
inactivated
with
ultraviolet
radiation.
Journal
of
Bacteriology
59:
329­
347.

Gates,
F.
L.
1929.
A
Study
of
the
Bactericidal
Action
of
Ultraviolet
Light.
J.
Gen.
Physiol.
13:
231­
260.

GE
Quartz.
2001.
Available
from
Internet
location
http://
www.
gequartz.
com/
en/
lamptube.
htm.
Internet.

Gump,
D.
2002.
Severn
Trent.
Personal
communication
with
Jennifer
Hafer,
Malcolm
Pirnie,
Inc.,
regarding
UV
reactors.
Tucson
AZ,
November
6.

UV
Disinfection
Guidance
Manual
6­
1
June
2003
Proposal
Draft
Chapter
6.
References
UV
Disinfection
Guidance
Manual
6­
2
June
2003
Jagger,
J.
1967.
Introduction
to
research
in
ultraviolet
photobiology.
Englewood
Cliffs,
N.
J:
Prentice­
Hall,
Inc.

Kashinkunti,
R.,
K.
G.
Linden,
G.
Shin,
D.
H.
Metz,
M.
D.
Sobsey,
M.
Moran,
and
A.
Samuelson.
2003.
Achieving
multi­
barrier
inactivation
in
Cincinnati:
UV,
byproducts,
and
biostability.
JAWWA.
Submitted
January
9,
2003.

Kelner,
A.
1950.
Light­
induced
recovery
of
microorganisms
from
ultraviolet
radiation
injury,
with
special
reference
to
Escherichia
coli.
Bulletin
of
the
New
York
Academy
of
Medicine
26:
189­
199.

Knudson,
G.
B.
1985.
Photoreactivation
of
UV­
irradiated
Legionella
Pneumoplila
and
other
Legionella
species.
Applied
and
Environmental
Microbiology
49,
no.
4:
975­
980.

Kruithof,
J.
C.,
and
R.
C.
van
der
Leer.
1990.
Practical
experiences
with
UV­
disinfection
in
the
Netherlands.
Proceedings
of
the
AWWA
seminar
on
emerging
technologies
in
practice,
AWWA
Annual
Conference,
June
17­
21,
Cincinnati,
OH.

Lienberger,
J.
2002.
Trojan
Technologies.
Personal
communication
with
Jennifer
Hafer,
Malcolm
Pirnie,
Inc,
regarding
UV
reactors.
November
5.

Lin,
L.
S.,
C.
T.
Johnston,
and
E.
R.
Blatchley
III.
1999a.
Inorganic
fouling
at
quartz:
water
interfaces
in
ultraviolet
photoreactors
I:
chemical
characterization.
Water
Research
33,
no.
15:
3321­
3329.

Lin,
L.
S.,
C.
T.
Johnston,
and
E.
R.
Blatchley.
1999.
Inorganic
fouling
at
quartz:
water
interfaces
in
ultraviolet
photoreactors
 
II
Temporal
And
Spatial
Distribution.
Water
Research
33,
no.
15:
3330­
3338.

Linden,
K.
G.,
G.
A.
Shin,
and
M.
D.
Sobsey.
2001.
Comparative
effectiveness
of
UV
wavelengths
for
the
inactivation
of
Cryptosporidium
parvum
oocysts
in
water.
Water
Science
&
Technology
43,
no.
12:
171­
174.

Linden,
K.
G.,
G.
A.
Shin,
G.
Faubert,
W.
Cairns,
and
M.
D.
Sobsey.
2002a.
UV
disinfection
of
Giardia
lamblia
cysts
in
water.
Environmental
Science
and
Technology
36,
no.
11:
2519­
2522.

Linden,
K.
G.,
L.
Batch,
and
C.
Schulz.
2002b.
UV
Disinfection
of
filtered
water
supplies:
water
quality
impacts
on
MS2
dose­
response
curves.
Proceedings
of
the
AWWA
Annual
Conference,
June
16­
20,
New
Orleans,
LA.

Linden,
K.
G.
2002.
Personal
email
communication
by
Laurel
Passantino.
April
24­
May
6.

Mackey,
E.
D.,
R.
S.
Cushing,
and
H.
B.
Wright.
2001.
Effect
of
water
quality
on
UV
disinfection
of
drinking
water.
Proceedings
of
the
First
International
Ultraviolet
Association
Congress,
June
14­
16,
Washington,
D.
C.

Proposal
Draft
Chapter
6.
References
UV
Disinfection
Guidance
Manual
6­
3
June
2003
Mackey,
E.
D.,
R.
S.
Cushing,
M.
L.
Janex,
N.
Picard,
J.
M.
Laine,
and
J.
P.
Malley.
2003.
Bridging
Pilot­
Scale
Testing
to
Full­
Scale
Design
of
UV
Disinfection
Systems,
Draft
Report.
Final
report
to
be
published
in
2004.
Denver,
CO:
AWWA
Research
Foundation.

Malley,
J.
P,
J.
P.
Shaw
and
J.
R.
Ropp.
1995.
Evaluation
of
by­
products
produced
by
treatment
of
groundwaters
with
ultraviolet
irradiation.
Denver,
CO.:
AWWA
Research
Foundation.

Malley,
J.
P.
2002.
Historical
perspective
of
UV
use.
Presented
at
the
AWWA
Water
Quality
Technology
Conference,
November
10­
14,
Seattle,
W.
A.

Meulemans,
C.
C.
E.
1986.
The
basic
principles
of
UV­
sterilization
of
water.
In:
Ozone
+
Ultraviolet
Water
Treatment,
Aquatec
Amsterdam,
Paris:
International
Ozone
Association,
B.
1.1­
B.
1.13.

Munakata,
N.,
M.
Daito,
and
K.
Hieda.
1991.
Inactivation
spectra
of
Bacillus
Subtilis
spores
in
extended
ultraviolet
wavelengths
(
50­
300nm)
obtained
with
Synchrotron
radiation.
Photochemistry
and
Photobiology
54,
no
5:
761­
768.

NWRI.
2000.
Ultraviolet
Disinfection:
Guidelines
for
Drinking
Water
and
Water
Reuse.
National
Water
Research
Institute,
Fountain
Valley,
CA.

Oguma
K.,
H.
Katayama,
H.
Mitani,
S.
Morita,
T.
Hirata,
and
S.
Ohgaki.
2001.
Determination
of
pyrimidine
dimers
in
Escherichia
coli
and
Cryptosporidium
parvum
during
UV
light
inactivation,
photoreactivation
and
dark
repair.
Applied
and
Environmental
Microbiology
67,
no
10:
4630
 
4637.

Passantino,
L.
B.
and
J.
P.
Malley.
2001.
Impacts
of
turbidity
and
algal
content
of
unfiltered
drinking
water
supplies
on
the
ultraviolet
disinfection
process.
Proceedings
of
the
2001
AWWA
Annual
Conference,
Washington
D.
C.

Powell,
W.
F.
1959.
Radiosensitivity
as
an
index
of
Herpes
Simplex
virus
development.
Virology
9:
1­
19.

Rauth,
A.
M.
1965.
The
physical
state
of
viral
nucleic
acid
and
the
sensitivity
of
viruses
to
ultraviolet
light.
Biophysical
Journal
5:
257­
273.

Roberts,
Andrew.
2000.
Letter
to
Dan
Schmelling
of
the
USEPA,
April
10.

Sharpless
C.
M.,
and
K.
G.
Linden.
2001.
UV
photolysis
of
nitrate:
effects
of
natural
organic
matter
and
dissolved
inorganic
carbon,
and
implications
for
UV
water
disinfection.
Environmental
Science
and
Technology
35,
no
14:
2949­
2955.

Shin
G.
A.,
K.
G.
Linden,
M.
J.
Arrowood,
G.
Faubert,
and
M.
D.
Sobsey.
2001.
DNA
repair
of
UV­
irradiated
Cryptosporidium
parvum
oocysts
and
Giardia
lamblia
cysts.
Proceedings
of
the
First
International
Ultraviolet
Association
Congress,
June
14­
16,
Washington,
D.
C.

Proposal
Draft
Chapter
6.
References
UV
Disinfection
Guidance
Manual
6­
4
June
2003
Snowball,
M.
R.,
and
I.
S.
Hornsey.
1988.
Purification
of
water
supplies
using
ultraviolet
light.
Developments
in
food
microbiology,
edited
by
R.
K.
Robinson,
171­
191.
NY:
Elsevier
Applied
Science.

Swaim,
P.
D.,
M.
A.
Morine,
R.
G.
Brauer,
M.
A.
Neher,
J.
L
Gebhart,
W.
D.
Bellamy.
2002.
Operating
data
from
Henderson's
full­
scale
UV
disinfection
facility.
Presented
at
the
water
quality
technology
conference,
November
2002.

Toivanen,
E.
2000.
Experiences
with
UV
disinfection
at
Helsinki
water.
IUVA
News
2,
no.
6:
4­
8
U.
S.
Environmental
Protection
Agency.
2000.
Technical
memorandum:
Worldwide
ultraviolet
disinfection
installations
for
drinking
water.
Office
of
Groundwater
and
Drinking
Water,
Washington,
DC.

U.
S.
Environmental
Protection
Agency.
2000.
Technological
costs
for
control
of
microbial
contaminants
and
disinfection
by­
products.
Office
of
Groundwater
and
Drinking
Water,
Washington,
DC.

U.
S.
Environmental
Protection
Agency.
2002.
Microbial
Toolbox
Guidance
Manual.
Office
of
Groundwater
and
Drinking
Water,
Washington,
DC.

U.
S.
Environmental
Protection
Agency.
2003.
Public
water
system
guidance
manual
for
source
water
monitoring
under
the
LT2ESWTR.
Office
of
Groundwater
and
Drinking
Water,
Washington,
DC.

von
Sonntag,
C.,
and
H.
P.
Schuchmann.
1992.
UV
disinfection
of
drinking
water
and
by­
product
formation
 
some
basic
considerations.
Journal
of
Water
Supply
Research
and
Technology
41,
no
2:
67­
74
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
This
appendix
supplements
Chapter
2,
Overview
of
UV
Disinfection,
with
an
additional
level
of
detail.
The
purpose
of
this
appendix
is
to
provide
technical
information
regarding
the
physical
mechanisms
of
UV
light
generation,
biological
reactions
causing
disinfection,
and
UV
reactor
equipment.
The
organization
of
this
appendix
is
presented
below,
including
the
key
questions
addressed
by
each
section.

 
How
is
UV
light
generated?.......................................................................
Section
A.
1.1
 
What
happens
to
UV
light
as
it
propagates
through
water?.......................
Section
A.
1.2
 
How
does
UV
light
inactivate
microorganisms?
.......................................
Section
A.
2.2
 
Can
microorganisms
undergo
repair
and
become
infectious
after
inactivation
by
UV
light?..................................................
Section
A.
2.3
 
How
is
UV
dose
determined
in
a
bench­
scale
(
batch)
system?....................................................................................................
Section
A.
2.4.1
 
How
does
UV
dose
vary
in
a
UV
reactor?..............................................
Section
A.
2.4.2
 
How
do
microbial
dose­
response
curves
differ?........................................
Section
A.
2.5
 
What
factors
influence
microbial
dose­
response?
.....................................
Section
A.
2.6
 
Do
all
microorganisms
have
the
same
sensitivity
to
UV
light?
..........................................................................................................
Section
A.
2.7
 
What
are
the
components
of
a
UV
installation?
...........................................
Section
A.
3
 
How
do
low
pressure,
low­
pressure
high­
output,
and
medium
pressure
lamps
differ?.................................................
Section
A.
3.1.2­
A.
3.1.4
 
What
happens
to
UV
lamps
as
they
age?................................................
Section
A.
3.1.6
 
How
are
UV
lamps
powered?
....................................................................
Section
A.
3.2
 
What
is
the
function
of
the
lamp
sleeve?
...................................................
Section
A.
3.3
 
How
are
lamp
sleeves
cleaned
and
why
is
it
necessary
to
clean
them?
............................................................................................
Section
A.
3.4
 
How
is
UV
light
monitored
in
a
reactor?...................................................
Section
A.
3.5
 
How
are
the
components
of
a
UV
reactor
arranged?.................................
Section
A.
3.8
UV
Disinfection
Guidance
Manual
A­
1
June
2003
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
2
June
2003
 
How
do
the
utility
and
the
State
know
the
UV
reactor
is
delivering
the
required
UV
dose?
..............................................................
Section
A.
3.9
 
What
are
the
impacts
of
water
quality
on
UV
disinfection?...............................................................................................
Section
A.
4.1
 
Do
any
disinfection
byproducts
form
as
a
result
of
UV
disinfection?...............................................................................................
Section
A.
4.2
A.
1
UV
Light
Generation
and
Propagation
Through
Liquid
Media
Using
UV
light
to
disinfect
drinking
water
involves
generating
UV
light
with
the
desired
germicidal
properties
and
subsequently
delivering
that
light
to
the
target
pathogens.
This
section
describes
fundamental
concepts
related
to
the
generation
and
transmission
of
UV
light.

A.
1.1
UV
Light
Generation
Atoms
and
ions
emit
light
when
they
change
from
a
higher
to
a
lower
energy
state.
An
atom
and
most
ions
consist
of
electrons
orbiting
a
nucleus
of
protons
and
neutrons.
The
electrons
in
each
orbital
occupy
a
unique
energy
state,
where
the
electrons
closest
to
the
nucleus
have
a
lower
energy
and
the
electrons
further
away
have
a
higher
energy.
When
an
electron
makes
a
transition
from
a
higher
energy
state
to
a
lower
energy
state,
a
discrete
amount
of
energy
is
released
as
photons
of
light
at
a
particular
wavelength
(
 )
according
to
Equation
A.
1.

 
hc
E
E
=
 
1
2
Equation
A.
1
where
E1
=
Lower
energy
state
(
J)
E2
=
Higher
energy
state
(
J)
h
=
Planck's
Constant
(
6.626
x
10­
34
J°
s)
c
=
Speed
of
light
(
2.997
x
108
m/
s)
 
=
Wavelength
(
m)

Energy
levels
of
a
given
atom
or
ion
are
unique
and
depend
on
the
number
of
electrons,
protons,
and
neutrons
within
that
atom
or
ion
and
their
interaction
with
external
force
fields.
As
such,
each
element
emits
a
unique
spectrum
of
light.
If
the
difference
between
energy
levels
is
appropriate,
the
light
emitted
is
in
the
UV
range.

A
transition
from
a
lower
to
a
higher
energy
state
requires
an
energy
input.
This
energy
may
be
derived
from
the
collision
of
the
atom
with
a
photon
of
light
of
wavelength
 
or
by
collision
with
other
atoms,
ions,
or
electrons.
Energy
transferred
to
the
atom
may
result
in
an
increase
in
the
atom's
kinetic
energy,
the
transfer
of
an
electron
to
a
higher
energy
level,
or
the
removal
of
an
electron
from
the
atom.
Removal
of
an
electron
from
the
atom
is
termed
ionization
and
results
in
a
positively
charged
cation
and
a
negatively
charged
free
electron.
The
energy
required
to
remove
an
electron
from
an
atom
is
termed
the
ionization
energy.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
3
June
2003
Recombination
of
a
free
electron
and
a
cation
may
result
in
the
emission
of
light.
Since
the
free
electron
and
cation
may
have
a
range
of
kinetic
energies,
the
wavelength
of
emitted
light
will
vary.
The
wavelength
range
will
be
bound
by
the
ionization
energy
of
the
atom,
and
there
will
be
a
peak
within
the
rage
that
depends
on
the
temperature
of
the
electrons
and
cations.
The
following
sections
discuss
the
relationship
between
atomic
energy
states
and
the
generation
of
UV
light
through
gas
and
mercury
discharges.

A.
1.1.1
Gas
Discharges
A
gas
discharge
is
a
mixture
of
non­
excited
atoms,
excited
atoms,
cations,
and
free
electrons
formed
when
a
sufficiently
high
voltage
is
applied
across
a
volume
of
gas.
The
wavelength
of
light
emitted
from
the
gas
discharge
depends
on
the
elemental
composition
of
the
gas
discharge
and
the
excitation,
ionization,
and
kinetic
energy
of
those
elements.

The
formation
of
the
gas
discharge
within
a
UV
lamp
involves
several
stages.
When
a
voltage
is
first
applied,
free
electrons
and
ions
present
in
the
gas
are
accelerated
by
the
electric
field
formed
between
two
electrodes.
Initially,
the
concentration
of
free
electrons
and
ions
arises
from
natural
radioactivity
and
is
very
low.
With
sufficient
voltage,
the
electrons
are
accelerated
to
high
kinetic
energies.
Collisions
of
the
free
electrons
with
atoms
result
in
a
transfer
of
energy
to
the
atoms.
If
the
energy
transferred
is
sufficient,
the
atoms
are
ionized.
This
ionization
provides
a
rapid
increase
in
the
number
of
free
electrons
and
cations,
a
corresponding
increase
in
lamp
current,
and
a
drop
in
the
voltage
across
the
lamp.

Cations
colliding
with
an
electrode
cause
electrons
to
be
emitted.
If
sufficient
electrons
are
emitted,
a
self­
sustaining
discharge
termed
a
glow
discharge
occurs.
Initially,
only
a
small
fraction
of
each
electrode
emits
electrons.
With
an
increase
in
current,
this
area
increases
until
the
entire
electrode
is
in
use.
To
increase
the
current
beyond
that
point,
the
voltage
is
increased
to
provide
more
kinetic
energy
to
the
cations.
High
energy
cations
that
collide
with
the
electrode
increase
the
electrode's
temperature.
At
sufficiently
high
temperatures,
the
electrode
begins
to
thermally
emit
electrons,
and
a
further
increase
in
current
reduces
the
voltage
requirement.
At
this
point,
the
electrode
discharge
is
termed
an
arc
discharge.

The
start
voltage,
which
is
the
voltage
required
to
start
the
gas
discharge,
is
typically
higher
than
the
ionization
potential
of
the
gas
unless
a
means
is
used
to
introduce
electrons.
Preheating
the
electrode
or
producing
a
strong
local
field
using
a
third
electrode
located
close
to
one
of
the
electrodes
can
be
used
to
introduce
electrons
and
aid
in
starting
the
gas
discharge.

A
gas
discharge
has
a
negative
impedance
that
is
intrinsically
unstable
unless
a
ballast
is
placed
in
series
to
provide
a
positive
impedance
to
the
power
supply.
With
a
direct
current
(
DC)
supply
powering
the
gas
discharge,
the
ballast
is
a
resistor.
With
an
alternating
current
(
AC)
supply,
the
ballast
is
either
an
inductor,
capacitor,
or
some
combination
of
those
components.
Inductors
and
capacitors
are
preferred
over
resistors
because
they
do
not
consume
power.
More
detail
on
ballasts
is
presented
in
section
A.
3.2.

The
frequency
of
the
AC
supply
impacts
the
performance
of
the
gas
discharge.
If
the
frequency
of
the
AC
supply
is
low
(<<
1
kHz),
electron­
cation
recombination
extinguishes
the
discharge
every
half
cycle
of
the
lamp
voltage.
Re­
ignition
during
the
next
half
cycle
is
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
4
June
2003
facilitated
by
electron
emission
from
the
still
warm
electrodes.
If
the
frequency
of
the
AC
supply
is
greater
than
1
kHz,
the
free
electrons
and
cations
do
not
have
sufficient
time
to
recombine
and
the
discharge
does
not
extinguish.

A.
1.1.2
Mercury
Discharges
Mercury
in
a
gas
discharge
is
used
to
generate
the
UV
light
produced
in
most
commercial
UV
lamps.
Mercury
is
an
advantageous
element
for
UV
disinfection
due
to
the
following
factors:

 
Electron
transitions
within
mercury
provide
electromagnetic
energy
in
the
germicidal
wavelength
range.

 
Mercury
at
low
vapor
pressure
and
near
room
temperature
produces
light
at
wavelength
253.7
nm
from
electrical
energy
with
high
efficiency.
This
wavelength
is
near
optimal
for
UV
disinfection
(
section
A.
2.2).

 
Mercury
at
high
vapor
pressures
produces
high
intensity
polychromatic
UV
light
with
reasonably
high
efficiency.

 
Mercury
has
a
low
ionization
energy;
therefore,
free
electrons
and
cations
required
for
the
formation
of
a
gas
discharge
are
easily
created
using
a
relatively
low
start
voltage.

 
Mercury
reacts
minimally
with
the
lamp
envelope
and
electrode
materials.

The
wavelength
and
magnitude
of
light
output
from
a
mercury
discharge
depend
on
the
concentration
of
mercury
atoms,
which
is
directly
related
to
the
mercury
pressure.
At
low
pressures
of
0.001
to
0.01
torr
(
2
x
10­
5
to
2
x
10­
4
psi),
the
concentration
of
mercury
is
low,
and
the
distance
electrons
travel
between
collisions
is
relatively
long.
Electrons
achieve
higher
kinetic
energies
with
the
longer
travel
distance.
Collisions
between
those
free
electrons
and
mercury
atoms
excite
mercury
to
the
first
energy
state
above
the
lowest
or
ground
state.
Transition
of
electrons
back
to
ground
state
results
in
the
emission
of
electromagnetic
energy
at
253.7
and
185
nm.
UV
lamps
with
this
type
of
mercury
discharge
are
commonly
referred
to
as
low
pressure
(
LP)
lamps.

At
higher
mercury
pressures
(
100
to
10,000
torr;
2
to
20
psi),
a
much
greater
collision
frequency
occurs
between
free
electrons
and
mercury.
This
increases
the
energy
state
of
the
mercury
atoms
and
cations
to
near
that
of
the
electrons
and
increases
the
temperature
of
the
gas
discharge
to
near
6,000
°
C.
When
the
atoms
return
to
lower
energy
states,
electromagnetic
energy
at
several
wavelengths
in
the
UV
light
and
visible
light
regions
is
produced.
Recombination
of
free
electrons
and
mercury
cations
produces
a
small
continuum
of
UV
light
between
200
and
245
nm.
UV
lamps
with
this
type
of
discharge
are
called
medium
pressure
(
MP)
lamps.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
5
June
2003
A.
1.2
UV
Light
Propagation
This
section
details
the
effects
that
the
UV
reactor
and
the
water
being
treated
have
on
the
propagation
of
UV
light.
As
UV
light
propagates,
it
interacts
with
the
materials
it
encounters
through
absorption,
reflection,
refraction,
and
scattering.

A.
1.2.1
Absorption
Absorption
is
the
transformation
of
light
to
other
forms
of
energy
as
it
passes
through
a
substance.
UV
absorbance
is
the
water
quality
parameter
that
measures
the
extent
to
which
the
intensity
of
UV
light
is
reduced
as
it
passes
through
water.
The
impact
of
absorption
on
the
intensity
light
as
it
travels
through
a
substance
is
calculated
as
follows:

d
A
d
a
d
c
e
i
i
e
I
I
 
 
 
 
 
 
=
=
=
 
=
254
10
10
10
10
1
2
Equation
A.
2
where
I1
=
Light
intensity
incident
on
a
cell
(
mW/
cm2)
I2
=
Light
intensity
passing
through
a
distance,
d,
in
the
cell
containing
a
solution
with
various
absorbing
components
(
mW/
cm2)
d
=
Distance
traveled
by
light
through
the
cell
(
cm)
 i
=
Molar
absorption
coefficient
of
component
i
(
L/
mol/
cm)
ci
=
Concentration
of
component
i
(
mol/
L)
a10
=
Decadic
(
base
10)
absorption
coefficient,
(
cm­
1)
A254
=
Decadic
(
base
10)
absorbance
(
unitless)
 e
=
Naperian
(
base
e)
absorption
coefficient
(
cm­
1)

When
UV
light
is
absorbed,
it
is
no
longer
available
to
disinfect
microorganisms.

A.
1.2.2
Refraction
Refraction
(
Figure
A.
1)
is
the
change
in
the
direction
of
light
propagation
as
it
passes
from
one
medium
to
another.

Figure
A.
1
Refraction
of
Light
Incident
Light
Medium
1
Medium
2
Refracted
Light
 1
 2
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
6
June
2003
Refraction
is
governed
by
Snell's
Law,
which
is
shown
in
Equation
A.
3:

2
2
1
1
sin
sin
 
 
n
n
=
Equation
A.
3
where
n1
=
Index
of
refraction
of
the
first
media
n2
=
Index
of
refraction
of
the
second
media
 1
=
Incident
angle
on
the
interface
 2
=
Exit
angle
from
the
interface
In
UV
reactors,
refraction
occurs
when
light
passes
from
the
lamp
through
an
air
gap,
through
the
lamp
sleeve,
and
into
the
water.
Although
refracted
light
is
still
available
for
disinfection,
refraction
changes
the
angle
that
the
light
strikes
target
pathogens.

A.
1.2.3
Reflection
Reflection
is
the
change
in
the
direction
of
light
propagation
when
it
is
deflected
by
the
interface
between
two
media
(
Figure
A.
2).
Reflection
may
be
classified
as
specular
or
diffuse.
Specular
reflection
occurs
at
smooth
polished
surfaces
where
the
roughness
of
the
surface
is
smaller
than
the
wavelength
of
light.
Reflection
from
specular
surfaces
follows
the
Law
of
Reflection,
which
states
that
the
angle
of
incidence
is
equal
to
the
angle
of
reflection.
Diffuse
reflection
occurs
at
rough
surfaces.
Light
is
scattered
in
all
directions
with
little
dependence
on
the
incident
angle.
The
intensity
of
diffuse
reflected
light
is
proportional
to
the
cosine
of
the
reflectance
angle.
Reflected
light
is
still
available
for
disinfection.

Figure
A.
2
Specular
and
Diffuse
Reflection
of
Light
Specular
Reflection
Reflected
Light
Incident
Light
Diffuse
Reflection
Incident
Light
 1
 1
Reflected
Light
In
a
UV
reactor,
reflection
will
take
place
at
UV­
transmitting
interfaces
like
an
air­
quartz
interface
and
at
also
interfaces
that
do
not
transmit
UV
light
like
the
reactor
wall.
The
intensity
of
reflected
light
from
a
UV­
transmitting
interface
is
governed
by
Fresnel's
Law,
which
is
shown
in
Equation
A.
4.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
7
June
2003
 
 

 
 

 
 

 
 

 

  
 

 

  
 
+
 
+

 

  
 

 

  
 
+
 
=
2
1
2
2
1
2
1
1
2
2
2
2
1
1
2
2
1
1
cos
cos
cos
cos
cos
cos
cos
cos
2
1
 
 
 
 
 
 
 
 
n
n
n
n
n
n
n
n
R
Equation
A.
4
where
R
=
the
ratio
of
reflected
intensity
to
incident
intensity
n1
=
Index
of
refraction
of
the
first
media
n2
=
Index
of
refraction
of
the
second
media
 1
=
Incident
angle
onto
the
interface
 2
=
Reflected
angle
from
the
interface
The
intensity
of
reflected
light
from
a
non­
transmitting
interface
depends
on
the
material,
incident
angle,
wavelength
of
light,
and
nature
of
the
surface.
Currently,
most
UV
reactors
are
constructed
of
stainless
steel,
which
reflects
24
percent
of
UV
light
at
254
nm
at
a
zero
degree
incident
angle
(
Jagger
1967).
This
indicates
that
76
percent
of
the
light
energy
reaching
the
reactor
wall
is
lost.
In
the
future,
UV
reactors
may
be
developed
using
materials
that
reflect
more
light,
which
may
improve
efficiency.

A.
1.2.4
Scattering
Scattering
of
light
is
the
change
in
direction
of
light
propagation
caused
by
interaction
with
a
particle
(
Figure
A.
3).
Scattered
light
is
still
capable
of
disinfecting
microorganisms.

