DISINFECTION BY-PRODUCT FORMATION BY ALTERNATIVE
DISINFECTANTS AND REMOVAL BY GRANULAR ACTIVATED CARBON
by
Wayne E. Koffskey
Department of Public Works
Jefferson Parish
Jefferson, Louisiana 70121
Cooperative Agreement No. CR814033
Project Officer
Ben Lykins, Jr.
Drinking Water Research Division
Water Engineering Research Laboratory
Cincinnati, Ohio 45268
WATER ENGINEERING RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U. S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
The information in this document has been funded wholly or in
part by the United States Environmental Protection Agency under
Assistance Agreement Number CR814033 to Jefferson Parish,
Louisiana. It has been subject to the Agency's Peer and
Administrative Review, and it has been approved for publication
as an EPA document. Mention of trade names or commercial
products does not constitute endorsement or recommendation for
use.
-------
FOREWORD
Today's rapidly developing and changing technologies and
industrial products and practices frequently carry with them the
increased generation of materials that, if improperly dealt with
can threaten both public health and the environment. The U. S.
Environmental Protection Agency is charged by Congress with
protecting the Nation's land, air, and water systems. Under a
mandate of national environmental laws, the agency strives to
formulate and implement actions leading to a compatible balance
between human activities and the ability of natural systems to
support and nurture life. These laws direct EPA to perform
research to define our environmental problems, measure the
impacts, and search for solutions.
The Risk Reduction Engineering laboratory is responsible for
planning, implementation, and management of research,
development, and demonstration programs to provide an
authoritive, defensible engineering basis in support of the
policies, programs, and regulations of the EPA with respect to
drinking water, wastewater, pesticides, toxic substances, solid
and hazardous wastes, and Superfund-related activities. This
publication is one of the products of that research and provides
a vital communication link between the researcher and the user
community.
This study reports on a cooperative research effort' that
examines the formation of halogenated disinfection by-products by
the alternative disinfectants and the removal of these
by-products and their precursors by granular activated carbon
filtration.
E. Timothy Oppelt, Director
Risk Reduction Engineering Laboratory
111
-------
ABSTRACT
The effects of the use of the alternative disinfectants on
the formation of halogenated disinfection by-products (DBFs)
including total organic halide, trihalomethanes, haloacetic
acids, haloacetonitriles, haloketones, chloral hydrate, and
chloropicrin, were examined along wi.th the ability of granular
activated carbon (GAG) to remove these by-products and their
precursors. Microbiological information was also obtained on the
operating system and included heterotropic plate count, total
coliform, and assimilable organic carbon (AOC). The ability of
the alternative disinfectants to inactivate MS2 coliphage was
also examined. One other aspect of the project was to provide
sampling sites for health effects research by EPA which will be
reported elsewhere.
The operating system was comprised of four parallel pilot
column process streams consisting of a 30 min contact chamber
followed by a sand column in series with a GAG column having a 20.
min empty bed contact time. One of four disinfectants, ozone,
chlorine dioxide, chlorine," or chloramines, was applied at the
beginning of each process stream. A fifth nondisinfected process
stream consisting of only a sand column in series with a GAG
column was used as a control.
The lowest levels of halogenated DBFs resulted from, the
combination of preozonation and postchloramination with annual
simulated distribution system averages of 27 ug Cl/L for TOX and
12 ug/L for the sum -of 18 DBFs when sand filtration was employed.
These respective concentrations were further reduced to 13 ug
Cl/L and 7 ug/L with subsequent GAC filtration having a 20 min
empty bed contact time. While ozonation produced significant
levels of AOC, sand filtration resulted in a 77 percent reduction
to 39 ug acetate C-eq/L with subsequent GAC filtration further
reducing AOC to 4 ug acetate C-eq/L.
This report was submitted in fulfillment of Cooperative
Agreement No. CR814033 by Jefferson Parish, LA under sponsorship
of the USEPA and covers an operational period from July 1989
through July 1990.
IV
-------
CONTENTS
Foreword .......... iii
Aos u r 3 c c • •»•••».•.•••.......... iv
Figures vi
Taoles xvi
Aooreviations xviii
Metric Conversion Taole xxi
AcKnowledger.enrs *....... xxii
I. Introduction 1
2. Conclusions 3
3. lAecommenaations 5
4. Pilot Column Configuration
and Operation 6
5. Sampling, Analysis, and Quality
Assurance 12
6. Results and Discussion
Disinfection By-Products
Total Organic Carbon 32
Total Organic halide 36
Total Trihalometharie 45
Kaloacetic Acids 58
Chloral Hydrate 96
Haloacetonitriles 102
Haloketones " 138
Chloropicrin 155
Summary of Disinfection
By-Products 155
TOX as a Surrogate for DBPs . . . 153
Microbiological Observations
Heterotrophic Plate Count .... 174
Total Coliform .......... 178
Assimilable Organic Carbon . . . 179
MS2 Coliphage ' 180
References „ 185
v
-------
FIGURES
Number Page
1. Pilot column system configuration. ...... 7
2. Water temperature of the pilot column
system influent 9
3. Method recoveries for di-and trichloro-
acetoni tr ile. 15
4. Comparison of the TOG levels in the
influent of each process stream 33
5. Comparison of the TOC levels in the sand
column effluent of each process stream. . . 34
6. Comparison of the TOC levels in the GAC
column effluent of each process stream. . . 35
7. Comparison of the TOX levels in the
disinfectant contact chamber effluents
of each process stream 37
8. Comparison of the TOX levels in the sand
column effluent of each process stream. . . 38
9. Comparison of the TOX-Cl- levels in the
sand column effluent of each process
stream 39
lij. Comparison of the TOX-NH2C1 levels in the
sand column effluent of each process
stream 40
11. Compariosn of the TOX levels in the GAC
column effluent of each process stream . . 41
12. Comparison of the TOX-C12 levels in the GAC
column effluent of each process stream . . 43
VI
-------
FIGURES
Number pac
13. Comparison of the TOX-NH2C1 levels in the
GAG column effluent of each process stream. 44
14. Comparison of the THM levels in the sand
column effluent of each process stream . . 46
15. Comparison of the THM-NH-C1 levels in the
sand column effluent of each process
stream 47
16. Comparison of the THM-Cl- levels in the
sand column effluent of each process
stream 48
17. Comparison of the THM levels in the GAC
column effluent of each process stream . . 49
18. THM breakthrough profile for the chlorine
dioxide GAC column 51
19. Comparison of the THM-C12 levels in the GAC
column effluent of each process stream . . 52
20. Comparison of the THM-NH2C1 levels in the
GAC column effluent or each process stream 53
21. Comparison of the DCAA levels in the sand
column effluent'of each process stream . . 55
22. Comparison of the influent and effluent
DCAA levels for the nondisinfected and
ozone sand columns 56
23. Comparison of the DCAA-C12 levels in the
sand column effluent of each process
stream 57
24. Comparison of the DCAA-NH-Cl levels in the
sand column -effluent or each process
stream 58
VI1
-------
FIGURES
Number Page
25. Comparison of the DCAA levels in the GAC
column effluent of each process stream . . 59
26. Comparison of the DCAA-C12 levels in the
GAC column effluent of each process stream 60
27. Comparison of the. DCAA-NH2C1 levels in the
GAC column effluent of each process stream 62
28. Comparison of the TCAA levels in the sand
column effluent of each process stream . . 63
29. Comparison-'of tne TCAA-C12 levels in the
sand column effluent or each process
stream 64
30. Comparison of the TCAA-NH2C1 levels in the
sand column effluent of each process
stream ...... 65
31. Comparison of the TCAA levels in the GAC
column effluent of each process stream . . 66
32. Comparison of the TCAA-C1- levels in the
GAC column effluent of each process
stream „ „ 57
33. Comparison of the TCAA-NH2C1 levels in the
GAC column effluent of each process stream 69
34. Comparison of the BCAA levels in the sand
column effluent of each process stream . . 70
35. Comparison of the BCAA-Cl- levels in the
sand column effluent of each process
stream 71
36. Comparison of the BCAA-NH2C1 levels in the
sand column effluent or each process
stream 72
Vlil
-------
FIGURES
Number Pac
'' -— ' ,_ _•
37. Comparison of the BCAA levels in the GAG
column effluent of each process stream . . 73
38. Comparison of the BCAA-Cl, levels in the GAG
column effluent of each process stream . . 74
39. Comparison of the BCAA-NH2C1 levels in the GAG
column effluent of each process stream . . 76
40. Comparison of the CAA levels in the sand
column effluent of each process stream . . 77
41. Comparison of the CAA-C12 levels in the
sand column effluent of each process stream 78
42. Comparison of the CAA-NH2C1 levels in the
sand column effluent of each process
stream 79
43. Comparison of the CAA levels in the GAC
column effluent of each process stream . . 80
44. Comparison of the CAA-C12 levels in the
GAC column effluent of each process
stream 81
.45. Comparison of the CAA-NH2C1 levels in the
GAC column effluent of each process
stream 83
46. Comparison of the DBAA levels in the
chlorine dioxide contact chamber and
sand column effluents 84
47. Comparison of the DBAA levels in the
nondisinfected and ozone contact chamber
and sand column effluents 85
48. Comparison of the DBAA levels in the sand
• column effluent of each process stream- . . 86
IX
-------
FIGURES
Mutnber Page
49. Comparison of the DBAA-C12 levels' in the
sand column effluent of eacn process
stream
50. Comparison of the DBAA-NH2C1 levels in the
sand column effluent of each process
stream .....