Figure
A.
3
Scattering
of
Light
Incident
Light
Forward
Scattered
Light
90
°
Scattered
Light
Back
Scattered
Light
Rayleigh
scattering
is
the
scattering
of
light
by
particles
that
are
smaller
than
the
wavelength
of
the
light.
With
Rayleigh
scattering,
light
is
scattered
uniformly
in
all
directions
at
an
intensity
inversely
proportional
to
the
wavelength
of
light
to
the
fourth
power
(
1/
 4).
As
such,
scattering
is
more
evident
at
shorter
wavelengths.
For
example,
the
intensity
of
scattered
light
at
200
nm
is
five
times
greater
than
at
300
nm
because
1/(
2004)
is
over
five
times
greater
than
1/(
3004).
Particles
in
water
that
cause
Rayleigh
scattering
of
UV
light
at
254
nm
include
small
viruses
and
large
molecules
(
25
to
300
nm).
With
larger
particles,
the
scattering
observed
is
nonuniform
and
more
light
is
scattered
in
the
forward
direction.
The
larger
particles
also
cause
backscattering,
which
is
nearly
independent
of
the
wavelength
of
light.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
8
June
2003
A.
1.2.5
UV
Absorbance
and
UV
Transmittance
UV
absorbance
(
A254)
is
a
commonly
used
water
quality
parameter
that
characterizes
the
decrease
in
the
amount
of
incident
light
as
it
passes
through
a
water
sample
over
a
specified
distance
or
pathlength.
If
the
measurement
is
made
according
to
a
modified
version
of
Standard
Method
5910B
(
APHA
et
al.
1998)
where
the
water
sample
is
not
filtered
or
pH
adjusted,
the
modified
measurement
accounts
for
scattering
and
some
absorption
from
particles
in
the
water
sample
that
may
interfere
with
UV
disinfection.
Although
the
Standard
Method
identifies
this
measurement
as
UV
absorption,
this
manual
will
refer
to
it
as
UV
absorbance
since
the
latter
term
is
widely
used
in
the
water
treatment
industry.

The
term
UV
transmittance
(
UVT)
has
also
been
used
extensively
in
the
literature
when
describing
the
behavior
of
UV
light.
UVT
is
the
percentage
of
light
passing
through
a
water
sample
over
a
specified
distance
and
is
related
to
UV
absorbance
by
Equation
A.
5:

254
10
100
%
A
UVT
 
 
=
Equation
A.
5
where
UVT
=
UV
Transmittance
at
specified
wavelength
(
e.
g.,
254
nm)
and
pathlength
(
e.
g.,
1
cm)
A254
=
UV
Absorbance
at
specified
wavelength,
based
on
1
cm
pathlength
(
unitless;
UV
absorption
as
measured
by
Standard
Method
5910B)

Since
UV
light
scattered
from
particles
is
capable
of
disinfecting
microorganisms,
it
should
be
considered
when
assessing
UVT.
Much
of
the
scattered
light
is
in
the
forward
direction
and
is
a
significant
portion
of
the
transmitted
UV
light.
Typically,
conventional
spectrophotometers
use
narrow
beams
of
light
and
small
detectors
that
will
not
measure
the
forward
scattered
light
and
therefore
underestimate
the
effective
UVT
of
the
water
sample
(
Jagger
et
al.
1975;
Linden
and
Darby
1998).
However,
spectrophotometers
can
be
equipped
with
integrating
spheres
(
Linden
and
Darby
1998)
or
detectors
capable
of
measuring
forward
scattered
light
(
Jagger
et
al.
1975)
in
order
to
provide
a
proper
assessment
of
the
UVT
of
water
samples
with
significant
scattering.

A.
1.2.6
Estimating
UV
Light
Intensity
Within
a
UV
Reactor
The
distribution
of
light
intensity
about
a
UV
lamp
is
influenced
by
the
shape
of
the
lamp
and
the
absorption,
refraction,
scattering,
and
reflection
of
light.
Complex
models
factoring
all
of
these
effects
can
be
used
to
determine
the
intensity
profile
about
a
lamp,
and
simplified
models
can
be
used
to
approximate
those
profiles.
These
models
are
useful
tools
for
understanding
the
impact
of
UV
absorbance,
UV
reactor
properties,
and
UV
reactor
dimensions
on
UV
dose
delivery
and
measurements
of
UV
intensity.

If
the
distance
from
the
lamp
is
greater
than
the
radius
of
the
arc
discharge,
the
lamp
can
be
treated
as
a
line
source
to
estimate
the
intensity.
For
LP
lamps,
since
the
arc
discharge
fills
the
entire
lamp,
the
radius
is
the
same
as
the
lamp
radius.
For
MP
lamps,
the
arc
discharge
is
much
smaller
than
the
radius
of
the
lamp.
There
are
two
approaches
commonly
used
for
modeling
a
line
source:
the
radial
model
and
the
point
source
summation
model.
If
the
distance
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
9
June
2003
from
the
lamp
is
smaller
than
the
radius
of
the
gas
discharge,
more
complex
modeling
tools
must
be
used.

The
radial
model
provides
a
two­
dimensional
representation
of
a
three­
dimensional
intensity
profile.
The
model
assumes
light
is
emitted
perpendicular
from
the
line
source
in
the
radial
direction
as
per
Equation
A.
6:

I
r
P
r
e
L
r
e
(
)
=
 

2
 
 
Equation
A.
6
where
PL
=
UV
power
emitted
per
unit
arc
length
of
the
line
source
(
mW/
cm)
r
=
Radial
distance
from
the
line
source
(
cm)
 e
=
Naperian
(
base
e)
absorption
coefficient
of
the
media
(
cm­
1)
I(
r)
=
UV
intensity
(
mW/
cm2)
at
a
distance
r
from
the
line
source
The
Point
Source
Summation
model
(
Jacob
and
Dranoff
1970)
treats
the
lamp
as
a
series
of
point
sources
radiating
uniformly
in
all
directions.
The
UV
intensity
at
a
point
within
the
reactor
is
the
sum
contribution
from
each
of
these
points
as
per
Equation
A.
7.

(
)
I
r
z
P
e
p
i
e
i
,
=
 
 
4
2
 
 
 
 
Equation
A.
7
where
Pp
=
UV
power
emitted
by
each
point
source
(
mW)
i
=
Number
of
point
sources
used
to
simulate
the
lamp
 i
=
Distance
from
the
ith
point
source
(
cm)
 e
=
Naperian
(
base
e)
absorption
coefficient
of
the
media
(
cm­
1)
r
=
Radial
distance
from
the
lamp
(
cm)
z
=
Axial
distance
along
the
lamp
(
cm)
I(
r,
z)
=
UV
intensity
(
mW/
cm2)
at
a
coordinate
position
(
r,
z)

The
radial
and
axial
distance
from
the
lamp
are
shown
in
Figure
A.
4.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
10
June
2003
Figure
A.
4
Radial
and
Axial
Distance
from
a
UV
Lamp
UV
Lamp
z
(
axial
distance)

r
(
radial
distance)
 
(
r,
z)

Note:
The
point
where
z
=
0
is
arbitrary.
It
can
be
at
the
lamp
ends
or
anywhere
along
the
lamp
length.

For
a
25
cm
long
UV
lamp
housed
in
a
lamp
sleeve
(
radius
=
4
cm)
and
immersed
in
water,
Figure
A.
5
presents
the
intensity
profile
predicted
using
Point
Source
Summation
as
a
function
of
radial
and
axial
distance
and
the
water
UV
absorbance.
For
a
given
radial
distance,
the
model
predicts
a
greater
UV
intensity
at
an
axial
position
corresponding
to
the
center
of
the
lamp
than
at
an
axial
distance
corresponding
to
the
lamp
ends.
The
model
also
demonstrates
that
UV
intensity
will
decrease
with
increased
distance
from
the
lamp
even
in
water
that
does
not
absorb
UV
light
(
i.
e.,
A254
=
0)
due
to
the
divergence
of
UV
light
from
the
source.
Last,
the
model
predictions
show
that
the
water
UV
absorbance
has
a
profound
impact
on
the
UV
intensity
profile
about
a
UV
source.

Figure
A.
5
UV
Intensity
Profile
of
a
25
cm
Medium­
Pressure
Mercury
Arc
Lamp
as
a
Function
of
(
a)
Radial
and
(
b)
Axial
Position
for
Waters
with
Different
UV
Absorbance
Y­
axis
=
UV
Intensity
10
cm
Radial
Distance
From
Lamp
mW/
cm2
Lamp
is
25
cm
long,
centered
at
x
=
0
0
50
100
150
­
30
­
20
­
10
0
10
20
30
Axial
Distance
along
Lamp
(
cm)
b.
0
100
200
300
400
500
0
5
10
15
20
Radial
Distance
from
Lamp
Center
(
cm)
UV
Intensity
(
mW/
cm2)

A254
=
0
cm­
1,
UVT
=
100
%

A254
=
0.046
cm­
1,
UVT
=
90
%

A254
=
0.097
cm­
1,
UVT
=
80
%

a.

More
advanced
models
of
the
intensity
profile
about
a
lamp
account
for
the
impacts
of
refraction
and
reflection
from
reactor
components
as
the
light
propagates
from
the
discharge
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
11
June
2003
(
Bolton
2000),
the
three­
dimensional
nature
of
the
gas
discharge
(
Irazoqui
et
al.
1973),
and
the
direction
of
light
emission
(
Phillips
1983).

A.
2
Microbial
Response
to
UV
Light
Disinfection
by
UV
light
differs
from
chemical
disinfectants
such
as
chlorine
and
ozone.
Chemical
disinfectants
inactivate
microorganisms
by
destroying
or
damaging
cellular
structures,
thereby
interfering
with
metabolism,
biosynthesis,
and
growth
(
Snowball
and
Hornsey
1988).
In
UV
disinfection,
microorganisms
are
inactivated
by
inducing
damage
to
their
nucleic
acid
such
that
they
can
no
longer
reproduce.
This
section
discusses
nucleic
acid
structure,
the
damage
that
causes
microbial
inactivation,
the
ability
of
microorganisms
to
repair
the
damage,
methods
for
determining
microbial
inactivation,
and
factors
that
affect
inactivation.

A.
2.1
DNA/
RNA
Structure
Nucleic
acid
is
a
fundamental
building
block
of
life
and
is
responsible
for
reproduction
and
defining
the
nature
of
life.
The
nucleic
acid
is
either
deoxyribonucleic
acid
(
DNA)
or
ribonucleic
acid
(
RNA).
The
nucleic
acid
within
the
nucleus
of
most
cells,
including
bacteria
and
protozoa,
is
composed
of
double
stranded
DNA.
DNA
contains
the
information
necessary
for
the
synthesis
of
ribosomal,
transfer,
and
messenger
RNA,
which
are
responsible
for
synthesizing
enzymes
that
drive
metabolic
processes
within
the
cell.
The
genetic
material
of
viruses
may
either
be
DNA
or
RNA
and
can
be
single
or
double
stranded.

DNA
and
RNA
are
long
polymers
comprised
of
combinations
of
four
nucleotides.
In
DNA,
the
nucleotides
are
purines
(
adenine
and
guanine)
and
pyrimidines
(
thymine
and
cytosine).
In
RNA,
the
nucleotides
are
the
same
except
that
uracil
replaces
thymine.
Each
nucleotide
can
be
broken
down
into
two
parts
 
a
sugar­
phosphate
backbone
and
a
nitrogenous
base
(
Figure
A.
6).
If
the
nucleic
acid
is
double­
stranded,
nucleotides
on
one
strand
will
compliment
those
on
the
other
strand.
Adenine
pairs
with
thymine
in
DNA
(
or
uracil
in
RNA)
while
guanine
pairs
with
cytosine.
Hydrogen
bonds
form
between
each
pair
(
Figure
A.
6).

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
12
June
2003
Figure
A.
6
Structure
of
DNA
and
Nucleotide
Sequences
Within
DNA
DNA
STRUCTURE
Sugar­
Phosphate
Backbone
DNA
SEQUENCE
A
=
Adenine
C
=
Cytosine
T
=
Thymine
G
=
Guanine
 
A
 
T
 
G
 
C
 
G
 
A
 
T
 
C
 
 
T
 
A
 
C
 
G
 
C
 
T
 
A
 
G
 
Hydrogen
Bonded
Nitrogenous
Base
Pairs
(
A,
T,
G,
C)

Purines
Pyrimidines
|
|
|
|
|
|
|
|

A.
2.2
Mechanism
of
Inactivation
by
UV
Light
UV
light
inactivates
microorganisms
by
damaging
DNA
or
RNA,
thereby
interfering
with
replication
of
the
microorganism.
Only
light
that
is
absorbed
by
a
system
can
induce
a
chemical
reaction
(
First
Law
of
Photochemistry).
As
shown
in
Figure
A.
7,
nucleotides
absorb
UV
light
in
from
200
to
300
nm,
which
enables
the
photobiological
effects
that
lead
to
nucleic
acid
damage.
The
UV
absorption
of
nucleic
acid
is
a
combination
of
the
absorbance
of
the
nucleotides
and
has
an
absorption
peak
near
260
nm
and
a
local
minimum
near
230
nm.

Figure
A.
7
UV
Absorbance
of
Nucleotides
and
Nucleic
Acid
at
pH
7
(
adapted
from
Jagger
1967)

200
260
300
Wavelength
(
nm)
UV
absorbance
(
Relative
Scale)

Adenine
Guanine
Thymine
Cytosine
220
240
280
0.0
0.2
0.4
0.6
0.8
1.0
0.0
0.2
0.4
0.6
0.8
1.0
DNA
200
260
300
Wavelength
(
nm)
220
240
280
UV
absorbance
(
Relative
Scale)

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
13
June
2003
While
both
purines
and
pyrimidines
strongly
absorb
UV
light,
the
rate
of
UV­
induced
damage
is
greater
with
pyrimidines
(
Jagger
1967).
Absorbed
UV
light
induces
six
types
of
damage
within
the
pyrimidines
of
nucleic
acid
(
Setlow
1967;
Snowball
and
Hornsey
1988;
Pfeifer
1997),
with
varying
levels
of
effectiveness
dependent
on
UV
dose:

 
Single
and
double
strand
breaks
are
only
significant
with
UV
doses
several
orders
of
magnitude
higher
than
those
practical
for
UV
disinfection.

 
DNA­
DNA
cross­
links
are
covalent
bonds
between
two
different
strands
of
DNA,
and
they
are
also
only
significant
with
UV
doses
orders
of
magnitude
higher
than
those
practical
for
UV
disinfection.

 
Protein­
DNA
cross­
links
are
covalent
bonds
between
a
protein
and
a
DNA
strand,
and
they
may
be
important
for
the
disinfection
of
certain
microorganisms
such
as
Micrococcus
radiodurans.

 
Pyrimidine
hydrates
do
not
contribute
to
UV
disinfection.

 
Pyrimidine
(
6
 
4)
pyrimidine
photoproducts
are
a
major
class
of
UV
damage.

 
Pyrimidine
dimers
are
covalent
bonds
between
two
pyrimidines
on
the
same
DNA
strand,
and
they
are
the
most
common
damage
resulting
from
UV
disinfection.

While
it
is
possible
for
thymine­
thymine,
cyctosine­
cytosine,
and
thymine­
cytosine
dimers
to
form
within
DNA,
thymine­
thymine
dimers
are
the
most
common.
However,
since
thymine
is
not
present
in
RNA,
uracil­
uracil
and
cytosine­
cytosine
dimers
are
formed.
Microorganisms
with
DNA
rich
in
the
thymine
tend
to
be
more
sensitive
to
UV
disinfection
(
Adler
1966).

Dimers
cause
faults
in
the
transcription
of
information
from
DNA
to
RNA,
which
in
turn
results
in
disruption
of
cell
metabolism.
However,
damage
to
nucleic
acid
does
not
prevent
the
cell
from
undergoing
metabolism
and
other
cell
functions.
As
discussed
in
the
next
section,
enzyme
mechanisms
within
the
cell
are
capable
of
repairing
some
of
the
damage
to
the
nucleic
acid.
To
directly
damage
the
internal
structure
of
the
cell,
UV
doses
much
higher
than
those
required
for
inactivation
are
necessary
(
Brandt
and
Giese
1956).

A.
2.3
Repair
Mechanisms
Because
microorganisms
that
have
been
exposed
to
UV
light
still
retain
metabolic
functions,
some
are
able
to
repair
the
damage
done
by
UV
light
and
regain
infectivity.
Repair
of
UV
light­
induced
DNA
damage
includes
photoreactivation
and
dark
repair
(
Knudson
1985).
At
the
doses
typically
used
in
UV
disinfection,
microbial
repair
can
be
controlled
and
accounted
for
as
discussed
in
section
3.1.1.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
14
June
2003
A.
2.3.1
Photoreactivation
Photreactivation
or
photorepair
is
the
cleaving
of
pyrimidine
dimers
by
the
enzyme
DNA
photolyase
(
Setlow
1967).
The
repair
mechanism
is
termed
photorepair
because
exposure
of
the
enzyme
to
light
between
310
and
490
nm
is
needed
to
activate
the
enzyme
and
provide
it
with
the
energy
necessary
to
split
the
paired
dimers.

Figure
A.
8
shows
the
difference
in
UV
dose
necessary
to
achieve
a
certain
log
inactivation
with
and
without
considering
photoreactivation
for
two
organisms.
Photorepair
varies
with
different
microorganism
types,
different
species,
and
different
strains
of
a
given
species.
The
extent
of
photorepair
depends
on
many
factors,
including
type
of
microorganism,
degree
of
inactivation,
time
between
exposure
to
UV
light
and
photoreactivating
light,
and
the
nutrient
state
of
the
microorganism.

Figure
A.
8
Repair
of
L.
Pneumophila
and
E.
Coli
(
adapted
from
Knudson
1985)

0
1
2
3
4
5
6
7
0
10
20
30
40
UV
Dose
(
mJ/
cm2)
Log
Inactivation
L.
pneumophila
­
No
Photorepair
L.
pneumophila
­
Photorepair
E.
coli
­
Photorepair
E.
coli
­
No
Photorepair
Photoreactivation
increased
the
UV
dose
necessary
to
achieve
3­
log
inactivation
of
seven
Legionella
species
between
1.1
and
6.3
fold
(
Knudson
1985).
Photoreactivation
also
increased
the
dose
necessary
for
4­
log
inactivation
of
twelve
species
of
bacteria
by
1.2
to
3.5
fold
(
Hoyer
1998).
However,
Shin
et
al.
(
2001)
did
not
observe
photorepair
with
Cryptosporidium
parvum.

Although
viral
DNA
does
not
have
the
necessary
enzymes
for
repair,
the
photorepair
of
viral
DNA
can
occur
using
the
enzymes
of
their
host
cells.
Lytle
(
1971)
reported
that
the
photorepair
of
Herpes
simplex
virus
by
mammalian
cells
varies
significantly,
depending
on
the
host
cell.
RNA
viruses
lack
the
ability
to
photorepair
in
a
host
cell
(
Rauth
1965).

Kelner
(
1950)
reported
that
the
ratio
of
UV
dose
necessary
to
achieve
a
certain
log
inactivation
with
and
without
considering
photorepair
is
independent
of
the
degree
of
inactivation.
However,
more
recent
research
by
Lindenauer
and
Darby
(
1994)
reported
that
the
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
15
June
2003
effect
of
photorepair
of
coliform
bacteria
in
wastewater
becomes
less
pronounced
as
UV
dose
increases.
Knudsen
(
1985)
also
found
a
slight
reduction
in
the
ability
of
Legionella
in
wastewater
to
repair
after
higher
inactivation
levels.

The
time
between
UV
light
exposure
and
exposure
to
photoreactivating
light
has
a
signific
of
a
he
rate
of
photorepair
is
constant
with
time
until
it
reaches
saturation,
where
saturation
is
defin
E.

was
sm.

he
nutrient
state
of
the
microorganism
also
impacts
the
ability
to
photorepair.
Giese
et
al.
(
195
.2.3.2
Dark
Repair
ark
repair
is
when
repair
processes
do
not
need
reactivating
light.
The
term
is
somew
e
does
ir,

.
Repair
endonnuclease
enzyme
recognizes
the
DNA
damage
and
cleaves
the
DNA
2.
xonuclease
enzyme
excises
the
damaged
section.

3.
NA
polymerase
rebuilds
the
removed
section
using
the
complementary
strand
as
a
4.
olynucleotide
ligase
rejoins
the
severed
strand.

One
study
(
Knudsen
1985)
examined
two
different
strains
of
E.
coli:
one
that
has
the
enzyme
cA­

the
ant
effect
on
the
ability
to
photoreactivate.
Dulbecco
(
1950)
reported
that
the
ability
T2
phage
to
repair
using
E.
coli
as
a
host
organism
decreases
as
the
time
between
exposure
to
UV
light
and
photoreactivating
light
increases.
Kelner
(
1950)
reported
that
E.
coli
at
37
°
C
in
nutrient
broth
lost
the
ability
to
photorepair
after
140
minutes
in
the
dark
after
exposure
to
UV
light
(
the
same
time
the
survivors
took
to
attempt
cell
division).
In
the
same
study,
cells
kept
at
colder
temperatures
maintained
their
ability
to
photorepair
for
several
hours
longer.

T
ed
as
the
maximum
amount
of
repair
possible
by
the
microorganism
given
its
repair
ability
and
the
extent
of
damage.
Kashimada
et
al.
(
1996)
reported
photorepair
saturation
of
coli
occurs
after
2
hours
of
exposure
under
fluorescent
lighting.
With
exposure
to
sunlight,
however,
they
reported
photorepair
saturation
after
15
minutes
followed
by
inactivation
that
attributed
to
the
UV
component
of
sunlight.
The
rate
of
repair
increases
with
temperature
(
Kelner
1950)
but
is
nearly
independent
of
the
reactivating
light
intensity
(
Setlow
1967),
suggesting
photorepair
is
rate
limited
by
the
enzyme
concentration
within
the
microorgani
T4
)
reported
that
a
starved
strain
of
paramecium,
Colpidium
colpoda,
needed
more
reactivating
light
to
reach
saturation
than
organisms
with
sufficient
nutrients.

A
D
hat
misleading
because
dark
repair
can
occur
in
the
presence
of
light
and
therefor
not
need
dark
conditions.
The
forms
of
dark
repair
include
excision
repair,
recombinational
repair,
and
inducible
error
prone
repair.
Excision
repair,
the
most
common
form
of
dark
repa
is
an
enzyme­
mediated
process
involving
four
steps:

1
strand.

E
D
template.

P
s
necessary
for
repair
(
B/
R
strain)
and
one
that
lacks
the
necessary
repair
enzymes
(
re
uvr­
strain).
The
difference
in
UV
dose
needed
for
1­
log
inactivation
of
the
strain
capable
of
repair
was
over
two
orders
of
magnitude
higher
than
the
dose
needed
for
1­
log
inactivation
of
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
16
June
2003
repair
deficient
strain,
indicating
that
dark
repair
impacts
the
UV
dose­
response
of
microorganisms.

Based
on
the
difference
in
UV
sensitivity
of
repair
proficient
and
deficient
bacteria,
Jagger
er,

zymes
A.
2.4
V
Dose
and
Dose
Distribution
V
dose
is
a
measurement
of
the
amount
of
the
energy
per
unit
area
that
is
incident
on
a
surface
.2.4.1
Calculation
of
UV
Dose
in
Bench­
Scale,
Batch
Systems
he
most
carefully
controlled
method
of
determining
UV
dose
is
in
a
batch
system
with
a
bench­
s
is
he
general
definition
of
UV
dose
is
the
product
of
UV
intensity
multiplied
by
the
exposu
Equation
A.
8
here
=
UV
intensity
(
mW/
cm2)

If
intensity
is
not
constant
with
respect
to
time,
the
integral
of
the
intensity
output
over
the
exp
dt
I
Dose
UV
0
Equation
A.
9
where
s
are
defined
as
in
Equation
A.
8.
(
1967)
discovered
that
roughly
99
percent
of
repair
that
occurs
is
dark
repair.
Howev
the
rate
at
which
dark
repair
occurs
is
unknown.
It
is
possible
that
microorganisms
have
dark
repaired
prior
to
the
microbial
assay,
and
dark
repair
is
not
detected.
Therefore,
the
effects
of
dark
repair
can
be
difficult
to
measure.
Unlike
bacteria,
viruses
do
not
have
the
enzymes
necessary
for
dark
repair.
However,
virus
can
repair
in
the
host
cell
using
the
host
cells'
en
(
Rauth
1965).

U
U
.
UV
dose
is
the
product
of
the
average
intensity
acting
on
a
microorganism
from
all
directions
and
the
exposure
time.
Units
commonly
used
for
UV
dose
are
J/
m2,
mJ/
cm2,
and
mWs/
cm2
(
10
J/
m2
=
1
mJ/
cm2
=
1
mWs/
cm2)
with
mJ/
cm2
being
the
most
common
units
in
North
America
and
J/
m2
being
the
most
common
in
Europe.
This
section
discusses
how
UV
dose
is
calculated
in
bench­
scale,
batch
systems
and
also
how
the
UV
dose
distribution
is
determined
in
continuous
flow
pilot­
or
full­
scale
UV
reactors.

A
Tc
ale
collimated
beam
apparatus.
Appendix
E
presents
procedures
for
collimated
beam
testing.
The
factors
impacting
UV
dose
calculation
in
collimated
beam
tests
are
described
in
th
section.

T
re
time.

t
I
Dose
UV
 
=

w
I
t
=
Exposure
time
(
s)

osure
time
should
be
used
in
place
of
intensity
as
in
Equation
A.
9.

t 
 
=

variable
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
17
June
2003
Due
to
several
conditions
present
in
collimated
beam
testing,
the
intensity
measured
by
the
radiometer
does
not
accurately
represent
the
intensity
of
light
that
reaches
the
target
organisms.
To
get
an
accurate
calculation
of
the
UV
dose
delivered
to
the
microorganisms,
the
following
factors
are
considered
as
shown
in
Equation
A.
10
(
Bolton
and
Linden
2003):
absorbance/
transmittance
of
the
water,
thickness
of
the
water
layer,
distribution
of
light
across
the
surface
of
the
suspension,
and
reflection
and
refraction
of
light
from
the
water.

t
d
a
R
P
I
t
I
Dose
UV
d
a
f
avg
 
 
 
 
=
 
=
 

10
ln
)
10
1
)(
1
(

10
0
254
10
Equation
A.
10
where
Iavg
=
Average
intensity
within
the
suspension
(
mW/
cm2)
t
=
Exposure
time
(
s)
I0
=
Intensity
measured
at
the
suspension's
surface
(
mW/
cm2)
R
=
Fraction
of
light
reflected
at
the
suspension's
surface
(
from
Fresnel's
Law)
a10
=
Decadic
(
base
10)
absorption
coefficient
of
the
suspension,
A254
(
cm­
1)
d
=
Thickness
of
water
layer
(
cm)
Pf
=
Petri
factor,
ratio
of
measured
intensity
at
the
center
of
the
exposure
dish
to
average
intensity
across
the
surface
area
of
the
exposure
dish
(
unitless)

Because
microorganisms
respond
differently
to
different
wavelengths
of
light,
if
a
polychromatic
light
source
(
e.
g.,
MP
lamp)
is
used,
it
is
also
critical
to
incorporate
the
light
intensity
and
the
inactivation
effectiveness
of
each
wavelength
in
the
germicidal
range
when
determining
UV
dose.
For
microorganisms
that
exhibit
inactivation
kinetics
that
are
independent
of
wavelength,
the
equivalent
dose
at
254
nm
from
a
polychromatic
source
is
calculated
as
follows
(
Meulemans
1986):

t
G
I
D
 
=

 =

300
200
254
)
(
)
(
 
 
 
Equation
A.
11
where
D254
=
UV
dose
equivalent
at
254
nm
 
=
Wavelength
of
light
(
nm)
I(
 )
=
Intensity
at
wavelength
 
over
1
nm
increments
G(
 )
=
Relative
action
spectrum
of
the
microorganism
defined
as
k /
k254
k 
=
First
order
inactivation
constant
at
wavelength
 
k254
=
First
order
inactivation
constant
at
254
nm
wavelength
t
=
Exposure
time
(
s)

However,
if
the
microorganism
does
not
exhibit
the
same
inactivation
kinetics
at
each
wavelength,
the
dose­
response
curve
may
be
characterized
by
a
shoulder
(
section
A.
2.5.2),
and
the
dose
equivalent
at
254
nm
is
calculated
using
Equation
A.
12
(
Cabaj
et
al.
2001):

 =
 
 
 
 
=
 
 
=
300
200
0
)
1
(
1
)
1
(
1
254
254
254
 
 
 
 
d
D
k
d
D
k
e
e
N
N
Equation
A.
12
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
18
June
2003
where
D 
=
Dose
delivered
at
wavelength
 
k 
=
First
order
inactivation
constant
at
wavelength
 
d 
=
Intercept
of
the
exponential
region
with
the
y­
axis
at
wavelength
 
A.
2.4.2
Dose
Distribution
in
Continuous
Flow
UV
Reactors
Determining
the
UV
dose
in
a
continuous
flow
pilot­
or
full­
scale
UV
reactor
is
complicated
by
hydrodynamics
(
particle
paths)
and
variations
in
UV
intensity
throughout
the
reactor.