51. Comparison of the DBAA levels in the GAC
column effluent of each process stream . . 89
52. Comparison of the DBAA-C12 levels in the
GAC column effluent of each process
stream .......... . - -
53. Comparison of the DBAA-NH2C1 levels in the
GAC column effluent of each process stream 92
54. Comparison of the' BAA levels in the sand
column effluent of each process stream . . 93
55. Comparison of the BAA levels in the ozone
contact chamber and sand column effluents. . 94
56. Comparison of the BAA levels in the GAC
column effluent of each process stream ... 95
57. Comparison of the CH levels in the sand
column effluent of each process stream ... 97
58. Comparison of the CH-C12 levels in the sand
column effluent of each process stream . .
59. Comparison of the CH-NH2C1 levels in the sand
column effluent of each process stream . .
60. Comparison of the CH-C12 levels in the GAC
column effluent of each process stream . .
61. Comparison of the CH-NH2C1 levels in the GAC '
column effluent of each process stream . .
98
99
100
101
-------
FIGURES
Number
62. Comparison of the DCAN levels in the sand
column effluent of each process stream . . . 103
63. Comparison of the DCAN-C12 levels in the
sand column effluent of each process
stream ..... .............. 104
64. Comparison of the DCAN-NH-Cl levels in the
sand column effluent or each process
stream . . . . ...... ......... 105
65. Comparison of the DCAN levels in the GAG
column effluent of each process stream . . . 107
66. Comparison of the DCAN-C12 levels in the
GAC column effluent of each process
stream
67. Comparison of the DCAN and DCAN-C12 levels
in the nondisinf ected process stream .... 109
68. Comparison of the DCAN and DCAN-C1 ' levels
in the chlorine process stream 7 . . . . . . 110
69. Comparison of the DCAN-NH2C1 levels in the
GAC column effluent of each process
stream ........... . ....... Ill
70. Comparison of the SCAN levels in the sand
column effluent of each process stream . . . 113
71. Comparison of the BCAN-C12 levels in the
sand column effluent or each process
stream '
72. Comparison of the BCAN-NH2C1 levels in tne
sand column effluent or each process
stream ............... -.... 115
73. Comparison of the BCAN levels in the GAC
column effluent of each process stream . . . 116
-------
FIGURES
Number page
74. Comparison of the BCAN-C12 levels in the
GAG column effluent of each process stream . 117
75. Comparison of the SCAN and BCAN-C12 levels
in the nondisinfected process scream .... 118
76. Comparison of the BCAN and BCAN-C1., levels
in the chlorine process stream 120
77. Comparison of the BCAN-NH-C1 levels in the
GAG column effluent of each process stream . 121
78. Comparison of the DBAN levels in the contact
chamber column effluent of each process
stream 122
79. , Comparison of the DBAN levels in the sand
column effluent of each process stream . . . 123
80. Comparison of the DBAN levels in the ozone
contact chamber and s-and column effluents . 124
81. Comparison of the DBAN-Cl- levels in the sand
column effluent of eacn process stream . . . 125
82. Comparison of the DBAN-NH-Cl levels in the
sand column effluent of each process stream . 126
83. Comparison of the DBAN and DBAN-NH2C1 levels
in the chloramine sand column effluent . . . 127
84. Comparison of the DBAN levels in the GAC column
effluent of each process stream 128
85. Comparison of the DBAN-C12 levels in the GAC
'column effluent of eacn process stream . . . 130
86. Comparison of the DBAN and DBAN-C12 levels in
the nondisinfected process stream • 131
87. Comparison of the DBAN and DBAN-C1? levels in
the chlorine process stream ......... 132
XI1
-------
FIGURES
Numbec ' Pag<
88. Comparison of the DBAN-NH2C1 levels in the GAG