In
an
ideal
reactor
that
has
plug­
flow
conditions
and
complete
mixing
perpendicular
to
the
flow,
all
microorganisms
entering
the
reactor
will
receive
the
same
UV
dose,
which
is
calculated
according
to
Equation
A.
13.

Q
V
I
t
I
Dose
UV
avg
r
avg
=
=
Equation
A.
13
where
Iavg
=
Volume­
averaged
UV
intensity
within
the
reactor
(
mW/
cm2)
tr
=
Theoretical
residence
time
of
the
reactor
(
s)
V
=
Volume
of
water
within
the
reactor
(
gal)
Q
=
Flowrate
through
the
reactor
(
gal/
s)

Equation
A.
13
calculates
the
maximum
UV
dose
possible
in
an
ideal
reactor.
However
in
practice,
microorganisms
take
different
paths
through
a
reactor
and
thus
do
not
all
receive
the
maximum
dose.
Instead,
the
UV
dose
delivered
to
the
organisms
is
best
described
using
a
dose
distribution
(
Figure
A.
9).
A
dose
distribution
is
a
curve
or
histogram
that
indicates
the
probability
of
a
microorganism
receiving
a
certain
dose
as
it
travels
through
the
UV
reactor.

Figure
A.
9.
Hypothetical
Dose
Distribution
Delivered
by
a
UV
Reactor
0
0.05
0.10
0.15
0.25
UV
Dose
(
mJ/
cm2)
Occurrence
Probability
0.20
0
10
20
30
40
50
60
70
80
90
100
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
19
June
2003
The
width
of
the
dose
distribution
is
an
indication
of
the
hydraulic
conditions
in
the
reactor
luge
and
he
dose
distribution
of
a
UV
reactor
cannot
be
measured
in
a
practical
manner
with
current
a
activation
achieved
by
a
reactor
with
a
modeled
dose
distribution
can
be
calculated
by
summi
.
The
more
narrow
the
distribution,
the
better
the
hydraulic
conditions
approximate
p
flow
with
complete
mixing.
However,
a
narrow
dose
distribution
does
not
always
imply
efficient
dose
delivery.
An
annular
reactor
with
a
thin
water
layer
between
the
lamp
sleev
the
reactor
wall
will
deliver
a
narrow
dose
distribution.
However,
if
the
reactor
wall
absorbs
UV
light,
energy
losses
at
the
wall
will
be
excessive
and
the
reactor
will
not
efficiently
utilize
the
UV
output
of
the
lamp.
The
most
cost
effective
design
of
a
UV
reactor
will
have
a
dose
distribution
that
reflects
a
compromise
between
inefficiency
due
to
energy
losses
at
the
reactor
wall
and
by
adjacent
lamps
as
well
as
inefficiency
due
to
hydrodynamics.

T
technology.
However,
by
predicting
particle
trajectories
through
the
intensity
field
of
UV
reactor
using
computational
fluid
dynamics
(
CFD),
dose
distributions
can
be
calculated
(
Wright
and
Lawryshyn
2000).

In
ng
the
inactivation
achieved
by
each
dose
in
the
dose
distribution
according
to
Equation
A.
14.

(
)
(
)
N
N
p
D
f
D
i
i
i
0
=
 
Equation
A.
14
here
=
Probability
of
dose
Di
occurring
microorganism
inactivation
as
a
function
of
sing
the
inactivation
kinetics
of
the
microorganism,
the
inactivation
is
related
to
a
single
dose
va
wp
i(
Di)
f(
Di)
=
Mathematical
function
describing
dose
U
l
ue
termed
the
reduction
equivalent
dose
(
RED)
by
Equation
A.
15.

)
(
)
(
)
(
0
RED
f
D
f
D
p
N
N
i
i
i
=
=
 
Equation
A.
15
here
=
Concentration
of
organisms
after
exposure
to
UV
light
t
i)
=
ibing
inactivation
as
a
function
of
dose
valent
dose
w
N
N0
=
Concentration
of
organisms
before
exposure
to
UV
ligh
p(
D
Probability
of
Di
occurring
f(
Di)
=
Mathematical
function
descr
Di
=
UV
Dose
RED
=
Reduction
equi
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
20
June
2003
If
microorganism
inactivation
can
be
described
using
first
order
kinetics
(
section
A.
2.5.1),
inactivation
is
related
to
RED
by
Equation
A.
16.

(
)
N
N
p
D
e
e
i
i
kD
k
RED
i
0
=
=
 
 
 
Equation
A.
16
where
variables
are
defined
as
in
Equation
A.
15
and
k
=
First
order
inactivation
coefficient
By
re­
arranging
Equation
A.
16,
the
reduction
equivalent
dose
is
calculated
according
to
Equation
A.
17.

(
)
RED
k
p
D
e
i
i
kDi
=
 

 
 
 

 
 

 
 
1
ln
Equation
A.
17
where
variables
are
defined
as
in
Equation
A.
15
and
A.
16
Because
UV
reactors
do
not
exhibit
ideal
dose
delivery,
the
RED
of
a
reactor
delivering
a
dose
distribution
depends
on
the
UV
sensitivity
of
the
microorganisms
used
to
calculate
RED.
The
RED
determined
when
using
a
challenge
microorganism
that
is
more
resistant
to
UV
disinfection
will
be
higher
compared
to
when
using
a
less
resistant
microorganism.
In
contrast,
the
RED
of
an
ideal
reactor
has
the
same
value
for
all
microorganisms.
Also,
the
RED
of
a
reactor
delivering
a
dose
distribution
will
vary
in
a
non­
linear
fashion
with
the
lamp
power
and
a
flowrate
while
the
ideal
reactor
model
predicts
a
proportional
relationship.
Lastly,
the
dependence
of
RED
on
UV
absorbance
of
the
water
will
be
more
pronounced
with
a
reactor
delivering
a
dose
distribution
than
an
ideal
reactor.
The
RED
will
decrease
with
increased
UV
absorbance
at
a
greater
rate
with
the
reactor
with
a
dose
distribution
than
is
predicted
by
ideal
models.

The
inactivation
of
a
microorganism
and
the
associated
RED
are
measured
using
biodosimetry
(
described
in
section
4.2).

A.
2.5
Dose­
Response
Relationships
UV
dose­
response
relationships
can
be
expressed
as
either
the
proportion
of
microorganisms
inactivated
(
log
inactivation,
results
in
dose­
response
curves
with
positive
slope)
or
the
proportion
of
microorganisms
remaining
(
log
survival;
results
in
dose­
response
curves
with
negative
slope)
as
a
function
of
UV
dose.
The
proportion
of
microorganisms
remaining
and
the
log
inactivation
are
typically
shown
on
a
logarithmic
(
base
10)
scale,
while
the
UV
dose
is
typically
shown
on
a
linear
scale.
This
manual
will
present
microbial
response
as
log
inactivation
since
the
terminology
is
widely
accepted
in
the
industry.
Therefore,
all
doseresponse
curves
presented
will
have
a
positive
slope.
The
log
inactivation
of
the
microorganisms
is
determined
by
measuring
the
concentration
of
replicating
microorganisms
after
exposure
to
a
measured
UV
dose
and
is
calculated
according
to
Equation
A.
18.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
21
June
2003
N
N
on
Inactivati
Log
0
10
log
=
Equation
A.
18
where
N0
=
Concentration
of
viable
microorganisms
before
exposure
to
UV
light
N
=
Concentration
of
viable
microorganisms
after
exposure
to
UV
light
Many
UV
dose­
response
curves
for
disperse
microorganisms
follow
first
order
inactivation,
but
in
some
cases,
the
dose­
response
curves
take
other
shapes
such
as
shoulders
or
tailing.
A
shoulder
is
characterized
by
a
period
of
very
little
inactivation
at
lower
doses
followed
by
linear
or
exponential
inactivation.
A
dose­
response
curve
that
exhibits
tailing
is
characterized
by
a
decrease
in
the
inactivation
rate
after
a
certain
degree
of
inactivation
has
been
observed.
Figure
A.
10
shows
various
shapes
of
dose­
response
curves.
Note
that
the
terms
"
shoulder"
and
"
tailing"
refer
to
the
shape
the
curves
take
when
the
y­
axis
of
the
dose­
response
curve
is
presented
as
log
survival
with
negative
slopes,
which
is
the
inverse
of
log
inactivation.

Figure
A.
10.
UV
Dose­
Response
Curves
(
adapted
from
Chang
et
al.
1985)

0
1
2
3
4
5
6
0
20
40
60
80
100
UV
Dose
(
mJ/
cm2)
Log
Inactivation
E.
coli
B.
subtilis
spores
Total
coliform­
wastewater
Rotavirus
A.
2.5.1
First
Order
Response
The
E.
coli
data
shown
in
Figure
A.
10
exhibit
first
order
dose­
response
behavior.
The
equation
for
first
order
inactivation
is
shown
in
Equation
A.
19:

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
22
June
2003
10
10
0
D
D
kD
e
N
N
 
 
=
=
Equation
A.
19
where
N0
=
Concentration
of
viable
microorganisms
before
UV
exposure
N
=
Concentration
of
viable
microorganisms
after
UV
exposure
k
=
First
order
inactivation
coefficient
of
the
microorganisms
(
cm2/
mJ)
D
=
UV
dose
delivered
to
the
microorganisms
(
mJ/
cm2)
D10
=
UV
dose
needed
to
inactivate
microorganisms
by
one
log
(
i.
e.,
90
percent
inactivation)
(
mJ/
cm2)

In
first­
order
response,
only
one
photon
of
light
is
needed
to
inactivate
a
microorganism.

A.
2.5.2
Shoulders
The
B.
subtilis
data
shown
in
Figure
A.
10
exhibit
a
shoulder
followed
by
first
order
doseresponse
behavior.
The
shoulder
is
attributed
to
a
delayed
response
of
a
microorganism
when
exposed
to
UV
light.
Unlike
first
order
inactivation,
more
than
one
photon
of
light
is
needed
to
inactivate
a
microorganism.
Although
the
number
of
photons
can
not
be
measured
directly,
it
can
be
related
to
first
order
response
through
curve
fits
of
empirical
equations.
Equation
A.
20
(
Cabaj
et
al.
2001)
is
one
of
the
many
equations
derived
from
empirical
curve
fits
that
can
be
used
to
model
inactivation
curves
with
shoulders.

(
N
N
e
kD
d
0
1
1
=
 
 
 
)
Equation
A.
20
where
N0
=
Concentration
of
viable
microorganisms
before
UV
exposure
N
=
Concentration
of
viable
microorganisms
after
UV
exposure
k
=
First
order
inactivation
coefficient
of
the
microorganisms
(
cm2/
mJ)
D
=
UV
dose
delivered
to
the
microorganisms
(
mJ/
cm2)
d
=
Intercept
of
the
exponential
region
of
the
dose­
response
with
the
y­
axis
Morton
and
Haynes
(
1969)
reported
a
decrease
in
the
shoulder
with
nutrient­
depleted
E.
coli
and
proposed
that
the
shoulder
was
associated
with
dark
repair.
Photoreactivation
significantly
increased
the
shoulder
observed
with
E.
coli
(
Hoyer
1998)
and
Legionella
(
Knudson
1985).

Note
that
the
equation
presented
is
only
one
of
the
many
equations
derived
from
empirical
curve
fits.
There
are
many
methods
to
model
UV
dose­
response
data
not
presented
here
that
may
better
describe
specific
UV
dose­
response
data
(
Severin
et
al.
1984).

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
23
June
2003
A.
2.5.3
Tailing
If
the
irradiated
microorganisms
are
a
mixture
of
disperse
microorganisms
and
clumped
or
particle­
associated
microorganisms,
the
UV
dose­
response
will
demonstrate
tailing,
or
a
flattening
of
the
curve
at
higher
UV
doses
(
Parker
and
Darby
1995).
With
wastewaters,
tailing
begins
after
2
to
3
log
of
disperse
microorganism
inactivation,
with
diminishing
inactivation
occurring
beyond
that
level
despite
increasing
UV
dose
(
Figure
A.
10,
Total
coliforms).
Doseresponse
with
tailing
can
be
modeled
using
Equation
A.
21.

D
k
p
kD
p
e
N
e
N
N
 
 
+
=
0
Equation
A.
21
where
N
=
Concentration
of
viable
microorganisms
after
UV
exposure
N0
=
Concentration
of
disperse
microorganisms
before
UV
exposure
k
=
First
order
inactivation
coefficient
of
the
microorganisms
(
cm2/
mJ)
D
=
UV
dose
delivered
to
the
microorganisms
(
mJ/
cm2)
Np
=
Concentration
of
particles
containing
the
microorganisms
kp
=
Pseudo
first
order
inactivation
constant
of
particle­
associated
microorganisms
(
cm2/
mJ)

A.
2.6
Factors
Impacting
Microbial
Response
Several
factors
impact
the
response
of
microorganisms
to
UV
light.
This
section
discusses
these
factors,
including
UV
intensity,
UV
absorbance,
temperature,
pH,
particles,
and
UV
wavelength.

A.
2.6.1
UV
Intensity
Oliver
and
Cosgrove
(
1975)
reported
that
UV
dose­
response
of
microorganisms
follows
the
Law
of
Reciprocity
over
an
intensity
range
of
1
to
200
mW/
cm2.
For
example,
the
inactivation
effectiveness
observed
with
UV
intensity
of
1
mW/
cm2
and
an
exposure
time
of
200
seconds
is
equivalent
to
the
inactivation
observed
with
an
exposure
time
of
1
second
and
UV
intensity
of
200
mW/
cm2
as
well
as
all
intensity­
time
combinations
between
1
and
200.

Exceptions
to
this
reciprocal
relationship
between
time
and
intensity
occur
at
very
low
and
high
intensities
(
Setlow
1967).
With
low
UV
intensities
and
long
exposure
times,
repair
may
compete
with
inactivation.
Sommer
et
al.
(
1998)
found
less
inactivation
of
E.
coli
at
a
given
dose
with
low
intensities
ranging
from
0.002
to
0.2
mW/
cm2.
However,
inactivation
of
B.
subtilis
spores,
MS2
bacteriophage
(
MS2),
 x174
phage,
and
B40­
8
phage
at
a
given
dose
was
the
same
regardless
of
UV
intensity.
At
UV
intensities
on
the
order
of
1010
to
1011
mW/
cm2
(
several
orders
of
magnitude
higher
than
the
intensity
from
lamps
used
for
UV
disinfection),
Gurzadyan
et
al.
(
1981)
reported
an
increase
in
single
strand
breaks
and
a
reduction
in
dimerization
in
the
nucleic
acid
of
 x174
phage.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
24
June
2003
A.
2.6.2
UV
Absorbance
Because
the
calculation
of
dose
delivered
to
a
microbial
suspension
(
Equation
A.
10,
section
A.
2.4.1)
accounts
for
UV
absorbance
in
bench­
scale
batch
experiments,
measured
UV
dose­
response
curves
like
those
presented
in
Figure
A.
10
are
independent
of
the
suspension's
UV
absorbance.
However,
as
the
UV
absorbance
of
the
suspension
increases,
the
UV
intensity
incident
on
the
sample
may
need
to
increase
in
order
to
deliver
a
given
dose.
In
bench­
scale,
batch
experiments,
there
are
several
ways
to
keep
the
UV
dose
constant
such
as
increasing
exposure
time
or
decreasing
the
depth
of
the
solution,
thereby
decreasing
the
pathlength.

A.
2.6.3
Temperature
Temperature
affects
the
configuration
of
nucleic
acid
and
the
activity
of
repair
enzymes;
however,
existing
research
shows
temperature
effects
on
UV
dose­
response
are
minimal
and
depend
on
the
microorganism.
Severin
et
al.
(
1983)
found
the
UV
dose
needed
for
a
given
log
reduction
of
E.
coli,
Candida
parapsilosis,
and
f2
phage
increases
slightly
as
temperature
decreased
(
Table
A.
1).
Malley
(
2000)
reported
the
dose­
response
of
MS2
is
independent
of
temperature
from
1
to
23
°
C
(
Figure
A.
11).

Table
A.
1
Impact
of
Temperature
on
UV
Disinfection
UV
dose
(
mJ/
cm2)
needed
to
achieve
2
log
inactivation
at
a
temperature
of
Microorganism
5
oC
20
o
C
35
o
C
E.
coli
11.8
11.2
10.7
C.
parapsilosis
30.9
28.8
28.0
f2
phage
72.5
65.6
60.7
Figure
A.
11
Impact
of
Temperature
on
MS2
UV
Dose­
Response
(
Malley
2000)

Delivered
UV
Dose
(
mJ/
cm2)
10
20
30
40
50
60
70
80
90
100
Log
Inactivation
of
MS2
0
1
2
3
4
5
pH
7
and
10
oC
pH
7
and
23
oC
pH
7
and
1
oC
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
25
June
2003
A.
2.6.4
pH
UV
dose­
response
is
usually
independent
of
the
pH
of
the
water.
The
UV
absorbance
of
nucleic
acid
and
repair
enzyme
activity
vary
with
pH
(
Jagger
1967).
However,
the
pH
within
a
cell
is
buffered
to
a
relatively
constant
value,
independent
of
water
pH.
For
example,
Malley
(
2000)
reported
the
dose­
response
of
MS2
is
independent
of
the
suspension
pH
from
pH
6
to
9
(
Figure
A.
12).

Figure
A.
12
Impact
of
pH
on
MS2
UV
Dose­
Response
(
Malley
2000)

Delivered
UV
Dose
(
mJ/
cm2)
10
20
30
40
50
60
70
80
90
100
Log
Inactivation
of
MS2
0
1
2
3
4
5
pH
9
and
1oC
pH
7
and
1oC
pH
6
and
1oC
A.
2.6.5
Particle
Association
To
date,
research
examining
the
effects
of
naturally
occurring
particles
and
microorganisms
is
limited
to
wastewater
studies.
Due
to
the
limited
concentration
of
microorganisms
in
drinking
water
sources,
methods
of
directly
examining
the
impact
of
particles
do
not
currently
exist.
However,
the
phenomena
observed
in
wastewater
studies
may
also
apply
to
particle
association
occurring
in
drinking
water.
The
effects
of
individual
particles
(
such
as
those
that
cause
turbidity)
on
UV
disinfection
are
discussed
in
section
A.
4.1.2.

Results
from
research
with
wastewaters
have
indicated
that
clumping
or
particle
association
will
shield
microorganisms
from
UV
light.
The
UVT
at
260
nm
through
10
microns
of
cell
tissue
is
roughly
10
percent
(
Jagger
1967),
suggesting
that
clumps
of
organisms
would
offer
protection.
The
water
content
of
cells
and
intracellular
material
and
the
porous
nature
of
flocculated
particles
will
influence
the
penetration
of
light
into
waterborne
particles.
Qualls
et
al.
(
1983)
reported
that
filtration
of
secondary
effluent
through
an
8
micron
filter
removes
the
particles
responsible
for
the
tailing
in
the
dose­
response
of
coliforms.
With
8
micron
filtration,
coliform
inactivation
at
12
mJ/
cm2
increased
from
3
log
to
over
4.5
log
inactivation.
Loge
et
al.
(
1999)
reported
the
UV
absorbance
of
wastewater
solids
varied
from
0.33
to
56.9
µ
m­
1
(
3,300
to
569,000
cm­
1)
with
the
high
absorbance
associated
with
activated
sludge
plant
using
iron
to
remove
phosphorus.
Petri
et
al.
(
2000)
reported
coagulation
of
MS2
by
iron
in
ground
water
increased
the
UV
dose
needed
to
inactivate
MS2
by
a
factor
of
2.5
to
3.5.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
26
June
2003
A.
2.6.6
Wavelength
Microbial
dose­
response
varies
with
the
wavelength
of
UV
light.
The
action
spectrum
(
also
called
UV
action)
of
a
microorganism
is
a
measure
of
inactivation
as
a
function
of
the
wavelength
for
a
given
UV
dose.
The
dependence
of
UV
action
on
wavelength
is
similar
to
the
dependence
of
the
UV
absorbance
of
DNA
on
wavelength
(
Figure
A.
13).
UV
action
peaks
near
260
nm,
has
a
local
minimum
near
230
nm,
and
drops
to
zero
near
300
nm.
While
it
is
generally
believed
that
microorganisms
are
most
sensitive
near
260
nm,
there
are
exceptions.
For
example,
the
UV
sensitivities
of
tobacco
mosaic
virus
(
Hollaender
and
Duggar
1936),
reovirus
(
Rauth
1965),
and
Herpes
simplex
virus
(
Powell
1959)
are
greater
below
230
nm.
Although
the
UV
action
increases
below
230
nm
for
most
microorganisms,
the
UV
absorbance
of
natural
waters
at
these
wavelengths
make
this
region
impractical
for
UV
disinfection
of
microorganisms
in
water.
Because
of
the
similarity
between
UV
action
and
DNA
absorbance,
and
because
DNA
absorbance
is
easier
to
measure
than
UV
action,
the
DNA
absorbance
spectrum
of
a
microorganism
is
often
used
as
a
surrogate
for
its
UV
action
spectrum.

Figure
A.
13
Comparison
of
Microbial
Action
to
DNA
UV
Absorbance
(
adapted
from
Rauth
1965
and
Linden
et
al.
2001)

0.0
0.5
1.0
1.5
2.0
200
300
Wavelength
(
nm)
UV
Action
or
DNA
Absorbance
Relative
to
254
nm
DNA
Cryptosporidium
MS2
Herpes
simplex
virus
220
240
260
280
A
plot
of
the
first
order
inactivation
coefficient
as
a
function
of
wavelength
can
be
used
to
show
the
action
spectrum
if
the
dose­
response
follows
first
order
inactivation.
Plots
of
two
kinetic
terms
as
a
function
of
wavelength
are
necessary
to
plot
the
action
spectrum
if
the
doseresponse
has
a
shoulder
(
Cabaj
et
al.
2001)
as
discussed
in
section
A.
2.5.2.
Plots
of
UV
action
spectra
are
often
presented
relative
to
some
wavelength,
typically
254
nm.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
27
June
2003
A.
2.7
UV
Dose­
Response
of
Differing
Microorganisms
The
UV
dose­
responses
of
microorganisms
have
been
tabulated
in
a
number
of
review
articles
and
are
summarized
in
this
section.

Data
presented
in
Tables
A.
2
and
A.
3
show
that
the
UV
sensitivity
of
microorganisms
varies
with
different
species.
Of
the
pathogens
of
interest
in
drinking
water,
viruses
are
most
resistant
to
UV
disinfection
followed
by
bacteria,
and
Cryptosporidium
oocysts
and
Giardia
cysts.
The
most
UV
resistant
viruses
of
concern
in
drinking
water
are
adenovirus
Type
40
and
41.
Appendix
B
provides
dose­
response
data
for
Giardia
cysts,
Cryptosporidium
oocysts,
and
viruses,
and
Chapter
1
(
Table
1.4)
contains
the
regulatory
UV
dose
requirements
for
inactivating
these
pathogens.

Table
A.
2
provides
average
dose
reported
without
photoreactivation
for
incremental
log
inactivation
of
various
pathogenic
bacteria,
virus,
and
protozoa
of
concern
in
drinking
water.
Table
A.
3
provides
similar
information
for
various
non­
pathogenic
indicator
bacteria,
spore
forming
bacteria,
and
bacteriophage.
All
data
in
Tables
A.
2
and
A.
3
are
for
microorganisms
suspended
in
water
and
irradiated
using
a
collimated
beam
apparatus
with
UV
light
at
254
nm.

Spore­
forming
and
gram­
positive
bacteria
are
more
resistant
to
UV
light
than
gram
negative
bacteria
(
Jagger
1967).
With
microorganisms
larger
than
1
micron,
the
absorption
of
UV
light
by
the
cytoplasm
can
be
significant,
depending
on
the
wavelength,
and
therefore
can
affect
UV
sensitivity.

Rauth
(
1965)
found
that
the
UV
sensitivity
of
virus
and
bacteriophage
varies
over
two
orders
of
magnitude
from
the
most
sensitive
to
the
most
resistant.
The
same
study
showed
viruses
with
single­
stranded
nucleic
acid
are
ten
times
more
sensitive
than
viruses
with
doublestranded
nucleic
acid.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
28
June
2003
Table
A.
2
UV
Sensitivity
of
Pathogenic
Microorganisms
in
Water1
UV
Dose
(
mJ/
cm2)
inactivation
indicated
Microorganism
Type
1­
log
2­
log
3­
log
4­
log
Reference
Aeromonas
hydrophila
Bacteria
1.1
2.6
3.9
5
Wilson
et
al.
1992
Campylobacter
jejuni
Bacteria
1.6
3.4
4
4.6
Wilson
et
al.
1992
Escherichia
coli
O157:
H7
Bacteria
1.5
2.8
4.1
5.6
Wilson
et
al.
1992
Legionella
pneumophila
Bacteria
3.1
5
6.9
9.4
Wilson
et
al.
1992
Salmonella
anatum
Bacteria
7.5
12
15
Tosa
and
Hirata
1998
Salmonella
enteritidis
Bacteria
5
7
9
10
Tosa
and
Hirata
1998
Salmonella
typhi
Bacteria
1.8
4.8
6.4
8.2
Wilson
et
al.
1992
Salmonella
typhimurium
Bacteria
2
3.5
5
9
Tosa
and
Hirata
1998
Shigella
dysenteriae
Bacteria
0.5
1.2
2
3
Wilson
et
al.
1992
Shigella
sonnei
Bacteria
3.2
4.9
6.5
8.2
Chang
et
al.
1985
Staphylococcus
aureus
Bacteria
3.9
5.4
6.5
10.4
Chang
et
al.
1985
Vibrio
cholerae
Bacteria
0.8
1.4
2.2
2.9
Wilson
et
al.
1992
Yersinia
enterocolitica
Bacteria
1.7
2.8
3.7
4.6
Wilson
et
al.
1992
Adenovirus
Type
40
2
Virus
30
59
90
120
Meng
and
Gerba
1996
Adenovirus
Type
41
2
Virus
22
50
80
Meng
and
Gerba
1996
Coxsackievirus
B5
Virus
6.9
14
21
Battigelli
et
al.
1993
Hepatitis
A
HM175
Virus
5.1
14
22
30
Wilson
et
al.
1992
Hepatitis
A
Virus
5.5
9.8
15
21
Wiedenmann
et
al.
1993
Hepatitis
A
HM175
Virus
4.1
8.2
12
16
Battigelli
et
al.
1993
Poliovirus
Type
1
Virus
4.0
8.7
14
21
Meng
and
Gerba
1996
Poliovirus
Type
1
Virus
6
14
23
30
Harris
et
al.
1987
Poliovirus
Type
1
Virus
5.6
11
16
22
Chang
et
al.
1985
Poliovirus
Type
1
Virus
5.7
11
18
13
Wilson
et
al.
1992
Rotavirus
SA11
Virus
7.6
15
23
Battigelli
et
al.
1993
Rotavirus
SA11
Virus
7.1
15
25
Chang
et
al.
1985
Rotavirus
SA11
Virus
9.1
19
26
36
Wilson
et
al.
1992
Cryptosporidium
parvum
2
Protozoa
<
2
<
3
<
5
Shin
et
al.
2001
Cryptosporidium
parvum
2
Protozoa
<
3
<
6
Clancy
et
al.
2000
Giardia
lamblia
2
Protozoa
<
1
<
2
Linden
et
al.
2002a
Giardia
lamblia
2
Protozoa
<
1
<
3
<
6
Mofidi
et
al.
2002
1
Adapted
from
Wright
and
Sakamoto
1999
2
Additional
data
for
adenovirus,
Cryptosporidium,
and
Giardia
are
in
Appendix
B.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
29
June
2003
Table
A.
3
UV
Sensitivity
of
Non­
Pathogenic
Bacteria,
Bacteriophage,
and
Spore­
Forming
Bacteria
in
Water1
UV
Dose
(
mJ/
cm2)
inactivation
indicated
Reference
Microorganism
Type
1­
log
2­
log
3­
log
4­
log
Escherichia
coli
Bacteria
2.5
3
3.5
5
Harris
et
al.
1987
Escherichia
coli
Bacteria
3
4.8
6.7
8.4
Chang
et
al.
1985
Escherichia
coli
Bacteria
4.0
5.3
6.4
7.3
Sommer
et
al.
1998
Escherichia
coli
Bacteria
4.4
6.2
7.3
8.1
Wilson
et
al.
1992
Streptococcus
faecalis
Bacteria
6.6
8.8
9.9
11
Chang
et
al.
1985
Streptococcus
faecalis
Bacteria
5.5
6.5
8
9
Harris
et
al.
1987
MS­
2
Phage
4
16
38
68
Wiedenmann
et
al.
1993
MS­
2
Phage
16
34
52
71
Wilson
et
al.
1992
MS­
2
Phage
12
30
Tree
et
al.
1997
MS­
2
Phage
21
36
Sommer
et
al.
1998
MS­
2
Phage
17
34
Rauth
1965
MS­
2
Phage
14
29
45
62
Meng
and
Gerba
1996
MS­
2
Phage
19
40
61
Oppenheimer
et
al.
1993
 X174
Phage
2.2
5.3
7.3
10
Sommer
et
al.
1998
 X174
Phage
2.1
4.2
6.4
8.5
Battigelli
et
al.
1993
 X174
Phage
4
8
12
Oppenheimer
et
al.
1993
PRD­
1
Phage
9.9
17
24
30
Meng
and
Gerba
1996
B­
40
Phage
12
18
23
28
Sommer
et
al.
1998
Bacillus
subtilis
spores
Spores
36
49
61
78
Chang
et
al.
1985
Bacillus
subtilis
spores
Spores
29
40
51
Sommer
et
al.
1998
1
Adapted
from
Wright
and
Sakamoto
1999.

A.
3
UV
Reactors
This
section
discusses
UV
reactor
components,
UV
reactor
configurations,
and
how
reactor
performance
is
monitored.
The
following
UV
reactor
components
are
discussed:

 
Mercury
lamps
 
Ballasts
and
power
supplies
 
Lamp
sleeves
 
Cleaning
systems
 
UV
intensity
sensors
 
UV
transmittance
monitors
 
Temperature
sensors
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
30
June
2003
A.
3.1
Mercury
Lamps
This
section
describes
mercury
lamps,
including
how
they
are
constructed,
their
components,
efficiency,
spectral
output,
and
aging.
A
majority
of
the
material
in
this
section
was
derived
from
Sources
and
Applications
of
Ultraviolet
Radiation
by
Roger
Phillips
(
1983).
Section
2.4.2
(
Table
2.1)
compares
the
operating
characteristics
of
LP,
LPHO,
and
MP
mercury
lamps.

A.
3.1.1
Lamp
Construction
Mercury
vapor
discharge
lamps
consist
of
a
UV­
transmitting
envelope
made
from
a
tube
of
quartz
sealed
at
both
ends
(
Figure
A.
14).
An
electrode
is
located
at
each
end
of
the
envelope
connected
to
the
outside
through
a
seal.
The
envelope
is
filled
with
mercury
and
an
inert
gas.

Figure
A.
14
Construction
of
a
UV
Lamp
LOW­
PRESSURE
MERCURY
LAMP
­
HOT
CATHODE
TYPE
Tungsten
Coil
Electrode
Electrical
Connection
Mercury
&
Inert
Gas
Fill
End
Seal
Envelope
LOW­
PRESSURE
HIGH­
OUTPUT
MERCURY
LAMP
­
AMALGAM
TYPE
MEDIUM­
PRESSURE
MERCURY
LAMP
Tungsten
Coil
Electrode
Electrical
Connection
Inert
Gas
Fill
Seal
Mercury
Amalgam
Envelope
Electrode
­
Tungsten
Coils
on
a
Tungsten
Rod
Electrical
Connection
Mercury
&
Inert
Gas
Fill
Seal
Envelope
Molybdenum
Foil
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
31
June
2003
Lamp
Envelope
The
envelope
of
the
lamp
should
transmit
germicidal
UV
light,
act
as
an
electrical
insulator,
and
not
react
with
the
lamp's
fill
gases.
A
non­
crystalline
form
of
quartz,
vitreous
silica,
is
often
used
for
the
lamp
envelope
because
of
its
high
UVT
and
its
resistance
to
high
temperatures.
However,
some
LP
lamps
use
UV­
transmitting
glass
instead
of
quartz.
Envelopes
are
approximately
1
to
2
mm
thick,
and
the
diameter
is
selected
to
optimize
the
UV
output
and
lamp
life.

As
quartz
is
exposed
to
high
temperatures,
it
begins
to
crystallize.
Crystallization
reduces
the
UVT
of
the
quartz
and
changes
its
coefficient
of
expansion,
which
causes
internal
stresses.
Envelopes
for
MP
lamps
must
be
able
to
withstand
thermal
shocks
associated
with
600
to
900
°
C
operating
temperatures
without
the
quartz
transforming
to
its
crystalline
form.
LP
lamps
have
lower
operating
temperatures
where
crystallization
is
not
a
concern,
which
is
why
some
LP
lamps
use
UV­
transmitting
glass
rather
than
quartz
for
the
lamp
envelope.

Envelopes
of
MP
lamps
may
be
covered
with
a
reflective
coating
at
the
ends.
This
is
to
keep
the
ends
warm
and
prevent
the
condensation
of
mercury
behind
the
electrodes.

The
UV
transmittance
of
the
envelope
affects
the
spectral
output
of
MP
mercury
lamps,
especially
at
lower
wavelengths.
Lamp
envelopes
can
be
made
from
doped
quartz,
or
quartz
that
is
altered
to
absorb
specific
wavelengths,
in
order
to
prevent
undesirable
non­
germicidal
photochemical
reactions.
If
the
lamp
envelope
is
not
made
from
doped
quartz,
the
lamp
sleeves
can
also
be
used
to
restrict
the
wavelengths
emitted
(
described
in
section
A.
3.3).

Electrodes
With
a
LP
mercury
lamp,
electrode
design
depends
on
whether
the
lamp
operates
with
a
glow
or
arc
discharge.
With
a
glow
discharge,
free
electrons
are
formed
from
the
bombardment
of
the
electrode
by
cations.
The
electrode
used
is
typically
a
cylinder
of
iron
or
nickel.
Lamps
of
this
type
of
electrode
operate
near
150
°
C
and
are
termed
cold­
cathode
lamps.
With
an
arc
discharge,
free
electrons
are
emitted
thermally
from
a
hot
electrode
operating
near
900
°
C
and
are
referred
to
as
hot­
cathode
lamps.
The
electrode
is
made
of
a
coil
of
tungsten
wire
embedded
with
oxides
of
calcium,
barium,
or
strontium.
The
high
melting
point
of
tungsten
prevents
evaporation
of
electrode
materials
that
could
coat
the
inside
of
the
lamp
and
reduce
output
of
UV
light.
The
oxides
embedded
within
the
tungsten
coil
reduce
the
temperature
needed
for
the
emission
of
electrons.
LP
and
LPHO
lamps
used
in
UV
disinfection
are
usually
hot­
cathode
lamps.

In
order
to
reduce
the
start
voltage
of
a
hot­
cathode
lamp,
each
electrode
may
have
two
electrical
connections
to
pass
current
through
the
electrode.
Resistive
heating
of
the
electrode
raises
the
electrode's
temperature,
thereby
facilitating
rapid
transition
from
a
glow
discharge
to
an
arc
discharge
at
a
lower
voltage.
Rapid
transition
from
a
glow
to
an
arc
discharge
reduces
electrode
sputtering
and
improves
lamp
life.
The
process
of
transitioning
from
glow
to
arc
discharges
and
how
it
produces
UV
light
is
described
in
section
A.
1.1.1.

Electrode
design
and
operation
is
critical
for
reliable
long
term
operation.
In
order
for
lamps
to
operate
at
an
optimal
temperature,
electrode
design
should
promote
heat
transfer.
The
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
32
June
2003
electrodes
of
a
MP
mercury
lamp
consist
of
a
tungsten
rod
wrapped
in
a
coil
of
tungsten
wire.
To
improve
thermal
emission
of
electrons,
thorium
or
alkaline­
earth
oxides
are
embedded
within
the
coils,
and
the
tungsten
rod
may
contain
thorium
oxide.
The
electrode
must
warm­
up
within
seconds
to
allow
transition
from
a
glow
to
an
arc
discharge
and
minimize
sputtering
of
tungsten
onto
the
envelope.
The
electrode
operating
temperature
must
be
high
enough
to
promote
thermal
emission
of
electrons
and
low
enough
to
prevent
the
evaporation
of
tungsten.

The
electrode
of
a
MP
lamp
is
connected
to
the
external
electrical
supply
via
a
thin
molybdenum
foil
sealed
in
the
quartz
at
the
lamp
ends.
The
molybdenum
foil
is
ductile
and
therefore
does
not
crack
the
quartz
when
the
lamp
expands
and
contracts
due
to
changes
in
operating
temperature.
If
the
temperature
of
the
seal
increases
beyond
350
°
C,
the
molybdenum
will
oxidize
and
the
seal
will
fail.
Because
the
seal
is
located
behind
the
electrode,
its
temperature
is
lower
than
the
temperature
of
the
arc.

Mercury
Fill
The
mercury
fill
present
in
UV
lamps
can
be
in
the
solid,
liquid,
or
vapor
phase.
At
typical
LP
and
LPHO
lamp
operating
temperatures,
mercury
remains
predominantly
in
the
liquid
or
solid
amalgam
phase
with
a
small
proportion
in
the
vapor
phase
(
which
is
responsible
for
producing
UV
light).
An
amalgam
is
an
alloy
of
elemental
mercury
with
another
metal
(
typically
indium
or
gallium
in
lamp
applications)
that
can
be
either
solid
or
liquid
at
room
temperature,
depending
on
the
relative
proportions
of
the
two
metals.
Amalgams
are
typically
used
in
LP
and
LPHO
lamps,
while
MP
lamps
contain
liquid
elemental
mercury.

Vapor
pressure
(
the
pressure
of
mercury
in
the
vapor
phase)
depends
on
the
temperature.
LP
lamps
operate
with
an
envelope
temperature
near
40
°
C,
resulting
in
a
mercury
vapor
pressure
near
0.007
torr
(
1.4
x
10­
4
psi),
which
is
optimal
for
the
production
of
UV
light
at
254
nm.
MP
lamps
operate
at
a
much
higher
envelope
temperature
(
600
to
900
°
C),
resulting
in
a
mercury
vapor
pressure
ranging
from
100
to
10,000
torr
(
2
to
200
psi).
In
MP
lamps,
the
concentration
of
mercury
in
the
vapor
phase
is
controlled
by
the
amount
of
mercury
in
the
lamp,
as
opposed
to
LP
and
LPHO
lamps
where
an
excess
of
mercury
is
placed
in
the
lamp
and
only
a
portion
of
the
elemental
mercury
enters
the
vapor
phase.

With
a
conventional
LP
lamp,
increasing
the
operating
current
will
not
produce
a
higher
UV
output.
Instead
the
operating
temperature
will
increase
causing
an
increase
in
vapor
pressure
and
the
UV
light
output
of
the
lamp
will
decrease.
LPHO
lamps
hold
the
mercury
vapor
pressure
constant
at
the
optimal
value,
allowing
the
UV
light
output
to
increase
as
current
increases
until
a
saturation
value
is
reached.
Methods
of
controlling
the
vapor
pressure
within
the
lamp
include
using
either
a
mercury
amalgam
attached
to
the
lamp
envelope,
a
cold
spot
on
the
lamp
wall,
or
a
mercury
condensation
chamber
located
behind
each
electrode.
With
each
method,
the
temperature
of
the
mercury
within
the
lamp,
and
hence
the
vapor
pressure,
is
controlled,
allowing
efficient
production
of
UV
light
at
higher
currents.

Inert
Gas
Fill
In
addition
to
mercury,
lamps
are
filled
with
an
inert
gas
(
typically
argon)
at
1
to
50
torr
(
0.02
to
1
psi).
The
inert
gas
aids
in
starting
the
gas
discharge
and
reduces
deterioration
of
the
electrode.
When
the
lamp
is
started
at
room
temperature,
the
concentration
of
mercury
atoms
is
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
33
June
2003
low
and
there
are
few
collisions
between
free
electrons
and
mercury.
However,
there
are
a
significant
number
of
collisions
between
free
electrons
and
argon
atoms.
These
collisions
excite
the
argon
atoms
to
a
metastable
higher
energy
state
that
does
not
return
quickly
to
a
ground
state.
Collisions
between
the
excited
metastable
argon
and
mercury
or
free
electrons
ionizes
the
mercury
and
argon,
respectively.
The
free
electrons
released
by
ionization
reduce
the
start
voltage
and
aid
in
the
formation
of
the
gas
discharge.
However,
if
lamps
are
manufactured
with
a
non­
ideal
argon
pressure
(>
50
torr;
1
psi),
the
collisions
between
electrons
and
argon
cause
energy
losses,
and
therefore
the
electrons
never
achieve
sufficient
kinetic
energy
to
excite
the
mercury
atoms.

A.
3.1.2
Low­
Pressure
Lamp
Efficiency
LP
lamps
are
designed
and
manufactured
to
efficiently
convert
electrical
energy
to
germicidal
UV
light.
An
optimal
LP
lamp
design
typically
includes
the
following
components:

 
3.6
cm
lamp
envelope
diameter
 
0.007
torr
mercury
fill
(
1.3
x
10­
4
psi)

 
3
torr
argon
fill
(
0.06
psi)

 
400
mA
operating
current
 
40
°
C
operating
temperature
 
0.5
W/
cm
power
input
per
arc
length
Under
such
conditions,
the
power
input
efficiency
is
as
follows:

 
60
percent
converted
to
UV
light
at
185
and
254
nm
 
3
percent
converted
to
other
wavelengths
 
15
percent
to
electrode
losses
 
22
percent
to
thermal
losses
from
the
arc
A.
3.1.3
Low­
Pressure
High
Output
Lamp
Efficiency
Theoretically,
LPHO
lamps
have
the
same
efficiency
as
LP
lamps
because
they
operate
at
similar
vapor
pressures.
However
in
practice,
LPHO
lamp
efficiency
can
be
significantly
lower,
depending
on
lamp
construction,
ballast
operation,
power
settings,
and
lamp
cooling.
The
energy
input
to
a
LPHO
lamp
can
be
converted
to
energy
in
the
following
forms:

 
30
to
45
percent
converted
to
UV
light
at
and
254
nm
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
34
June
2003
 
5
to
25
percent
converted
to
light
at
other
wavelengths,
primarily
185,
313,
365,
and
436
nm
 
50
to
65
percent
to
thermal
losses
from
the
arc
A.
3.1.4
Medium­
Pressure
Lamp
Efficiency
For
the
purposes
of
UV
disinfection,
the
efficiency
of
a
MP
lamp
can
be
defined
as
the
ratio
of
its
germicidal
output
to
its
electrical
input.
Equation
A.
22
defines
germicidal
efficiency
as
a
function
of
power
input,
lamp
output,
and
the
action
spectra
of
a
given
microorganism.

E
nm
nm
E
G
P
G
P
P
P
 
=
=
=
300
200
)
(
)
(
 
 
 
 
Equation
A.
22
where
 
=
Germicidal
efficiency
of
the
lamp
PG
=
Germicidal
lamp
output
(
W)
PE
=
Electrical
power
input
(
W)
 
=
Wavelength
(
nm)
P(
 )
=
Lamp
output
measured
over
1
nm
increments
at
wavelength
 
(
W)
G(
 )
=
Action
spectra
of
the
microorganism
at
wavelength
 
(
unitless)

Figure
A.
15
presents
the
output
versus
electrical
input
between
248
and
320
nm
for
three
MP
lamps
containing
different
mercury
doses.
For
the
lamps
considered,
lamp
efficiency
varied
slightly
with
input
power
to
the
lamp
but
did
not
vary
with
mercury
dose.
Lamp
efficiency
on
average
was
10
percent.
Because
lamp
data
used
to
generate
Figure
A.
15
were
based
only
on
lamp
output
from
248
to
320
nm,
the
lamp
efficiency
may
be
underestimated.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
35
June
2003
Figure
A.
15
Germicidal
Output
from
248
to
320
nm
of
Three
MP
Lamps
Calculated
for
MS2
as
a
Function
of
Electrical
Power
Input
(
adapted
from
Phillips
1983)

0
5
10
15
0
50
100
Electrical
Power
Consumed
(
W/
cm)
Germicidal
Output
(
W/
cm)

150
4.8
mg/
cm
Hg
dose
8.0
mg/
cm
Hg
dose
10.1
mg/
cm
Hg
dose
A.
3.1.5
Spectral
Output
of
Lamps
LP
lamps
emit
primarily
resonant
light
at
253.7
nm
(
Figure
A.
16a)
that
is
formed
from
electron
transitions
from
the
first
excited
states
to
the
ground
state
of
mercury.
They
also
emit
light
at
185
nm
with
intensity
varying
from
12
to
34
percent
of
the
UV
intensity
at
253.7
nm
depending
on
the
operating
current,
wall
temperature,
and
inert
gas
fill.
UV
light
at
185
nm
will
react
with
oxygen
and
promote
the
formation
of
ozone
within
the
lamp
sleeve.
Ozone
is
a
corrosive
and
toxic
compound
that
absorbs
UV
light.
As
such,
LP
lamps
for
UV
disinfection
applications
are
manufactured
to
reduce
or
eliminate
the
emission
of
UV
light
at
185
nm.
Other
wavelengths
of
light
including
313,
365,
405,
436,
and
546
nm
also
are
emitted
from
LP
lamps
at
low
intensities
due
to
higher
energy
electron
transitions
within
the
mercury.

The
spectral
output
of
LPHO
lamps
is
similar
to
LP
lamps.
Although
all
of
the
wavelengths
emitted
are
identical,
the
intensity
of
light
from
LPHO
lamps
is
higher.

The
spectral
output
of
MP
mercury
lamps
involves
peaks
overlying
a
continuum
(
Figure
A.
16b).
The
combination
of
free
electrons
and
mercury
cations
within
the
arc
creates
a
broad
continuum
of
UV
energy
lines
between
200
and
245
nm.
This
continuum
does
not
occur
with
LP
lamps,
where
non­
radiating
recombination
occurs
at
the
envelope
walls.
Electron
transitions
within
the
mercury
produce
numerous
narrow
peaks
of
electromagnetic
energy
in
the
visible
and
ultraviolet
range.
These
transitions
result
in
a
broadening
of
the
emitted
light
and
a
shift
in
its
peak,
usually
to
longer
wavelengths.
For
example,
the
peak
from
260
to
270
nm
arises
primarily
due
to
the
254
nm
electron
transition.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
36
June
2003
Figure
A.
16.
UV
Output
of
LP
(
a)
and
MP
(
b)
Mercury
Lamps
(
Sharpless
and
Linden
2001)

0.0
0.2
0.4
0.6
0.8
1.0
1.2
200
250
300
350
400
Wavelength
(
nm)
Relative
Lamp
Output
Lamp
Output
Relative
to
Maximum
Output
in
Range
b.
Medium
Pressure
Lamp
UV
Continuum
0.0
0.2
0.4
0.6
0.8
1.0
1.2
200
250
300
350
400
Wavelength
(
nm)
Relative
Lamp
Output
Lamp
Output
Relative
to
Maximum
Output
in
Range
a.
Low
Pressure
Lamp
A.
3.1.6
Lamp
Aging
Over
time,
UV
lamps
can
degrade,
resulting
in
a
reduction
in
output
where
lower
germicidal
wavelengths
degrade
faster
than
higher
wavelengths.
Lamp
output
will
decrease
over
time
as
a
function
of
lamp
hours
in
operation,
number
of
on/
off
cycles
and
power
applied
per
unit
(
lamp)
length.
The
rate
of
decrease
in
lamp
output
slows
as
the
lamp
ages
(
Figure
A.
17).

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
37
June
2003
Figure
A.
17.
Reduction
in
UV
Output
of
LP
and
MP
Lamps
Over
Time
(
adapted
from
Schenck
1981)

a.
Low­
Pressure
Mercury
Lamps
0
20
40
60
80
100
120
0
5000
10000
Time
(
Hrs)
UV
Output
(%)
b.
Medium­
Pressure
Mercury
Lamps
0
20
40
60
80
100
120
0
500
1000
Time
(
Hrs)
UV
Output
(%)

UV
A
UV
B
UV
C
Lamp
aging
can
be
affected
by
the
following
factors:

 
Ballast
operations,
including
power
setting,
frequency,
and
harmonic
distortion
of
the
voltage
and
current
driving
the
lamp
 
Water
temperature
and
heat
transfer
from
lamps
 
Vibration
of
the
lamp
sleeves
caused
by
water
flowing
through
the
reactor
 
The
frequency
of
on­
off
cycles
With
LP
and
MP
lamps,
sputtering
of
the
electrode
during
the
glow
phase
of
start­
up
can
coat
the
inside
surface
of
the
lamp
envelope
with
tungsten.
The
tungsten
coating
is
black
in
color,
non­
uniform,
concentrated
within
a
few
inches
of
the
electrode,
and
can
absorb
UV
light
(
Figure
A.
18).
Sputtering
from
the
electrode
can
be
reduced
by
the
following
conditions:

 
Pre­
heating
the
electrode
before
applying
the
start
voltage
 
Driving
the
lamp
with
a
sinusoidal
current
waveform
 
Using
a
lamp
with
a
higher
argon
content
 
Minimizing
the
number
of
lamp
starts
during
operation
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
38
June
2003
Figure
A.
18.
Aged
UV
Lamp
(
right)
in
Comparison
to
a
New
UV
Lamp
(
left)
(
Mackey
et
al.
2003)

If
a
MP
lamp
is
not
sufficiently
cooled
during
operation,
tungsten
and
oxides
between
the
tungsten
coils
may
evaporate
and
coat
the
inside
of
the
envelope.
LP
lamps
using
UV­
transmitting
glass
may
have
mercury
combine
with
sodium
in
the
glass
to
create
a
UV
absorbing
coating.
Any
deposits
on
the
inner
or
outer
surfaces
of
the
lamp
envelope
and
by
metallic
impurities
within
the
envelope
will
absorb
UV
light.
The
absorption
of
UV
light
can
raise
the
temperature,
which
may
lead
to
localized
overheating
of
the
lamp
envelope.
If
the
lamp
envelope
is
quartz,
the
increase
in
temperature
can
lead
to
devitrification
(
crystallization),
contributing
to
an
additional
decrease
in
UVT.

With
MP
lamps,
reaction
of
the
electrode
with
any
water
molecules
that
have
entered
the
lamp
envelope
as
a
result
of
lamp
seal
failure
will
form
an
oxide
and
hydrogen
and
also
increase
the
start
voltage.
The
molybdenum
seal
of
a
MP
lamp
will
oxidize
and
fail
if
the
seal
temperature
exceeds
350
°
C.
High
operating
temperatures
of
a
MP
lamp
can
also
lead
to
bubbles
and
distortion
of
the
lamp
envelope
materials
and
devitrification
(
crystallization),
which
leads
to
a
decrease
in
UVT.
The
coefficient
of
expansion
of
crystalline
quartz
is
higher
than
that
of
noncrystalline
quartz,
and
rapid
changes
in
temperature
will
also
stress
the
envelope,
which
may
lead
to
lamp
breakage.

A.
3.2
Lamp
Power
Supply
and
Ballasts
UV
lamps
are
typically
operated
with
an
AC
supply.
Unlike
an
incandescent
lamp,
a
mercury
vapor
lamp
cannot
be
connected
directly
to
the
electrical
service
because
it
has
a
nonlinear
voltage
to
ampere
characteristic
(
Persson
and
Kuusisto
1998).
In
order
for
the
mercury
vapor
lamp
to
function
properly,
a
ballast
must
be
inserted
into
the
circuit
to
limit
the
current
flow
through
the
lamp.
When
placed
in
series
with
the
lamp,
the
ballast
provides
an
impedance
to
the
power
supply
with
a
positive
voltage­
current
characteristic.
The
power
supply
and
ballast
are
designed
to
provide
the
following
features:

 
Reliable
and
rapid
starting
of
the
gas
discharge
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
39
June
2003
 
Re­
ignition
of
the
gas
discharge
every
half
cycle
of
the
power
supply
ductors,
or
combinations
of
these
can
be
used
as
ballasts;
owever,
resistors
are
not
used
because
they
consume
power
and
therefore
reduce
electrical
efficien
ic)
or
.

.3.2.1
Magnetic
Ballasts
two
types
of
magnetic
ballasts:
capacitive
(
those
with
capacitors)
and
inductive
(
those
with
inductors).
Each
ballast
type
is
designed
to
control
the
current
to
the
lamp.

e
capacitance
used
and
does
not
vary
significantly
with
the
applied
voltage
or
the
lamp
properties.
An
adv
asts,

the
applied
voltage,
and
the
lamp
properties.
As
electrical
current
flows
through
the
inductor,
it
generat
f
er
tion
netic
ballasts
are
currently
the
most
common
type
of
ballast
used
for
medium
pressur
ps
due
to
their
durability
and
proven
operating
stability
in
the
higher
power
applica
ent
to
agnetic
 
An
appropriate
current
waveform
 
A
high
power
factor
 
Stable
light
output
Resistors,
capacitors,
in
h
cy.
Lamp
ballasts
are
often
termed
either
magnetic
(
also
known
as
electromagnet
electronic.
Magnetic
ballasts
can
be
inductive
or
capacitive
and
operate
at
the
line
frequency
Electronic
ballasts
operate
at
frequencies
higher
than
that
of
the
line
voltage
and
involve
solid
state
devices
or
a
mixture
of
solid
state
devices,
inductors,
and
capacitors.

A
There
are
With
a
capacitive
ballast,
the
current
through
the
lamp
is
primarily
a
function
of
th
antage
of
the
capacitive
ballast
is
that
the
power
delivered
to
the
lamp
and
the
lamp
output
are
independent
of
line
voltage.
A
disadvantage
is
that
electrode
sputtering
can
increase,
which
accelerates
electrode
aging.
Capacitive
ballasts
are
more
efficient
than
inductive
ball
but
less
efficient
than
electronic
ballasts.
Because
of
the
stored
energy
in
the
capacitor
and
the
coil,
capacitive
ballasts
are
less
prone
to
failure
as
a
result
of
small
fluctuations
in
power
quality.

With
the
inductive
ballast,
the
current
through
the
lamp
is
a
function
of
the
inductance,

es
a
magnetic
field.
The
magnetic
field
opposes
the
electrical
current,
and
the
strength
o
the
field
is
proportional
to
the
current
passing
through
the
inductor.
Therefore,
as
the
current
increases,
so
does
the
resistance
to
the
current.
This
interaction
limits
the
total
current
flow
to
the
lamps
to
a
specific
amperage.
The
highest
power
achieved
with
the
inductive
ballast
is
low
than
with
the
capacitive
ballast.
However,
electrode
sputtering
is
less
than
with
capacitive
ballasts,
leading
to
extended
electrode
life.
With
capacitive
ballasts,
the
UV
lamp
output
varies
with
the
line
voltage.
Inductive
ballasts
provide
more
stable
current
output
and
better
resolu
and
control
than
capacitive
ballasts,
but
are
generally
less
efficient,
larger,
heavier,
and
more
expensive.

Mag
e
lam
tions.
Medium
pressure
reactors
typically
incorporate
some
form
of
power
adjustm
optimize
energy
efficiency
and
control
dose
delivery.
Because
of
the
manner
in
which
m
ballasts
operate,
power
can
only
be
adjusted
by
incorporating
capacitors
or
inductors
into
the
circuit.
Adjustment
occurs
in
a
series
of
steps,
and
the
number
of
steps
is
limited
by
the
number
of
capacitors
or
inductors
that
are
included
in
the
ballast.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
40
June
2003
A.
3.2.2
Electronic
Ballasts
Electronic
ballasts,
sometimes
referred
to
as
solid
state
ballasts,
contain
semiconductors
and
other
electronic
components
such
as
low­
pass
filters,
rectifiers,
buffer
capacitors,
and
high
frequen
lamp
allasts
and
are
therefore
less
proven.
Although
they
have
limited
operational
experience,
electronic
ballasts
offer
increased
efficien
d
.3.2.3
Comparison
of
Ballast
Types
advantages
and
disadvantages.
Manufacturers
consider
these
advantages
and
disadvantages
when
determining
the
technology
to
incorpo
r
cy
oscillators
that
allow
the
ballast
to
behave
like
a
switching
power
supply.
A
chopped
electrical
current
with
up
to
50,000
pulses
per
second
of
electricity
is
supplied
to
the
lamp,
whereas
a
magnetic
ballast
typically
produces
only
100
to
120
pulses
per
second.
With
an
electronic
ballast,
the
frequency
of
electrical
pulses
supplied
to
the
lamp
is
longer
when
the
is
cold.
As
the
lamp
approaches
its
optimum
operating
temperature,
the
electronic
ballast
provides
shorter
and
less
frequent
pulses
of
current
to
the
lamp.

Electronic
ballasts
are
a
newer
technology
than
magnetic
b
cy,
smaller
size
and
weight,
and
the
opportunity
for
nearly
continuous
power
adjustment
over
a
wide
range
of
settings.
Reliability
has
improved
significantly
since
electronic
ballasts
were
initially
developed.
Currently,
manufacturers
of
low
pressure
reactors
and
smaller
medium
pressure
reactors
often
use
electronic
ballasts
in
their
design.
Because
of
the
reduction
in
store
energy,
electronic
ballasts
are
more
susceptible
to
failure
due
to
power
inconsistencies;
however,
by
incorporating
a
buffer
capacitor,
minor
power
disturbances
can
be
smoothed
out,
reducing
the
occurrence
of
lamp
failure.

A
Electronic
and
magnetic
ballasts
each
have
specific
rate
into
their
equipment
designs.
The
final
selection
takes
into
account
the
relative
importance
of
each
of
the
advantages
and
disadvantages
for
a
given
application.
A
single
manufacturer
may
have
equipment
designs
based
on
both
ballast
types.
For
example,
one
UV
manufacturer
uses
electronic
ballasts
for
its
smaller
units
and
magnetic
ballasts
for
its
large
units.
A
summary
of
some
of
the
advantages
and
disadvantages
of
each
ballast
technology
is
shown
in
Table
A.
4.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
41
June
2003
Table
A.
4.
Comparison
of
Magnetic
and
Electronic
Ballasts.

Magnetic
Ballast
Electronic
Ballast
Comparative
Advantages
 
Less
potential
for
power
interference
due
to
stored
energy
 
More
resistant
to
power
surges
 
More
resistant
to
high
temperatures.
 
Less
prone
to
interference
with
electronic
devices
 
Less
prone
to
sputtering
(
inductive
less
than
capacitive)
 
Proven
technology
(
in
use
for
nearly
70­
years)
 
Less
expensive
 
More
efficient
 
Lighter
weight
 
Smaller
size
 
Less
potential
for
heat
generation
 
Less
potential
for
noise
 
Continuous
power
adjustment
 
Longer
lamp
operating
life
Comparative
Disadvantages
 
Less
efficient
(
capacitive
more
efficient
than
inductive)
 
Heavier
weight
 
Larger
size
 
More
potential
for
heat
generation
 
More
potential
for
noise.
 
Step­
function
power
adjustment
(
number
of
steps
proportional
to
number
of
inductors/
capacitors)
 
Shorter
lamp
operating
life
 
More
potential
for
power
interference
due
to
stored
energy
(
can
be
minimized
by
incorporating
a
capacitor)
 
Less
resistant
to
power
surges
 
Less
resistant
to
high
temperatures
 
More
prone
to
interference
with
electronic
devices
 
More
potential
for
sputtering
 
Newer
technology
(
limited
operating
experience,
especially
in
larger
sizes)
 
More
expensive
A.
3.2.4
Lamp
Startup
The
voltage
applied
to
the
lamps
must
be
sufficiently
high
to
start
and
operate
the
lamps.
Step­
up
transformers
are
needed
to
increase
the
voltage
above
the
mains
to
start
cold­
cathode
lamps.
Hot­
cathode
lamps
are
classified
as
either
instant
or
switch
start.
Instant­
start
lamps
have
a
single
connection
with
each
electrode.
Starting
instant­
start
lamps
needs
the
application
of
a
high
voltage.
As
the
electrodes
warm­
up,
the
needed
voltage
drops.
Switch­
start
lamps
have
two
electrical
connections
with
each
electrode,
and
the
electrodes
are
preheated
for
1
to
2
seconds
before
the
start
voltage
is
applied.
This
reduces
the
start
voltage
and
lengthens
the
lamp
life.
Because
of
their
relatively
high
impedance,
MP
lamps
typically
need
a
higher
voltage
than
LP
lamps
for
starting
and
stable
operation.
Operating
voltage
ranges
from
5
to
30
volts/
cm,
depending
on
arc
length,
mercury
dose,
lamp
diameter,
and
electrode
losses.
With
the
exception
of
short
lamps,
step­
up
transformers
are
needed
to
operate
MP
lamps
and
high
voltage
pulses
are
used
to
start
them.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
42
June
2003
A.
3.2.5
Voltage
Frequency
Converters
With
LP
and
LPHO
lamps,
frequency
converters
may
be
used
to
increase
the
voltage
frequency
from
that
of
the
mains
(
typically
60
Hz)
to
20
to
100
kHz.
Typically,
the
efficiency
of
UV
light
output
from
electrical
power
increases
by
as
much
as
10
percent
as
the
frequency
increases
above
500
Hz.
Furthermore,
the
higher
frequency
reduces
electrode
deterioration,
makes
the
lamps
easier
to
start,
and
extends
the
lamp
life.
These
benefits,
however,
can
be
offset
by
power
losses
associated
with
the
frequency
converter.

A.
3.3
Lamp
Sleeves
In
UV
reactors,
lamps
are
housed
within
lamp
sleeves.
Sleeve
length
is
sufficient
to
include
the
lamp
and
associated
electrical
connections.
Sleeve
diameter
is
typically
1
inch
(
2.5
cm)
for
LP
mercury
lamps
and
2
to
4
inches
(
5
to
10
cm)
for
MP
lamps.
Sleeve
walls
are
typically
2
to
3
mm
thick
and
absorb
some
UV
light.
Sleeves
made
of
doped
quartz
are
used
to
prevent
the
transmission
of
low­
wavelength
UV
light,
thereby
reducing
undesirable
photochemical
reactions.

Lamp
sleeves
have
several
functions
other
than
housing
the
lamps.
They
maintain
the
lamp
temperature
at
an
optimal
value
and
control
heat
transfer
from
the
lamps.
Heat
transfer
from
the
MP
lamp
prevents
failure
of
the
molybdenum
seal,
distortion
of
the
lamp
envelope,
and
evaporation
of
the
tungsten
electrode.
Also,
lamp
sleeves
isolate
the
lamp
and
its
electrical
connections
from
the
water.
Lastly,
they
protect
the
lamp
from
mechanical
forces
such
as
water
hammer
and
protect
the
lamp
from
thermal
shock
arising
from
differences
in
water
and
lamp
envelope
temperature.

Typically,
LP
lamps
are
centered
using
Teflon
®
rings,
and
MP
lamps
are
centered
using
ceramic
or
metal
disks.
The
positioning
of
the
lamp
along
the
length
of
the
sleeve
can
influence
dose
delivery
by
the
reactor.

Sealing
the
lamp
sleeve
assembly
prevents
water
condensation
within
the
sleeve
and
contains
any
ozone
formed
between
the
lamp
envelope
and
lamp
sleeve.
Components
within
the
sleeve
should
withstand
exposure
to
UV
light,
ozone,
and
high
temperatures.
If
the
components
are
not
made
of
the
appropriate
material,
exposure
can
cause
component
deterioration
and
offgassing
of
any
impurities
present
in
the
quartz
from
manufacturing.
Off­
gassed
materials
can
form
UV­
absorbing
deposits
on
the
inner
surfaces
of
the
lamp
sleeve.
Off­
gassing
and
ozone
formation
will
be
a
greater
issue
with
MP
lamps
because
they
operate
at
a
higher
temperature
and
emit
low­
wavelength
ozone­
forming
UV
light.
Off­
gassing
can
be
minimized
through
proper
manufacturing
of
the
lamp
sleeves.

The
UVT
of
a
lamp
sleeve
influences
the
intensity
of
UV
light
transmitted
from
the
lamp
into
the
water.
The
UVT
is
a
function
of
the
reflectance
and
absorbance
of
UV
light
by
the
sleeve,
as
per
Equation
A.
23.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
43
June
2003
)
)
(
(
)]
(
1
)][
(
1
[
)
(
L
SW
AS
s
e
R
R
UVT
 
 
 
 
 
 
 
 
=
Equation
A.
23
where
UVTS(
 )
=
Sleeve
UVT
at
wavelength
 
RAS(
 )
=
Reflectance
of
UV
light
at
the
air­
sleeve
interface
at
wavelength
 
RSW(
 )
=
Reflectance
of
UV
light
at
the
sleeve­
water
interface
at
wavelength
 
 (
 )
=
Sleeve
absorption
(
Base
e)
at
wavelength
 
L
=
Pathlength
of
light
through
the
sleeve
Because
the
refractive
indices
of
the
lamp
sleeve
and
water
are
similar,
the
reflectance
of
UV
light
at
the
sleeve­
water
interface
(
RSW)
is
often
considered
negligible
in
this
equation.
The
absorption
coefficient
of
the
sleeve
varies
strongly
with
wavelength
and
the
material
of
the
sleeve.
For
a
zero
degree
incidence
angle,
Figure
A.
19
presents
the
UVT
over
the
germicidal
range
of
two
types
of
quartz:
standard
and
wavelength­
selective.
Quartz
can
be
manufactured
to
select
for
a
variety
of
wavelengths
depending
on
the
desired
application.
For
UV
disinfection
applications,
wavelength­
selective
quartz
is
primarily
used
to
prevent
the
transmission
of
low
wavelength
(<
200
nm)
UV
light
into
the
water.

Figure
A.
19.
UV
Transmittance
of
Two
Types
of
Quartz
Commonly
Used
to
Make
Lamp
Sleeves
(
GE
Quartz
2001)

Standard
Quartz
Wavelength­
Selective
Quartz
100
80
60
40
20
0
UV
Transmittance
(%)

200
220
240
260
280
300
320
Wavelength
(
nm)

In
order
to
reduce
fouling
on
the
sleeve
surfaces,
some
UV
reactors
using
LP
lamps
have
sleeves
made
of
Teflon
®
or
Teflon
®
­
coated
quartz.
However,
Teflon
®
sleeves
have
a
lower
UV
transmittance,
and
their
transmittance
degrades
faster
than
conventional
quartz.

Failure
mechanisms
for
sleeves
include
fractures
and
fouling.
Fractures
arise
from
internal
stresses
created
during
the
production
of
the
quartz
and
external
mechanical
forces.
Annealing
the
quartz
at
high
temperatures
during
production
removes
internal
stresses.
Visual
inspection
using
polarized
light
can
also
reveal
whether
or
not
sleeves
are
stress
free.
Fractures
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
44
June
2003
may
arise
from
mechanical
forces
such
as
wiper
jams,
water
hammer,
resonant
vibration,
and
impact
by
foreign
objects.
Fouling
may
occur
on
both
internal
and
external
surfaces
and
is
discussed
in
more
detail
in
section
A.
4.1.4.
Exposure
of
quartz
contaminated
with
metal
cations
from
the
manufacturing
process
can
cause
solarization
and
an
increase
in
UV
absorption.

A.
3.4
Cleaning
Systems
Due
to
fouling
on
the
lamp
sleeves,
cleaning
the
external
surface
of
sleeves
is
important
to
maintain
dose
delivery.
UV
reactor
manufacturers
have
developed
different
approaches
for
cleaning
lamp
sleeves,
depending
on
the
application.
Both
manual
and
automatic
cleaning
regimes
are
used.
A
reactor
must
be
shut
down
and
drained
prior
to
manual
cleaning.
The
sleeves
are
removed
once
the
reactor
is
drained
and
wiped
with
a
cloth
and
cleaning
solution.
Manual
cleaning
is
primarily
used
for
smaller
systems
with
relatively
few
sleeves
and
lower
fouling
potential.

Automatic
cleaning
approaches
are
typically
used
for
larger
systems.
They
may
be
classified
as
off­
line
chemical
cleaning
(
OCC)
or
on­
line
mechanical
cleaning
(
OMC).
OCC
systems,
also
referred
to
as
flush
and
rinse
systems,
involve
a
sequence
of
events
controlled
by
the
UV
reactor.
In
OCC
systems,
the
reactor
is
shut
down,
drained,
and
flushed
with
a
cleaning
solution.
Solutions
used
to
clean
sleeves
include
citric
acid,
phosphoric
acid,
or
a
food
grade
proprietary
solution
provided
by
the
UV
reactor
manufacturer.
The
reactor
is
rinsed
and
returned
to
operation
after
sufficient
time
to
dissolve
the
substances
fouling
the
sleeves
is
allowed.
OCC
cleaning
approaches
are
typically
used
by
reactors
with
LPHO
lamps.

In
OMC
systems,
the
UV
reactor
remains
on­
line
while
the
lamp
sleeves
are
cleaned.
OMC
systems
have
mechanical
or
physical­
chemical
wipers
that
are
built­
in
to
the
UV
reactor.
The
wipers
are
either
driven
by
screws
attached
to
electric
motors
or
pneumatic
pistons.
Mechanical
wipers
may
consist
of
steel
brush
collars
or
Teflon
®
rings
that
move
along
the
lamp
sleeve.
Physical­
chemical
wipers
have
a
collar
filled
with
cleaning
solution
that
move
along
the
lamp
sleeve.
The
wiper
physically
removes
fouling
on
the
lamp
sleeve
surface
while
the
cleaning
solution
within
the
collar
dissolves
fouling
materials.
UV
reactors
with
MP
lamps
typically
use
wipers
because
the
higher
lamp
temperatures
accelerate
fouling
under
certain
water
qualities.

The
time
between
sleeve
cleaning
will
depend
on
the
rate
of
fouling.
Sleeve
cleaning
can
be
initiated
manually,
at
regular
intervals,
or
triggered
by
a
calculated
UV
dose
or
measured
UV
intensity,
depending
on
the
reactor
control
logic.
In
physical­
chemical
wipers,
solution
replacement
varies
with
the
rate
of
fouling
and
is
on
the
order
of
months.
Replacing
the
cleaning
solution
is
necessary
because
reaction
with
the
foulant
and
dilution
with
water
reduces
the
ability
of
the
cleaning
solution
to
dissolve
the
foulant.

A.
3.5
UV
Intensity
Sensors
UV
intensity
sensors
are
photosensitive
detectors
that
are
used
to
indicate
dose
delivery
by
providing
information
related
to
UV
intensity
at
different
points
in
the
UV
reactor.
UV
intensity
sensors
include
the
following
components
arranged
as
shown
in
Figure
A.
20.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
45
June
2003
 
Monitoring
windows
and
light
pipes
deliver
light
to
the
photodetector.
Monitoring
windows
are
typically
quartz
discs
and
light
pipes
are
cylindrical
probes
made
of
quartz
(
quartz
silica
probe).

 
Diffusers
and
apertures
reduce
the
UV
light
incident
on
the
photodetector,
thereby
reducing
UV
intensity
sensor
degradation.
Diffusers
also
modify
the
UV
intensity
sensor's
angular
response.

 
Filters
limit
the
light
delivered
to
the
diode,
often
restricting
it
to
germicidal
wavelengths.

 
Photodetectors
are
solid­
state
devices
that
produce
a
current
proportional
to
the
irradiance
on
the
detector's
active
surface.
The
responsivity
of
typical
photodetector
to
UV
light
is
on
the
order
of
0.1
to
0.4
milliamps/
mW
(
mA/
mW).

 
Amplifiers
convert
the
output
of
the
photodetector
from
a
low­
level
current
to
a
standardized
output
proportional
to
the
irradiance
(
e.
g.,
converts
intensity
to
a
4
to
20
mA
output
for
use
in
process
control
interfaces).

 
The
housing
of
the
UV
intensity
sensor
protects
the
components
from
the
external
environment.
The
housing
should
be
electrically
grounded
to
shield
the
photodetector
and
amplifier,
thereby
reducing
electrical
noise
and
bias.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
46
June
2003
Figure
A.
20.
Interior
UV
Intensity
Sensor
Schematics
(
courtesy
of
(
a)
Severn
Trent
Services
and
(
b)
Wedeco­
Ideal
Horizons)

Quartz
Window
Filter
Photodetector
Housing
Signal
Output
UV
Light
UVC
Filter
Signal
Amplifier
Printed
Circuit
Boards
Quartz
Silica
Probe
Digital
Signal
Output
Photodetector
UVC
Light
Housing
a.
b.

Note
that
the
sensor
shown
in
Figure
A.
20b
is
cylindrical
in
shape.
All
dimensions
are
standardized
as
detailed
in
the
German
standards
for
UV
disinfection.

A.
3.5.1
UV
Intensity
Sensor
Properties
UV
intensity
sensor
properties
that
impact
the
measurement
of
UV
intensity
and
dose
delivery
monitoring
include
angular
response,
acceptance
angle,
spectral
response,
working
range,
detection
limit
and
resolution,
linearity,
temperature
response,
long
term
drift,
calibration
factor,
and
measurement
uncertainty.
An
ideal
UV
intensity
sensor
will
have
a
linear
response
over
the
working
range,
provide
a
response
unaffected
by
ambient
temperature,
be
stable
over
time,
have
zero
measurement
noise
and
bias,
respond
only
to
germicidal
UV
light,
and
have
zero
measurement
uncertainty.

Angular
response
is
a
plot
of
the
sensor
measurement
as
a
function
of
the
incident
angle
of
UV
light
at
the
sensor's
window.
Angular
response
is
affected
by
the
UV
intensity
sensor's
aperture
size,
the
size
of
the
photodetector's
active
surface,
the
distance
between
the
aperture
and
the
active
surface,
and
the
impact
of
any
diffusers
and
reflecting
surfaces
within
the
UV
intensity
sensor.
An
ideal
sensor
has
a
cosine
response
(
Equation
A.
24)
because
a
cosine
response
results
in
an
accurate
measure
of
the
light
incident
on
the
surface
of
the
photodetector.

(
)
I
I
m
i
=
cos
 
Equation
A.
24
where
Im
=
Intensity
measured
by
photodetector
Ii
=
Intensity
incident
on
photodetector's
surface
 
=
Incident
angle
at
the
photodetector
surface
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
47
June
2003
In
practice,
sensors
deviate
from
cosine
response;
some
potential
responses
are
shown
in
Figure
A.
21.

Figure
A.
21.
Angular
Response
of
Two
UV
Intensity
Sensors
Relative
to
Ideal
Cosine
Response
0.0
0.2
0.4
0.6
0.8
1.0
­
90
­
60
­
30
0
30
60
90
Incident
Angle
(
degrees)
Angular
Response
(
Relative
Scale)
Cosine
Sensor
1
Sensor
2
The
opening
or
acceptance
angle
of
the
UV
intensity
sensor
is
the
angle
over
which
the
sensor
detects
UV
light.
The
opening
angle
is
typically
measured
by
either
the
threshold
detection
of
UV
light
or
detection
at
some
percentage
of
the
maximum
detection
value
(
e.
g.,
50
percent).
The
acceptance
angle
is
a
characteristic
of
the
sensor
but
does
not
affect
sensor
performance.

The
spectral
response
is
a
measure
of
the
output
of
the
UV
intensity
sensor
as
a
function
of
wavelength.
The
sensor
spectral
response
depends
on
the
response
of
the
photodetector
and
filters
and
the
UVT
of
the
monitoring
windows,
light
pipes,
and
filters.
Some
sensors
use
filters
to
limit
the
spectral
response
to
the
wavelengths
within
the
germicidal
range
(
200
to
300
nm)
because
it
can
be
advantageous
for
sensors
to
only
respond
to
UV
light
that
causes
damage
to
microorganisms.

The
working
range
of
the
UV
intensity
sensor
is
the
range
that
the
sensor
is
able
to
measure.
The
low
end
of
the
working
range
is
defined
by
the
detection
limit
of
the
measurement.
The
high
end
of
the
measurement
range
is
limited
by
the
saturation
of
the
photodetector
and
the
amplifier.
Saturation
is
the
point
at
which
the
sensor
can
no
longer
respond
to
an
increase
in
intensity.

The
detection
limit
of
the
UV
intensity
sensor
is
the
lowest
UV
intensity
that
can
be
detected
and
quantified
at
a
known
confidence
level.
The
detection
limit
is
calculated
as
a
confidence
of
repeated
measurements
of
low
intensity
UV
light,
usually
at
a
specific
percentage
confidence
interval.
The
measurement
resolution
is
the
smallest
difference
in
UV
intensity
that
can
be
differentiated
at
a
given
confidence
limit.
The
detection
limit
and
the
resolution
depend
on
the
measurement
noise
and
on
any
digitalization
of
the
analog
output
from
the
UV
intensity
sensor
by
the
system's
electronics.
The
measurement
noise
is
the
root­
mean­
square
(
RMS)
of
the
random
variation
in
the
sensor
measurement
over
time.
The
measurement
bias
is
the
timeaveraged
sensor
measurement
obtained
with
no
incident
light.
The
measurement
bias
and
noise
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
48
June
2003
of
a
photodetector
are
increased
by
electromagnetic
fields
within
the
UV
reactor
if
the
sensor
is
not
properly
shielded
and
grounded.

An
ideal
UV
intensity
sensor
responds
proportionally
to
the
intensity
incident
on
the
sensor
(
Figure
A.
22).
The
linearity
of
the
UV
intensity
sensor
is
a
measure
of
the
deviation
of
the
sensor
response
from
that
proportional
relationship.
Linearity
is
affected
by
bias
and
saturation.
The
linearity
is
reported
as
the
ratio
of
the
measured
response
to
the
known
incident
intensity,
usually
at
a
specific
percentage
confidence
interval.

Figure
A.
22
Example
of
Sensor
Linearity
0
1
2
3
4
5
6
7
0
1
2
3
4
5
6
7
Incident
Intensity
(
mW/
cm2)
Sensor
Reading
(
mW/
cm2)

Ideal
Response
Region
of
Saturation
Bias
Region
of
Linear
Response
0
1
2
3
4
5
6
7
0
1
2
3
4
5
6
7
Incident
Intensity
(
mW/
cm2)
Sensor
Reading
(
mW/
cm2)

Ideal
Response
Region
of
Saturation
Bias
Region
of
Linear
Response
UV
intensity
sensor
measurement
is
also
affected
by
ambient
temperature.
The
changes
in
sensor
response
arise
from
thermal
expansion
of
the
optical
components,
the
photodetector,
and
the
amplifier.
UV
intensity
sensor
electronics
can
compensate
to
reduce
the
effects
of
temperature.

The
long­
term
drift
of
the
UV
intensity
sensor
is
the
change
in
response
as
a
function
of
time.
Exposure
to
UV
light
damages
optical
and
electronic
components
within
the
UV
intensity
sensor.
The
damage
caused
by
UV
light
is
typically
greater
at
higher
UV
intensities
and
lower
wavelengths.
Degradation
of
the
filter
can
increase
the
filter's
bandwidth
(
the
wavelength
range
passed
by
the
filter),
thereby
increasing
the
UV
intensity
sensor
measurement
even
though
the
UV
lamp
output
has
not
increased.
Degradation
of
the
monitoring
windows
and
light
pipes
may
cause
a
decrease
in
UVT
due
to
solarization.
Off­
gassing
from
damaged
components
can
coat
optical
components,
reducing
the
measured
intensity.

The
calibration
factor
of
the
UV
intensity
sensor
is
a
value
used
to
convert
the
standard
electrical
output
of
the
UV
intensity
sensor
(
mA
or
volts)
to
UV
intensity
(
mW/
cm2
or
W/
m2).
The
calibration
factor
is
the
ratio
of
the
known
intensity
of
the
UV
light
to
the
electrical
output
of
the
sensor.
Sensors
used
in
UV
reactors
equipped
with
LP
or
LPHO
lamps
are
calibrated
with
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
49
June
2003
UV
light
at
254
nm.
Sensors
used
in
UV
reactors
equipped
with
MP
lamps
can
either
be
calibrated
with
light
only
at
254
nm
or
can
be
calibrated
with
polychromatic
UV
light
from
lamp.
a
MP
he
uncertainty
of
a
UV
intensity
sensor
represents
the
difference
in
intensity
between
that
me
A.
3.6
UV
Transmittance
Monitors
As
stated
previously,
UVT
is
an
important
parameter
in
determining
the
efficiency
of
UV
disinfec
general,
commercial
on­
line
UVT
monitors
calculate
transmittance
by
measuring
UV
intensi
P
ce
Figure
A.
23.
UV
Transm
T
asured
by
the
sensor
and
an
accepted
reference
sensor.
This
uncertainty
incorporates
the
uncertainty
that
arises
due
to
variability
in
calibration,
linearity,
spectral
response,
angular
response,
temperature
response,
and
long­
term
drift.

tion.
Therefore,
monitoring
UVT
(
or
UV
absorbance,
A254,
to
calculate
UVT)
is
critical
to
the
success
of
a
UV
disinfection
application.
UVT
can
be
determined
either
by
grab
samples
with
a
laboratory
instrument
or
by
an
on­
line
instrument.
Several
commercial
UV
reactors
use
the
measurement
of
UVT
to
help
monitor
and
control
the
calculated
UV
dose
in
the
reactor.

In
ty
at
various
distances
from
a
lamp.
One
such
monitor
is
schematically
displayed
in
Figure
A.
23.
In
this
monitor,
a
stream
of
water
passes
through
a
cavity
containing
a
short
L
lamp
with
three
UV
intensity
sensors
located
at
various
distances
from
the
lamp.
The
differen
in
sensor
readings
is
used
to
calculate
UVT.

ittance
Monitor
Design
(
courtesy
of
Severn
Trent
Services)

UV
Lamp
UV
Intensity
Sensor
Inlet
Outlet
A
B
C
UV
Lamp
UV
Intensity
Sensor
Inlet
Outlet
A
B
C
UV
Lamp
UV
Intensity
Sensor
Inlet
Outlet
A
B
C
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
50
June
2003
A.
3.7
Temperature
Sensors
Energy
input
per
unit
volume
is
relatively
high
for
a
UV
reactor.
The
water
flowing
through
a
reactor
efficiently
absorbs
the
waste
heat
and
maintains
operating
temperatures
within
a
desirable
range.
Nevertheless,
temperatures
can
become
elevated
under
the
following
circumstances:

 
Water
level
in
the
reactor
drops
and
lamps
are
exposed
to
air.

 
Water
stops
flowing
in
the
reactor.

Most
temperature
sensors
are
located
at
the
top
of
the
UV
reactor.
The
temperature
sensor
can
either
measure
the
water
temperature
or
the
reactor
shell
temperature.
In
either
case,
if
the
temperature
exceeds
a
setpoint
value,
it
will
register
a
high­
temperature
alarm.
The
temperature
alarms
can
be
integrated
into
a
supervisory
control
and
data
acquisition
(
SCADA)
system
such
that
the
alarm
results
in
an
operations
change
to
reduce
the
potential
for
lamp
breakage.
For
instance,
the
reactor
can
be
shut
down
or
valves
can
open
or
close
to
change
the
flow
of
water
to
the
reactor.

A.
3.8
Reactor
Configuration
This
section
describes
the
configuration
of
UV
lamps
and
UV
intensity
sensors
as
well
as
the
hydraulic
considerations
of
the
overall
reactor
design.

A.
3.8.1
Lamp
Placement
The
lamp
configuration
in
a
reactor
is
designed
to
optimize
dose
delivery.
UV
lamps
may
be
oriented
parallel,
perpendicular,
or
diagonal
to
flow.
Depending
on
the
installation
of
the
reactor,
this
can
result
in
lamps
oriented
horizontally,
vertically,
or
diagonally
relative
to
the
ground.
Orienting
MP
lamps
horizontal
relative
to
the
ground
prevents
overheating
at
the
top
of
the
lamps
and
reduces
the
potential
for
lamp
breakage
due
to
temperature
differentials.

In
a
reactor
with
a
square­
cross
section,
typically
lamps
are
placed
with
lamp
arrays
perpendicular
to
flow.
This
pattern
may
be
staggered
to
improve
disinfection
efficiency.
With
a
circular
cross­
section,
lamps
typically
are
evenly
spaced
on
one
or
more
concentric
circles
parallel
to
flow.
The
water
layer
between
lamps
and
between
the
lamps
and
the
reactor
wall
influences
dose
delivery.
If
the
water
layer
is
too
small,
the
reactor
wall
and
adjacent
lamps
will
absorb
UV
light.
If
the
water
layer
is
too
large,
water
will
pass
through
regions
of
lower
UV
intensity
and
experience
a
lower
UV
dose.
The
optimal
spacing
between
lamps
depends
on
the
UVT
of
the
water,
the
output
of
the
lamp,
and
the
degree
of
hydraulic
mixing
within
the
reactor.

A.
3.8.2
UV
Intensity
Sensor
Placement
UV
intensity
sensors
may
be
located
to
view
either
one
or
more
lamps.
The
measurement
of
UV
intensity
from
a
given
lamp
depends
on
the
following
conditions:

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
51
June
2003
 
Output
of
the
lamp
 
UVT
of
the
water
 
Distance
from
the
lamp
to
the
UV
intensity
sensor
 
Incident
angle
of
the
light
on
the
UV
intensity
sensor
As
such,
a
given
measurement
by
a
UV
intensity
sensor
viewing
more
than
one
lamp
can
have
many
interpretations,
and
such
measurements
should
be
properly
understood
to
avoid
misinterpretation.
Also,
UV
intensity
sensors
may
be
located
to
view
the
output
from
the
center
or
ends
of
the
lamp.
The
optimal
sensor
placement
will
give
a
representative
or
conservative
measure
of
the
lamp
output,
given
that
lamp
aging
and
sleeve
fouling
is
non­
uniform
along
the
length
of
the
lamp.

The
number
of
UV
intensity
sensors
used
in
a
reactor
can
vary
from
one
per
lamp
to
one
per
reactor.
The
appropriate
number
of
sensors
will
depend
on
the
type
of
lamp
used,
the
variance
in
lamp­
to­
lamp
output
(
especially
after
the
lamps
have
aged),
and
the
impact
of
that
variance
on
dose
delivery
and
dose
monitoring.
The
implications
of
the
number
of
sensors
used
per
reactor
are
discussed
in
the
background
to
the
validation
protocol
(
section
F.
3.5)

UV
intensity
sensors
may
view
the
lamps
either
from
a
UV
intensity
sensor
port
located
on
the
reactor
wall
or
through
a
lamp
sleeve
located
within
the
reactor.
UV
intensity
sensors
are
classified
as
wet
or
dry.
A
dry
UV
intensity
sensor
views
the
UV
light
through
a
monitoring
window
as
shown
in
Figure
A.
20b.
A
wet
UV
intensity
sensor
is
in
direct
contact
with
the
water
flowing
through
the
reactor
and
is
shown
in
Figure
A.
20a.
While
checking
the
on­
line
UV
intensity
sensor
with
a
reference
UV
intensity
sensor
is
easier
with
a
separate
monitoring
window,
condensation
on
the
window
can
interfere
with
the
measurement
of
UV
intensity.

A.
3.8.3
Hydraulic
Considerations
The
flow
through
UV
reactors
is
turbulent
with
residence
times
on
the
order
of
tenths
of
a
second
for
MP
lamps
or
seconds
for
LP
lamps.
In
theory,
optimal
dose
delivery
by
a
UV
reactor
is
obtained
with
plug
flow
hydrodynamics
and
complete
mixing
perpendicular
to
the
flow.
In
practice,
however,
UV
reactors
do
not
have
these
ideal
hydrodynamics.

Lamp
placement,
inlet
and
outlet
conditions,
baffles,
and
mixers
all
affect
hydrodynamics
within
a
reactor.
Turbulence
and
eddies
form
in
the
wake
behind
lamp
sleeves
oriented
perpendicular
to
flow.
Staggered
lamp
arrays
promote
mixing
within
the
reactor,
thereby
minimizing
short­
circuiting
of
flow.

Inlet
and
outlet
conditions
can
have
a
significant
impact
on
reactor
hydrodynamics.
Ninety­
degree
inlet
and
outlets
promote
short­
circuiting,
eddies,
and
dead
zones
within
the
reactor.
Straight
inlet
conditions
with
gradual
changes
in
cross
sectional
area
can
be
used
to
develop
flow
for
optimal
dose
delivery.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
52
June
2003
Some
manufacturers
insert
baffles
to
improve
hydrodynamics
in
the
reactor.
Perforated
plates
can
be
used
to
even
the
flow
throughout
the
reactor's
cross­
section.
Plates
with
a
single
opening
are
used
to
direct
flow
towards
high
intensity
regions
within
the
reactor.
Mixers
used
within
reactors
are
designed
to
promote
either
turbulent
or
vortex
mixing.

Improvements
to
the
hydrodynamics
through
the
reactor
are
often
obtained
at
the
expense
of
headloss.
Perforated
baffle
plates
and
turbulent
mixers
can
increase
dose
delivery
but
will
significantly
increase
headloss.
However,
inlet
and
outlet
conditions
surrounding
the
reactor
can
be
changed
to
reduce
headloss
without
changing
the
disinfection
effectiveness
within
the
reactor.
Also,
using
vortex
mixers
allows
the
spacing
between
lamps
to
increase,
thereby
reducing
headloss
through
the
reactor.

A.
3.9
Monitoring
UV
Disinfection
Performance
Some
method
of
monitoring
the
performance
of
an
operating
UV
installation
is
required
to
demonstrate
to
the
utility
and
primacy
agency
that
adequate
disinfection
is
being
achieved
(
40
CFR
141.729(
d)).
Because
the
concentration
of
pathogenic
organisms
cannot
be
measured
continuously
in
the
UV­
treated
water
and
the
dose
cannot
be
measured
directly
in
real
time,
various
strategies
have
been
developed
to
demonstrate
adequate
dose
delivery.
Any
dose
monitoring
method
selected
must
be
evaluated
during
reactor
validation
(
described
in
Appendix
C)
and
the
outputs
measured
during
validation
will
be
part
of
the
monitoring
setpoints.

Currently,
there
are
three
fundamental
approaches
to
monitor
UV
disinfection
performance
in
a
UV
reactor:

1.
UV
Intensity
Setpoint
Approach.
In
this
approach,
measurements
made
by
the
UV
intensity
sensor
are
used
to
control
the
UV
reactor.
The
UV
intensity
sensor
is
located
in
a
position
that
allows
it
to
properly
respond
to
both
changes
in
UV
intensity
output
of
the
lamps
and
also
UVT
of
the
water.
The
UV
intensity
sensor
output
and
the
flowrate
are
used
to
monitor
dose
delivery.
The
setpoint
value
for
UV
intensity
over
a
range
of
flowrates
is
determined
during
validation
(
see
Chapter
4).

2.
UV
Intensity
and
UVT
Setpoint
Approach.
This
approach
is
similar
to
the
UV
intensity
sensor
setpoint
approach,
except
that
the
UV
sensor
is
placed
close
to
the
lamp
such
that
it
only
responds
to
changes
in
UV
lamp
output.
UVT
is
monitored
separately.
For
a
specific
flowrate,
the
UV
intensity
and
UVT
measurements
are
used
to
monitor
dose
delivery.
The
setpoints
for
UV
intensity
and
UVT
over
a
range
of
flowrates
are
determined
during
validation
(
see
Chapter
4).

3.
Calculated
UV
Dose
Approach.
In
this
approach,
the
UV
intensity
sensor
is
placed
close
to
the
lamp,
which
is
similar
to
the
UV
intensity
and
UVT
setpoint
approach.
Flowrate,
UVT,
and
UV
intensity
are
all
monitored,
and
the
outputs
are
used
to
calculate
UV
dose
via
a
validated
computational
algorithm
developed
by
the
UV
reactor
manufacturer.

The
strategy
for
dose
monitoring
depends
on
the
manufacturer
and
is
typically
proprietary.
Dose
monitoring
recommendations
are
discussed
in
section
5.4.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
53
June
2003
A.
4
Water
Quality
Impacts
and
Byproduct
Formation
Constituents
in
the
water
affect
the
performance
of
UV
reactors.
In
addition,
most
disinfectants
form
byproducts,
and
the
goal
of
the
overall
disinfection
process
is
to
maximize
disinfection
while
minimizing
byproduct
formation.

A.
4.1
Water
Quality
Impacts
UV
absorbance,
particle
content,
and
constituents
that
foul
lamp
sleeves
and
other
wetted
components
are
the
most
significant
water
quality
factors
impacting
UV
disinfection
effectiveness.
In
spite
of
these
effects,
the
impact
of
water
quality
on
dose
delivery
can
be
adequately
addressed
in
virtually
all
drinking
water
applications
if
carefully
considered
during
the
design
of
the
UV
reactors.

A.
4.1.1
UV
Absorbance
The
most
important
water
quality
parameter
affecting
reactor
performance
is
UV
absorbance.
As
UV
absorbance
increases,
the
intensity
throughout
the
reactor
decreases
for
a
given
lamp
output.
This
results
in
a
reduction
in
UV
dose
delivered
to
the
microorganism
and
measured
UV
intensity.
Section
3.1.3.1
discusses
how
to
incorporate
the
impact
of
UV
absorbance
into
UV
reactor
design.

UV
absorbers
in
typical
source
waters
include
humic
and
fulvic
acids,
other
aromatic
organics
(
e.
g.,
phenols),
metals
(
e.
g.,
iron),
and
anions
(
e.
g.,
nitrates
and
sulfites)
(
Yip
and
Konasewich
1972;
DeMers
and
Renner
1992).
Both
soluble
and
particulate
forms
of
these
compounds
will
absorb
UV
light.
UV
absorbance
will
vary
over
time
due
to
changing
concentrations
of
these
compounds.
Temporal
variability
in
UV
absorbance
is
greater
in
rivers
and
small
lakes
than
in
large
lakes
and
reservoirs.
UV
absorbance
will
vary
seasonally
due
to
rainfall,
lake
stratification
and
destratification
(
turnover),
and
changes
in
biological
activity
of
organisms
within
the
water
source.

Water
treatment
processes
can
reduce
the
UV
absorbance
of
water.
Coagulation,
flocculation,
and
sedimentation
remove
soluble
and
particulate
material,
and
filtration
removes
particles.
Oxidants
such
as
chlorine
and
ozone
reduce
soluble
material,
precipitate
metals,
and
reduce
UV
absorbance.
Activated
carbon
absorption
also
reduces
soluble
organics.
Because
these
treatment
processes
reduce
UV
absorbance,
the
lowest
UV
absorbance
occurs
at
the
end
of
the
treatment
train,
and
therefore
UV
disinfection
is
most
effective
when
applied
after
filtration.
Chemicals
used
in
the
water
treatment
process
can
also
increase
the
UV
absorbance
of
the
water,
and
their
impacts
are
discussed
in
section
A.
4.1.3.

A.
4.1.2
Particles
Particle
content
can
also
impact
UV
disinfection
performance.
Particles
may
absorb
and
scatter
light,
thereby
reducing
the
UV
intensity
delivered
to
the
organisms.
Particle­
associated
microorganisms
also
may
be
shielded
from
UV
light,
effectively
reducing
disinfection
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
54
June
2003
performance
as
discussed
in
section
A.
2.6.5
and
causing
a
tailing
or
flattening
of
the
doseresponse
curve
when
higher
inactivation
levels
are
desired.
Particles
in
source
waters
are
diverse
in
composition
and
size
and
include
large
molecules,
microorganisms,
clay
particles,
algae,
and
flocs.
Sources
of
particles
include
wastewater
discharges,
erosion,
runoff,
microbial
growth,
and
animal
waste.
The
particle
concentration
will
vary
over
time
both
seasonally
and
over
the
short
term.
Storm
events,
lake
turn
over,
and
spring
runoff
are
some
events
that
increase
the
concentration
of
particles.

Recent
research
by
Linden
et
al.
(
2002b)
indicated
that
the
UV
dose­
response
of
microorganisms
added
to
filtered
drinking
waters
is
not
altered
by
variation
in
turbidity
of
filtered
water
that
met
regulatory
requirements
(
40
CFR
141.73).
For
unfiltered
raw
waters,
Passantino
and
Malley
(
2001)
found
that
source
water
turbidity
up
to
10
NTU
does
not
impact
the
UV
dose­
response
of
separately
added
(
seeded)
organisms.
In
these
experiments,
however,
organisms
were
added
to
waters
containing
various
levels
of
treated
or
natural
turbidity.
Therefore,
it
was
not
possible
to
examine
microorganisms
associated
directly
with
particles
in
their
natural
or
treated
states.
Consequently,
these
drinking
water
studies
can
only
suggest
the
impact
of
turbidity
on
dose­
response
as
it
relates
to
the
impact
of
UV
light
scattering
by
particles
rather
than
particle­
association
or
clumping
of
microorganisms.

Water
treatment
unit
processes
such
as
coagulation,
flocculation,
sedimentation,
and
filtration
are
designed
to
remove
particles
from
water.
Organisms
within
coagulated
and
flocculated
particles
will
be
more
difficult
to
inactivate;
however,
they
will
typically
be
removed
during
filtration.

A.
4.1.3
Water
Treatment
Chemicals
Water
treatment
chemicals
affect
the
UVT
of
the
water,
the
formation
of
conglomerate
particles,
and
the
fouling
potential
of
the
water.

Water
treatment
processes
upstream
of
the
UV
reactors
can
be
operated
to
control
and
increase
UVT,
thereby
optimizing
the
design
and
costs
of
the
UV
reactor.
Chemicals
such
as
chlorine,
ozone,
and
hydrogen
peroxide
oxidize
UV­
absorbing
compounds
but
may
also
absorb
UV
light
with
ozone
showing
the
most
pronounced
effect
on
UV
absorbance.
Oxidant
residuals
can
be
quenched
with
chemicals
such
as
sodium
thiosulfate
or
sodium
bisulfite.
However,
the
use
of
these
chemicals
can
also
increase
the
UV
absorbance
of
water.

Table
A.
5
lists
the
UV
absorption
coefficients
of
common
water
treatment
chemicals
and
their
"
impact
threshold
concentration",
defined
as
the
concentration
that
will
decrease
the
UVT
from
91
to
90
percent.
Of
these
chemicals,
ozone
and
ferric
iron
have
the
greatest
potential
of
impacting
the
UV
absorbance
of
water.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
55
June
2003
Table
A.
5
UV
Absorbance
Characteristics
of
Common
Water
Treatment
Chemicals
(
Adapted
from
Bolton
et
al.
2001)

Compound
Molar
Absorbtion
Coefficient
(
M­
1
cm­
1)
Mass­
based
Absorbance
(
L/
mg
cm­
1)
Impact
Threshold
Concentration
1
(
mg/
L)
Ozone
(
O3)
(
aqueous)
3,250
0.0677
0.071
Ferric
iron
(
Fe3+)
4,716
0.0844
0.057
Permanganate
(
MnO4
­)
657
0.0055
0.91
Thiosulfate
ion
(
S2O3
2­)
201
0.00178
2.7
Hypochlorite
ion
(
ClO­)
29.5
0.000573
8.4
Hydrogen
peroxide
(
H2O2)
18.7
0.00055
8.7
Ferrous
iron
(
Fe2+)
28
0.0005
9.6
Sulfite
ion
(
SO3
2­)
16.5
0.000206
23
Zinc
ion
(
Zn2+)
1.7
0.000026
187
Ammonia
(
NH3)
NSA
NSA
N/
A
Ammonium
ion
(
NH4
+)
NSA
NSA
N/
A
Calcium
ion
(
Ca2+)
NSA
NSA
N/
A
Hydroxide
ion
(
OH­)
NSA
NSA
N/
A
Magnesium
ion
(
Mg2+)
NSA
NSA
N/
A
Manganese
ion
(
Mn2+)
NSA
NSA
N/
A
Phosphate
species
NSA
NSA
N/
A
Sulfate
ion
(
SO4
2­)
NSA
NSA
N/
A
NSA
No
significant
absorbance
N/
A
Not
applicable
1
Concentration
in
mg/
L
resulting
in
UVT
decrease
from
91
percent
to
90
percent
(
A254
increase
from
0.041
cm­
1
to
0.046
cm­
1)

A.
4.1.4
Fouling
Potential
Wetted
components
within
a
UV
reactor
can
become
fouled
over
time.
Fouling
on
the
external
surfaces
of
the
lamp
sleeve
reduces
the
transmittance
of
UV
light
from
the
lamps
into
the
water,
thereby
reducing
dose
delivery.
Fouling
on
UV
intensity
sensor
windows
reduces
the
intensity
of
UV
light
measured
by
the
sensors,
resulting
in
under
prediction
of
dose
delivery.
Fouling
on
the
inside
surfaces
of
the
reactor
reduces
reflection
of
UV
light
from
those
surfaces,
which
reduces
the
amount
of
UV
light
available
for
disinfection.

Fouling
on
the
wetted
surfaces
of
a
UV
reactor
has
been
attributed
to
the
following
events:

 
Compounds
whose
solubility
decreases
as
temperature
increases
will
precipitate
(
e.
g.,
CaCO3,
CaSO4,
MgCO3,
MgSO4,
FePO4,
FeCO3,
Al2(
SO4)
3).
These
compounds
will
foul
MP
lamps
faster
than
LP
lamps
due
to
differences
in
operating
temperature.

 
Compounds
with
low
solubility
will
precipitate
(
e.
g.,
Fe(
OH)
3,
Al(
OH)
3).

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
56
June
2003
 
Particles
will
deposit
on
the
lamp
sleeve
surface
due
to
gravity
settling
and
turbulence­
induced
collisions
(
Lin
et
al.
1999a).

Precipitation
will
depend
on
the
water
temperature,
pH,
alkalinity,
ion
concentration,
total
hardness,
and
the
particle
concentration.
Residual
concentrations
of
coagulants
like
ferric
sulfate
can
also
affect
fouling.
The
fouling
will
vary
spatially
along
and
around
the
lamp
sleeve,
and
will
depend
on
the
operating
temperature
of
the
lamp.
Precipitation
of
compounds
whose
solubility
decreases
with
increasing
temperature
is
more
notable
with
lamps
operating
at
higher
temperatures
(
e.
g.,
MP
lamps;
Sheriff
and
Gehr
2001).
Organic
fouling
can
occur
when
a
reactor
is
left
off
and
full
of
water
for
an
extended
period
of
time
(
Toivanen
2000).

Fouling
rates
on
lamp
sleeves
are
reported
to
follow
first
order
kinetics
after
an
initial
induction
period
(
Lin
et
al.
1999b).
Currently,
there
is
not
sufficient
information
to
predict
quantitatively
the
fouling
based
on
water
quality.
The
potential
for
fouling
and
the
frequency
of
sleeve
cleaning
will
be
site
and
equipment
specific.
The
fouling
observed
during
several
pilot­
and
full­
scale
UV
facilities
is
shown
in
section
2.5.1
(
Table
2.3).

The
Langelier
Saturation
Index
(
LSI)
or
the
calcium
carbonate
precipitation
potential
(
CCPP)
can
be
used
to
help
indicate
fouling
potential.
The
LSI
is
defined
as
the
difference
between
the
pH
of
the
water
and
the
pH
at
which
calcium
and
carbonate
are
at
equilibrium
with
solid
CaCO3.
The
CCPP
is
the
amount
of
calcium
carbonate
that
will
precipitate
when
equilibrium
conditions
in
the
water
have
been
reached.
Both
the
LSI
and
CCPP
are
functions
of
temperature,
pH,
calcium
hardness,
total
dissolved
solids
(
TDS),
and
alkalinity.
For
UV
disinfection,
the
temperature
of
the
lamp
sleeve
surface
should
be
used
to
calculate
the
LSI
and
CCPP.
The
LSI
and
CCPP
will
depend
on
upstream
processes,
such
as
pH
adjustment
and
lime
softening,
and
may
vary
daily
or
seasonally.

A.
4.1.5
Algae
Growth
Visible
light
emitted
from
UV
lamps
may
promote
algae
growth
in
UV
reactors
and
the
surrounding
piping.
Depending
on
the
species,
algae
growth
could
cause
taste
and
odor
problems
in
the
finished
water.
Algae
growth
is
a
greater
issue
with
MP
lamps
than
LP
lamps
because
MP
lamps
produce
more
light
in
the
visible
range.
Algae
growth
also
depends
on
water
temperature,
pH,
and
nutrient
concentration
(
Sterner
and
Grover
1998).

A.
4.2
Disinfection
Byproducts
UV
disinfection
byproducts
(
DBPs)
arise
either
directly
through
photochemical
reactions
or
indirectly
through
reactions
with
products
of
photochemical
reactions.
Photochemical
reactions
will
only
take
place
if
a
chemical
species
absorbs
UV
light,
and
the
resulting
excited
state
reacts
to
form
a
new
species.
The
resulting
concentration
of
new
species
will
depend
on
the
concentration
of
the
reactants
and
the
UV
dose.

When
UV
light
is
absorbed
by
an
atom,
electrons
within
the
atom
are
excited
to
higher
energy
states.
An
excited
atom
may
return
to
its
original
ground
state
releasing
the
absorbed
energy
as
light,
or
it
may
interact
with
other
atoms
forming
or
breaking
bonds.
The
formation
or
Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
57
June
2003
breaking
of
bonds
between
atoms
results
in
the
formation
of
a
new
chemical
species.
Chemical
reactions
promoted
by
light
are
termed
photochemical
reactions.
In
some
cases,
the
products
of
photochemical
reactions
are
radical
species.
Radical
species
may
react
with
other
chemicals
to
form
new
chemical
species
(
i.
e.,
UV
DBPs).

In
drinking
water,
research
has
focused
on
the
impact
of
UV
light
on
the
formation
of
halogenated
DBPs
following
subsequent
chlorination
and
the
transformation
of
organic
material
to
more
degradable
components.
For
ground
water
and
filtered
drinking
water,
UV
disinfection
at
typical
doses
is
not
shown
to
impact
the
formation
of
trihalomethanes
(
THM)
or
haloacetic
acids
(
HAA),
two
categories
of
DBPs
currently
regulated
by
EPA
(
Malley
et
al.
1995;
Kashinkunti
et
al.
2003).
Several
studies
have
shown
low­
level
formation
of
degradable,
nonregulated
DBPs
(
e.
g.,
aldehydes)
as
a
result
of
applying
UV
light
to
wastewater
and
raw
drinking
water
sources.
However,
a
study
performed
with
filtered
drinking
water
indicates
no
significant
change
in
aldehydes,
carboxylic
acids,
or
total
organic
halides
(
TOX)
(
Kashinkunti
et
al.
2003).
The
difference
in
results
can
be
attributed
to
the
difference
in
water
quality,
most
notably
the
higher
concentration
of
organic
material
in
raw
waters
and
wastewaters.

Akhlaq
et
al.
(
1990)
reported
that
UV
doses
of
250
mJ/
cm2
from
an
LP
lamp
do
not
break
down
alginic
acid,
a
model
compound
for
polysaccharides
in
drinking
water.
They
concluded
that
UV
disinfection
does
not
increase
the
assimilable
organic
carbon
(
AOC)
of
drinking
water.
With
UV
doses
ranging
from
18
to
161
mJ/
cm2,
Kruithof
et
al.
(
1989)
reported
no
increase
in
AOC
or
mutagenicity
of
a
granular
activated
carbon
(
GAC)
filtrate.

Malley
et
al.
(
1995)
evaluated
the
impact
of
UV
doses
of
60,
130,
and
200
mJ/
cm2
on
DBP
formation
in
ground
waters
and
treated
surface
waters.
They
reported
no
change
in
pH,
turbidity,
dissolved
organic
carbon,
A254,
color,
nitrate,
nitrite,
bromide,
iron,
or
manganese.
Formaldehyde
increased
from
1.2
to
12.1
µ
g/
L
with
one
highly
colored
ground
water.
Formaldehyde
increased
from
less
than
2
µ
g/
L
up
to
14
µ
g/
L
with
untreated
surface
waters
but
only
2
to
3
µ
g/
L
with
treated
surface
waters.
A
small
but
insignificant
increase
in
AOC
was
observed
with
all
waters.

Zheng
et
al.
(
1999)
observed
an
8
to
17
percent
decrease
in
THM
and
a
9
to
19
percent
increase
in
HAA
when
MP
UV
light
was
applied
at
a
dose
of
2000
mJ/
cm2
after
chlorination.
However,
at
a
lower
dose
of
100
mJ/
cm2,
they
observed
a
1
to
7
percent
decrease
in
THM
and
no
change
in
HAA.

A
low
conversion
of
nitrate
to
nitrite
by
UV
light
has
been
observed
(
approximately
1
percent;
Sharpless
and
Linden
2001).
Von
Sonntag
and
Schuchmann
(
1992)
also
reported
0.001
and
0.072
mg/
L
nitrite
formed
from
50
mg/
L
nitrate
exposed
to
25
mJ/
cm2
from
LP
and
MP
lamps,
respectively.
Conversion
is
lower
with
LP
lamps
than
MP
lamps
because
the
UV
absorbance
of
nitrate
is
higher
below
240
nm
than
it
is
at
254
nm.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
58
June
2003
A.
5
References
Adler,
H.
I.
1966.
The
genetic
control
of
radiation
sensitivity
in
microorganisms.
Advances
in
Radiation
2:
167­
191.

Akhlaq,
M.
S.,
H.
P.
Schuchmann,
and
C.
von
Sonntag.
1990.
Degradation
of
the
polysaccharide
alginic
acid:
a
comparison
of
the
effects
of
UV
light
and
ozone.
Environmental
Science
and
Technology
24:
379­
383.

APHA­
AWWA­
WEF.
1998.
Standard
methods
for
the
examination
of
water
and
wastewater,
20th
ed.
Washington
D.
C.

Battigelli,
D.
A.,
M.
D.
Sobsey,
and
D.
C.
Lobe.
1993.
The
inactivation
of
Hepatitis
A
virus
and
other
model
viruses
by
UV
irradiation.
Water
Science
&
Technology
27,
no.
3­
4:
339­
342.

Bolton,
J.
R.
2000.
Calculation
of
ultraviolet
fluence
rate
distributions
in
an
annular
reactor:
significance
of
refraction
and
reflection.
Water
Research
34,
no.
13:
3315­
3324.

Bolton,
J.
R.
and
K.
G.
Linden.
2003.
Standardization
of
Methods
for
Fluence
(
UV
Dose)
Determination
in
Bench­
scale
UV
Experiments.
ASCE
Journal
of
Environmental
Engineering
129,
no.
3:
209­
215.

Bolton,
J.
R.,
M.
I.
Stefan,
R.
S.
Cushing,
and
E.
Mackey.
2001.
The
importance
of
water
absorbance
on
the
efficiency
of
ultraviolet
disinfection
reactors.
Proceedings
of
the
IUVA
1st
International
Congress,
June
14­
16,
Washington,
D.
C.

Brandt,
C.
L.
and
A.
C.
Giese.
1956.
Photoreversal
of
nuclear
and
cytoplasmic
effects
of
short
ultraviolet
radiation
on
Paramecium
caudatum.
Journal
of
General
Physiology
39,
no.
5:
735­
751.

Cabaj,
A.,
R.
Sommer,
W.
Pribil,
and
T.
Haider.
2001.
What
means
"
dose"
in
UV­
disinfection
with
medium
pressure
lamps.
Ozone
Science
&
Engineering
23:
239­
244.

Chang,
J.
C.
H.,
S.
F.
Osoff,
D.
C.
Lobe,
M.
H.
Dorfman,
C.
M.
Dumais,
R.
G.
Qualls,
and
J.
D.
Johnson.
1985.
UV
inactivation
of
pathogenic
and
indicator
microorganisms.
Applied
and
Environmental
Microbiology
49,
no.
6:
1361­
1365.

Clancy,
J.
L.,
Z.
Bukhari,
T.
M.
Hargy,
J.
R.
Bolton,
B.
W.
Dussert,
and
M.
M.
Marshall.
2000.
Using
UV
to
Inactivate
Cryptosporidium.
Journal
of
the
American
Water
Works
Association
92,
no.
9:
97­
104.

DeMers,
L.
D.
and
R.
C.
Renner.
1992.
Alternative
Disinfection
Technologies
for
Small
Drinking
Water
Systems.
Chapter
5.
American
Water
Works
Association
Research
Foundation,
Denver
Colorado.

Dulbecco,
R.
1950.
Experiments
on
photoreactivation
of
bacteriophages
inactivated
with
ultraviolet
radiation.
Journal
of
Bacteriology
59:
329­
347.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
59
June
2003
GE
Quartz.
2001.
Available
from
Internet
location
http://
www.
gequartz.
com/
en/
lamptube.
htm.
Internet.

Giese
A.
C.,
C.
L.
Brandt,
C.
Jacobson,
D.
C.
Shephard,
and
R.
T.
Sanders.
1954.
The
effects
of
starvation
on
photoreactivation
in
Colpidium
colpoda.
Physiological
Zoology
71:
71­
78.

Gurzadyan,
G.
G.,
D.
N.
Nikogosyan,
P.
G.
Kryukov,
V.
S.
Letokhov,
T.
S.
Balmukhanov,
A.
A.
Belogurov,
and
G.
B.
Zavilgelsky.
1981.
Mechanism
of
high
power
picosecond
laser
UV
inactivation
of
viruses
and
bacterial
plasmids.
Photochemistry
and
Photobiology
33:
835­
838.

Harris,
G.
D.,
V.
D.
Adams,
D.
L.
Sorensen,
and
M.
S.
Curtis.
1987.
Ultraviolet
inactivation
of
selected
bacteria
and
virus
with
photoreactivation
of
the
bacteria.
Water
Research
21,
no.
6:
687­
692.

Hollaender,
A.
and
B.
M.
Duggar.
1936.
Irradiation
of
plant
viruses
and
of
microorganisms
with
monochromatic
light.
III.
resistance
of
the
virus
of
typical
tobacco
mosaic
and
Escherichia
Coli
to
radiation
from
L
3000
To
L
2250
A.
Proc.
N.
A.
S
.22:
19­
24.

Hoyer,
O.
1998.
Testing
performance
and
monitoring
of
UV
systems
for
drinking
water
disinfection.
Water
Supply
16,
no.
1/
2:
424­
429.

Irazoqui,
H.
A.,
J.
Cerda,
and
A.
E.
Cassano.
1973.
Radiation
profiles
in
an
empty
annular
photoreactor
with
a
source
of
finite
spatial
dimensions.
AIChE
Journal
19,
no.
3:
460­
467.

Jacob,
S.
M.
and
J.
S.
Dranoff.
1970.
Light
intensity
profiles
in
a
perfectly
mixed
photoreactor.
AIChE
Journal
16,
no.
3:
359­
363.

Jagger,
J.
1967.
Introduction
to
research
in
ultraviolet
photobiology.
Englewood
Cliffs,
NJ:
Prentice
Hall,
Inc.

Jagger,
J.,
T.
Fossum,
and
S.
McCaul.
1975.
Ultraviolet
irradiation
of
suspensions
of
microorganisms
possible
errors
involved
in
the
estimation
of
average
fluence
per
cell.
Photochemistry
and
Photobiology
21:
379­
382.

Kashimada,
K.,
N.
Kamiko,
K.
Yamamoto,
and
S.
Ohgaki.
1996.
Assessment
of
photoreactivation
following
ultraviolet
light
disinfection.
Water
Science
&
Technology
33,
no.
10­
11:
261­
269.

Kashinkunti,
R.,
K.
G.
Linden,
G.
Shin,
D.
H.
Metz,
M.
D.
Sobsey,
M.
Moran,
and
A.
Samuelson.
2003.
Achieving
multi­
barrier
inactivation
in
Cincinnati:
UV,
byproducts,
and
biostability.
Journal
of
the
American
Water
Works
Association.
Submitted
January
9,
2003.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
60
June
2003
Kelner,
A.
1950.
Light­
induced
recovery
of
microorganisms
from
ultraviolet
radiation
injury,
with
special
reference
to
Escherichia
coli.
Bulletin
of
the
New
York
Academy
of
Medicine
26:
189­
199.

Knudson,
G.
B.
1985.
Photoreactivation
of
UV­
irradiated
Legionella
Pneumoplila
and
other
Legionella
species.
Applied
and
Environmental
Microbiology
49,
no.
4:
975­
980.

Kruithof,
J.
C.,
R.
C.
van
der
Leer,
W.
A.
M.
Hijnen,
P.
A.
N.
M.
Nuhn,
F.
A.
P.
Houtepen,
and
L.
A.
C.
Feij.
1989.
Ultra
violet
disinfection
of
carbon
filtered
drinking
water.
Ozone
+
UV
in
the
treatment
of
water
and
other
liquids.
Edited
by
W.
J.
Masschelein.
III
3.1
 
III3.15.
Berlin
Lin,
L.
S.,
C.
T.
Johnston,
and
E.
R.
Blatchley
III.
1999a.
Inorganic
fouling
at
quartz:
water
interfaces
in
ultraviolet
photoreactors
I:
chemical
characterization.
Water
Research
33,
no.
15:
3321­
3329.

Lin,
L.
S.,
C.
T.
Johnston,
and
E.
R.
Blatchley
III.
1999b.
Inorganic
fouling
at
quartz:
water
interfaces
in
ultraviolet
photoreactors
 
II
Temporal
And
Spatial
Distribution.
Water
Research
33,
no.
15:
3330­
3338.

Linden,
K.
G.
and
J.
L.
Darby.
1998.
UV
disinfection
of
marginal
effluents:
determining
UV
absorbance
and
subsequent
estimation
of
UV
intensity.
Water
Environment
Research
70,
no.
2:
214­
223.

Linden,
K.
G.,
G.
A.
Shin,
and
M.
D.
Sobsey.
2001.
Comparative
effectiveness
of
UV
wavelengths
for
the
inactivation
of
Cryptosporidium
parvum
oocysts
in
water.
Water
Science
&
Technology
43,
no.
12:
171­
174.

Linden,
K.
G.,
G.
A.
Shin,
G.
Faubert,
W.
Cairns,
and
M.
D.
Sobsey.
2002a.
UV
disinfection
of
Giardia
lamblia
cysts
in
water.
Environmental
Science
&
Technology
36,
no.
11:
2519­
2522.

Linden,
K.
G.,
L.
Batch,
and
C.
Schulz.
2002b.
UV
Disinfection
of
filtered
water
supplies:
water
quality
impacts
on
MS2
dose­
response
curves.
Proceedings
of
the
AWWA
Annual
Conference,
June
16­
20,
New
Orleans,
LA.

Lindenauer,
K.
G.
and
J.
L.
Darby.
1994.
Ultraviolet
disinfection
of
wastewater:
effect
of
UV
dose
on
subsequent
photoreactivation.
Water
Research
28:
805.

Loge,
F.
J.,
R.
W.
Emerick,
D.
E.
Thompson,
D.
C.
Nelson,
and
J.
L.
Darby.
1999.
Factors
influencing
ultraviolet
performance.
Part1:
Light
penetration
to
wastewater
particles.
Water
Research
71,
no.
6:
377­
381.

Lytle,
C.
D.
1971.
Host­
cell
reactivtaion
in
mammalian
cells
I.
Survival
of
ultra­
violet­
irradiated
Herpes
virus
in
different
cell­
lines.
International
Journal
of
Radiation
Biology
19,
no.
4:
329­
337.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
61
June
2003
Mackey,
E.
D.,
R.
S.
Cushing,
M.
L.
Janex,
N.
Picard,
J.
M.
Laine,
and
J.
P.
Malley.
2003.
Bridging
Pilot­
Scale
Testing
to
Full­
Scale
Design
of
UV
Disinfection
Systems,
Draft
Report.
Final
report
to
be
published
in
2004.
Denver,
CO:
AWWA
Research
Foundation.

Malley,
J.
P.
2000.
The
state
of
the
art
in
using
UV
disinfection
for
waters
and
wastewaters
in
North
America.
Proceedings
of
ENVIRO
2000
­
Australian
Water
Association
(
AWA),
April
9­
13,
Sydney,
Australia.

Malley,
J.
P,
J.
P.
Shaw,
and
J.
R.
Ropp.
1995.
Evaluation
of
by­
products
produced
by
treatment
of
groundwaters
with
ultraviolet
irradiation.
Denver,
CO.:
AWWA
Research
Foundation.

Meng
Q.
S.
and
C.
P.
Gerba.
1996.
Comparative
inactivation
of
enteric
adenovirus,
poliovirus
and
coliphages
by
ultraviolet
irradiation.
Water
Research
30,
no.
11:
2665­
2668.

Meulemans,
C.
C.
E.
1986.
The
basic
principles
of
UV­
sterilization
of
water.
In:
Ozone
+
Ultraviolet
Water
Treatment,
Aquatec
Amsterdam,
Paris:
International
Ozone
Association,
B.
1.1­
B.
1.13.

Mofidi,
A.
A.,
E.
A.
Meyer,
P.
M.
Wallis,
C.
I.
Chou,
B.
P.
Meyer,
S.
Ramalingam,
and
B.
M.
Coffey.
2002.
The
effect
of
UV
light
on
the
inactivation
of
Giardia
lamblia
and
Giardia
muris
cysts
as
determined
by
animal
infectivity
assay
(
P­
2951­
01).
Water
Research
36:
2098­
2108.

Morton,
R.
A.
and
R.
H.
Haynes.
1969.
Changes
in
the
ultraviolet
sensitivity
of
Escherichia
coli
during
growth
in
batch
cultures.
Journal
of
Bacteriology
97(
3):
1379­
1385.

Oliver,
B.
G.
and
E.
G.
Cosgrove.
1975.
The
disinfection
of
sewage
treatment
plant
effluents
using
ultraviolet
light.
Canadian
Journal
of
Chemical
Engineering
53:
170­
174.

Oppenheimer,
J.
A.,
J.
E.
Hoagland,
J.­
M.
Laine,
J.
G.
Jacangelo,
and
A.
Bhamrah.
1993.
Microbial
inactivation
and
characterization
of
toxicity
and
by­
products
occurring
in
reclaimed
wastewater
disinfected
with
UV
radiation.
Planning,
design
&
operation
of
effluent
disinfection
systems,
13­
24.
Whippany,
N.
J.:
the
Water
Environment
Federation.

Parker,
J.
A.
and
J.
L.
Darby.
1995.
Particle­
associated
coliform
in
secondary
effluents:
shielding
from
ultraviolet
light
disinfection.
Water
Environment
Research
67,
no
7:
1065­
1075.

Passantino,
L.
B.
and
J.
P.
Malley.
2001.
Impacts
of
turbidity
and
algal
content
of
unfiltered
drinking
water
supplies
on
the
ultraviolet
disinfection
process.
Proceedings
of
the
2001
AWWA
Annual
Conference,
Washington
D.
C.

Persson,
E.
and
D.
Kuusisto.
1998.
A
performance
comparison
of
electronic
vs.
magnetic
ballast
for
powering
gas­
discharge
UV
lamps.
Proceedings
of
the
RadTech
Conference,
RadTech
International,
Chicago,
IL.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
62
June
2003
Petri,
M.
P.,
G.
Fang,
J.
P.
Malley,
D.
C.
Moran,
and
H.
Wright.
2000.
Groundwater
UV
disinfection:
challenges
and
solutions.
Proceedings
of
the
Water
Quality
Technology
Conference,
AWWA,
November
5­
9.
Salt
Lake
City,
UT.

Pfeifer,
G.
P.
1997.
Formation
and
processing
of
UV
photoproducts:
effects
of
DNA
sequence
and
chromatin
environment.
Photochemistry
&
Photobiology
65(
2):
270­
283.

Phillips,
R.
1983.
Sources
and
application
of
ultraviolet
radiation.
New
York:
Academic
Press.

Powell,
W.
F.
1959.
Radiosensitivity
as
an
index
of
Herpes
Simplex
virus
development.
Virology
9:
1­
19.

Qualls,
R.
G.,
M.
P.
Flynn,
and
J.
D.
Johnson.
1983.
The
role
of
suspended
particles
in
ultraviolet
disinfection.
Journal
WPCF
55,
no
10:
1280­
1285.

Rauth,
A.
M.
1965.
The
physical
state
of
viral
nucleic
acid
and
the
sensitivity
of
viruses
to
ultraviolet
light.
Biophysical
Journal
5:
257­
273.

Schenck,
G.
O.
1981.
Ultraviolet
sterilization.
Handbook
of
water
purification.
2nd
ed.
Edited
by
W.
Lorch.
New
York:
John
Wiley
&
Sons.

Setlow,
J.
K.
1967.
Comprehensive
biochemistry.
Chapter
5
in
The
effects
of
ultraviolet
radiation
and
photoreactivation
27:
157­
209.

Severin,
B.
F.,
M.
T.
Suidan,
and
R.
S.
Engelbrecht.
1983.
Effects
of
temperature
on
ultraviolet
light
disinfection.
Environmental
Science
&
Technology
17(
12):
717­
721.

Severin,
B.
F.,
M.
T.
Suidan,
and
R.
S.
Engelbrecht.
1984.
Mixing
effects
in
UV
disinfection.
Journal
WPC.
56,
no
7:
881­
888.

Sharpless
C.
M.
and
K.
G.
Linden.
2001.
UV
photolysis
of
nitrate:
effects
of
natural
organic
matter
and
dissolved
inorganic
carbon,
and
implications
for
UV
water
disinfection.
Environmental
Science
&
Technology
35,
no
14:
2949­
2955.

Sheriff,
M.
and
R.
Gehr.
2001.
Laboratory
investigation
of
inorganic
fouling
of
low
pressure
UV
disinfection
lamps.
Water
Quality
Research
Journal
of
Canada
36,
no
1:
71­
92.

Shin,
G­
A.,
K.
G.
Linden,
M.
J.
Arrowood,
and
M.
D.
Sobsey.
2001.
Low­
pressure
UV
inactivation
and
DNA
repair
potential
of
Cryptosporidium
parvum
oocysts.
Applied
&
Environmental
Microbiology
67
no.
7:
3029­
3032.

Snowball,
M.
R.
and
I.
S.
Hornsey.
1988.
Purification
of
water
supplies
using
ultraviolet
light.
Developments
in
food
microbiology,
edited
by
R.
K.
Robinson,
171­
191.
NY:
Elsevier
Applied
Science.

Sommer,
R.,
T.
Haider,
A.
Cabaj,
W.
Pribil,
and
M.
Lhotsky.
1998.
Time
dose
reciprocity
in
UV
disinfection
of
water.
Water
Science
&
Technology
38:
145­
150.

Proposal
Draft
Appendix
A.
Fundamentals
of
UV
Disinfection
UV
Disinfection
Guidance
Manual
A­
63
June
2003
Sterner,
R.
W.
and
J.
P.
Grover.
1998.
Algal
growth
in
warm
temperate
reservoirs:
kinetic
examination
of
nitrogen,
temperature,
light,
and
other
nutrients.
Water
Research
32,
no
12:
3539­
3548.

Toivanen,
E.
2000.
Experiences
with
UV
disinfection
at
Helsinki
water.
IUVA
News
2,
no.
6:
4­
8.

Tosa,
K.
and
T.
Hirata.
1998.
HRWM­
39:
Photoreactivation
of
Salmonella
following
UV
disinfection.
Vol.
10
Health­
Related
Water
Microbiology.
IAWQ
19th
Biennial
International
Conference,
Jun21­
26,
168­
174.

Tree,
J.
A.,
M.
R.
Adams,
and
D.
N.
Lees.
1997.
Virus
inactivation
during
disinfection
of
wastewater
by
chlorination
and
UV
irradiation
and
the
efficacy
of
F+
bacteriophage
as
a
viral
indicator.
Water
Science
&
Technology
35(
11­
12):
227­
232.

von
Sonntag,
C.
and
H­
P.
Schuchmann.
1992.
UV
disinfection
of
drinking
water
and
by­
product
formation
 
some
basic
considerations.
Journal
of
Water
Supply
Research
and
Technology
41,
no
2:
67­
74
.

Wiedenmann,
A.,
B.
Fischer,
U.
Straub,
C.­
H.
Wang,
B.
Flehmig,
and
D.
Schoenen.
1993.
Disinfection
of
Hepatitis
A
virus
and
MS­
2
coliphage
in
water
by
ultraviolet
irradiation:
Comparison
of
UV­
susceptibility.
Water
Science
&
Technology
27,
no
3­
4:
335­
338.

Wilson,
B.
R.,
P.
F.
Roessler,
E.
Van
Dellen,
M.
Abbaszadegan,
and
C.
P.
Gerba.
1992.
Coliphage
MS­
2
as
a
UV
water
disinfection
efficacy
test
surrogate
for
bacterial
and
viral
pathogens,
219­
235.
Proceedings
of
the
water
quality
technology
conference,
Nov.
15­
19,
Toronto.

Wright,
H.
B.
and
G.
Sakamoto.
1999.
UV
dose
required
to
achieve
incremental
log
inactivation
of
bacteria,
virus,
and
protozoa.
Trojan
Technologies,
Inc.,
London,
Ontario,
Canada.

Wright,
H.
B.
and
Y.
A.
Lawryshyn.
2000.
An
assessment
of
the
bioassay
concept
for
UV
reactor
validation.
Disinfection
of
Wastes
in
the
New
Millenium,
New
Orleans,
Louisiana,
March
15­
18,
2000;
Water
Environment
Federation,
Alexandria,
Virginia
Yip,
R.
W.
and
D.
E.
Konasewich.
1972.
Ultraviolet
sterilization
of
water
­
its
potential
and
limitations.
Water
and
Pollution
Control.
14:
14­
18.

Zheng,
M,
S.
A.
Andrews,
and
J.
R.
Bolton.
1999.
Impacts
of
medium
pressure
UV
on
THM
and
HAA
formation
in
pre­
UV
chlorinated
drinking
water.
Water
Quality
Technology
Conference,
October
31­
November
3,
Tampa,
F.
L.

Proposal
Draft
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
In
support
of
the
Long
Term
2
Enhanced
Surface
Water
Treatment
Rule
(
LT2ESWTR),
the
U.
S.
Environmental
Protection
Agency
(
EPA)
developed
UV
dose
requirements
for
Cryptosporidium,
Giardia,
and
virus
inactivation.
The
requirements
represent
the
UV
dose
necessary
to
achieve
a
given
inactivation
level,
similar
to
the
concentration
*
time
(
CT)
requirements
for
chemical
disinfectants.

The
UV
dose
requirements
were
developed
to
account
for
uncertainty
associated
with
the
dose­
response
of
microorganisms
(
Cryptosporidium,
Giardia,
and
virus)
in
controlled
experimental
conditions.
In
practical
application,
other
sources
of
variability
and
uncertainty
are
introduced
due
to
the
hydraulic
effects
of
the
UV
installation,
UV
reactor,
and
UV
intensity
sensors.
The
validation
protocol,
as
described
in
Chapter
4
and
Appendix
C,
addresses
these
and
other
areas
of
variability
and
uncertainty
by
applying
safety
factors
to
the
UV
dose
requirements
derived
in
this
appendix.
Therefore,
the
dose
requirements
presented
in
this
appendix
are
not
the
actual
dose
levels
at
which
utilities
will
be
required
to
validate
and
operate
UV
reactors
for
a
given
log
inactivation.

This
appendix
explains
the
derivation
of
the
UV
dose
requirements
through
a
three­
step
process
of
data
collection,
qualitative
review
to
establish
working
data
sets,
and
mathematical
analyses.

B.
1
Data
Collection
EPA
collected
UV
dose­
response
research
data
for
adenovirus,
Giardia
lamblia,
Giardia
muris,
and
Cryptosporidium
parvum.
Adenovirus
was
evaluated
because,
of
the
data
available,
it
is
considered
the
most
resistant
to
inactivation
by
UV
light
of
the
pathogenic
waterborne
viruses.
In
compiling
data,
EPA
reviewed
published
and
unpublished
studies
conducted
over
the
past
50
years
as
provided
in
published
literature,
electronic
databases,
research
reports,
and
conference
proceedings.
The
experimental
conditions
varied
among
batch
and
continuous
flow
UV
apparatuses,
types
of
UV
lamps,
and
water
quality
conditions.
Table
B.
1
summarizes
these
studies.

UV
Disinfection
Guidance
Manual
B­
1
June
2003
Proposal
Draft
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
Table
B.
1
Summary
of
Data
Collected
(
cont)

General
Microbial
Information
Experimental
Information
Reference
Peer
Reviewed
Literature
(
Y/
N)
Organism
Species
Host
Strain
Type
Assay
Used
Experiment
Type
Lamp
Type
UV
Dose
Measurement
Water
Quality1
Gerba
2000
No
Adenovirus
N/
A
Human
N/
A
2
Cell
Culture
(
PLC/
PRF/
5)
Batch
LP
Radiometer
Lab
Gilead
and
Ginsberg
1966
Yes
Adenovirus
N/
A
Human
N/
A
12
Cell
Culture
(
KB
cells)
Batch
LP
None
Lab
Hara
et
al.

1990
Yes
Adenovirus
N/
A
Human
N/
A
19
Cell
Culture
(
Vero)
Batch
LP
Not
given
Lab
Malley
2000b
No
Adenovirus
N/
A
Human
N/
A
41
Cell
Culture
(
Hep­
2)
and
RT­
PCR
Batch
MP
MP­
Calculated
(
DNA),
Radiometer
Lab
Meng
and
Gerba
1996
Yes
Adenovirus
N/
A
Not
given
N/
A
40
41
Cell
Culture
(
PLC/
PRF/
5)
Batch
LP
Radiometer
Lab
Shin
et
al.

2001a
No
Adenovirus
N/
A
Human
N/
A
5
Cell
Culture
(
A549)
Batch
LP
Radiometer
Lab
Thompson
et
al.
2002
Yes
Adenovirus
N/
A
Human
N/
A
2
15
Cell
Culture
(
A­
549)
Batch
LP
Radiometer
Low
Turbidity
Reclaimed
Wastewater
Thurston
et
al.
2002
No
Adenovirus
N/
A
Human
N/
A
40
Cell
Culture
(
PLC/
PRF/
5)
Batch
LP
Radiometer
Lab
and
Groundwater
UV
Disinfection
Guidance
Manual
B­
2
June
2003
Proposal
Draft
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
Table
B.
1
Summary
of
Data
Collected
(
cont)

General
Microbial
Information
Experimental
Information
Reference
Peer
Reviewed
Literature
(
Y/
N)
Organism
Species
Host
Strain
Type
Assay
Used
Experiment
Type
Lamp
Type
UV
Dose
Measurement
Water
Quality1
Craik
et
al.

2000
Yes
Giardia
muris
Bovine
N/
A
N/
A
Mouse
Infectivity
(
C3H/
HeN)
Batch
MP
MP­
Calculated
(
DNA)
WTP
Filtered
Water
Danielson
et
al.
2001
No
Giardia
muris
Bovine
N/
A
N/
A
Mouse
Infectivity
Batch
LP/
LPHO
Radiometer
Lab
Hayes
et
al.

2001
No
Giardia
muris
Hamster
N/
A
N/
A
Mouse
Infectivity
Batch
LP
Radiometer
Lab
Oppenheimer
et
al.
2002
No
Giardia
muris
Mouse
N/
A
N/
A
Mouse
Infectivity
Batch
LP/
MP
Not
Given
Unfiltered
Campbell
and
Wallis
2002
Yes
Giardia
lamblia
Human
N/
A
N/
A
Gerbil
Infectivity
Batch
LP
Radiometer
Lab
Linden
et
al.

2002
No
Giardia
lamblia
Bovine
N/
A
N/
A
Gerbil
Infectivity
Batch
LP
Radiometer
Lab
Malley
2000a
No
Giardia
lamblia
Bovine
N/
A
N/
A
Gerbil
Infectivity
Batch
P­
UV
Bioassay
Reclaimed
Wastewater
Mofidi
et
al.

2002
Yes
Giardia
lamblia
Bovine
N/
A
N/
A
Mouse
Infectivity
Gerbil
Infectivity
Batch
LP
Radiometer
WTP
Filtered
Water
UV
Disinfection
Guidance
Manual
B­
3
June
2003
Proposal
Draft
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
Table
B.
1
Summary
of
Data
Collected
(
cont)

General
Microbial
Information
Experimental
Information
Reference
Peer
Reviewed
Literature
(
Y/
N)
Organism
Species
Host
Strain
Type
Assay
Used
Experiment
Type
Lamp
Type
UV
Dose
Measurement
Water
Quality1
Bukhari
et
al.

1999
Yes
Crypto.
parvum
Bovine
Iowa
N/
A
Mouse
Infectivity
(
CD­
1)
Continuous
Flow
MP
Math
Model
(
PSS)
WTP
Filtered
Water
Clancy
et
al.

2000
Yes
Crypto.
parvum
Bovine
Iowa
N/
A
Mouse
Infectivity
(
CD­
1)
Batch
MP/

LP
MP­
Calculated
(
DNA)
Radiometer
Lab
/
Backwash
supernatant
recycle
Clancy
Env
2000
No
Crypto.
parvum
Bovine
Iowa
N/
A
Mouse
Infectivity
(
CD­
1)
Batch
LP/
MP
Radiometer
MP­
Calculated
(
DNA)
Unfiltered
Clancy
et
al.

2002
Yes
Crypto.
parvum
Bovine
TAMU
Moredum
Iowa
Maine
Glasgow
N/
A
Mouse
Infectivity
Batch
LP
Radiometer
Lab
Craik
et
al.

2001
Yes
Crypto.
parvum
Bovine
Iowa
N/
A
Mouse
Infectivity
(
CD­
1)
Batch
LP/

MP
Radiometer
­
calculated
Lab
or
WTP
Filtered
Water
Hargy
et
al.

2000
Yes
Crypto.
parvum
Bovine
Iowa
N/
A
Mouse
Infectivity
(
CD­
1)
Continuous
Flow
MP
Math
Model
(
PSS)
Untreated
Surface
Water
Landis
et
al.

2000
No
Crypto.
parvum
Bovine
Iowa
N/
A
Cell
Culture
(
HCT­
8)
Batch
LP
Radiometer
Lab
Mackey
et
al.

2000
No
Crypto.
parvum
Bovine
Iowa
N/
A
Mouse
Infectivity
(
CD­
1)
Continuous
Flow
LPHO
Bioassay
WTP
Filtered
Water
UV
Disinfection
Guidance
Manual
B­
4
June
2003
Proposal
Draft
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
Table
B.
1
Summary
of
Data
Collected
(
cont)

General
Microbial
Information
Experimental
Information
Reference
Peer
Reviewed
Literature
(
Y/
N)
Organism
Species
Host
Strain
Type
Assay
Used
Experiment
Type
Lamp
Type
UV
Dose
Measurement
Water
Quality1
Mofidi
et
al.

1999
No
Crypto.
parvum
Bovine
Iowa
N/
A
Cell
Culture
(
HCT­
8)
and
RT­
PCR
Batch
MP/

P­
UV
MP
­
calculated
Joulemeter
WTP
Filtered
Water
Oppenheimer
et
al.
2002
No
Crypto.
parvum
Bovine
Iowa
N/
A
Mouse
Infectivity
(
CD­
1)
Batch
LP/
MP
Not
Given
Unfiltered
Shin
et
al.

2001b
Yes
Crypto.
parvum
Bovine
Iowa
N/
A
Cell
Culture
(
MDCK)
Batch
LP
Radiometer
Lab
Kashinkunti
et
al.
2002
No
Crypto.
parvum
Bovine
Iowa
N/
A
Cell
Culture
(
MDCK)
Batch
LP
Radiometer
WTP
Filtered
Water
Sommer
et
al.
2001
No
Crypto.
parvum
Bovine
Iowa
N/
A
Cell
Culture
(
HCT­
8)
Batch
LP
Radiometer
WTP
Filtered
Water
N/
A
 
Not
applicable;
LP
 
low
pressure
lamp;
LPHO
 
low
pressure­
high
output
lamp;
MP
 
medium
pressure
1
Water
Quality
Definitions:

Lab
­
Tap
water
treated
in
the
lab
by
de­
ionization
and
buffering
(
in
some
cases).

Reclaimed
Wastewater
­
Tertiary
treated
wastewater.

Low
Turbidity
Wastewater
­
Tertiary
treated
wastewater
with
turbidity
less
than
1
NTU.

Unfiltered
­
Water
that
meets
EPA's
filtration
avoidance
criteria.

Untreated
Surface
Water
­
Water
from
an
untreated
surface
water
(
e.
g.,
lake,
river).

WTP
Filtered
Water
 
Filtered
water
from
a
water
treatment
plant.

Lab
Filtered
Surface
Water
­
Filtered
water
from
a
water
treatment
plant
that
is
filtered
subsequently
in
the
lab
UV
Disinfection
Guidance
Manual
B­
5
June
2003
Proposal
Draft
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
UV
Disinfection
Guidance
Manual
B­
6
June
2003
B.
2
Data
Review
 
Criteria
for
Inclusion
in
Statistical
Analysis
EPA
evaluated
the
data
presented
in
Table
B.
1
to
determine
the
data
sets
to
be
used
in
analyzing
dose­
response
for
each
target
microorganism.
To
be
included
in
the
statistical
analysis,
the
experimental
design
had
to
be
sufficiently
documented
with
respect
to
experimental
conditions,
methodology,
and
calculation
of
results
to
allow
an
accurate
assessment
of
UV
dose­
response.
For
instance,
studies
were
not
included
if
the
report
did
not
provide
sufficient
information
to
determine
the
UV
dose
measurement
method
or
whether
the
reported
UV
dose
accounted
for
appropriate
parameters
(
e.
g.,
UV
absorbance).
The
statistical
dose­
response
analysis
combines
data
across
different
experimental
designs
and
conditions;
therefore,
it
is
important
to
ensure
the
differences
between
studies
do
not
affect
the
UV
dose­
response
relationship.

B.
2.1
Appropriate
Experimental
Design
and
Conditions
Research
studies
with
the
following
criteria
were
selected
for
the
statistical
analyses:

 
Batch
experimental
design
 
Low
pressure
(
LP)
lamps
as
the
UV
light
source
 
Filtered
water,
high
quality
unfiltered
water,
laboratory
water,
or
low
turbidity
reclaimed
wastewater
 
UV
dose
of
the
target
microorganism
inactivation
directly
measured
and
not
derived
from
the
inactivation
response
of
another
microorganism
Data
from
continuous
flow
studies
were
not
included
in
the
analyses
because
flowthrough
UV
reactors
apply
a
distribution
of
UV
doses
as
opposed
to
a
single
dose.
Moreover,
UV
dose
in
a
reactor
is
difficult
to
calculate
precisely
due
to
the
variability
in
hydraulic
detention
time
and
UV
intensity
distributions
in
reactors.

Studies
were
not
included
if
the
researchers
utilized
a
UV
light
source
that
did
not
have
a
widely
accepted
dose
measurement
methodology,
such
as
pulsed
UV
lamps.
Medium
pressure
(
MP)
lamps
pose
a
challenge
of
dose
measurement
due
to
the
polychromatic
nature
of
the
MP
UV
light
and
the
absence
of
a
standard
method
for
calculating
dose
from
MP
lamps.
The
results
of
a
t­
test
indicated
the
LP
and
MP
UV
dose­
response
data,
as
reported,
were
statistically
different1;
therefore,
only
LP
lamp
data
were
used
in
the
statistical
analyses.

Given
the
potential
interference
of
particles
in
water
and
the
fact
that
utilities
installing
UV
disinfection
would
need
to
meet
finished
water
turbidity
levels,
EPA
restricted
media
to
water
with
turbidity
values
less
than
or
equal
to
1
nephelometric
turbidity
unit
(
NTU).

1
The
t­
test
was
calculated
with
Cryptosporidium
data
at
low
doses.
The
Giardia
data
and
higher
dose
Cryptosporidium
data
had
too
many
data
reported
as
"
greater
than
a
value"
(
referred
to
as
censored
data)
and
thus,
could
not
be
used
in
a
t­
test.
The
adenovirus
data
had
too
few
MP
data
to
conduct
a
t­
test.

Proposal
Draft
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
UV
Disinfection
Guidance
Manual
B­
7
June
2003
Studies
utilizing
non­
standard
microbial
assay
methods
(
i.
e.,
not
generally
accepted
in
standard
microbiological
methods
references)
or
studies
not
providing
an
evaluation
of
pathogen
infectivity
were
not
included.

Note
that
many
research
studies
evaluated
multiple
experimental
conditions,
but
only
the
subset
of
data
meeting
the
criteria
specified
for
this
statistical
analysis
were
used.

B.
2.2
Research
Studies
and
Data
Included
in
Statistical
Analysis
UV
dose­
response
data
sets
for
adenovirus,
Cryptosporidium
parvum,
and
Giardia
lamblia
and
Giarida
muris
that
met
the
criteria
specified
previously
are
presented
in
this
section.

B.
2.2.1
Viruses
For
adenovirus,
4
of
the
9
studies
met
the
criteria
discussed
for
inclusion
in
the
statistical
analysis.
Figure
B.
1
shows
the
data
of
the
selected
studies.

Figure
B.
1
Observed
Adenovirus
Data
from
Selected
Research
Studies
0
1
2
3
4
5
6
7
0
50
100
150
200
250
300
UV
Dose
(
mJ/
cm2)
Log
Inactivation
Gerba
2000
(
Type
2)
Meng
and
Gerba,
1996
(
Type
40
and
41)
Shin
et
al.,
2001a
(
Type
5)
Thompson
et
al.,
2002
(
Type
2)
Thompson
et
al.,
2002
(
Type
15)
Thompson
et
al.,
2002
(
Type
Not
Avail.)
Thurston
et
al.,
2002
(
Type
40)

Proposal
Draft
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
UV
Disinfection
Guidance
Manual
B­
8
June
2003
B.
2.2.2
Protozoa
For
Cryptosporidium
parvum,
9
of
the
13
studies
met
the
criteria
for
inclusion
in
the
statistical
analysis.
For
Giardia
(
including
both
lamblia
and
muris),
6
of
the
8
studies
were
included.
Figures
B.
2
and
B.
3
show
the
Cryptosporidium
parvum
and
Giardia
data
of
the
selected
studies,
respectively.
The
data
are
both
censored
and
uncensored
and
noted
as
such
on
each
graph.
Censored
data
are
those
with
log
inactivation
of
"
greater
than"
a
particular
value
rather
than
an
absolute
value
(
termed
uncensored).

Figure
B.
2
Cryptosporidium
Data
from
Selected
Research
Studies
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
0
5
10
15
20
25
30
35
40
45
50
UV
Dose
(
mJ/
cm2)
Log
Inactivation
Clancy
et
al.,
2000
Clancy
et
al.,
2000
Clancy
et
al.,
2002
Clancy
et
al.,
2002
Craik
et
al.,
2001
Craik
et
al.,
2001
Landis
et
al.,
2000
Landis
et
al.,
2000
Clancy
Envionmental,
2002
Clancy
Envionmental,
2002
Oppenheimer
et
al.,
2002
Oppenheimer
et
al.,
2002
Shin
et
al.,
2001b
Shin
et
al.,
2001b
Kashinkunti
et
al.,
2002
Kashinkunti
et
al.,
2002
Sommer
et
al.,
2001
Open
symbols
report
Log
Inactivation
Closed
symbols
report
>
Log
Proposal
Draft
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
UV
Disinfection
Guidance
Manual
B­
9
June
2003
Figure
B.
3
Giardia
Data
from
Selected
Research
Studies
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
0
5
10
15
20
25
30
35
40
45
50
UV
Dose
(
mJ/
cm2)
Log
Inactivation
Mofidi
et
al.,
2002
­
G.
lamblia
Mofidi
et
al.,
2002
­
G.
lamblia
Linden
et
al.,
2002
­
G.
lamblia
Linden
et
al.,
2002
­
G.
lamblia
Hayes
et
al.,
2001
­
G.
muris
Hayes
et
al.,
2001
­
G.
muris
Mofidi
et
al.,
2002
­
G.
muris
Mofidi
et
al.,
2002
­
G.
muris
Campbell
and
Wallis,
2002
­
G.
lamblia
Oppenheimer
et
al.,
2002
­
G.
muris
Open
symbols
report
Log
Inactivation
Closed
symbols
report
>
Log
Inactivation
B.
3
Statistical
Analysis
To
determine
the
relationships
between
UV
dose
and
log
inactivation
of
Cryptosporidium,
Giardia,
and
virus,
a
mathematical
model
with
hierarchical
Bayesian
parameter
estimation
techniques
was
used.
This
model
performs
a
meta­
analysis
that
summarizes
and
integrates
the
findings
of
multiple
research
studies.
It
can
be
considered
as
a
compromise
of
two
extreme
methods
of
combining
data
from
different
sources.
One
extreme
method
treats
the
data
from
different
sources
as
identical
replications
and
computes
a
regression
as
if
the
data
were
from
a
single
source.
The
second
extreme
method
treats
each
individual
study
as
totally
unrelated
to
other
studies.
In
this
second
method,
the
separately
estimated
regression
coefficients
are
pooled
only
to
reflect
the
possible
range.
The
Bayesian
meta­
analysis
treats
the
studies
as
exchangeable,
but
not
identical
or
completely
unrelated
(
Hedges
1997).
Regression
coefficients
for
each
study
are
estimated
using
the
same
calculations
and
allowed
to
differ
between
studies.
A
Bayesian
hierarchical
modeling
approach
represents
a
more
general
and
reasonable
approach
for
combining
information
(
Gelman
et
al.
1995;
Condon
2001).

Proposal
Draft
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
UV
Disinfection
Guidance
Manual
B­
10
June
2003
B.
3.1
Model
Description
The
model
used
to
relate
UV
dose
to
Cryptosporidium,
Giardia,
and
virus
log
inactivation
is
described
by
Equation
B.
1.
Qian
et
al.
(
2003)
provides
a
complete
description
of
the
model
and
further
statistical
analyses.

(
)
(
)
ijk
ij
ijk
C
I
N
Y
1
,
~
 
µ
Equation
B.
1
(
)
(
)
(
)
(
)
0001
.
0
,
0
~
001
.
0
,
001
.
0
~
,
~
1
log
)
exp(

2
,
1
2
10
N
gamma
N
X
i
ij
i
ij
 
 
 
 
 
 
µ
+
=

where
Yijk
=
Log
inactivation
of
the
kth
observation
exposed
to
the
jth
UV
dose
level
in
the
ith
study
N(
µ
ij,
 1)=
Normal
distribution
with
mean
µ
and
precision
 
I(
Cijk)
=
Censor
operator
with
Cijk
as
the
estimated
lower
bound
of
the
log
inactivation
value
for
the
kth
observation
exposing
to
the
jth
UV
dose
level
in
the
ith
study
Xij
=
jth
dose
level
of
study
i,
 i
=
Regression
coefficient
for
study
i
 
=
Integrated
regression
coefficient,
combining
information
from
all
studies
When
an
observation
is
known
to
be
greater
than
a
value
(
right­
censored),
the
reported
value
is
used
as
a
lower
bound
value
(
Cijk.).
The
prior
distributions
on
precision
(
inverse
of
variance),
 1,2,
are
modeled
using
gamma(
0.001,
0.001),
which
is
considered
"
non­
informative"
(
the
log
variance
is
almost
uniform).
The
prior
distribution
on
 
is
N(
0,
0.0001),
a
practically
flat
distribution.

One
of
the
benefits
of
using
a
Bayesian
modeling
approach
is
it
allows
known
information
that
can
better
explain
the
data
relationships
to
be
incorporated
into
the
model.
In
this
model,
two
known
pieces
of
information
were
incorporated:
(
1)
as
UV
dose
increases
the
number
of
microorganisms
inactivated
increases
 
incorporated
by
taking
the
exponential
of
 i
in
the
second
line
of
Equation
B.
1,
which
restricts
the
slope
of
the
regression
between
log
inactivation
and
UV
dose
to
a
positive
value;
and
(
2)
when
the
UV
dose
is
zero,
no
microorganism
inactivation
due
to
UV
light
occurs
 
incorporated
by
setting
the
intercept
of
the
regression
line
to
zero
(
the
second
line
in
Equation
B.
1
has
no
intercept
term).

A
Markov
Chain
Monte
Carlo
simulation
method
is
used
for
estimating
the
model
parameters.
To
impute
the
censored
data,
an
iterative
procedure
is
used.
At
a
given
iteration,
a
random
sample
of
log
inactivation
is
taken
from
a
normal
distribution
with
the
mean
and
variance
calculated
by
the
current
estimates
of
 i
and
 1:
2.
If
the
generated
value
is
less
than
the
reported
value
(
the
lower
bound),
it
is
not
used
and
a
new
value
is
generated
until
one
that
is
larger
than
the
reported
value
is
found.
The
model
is
then
refitted
with
new
estimates
of
 i
and
 1,2.
This
process
is
repeated
many
times
(
200,000
in
this
case).
Mathematical
theories
indicate
that
the
effect
of
a
set
of
random
initial
values
for
all
model
coefficients
and
the
censored
values
Proposal
Draft
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
UV
Disinfection
Guidance
Manual
B­
11
June
2003
will
gradually
disappear,
and
the
samples
will
converge
to
their
respective
posterior
marginal
distributions
after
a
number
of
iterations.
In
this
case,
the
first
140,000
iterations
were
discarded
and
1,000
samples
for
each
of
the
unknown
quantities
(
i.
e.,
coefficients,
predictions,
and
censored
values)
were
taken
from
the
remaining
60,000
iterations.
The
computation
is
implemented
under
WinBUGS
(
Spiegelhalter
et
al.
1996).

The
Bayesian
hierarchical
model
of
Equation
B.
1
estimates
the
integrated
model
coefficients
using
the
coefficient
estimates
from
each
study.
As
the
model
indicated,
 
is
assumed
to
be
the
mean
of
the
parent
distribution
of
 i.
This
integration
accounts
for
the
uncertainty
of
each
study
and
"
weights"
each
study
accordingly.

B.
3.2
Cryptosporidium
and
Giardia
Modeled
Results
The
modeled
results
for
Cryptosporidium
and
Giardia
are
shown
graphically
in
Figures
B.
4
and
B.
5,
respectively.
The
graphs
show
the
estimated
regression
for
each
study.
The
model
incorporates
the
coefficients
from
each
study
and
calculates
the
predicted
median
and
80
percent
credible
intervals,
shown
by
the
black
solid
line
(
median)
and
dark
dotted
lines
(
credible
intervals).

Proposal
Draft
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
UV
Disinfection
Guidance
Manual
B­
12
June
2003
Figure
B.
4
Cryptosporidium
Modeled
Data
and
Predictive
Credible
Intervals
UV
Dose
Log
Inactivation
o
Uncensored
data
x
Censored
data
+
Predicted
Censored
Predicted
Median
80%
Credible
Interval
Clancy
et
al.
2000
Clancy
et
al.
2002
Craik
et
al.
2001
Landis
et
al.
2000
Oppenheimer
et
al.
2002
Sommer
et
al.
2001
Kashinkunti
et
al.
2002
Shin
et
al.
2001b
Clancy
Environmental
2002
0
5
10
15
0
1
2
3
4
6
8
11
30
45
75
120
195
300
+

+
+
+

+
+

+

+
++
+
+
+
+
+
+

+
++
+
+
+
+

o
o
o
o
o
o
o
oo
o
ooooo
o
o
o
o
o
oo
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
x
x
x
x
x
x
x
x
x
xx
x
x
x
xx
xx
x
x
x
x
xx
x
x
x
x
x
3
log
inactivation
Lower
bound
credible
interval
Proposal
Draft
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
UV
Disinfection
Guidance
Manual
B­
13
June
2003
Proposal
Draft
Figure
B.
5
Giardia
Modeled
Data
and
Predictive
Credible
Intervals
UV
Dose
Log
Inactivation
o
Uncensored
data
x
Censored
data
+
Predicted
Censored
Predicted
Median
80%
Credible
Interval
Campbell
and
Wallis
2002
Mofidi
et
al.
2001
Hayes
et
al.
2001
Shin
et
al.
2000
Oppenheimer
et
al.
2002
0
10
20
30
40
50
60
0
1
2
3
4
5
7
9
12
30
45
60
90
135
+
+

+
+
++
+++
+

+
+
+
+

++
+
+
+
+
+
+

o
o
o
o
o
oo
o
o
o
o
o
o
o
o
o
oo
oo
o
o
o
o
o
o
o
o
o
o
o
o
o
x
x
x
x
xx
xxx
x
x
x
x
x
x
x
x
x
x
x
x
x
x
3
log
inactivation
Lower
bound
credible
interval
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
UV
Disinfection
Guidance
Manual
B­
14
June
2003
B.
3.3
Virus
Modeled
Results
The
model
for
the
virus
data
is
slightly
different
from
the
Cryptosporidium
and
Giardia
model
described
in
Equation
B.
1.
First,
there
were
no
censored
data
points;
as
a
result,
the
term
I(
Cijk)
is
not
included.
Second,
based
on
the
data,
a
log
transformation
on
the
UV
dose
is
not
necessary,
i.
e.,
the
mean
is
modeled
by
 i
Xij.
Figure
B.
6
displays
the
modeled
results
for
adenovirus.

Proposal
Draft
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
UV
Disinfection
Guidance
Manual
B­
15
June
2003
Proposal
Draft
Figure
B.
6
Virus
Modeled
Data
and
Predictive
Credible
Intervals
UV
Dose
Log
Inactivation
o
Data
Predicted
Median
80%
Credible
Interval
Gerba,
2000
Meng
et
al.,
1996
Shin
2001a
Thompson
et
al.,
2000
Thompson
et
al.,
2002
(
Type
15)

Thompson
et
al.,
2002
(
Type
2)

Thurston,
2002
(
Type
40)

0
2
4
6
8
0
50
100
150
200
250
o
oo
oo
o
o
oo
o
o
o
o
o
o
o
o
o
o
o
o
o
oo
oo
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
oo
o
oo
o
oo
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
oo
o
o
o
o
o
o
o
o
o
oo
o
o
o
o
o
o
o
o
4
log
inactivation
Lower
bound
credible
interval
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
UV
Disinfection
Guidance
Manual
B­
16
June
2003
B.
3.4
Calculating
UV
Dose
Requirements
from
Modeled
Results
Table
B.
2
presents
the
UV
dose
requirements
for
Cryptosporidium,
Giardia,
and
viruses.
Each
of
the
graphs
presented
in
Figures
B.
4
through
B.
6
show
the
80
percent
credible
interval.
The
UV
dose
requirements
for
given
log
inactivation
levels
were
calculated
from
the
fitted
model's
lower
bound
of
the
credible
interval
(
as
called
out
in
Figures
B.
4­
B.
6).
Using
the
lower
bound
means
that
at
a
given
UV
dose,
the
corresponding
log
inactivation
is
expected
to
be
achieved
90
percent
of
the
time.

Table
B.
2
UV
Dose
Requirements
for
Inactivation
of
Cryptosporidium,
Giardia
and
Viruses
During
Validation
Testing
Log
Inactivation
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
Cryptosporidium
1.6
2.5
3.9
5.8
8.5
12
­
­
Giardia
1.5
2.1
3.0
5.2
7.7
11
­
­
Virus
39
58
79
100
121
143
163
186
B.
4
References
Bukhari,
Z.,
T.
M.
Hargy,
J.
R.
Bolton,
B.
Dussert,
and
J.
L.
Clancy.
1999.
Medium­
pressure
UV
for
oocyst
inactivation.
JAWWA.
91(
3):
86­
94.

Campbell,
A.
T.
and
P.
Wallis.
2002.
The
effect
of
UV
irradiation
on
human
derived
Giardia
lamblia
cysts.
Water
Research.
36(
4):
963­
969.

Clancy
Environmental
Consultants.
2000.
UV
Disinfection
of
Cryptosporidium
and
MS2
Coliphage
in
Catskill­
Delaware
Water.
Report
to
Hazen
and
Sawyer.

Clancy,
J.
L.,
Z.
Bukhari,
T.
M.
Hargy,
J.
R.
Bolton,
B.
W.
Dussert,
and
M.
M.
Marshall.
2000.
Using
UV
to
inactivate
Cryptosporidium.
Journal
AWWA.
92(
9):
97­
104.

Clancy,
J.
L.,
T.
M.
Hargy,
D.
A
Battigelli,
M.
M.
Marshall,
D.
Korich,
and
W.
L.
Nicholson.
2002.
Susceptibility
of
Multiple
Strains
of
C.
parvum
oocysts
to
UV
Light.
Final
report
to
AWWA
Research
Foundation
And
American
Water
Works
Association.

Condon,
P.
2001.
Bayesian
Statistical
Modeling.
Wiley,
London.

Craik,
S.
A.,
G.
R.
Finch,
J.
R.
Bolton,
and
M.
Belosevic.
2000.
Inactivation
of
Giardia
muris
cysts
using
medium­
pressure
ultraviolet
radiation
in
filtered
drinking
water.
Water
Research.
34(
18):
4325­
4332.

Craik,
S.
A.,
D.
Weldon,
G.
R.
Finch,
J.
R.
Bolton,
and
M
Belosevic.
2001.
Inactivation
of
Cryptosporidium
parvum
oocysts
using
medium­
and
low­
pressure
ultraviolet
radiation.
Water
Research.
35(
6):
1387­
1398.

Proposal
Draft
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
UV
Disinfection
Guidance
Manual
B­
17
June
2003
Danielson,
R.
E.,
J.
Phillips,
and
F.
Soroushian.
2001.
Inactivation
of
Giardia
muris,
Bacillus
subtilis
spores
and
MS2
coliphage
by
ultraviolet
light.
Proceedings
IUVA
Congress,
Washington
D.
C.

Gelman,
A.,
J.
B.
Carlin,
H.
S.
Stern,
and
D.
B.
Rubin.
1995.
Bayesian
Data
Analysis.
Chapman
&
Hall.
London.

Gerba,
C.
P.
et
al.
2000.
Treatability
Screening
for
CCL
Microbial
Contaminants
­
Interim
DRAFT
Report
to
US
EPA
Office
of
Water.

Gilead,
Z.
and
H.
S.
Ginsberg.
1966.
Comparison
of
the
rates
of
ultraviolet
inactivation
of
the
capacity
of
Type
12
Adenovirus
to
infect
cells
and
to
induce
T
antigen
formation.
J.
Bacteriol.
92(
6):
1853­
1854.

Hara,
J.
et
al,
1990,
Survival
and
Disinfection
of
adenovirus
type
19
and
enterovirus
70
in
ophthalmic
practice.
Jpn
J
ophthamol.
34:
421­
427.

Hargy,
T.
M.,
J.
L.
Clancy,
Z.
Bukhari,
and
M.
M.
Marshall.
2000.
Shedding
UV
Light
on
the
Cryptosporidium
Threat.
Journal
of
Environmental
Health,
63:
19­
22.

Hayes,
S.
L.,
C.
D.
Barnes,
E.
W.
Rice,
M.
W.
Ware,
and
F.
W.
Schaefer
III.
2001.
Low
Pressure
Ultraviolet
Studies
for
Inactivation
for
Giardia
muris
cysts,
Proceedings
IUVA
Congress,
Washington
D.
C.

Hedges,
L.
1997.
Bayesian
meta­
analysis.
In
Statistical
Analysis
of
Medical
Data,
eds
B.
Everitt
and
I.
Dunn.
Arnold.
London.
pp.
251­
275.

Kashinkunti,
R.
D.,
D.
H.
Metz,
K.
G.
Linden,
M.
D.
Sobsey,
M.
Moran,
and
A.
Samuelson.
2002.
Microbial
Inactivation
Strategies
for
the
Future:
UV,
Chlorine,
and
DBPs.
Proceedings
WQTC
2002,
Seattle,
November
10­
14,
2002.

Landis,
H.
E.,
Thompson,
J.
E.,
Robinson,
J.
P.,
Blatchley,
E.
R.
2000.
Inactivation
responses
of
Cryptosporidium
parvum
to
UV
radiation
and
gamma
radiation.
Proceedings
WQTC
2000,
Salt
Lake
City,
WA.
November
5­
9,
2000.

Linden,
K.
G.,
G­
A.
Shin,
G.
Faubert,
W.
Cairns,
and
M.
D.
Sobsey.
2002.
UV
disinfection
of
Giardia
lamblia
in
water.
Environmental
Science
and
Technology.
36(
11):
2519­
2522.

Mackey,
E.
D.,
J.
P.
Malley,
T.
M.
Hargy,
and
R.
S.
Cushing.
2000.
MS­
2
Bioassays
and
Cryptosporidium
Challenges:
Comparing
and
Contrasting
UV
Reactor
Validation
Techniques.
Proceedings
AWWA
Water
Quality
Technical
Conference,
Salt
Lake
City,
UT.

Malley,
J.
P.
2000a.
Reduction
of
Microbial
and
Chemical
Pollution
of
Coastal
Waters
Using
Pulsed­
UV
Disinfection,
Interim
Report,
NOAA.

Malley,
J.
P.
2000b.
AWWARF
Project
2593:
"
Inactivation
of
Pathogens
by
Innovative
UV
Technologies"
Second
Quarterly
Report.

Proposal
Draft
Appendix
B.
Derivation
of
UV
Dose­
Response
Requirements
UV
Disinfection
Guidance
Manual
B­
18
June
2003
Meng,
Q.
S.
and
C.
P.
Gerba.
1996.
Comparative
inactivation
of
enteric
adenovirus,
poliovirus
and
coliphages
by
ultraviolet
irradiation.
Water
Research.
30(
11):
2665­
2668.

Mofidi,
A.
A.,
H.
Baribeau
and
J.
F.
Green.
1999.
Inactivation
of
Cryptosporidium
parvum
with
polychromatic
systems.
Proceedings
WQTC,
Tampa,
Florida,
October
31­
November
4,
1999.

Mofidi,
A.
A.,
E.
A.
Meyer,
P.
M.
Wallis,
C.
I.
Chou,
Meyer,
B.
P.,
S.
Ramalingam,
and
B.
M.
Coffey.
2002.
The
effect
of
UV
light
on
the
inactivation
of
Giardia
lamblia
and
Giardia
muris
cysts
as
determined
by
animal
infectivity
assay.
Water
Research.
36:
2098­
2108.

Oppenheimer,
J.,
T.
Gillogly,
G.
Stolarik,
and
R.
Ward.
2002.
Comparing
the
Efficiency
of
Low
and
Medium
Pressure
UV
Light
for
Inactivating
Giardia
muris
and
Cryptosporidium
parvum.
Proceedings
AWWA
Annual
Conference,
New
Orleans,
LA.

Qian,
S.
S.,
M.
Donnelly,
D.
C.
Schmelling,
M.
Messner,
K.
G.
Linden,
and
C.
Cotton.
2003.
Ultraviolet
light
inactivation
of
Cryptosporidium
and
Giardia
in
drinking
water:
A
Bayesian
meta­
analysis.
Manuscript
submitted
to
Water
Research.

Shin,
G­
A,
K.
G.
Linden,
and
M.
D.
Sobsey.
2001a.
UV
Disinfection
of
Adenovirus.
WERF
Progress
Report
6,
98­
HHE­
2.

Shin,
G­
A,
K.
G.
Linden,
M.
Arrowood,
and
M.
D.
Sobsey.
2001b.
Low
pressure
UV
inactivation
and
subsequent
DNA
repair
potential
of
Cryptosporidium
parvum
oocysts.
Appl.
Environ.
Micro.
67(
7):
3029­
3032.

Sommer,
R.,
S.
Appelt,
W.
Pribil,
and
T.
Slifko.
2001.
UV
inactivation
of
C.
parvum
evaluated
in
cell
culture.
Unpublished
data,
personal
communication.

Spiegelhalter,
D.
J.,
A.
Thomas,
A.,
and
W.
R.
Best.
2000.
WinBUGS
Version
1.3:
User
Manual.
Medical
Research
Council,
Biostatistics
Unit,
Institute
of
Public
Health,
Cambridge,
UK.

Thompson,
S.
S.,
J.
L.
Jackson,
M.
Suva­
Castillo,
W.
A.
Yanko,
Z.
El
Jack,
J.
Kuo,
C.
L.
Chen,
F.
P.
Williams,
and
D.
P.
Schnurr.
2002.
Detection
of
Infectious
Human
Adenoviruses
in
Tertiary
Treated
and
UV
Disinfected
Wastewater.
Accepted
for
publication,
Water
Environment
Research.

Thurston­
Enriquez,
J.,
C.
Hass,
J.
Jacangelo,
K.
Riley,
and
C.
P.
Gerba.
2002.
Inactivation
of
Feline
Calicivirus
and
Adenovirus
Type
40
by
Ultraviolet
Irradiation.
Proceedings,
WEF
Disinfection
2002,
St.
Petersburg
FL,
February
17­
20,
2002.

Proposal
Draft