column effluent of eacn process stream . . . 133
89. Comparison of the TCAN levels in the sand
column effluent of each process stream . . . 134
90. Comparison of the TCAN-C12 levels in the sand
column effluent of eacn process stream . . . 135
91. Comparison of the TCAN levels in the GAG
column effluent of each process stream . . . 136
92. Comparison of the TCAN-C12 levels in the GAG
column effluent of eacn process stream . . . 137
93. Comparison of the TCP levels in the sand
column effluent of each process stream . . . 139
94. Comparison of the TCP levels in the ozone
contact chamber and sand column effluents . 140
95. Comparison of the TCP-C12 levels in the sand
column effluent of each process stream . . . 141
96. Comparison of the TCP-NH2C1 levels in the sand
column effluent of each process stream . . . 142
97. TCP reductions in the terminal chloramine
sand column effluents of the chlorine
dioxide and c'hloramine process stream . . . 143
98. Comparison of the TCP levels in the GAG
column effluent of each process stream . . . 144
99. Comparison of the TCP-C12 levels in the GAG
column effluent of each process stream . . . 145
Comparison of the TCP-NH2C1 levels in the GAG
column effluent of each process stream . .
100,
*. --- _ ___»i «. ,^, fc !•»**« \*r .*, _l_ «—• V Vrf. .J- hJ J, 1,1 l_X J ^ \J/1\^
147
101. Comparison of the DCP levels in the sand
column effluent of each process stream . . . 148
Kill
-------
FIGURES
Number
182.
IDS.
134.
14)5.
106.
107.
108.
109.
111.
112.
113.
114.
Comparison of the DCP-C12 levels in the sand
column effluent of each process stream . . .
Comparison of tne DCP-NH2C1 levels in the sand
column effluent of each process stream . . .
Comparison of the DCP and DCP-NH2C1 levels
in the sand column effluents of the
chlorine dioxide and chloramine process
stream ...... <
Comparison of the DCP levels in the GAC
column effluent of each process stream . . ,
Comparison of the DCP-C12 levels in the GAC
column effluent of each process stream . .
149
150
151
152
153
Comparison of the DCP-NhUCl levels in the
GAC column effluent of each process stream . 154
Comparison of the CP levels in the sand
column effluent of each process stream ... 156
Comparison of the CP-C12 levels in the sand
column effluent of each process stream . . . 157
Comparison of the CP-NH2C1 levels in the
sand column effluent of each process
stream 158
Comparison of the CP levels in the GAC
Column effluent of each process stream .
Comparison of the CP-C12 levels in the GAC
column effluent of each process stream ,
Comparison of the CP-NH2C1 levels in the
GAC column effluent of each process
stream
Comparison of total DBP levels in the
effluent of each GAC column . . .
159
160
161
165
xiv
-------
FIGURES
J.NJ umber Page
115. Regression analysis of TOX and the DBPs
at all locations by process stream 169
116. Regression analysis of TOX and the DBPs
at all locations by sample type 170
117. Expansion of regression analysis of TOX
and the DSPs at all locations by
process stream 171
118. Expansion of ' regression analysis of TOX
and the DBPs at all locations by
process stream 172
119. TOX vs DBPs for the chloramine process
stream . . . 173
120. Regression of all terminal chlorine and
chloramine TOX and DBPs for all locations . 175
121. Regression of all terminal chlorine TOX
and DBPs for all locations 17-6
122. Regression of all terminal chloramine TOX
and DBPs for all locations 177
123. Comparison of AOC levels across the ozone
and chlorine process streams ... 181
xv
-------
TABLES
Number p
1. Physical pilot column operating conditions 8
2. Average nondisinfected influent water quality .. 10
3. Average disinfectant contact chamber
demands and residuals 10
4. Average disinfectant contact chamber and sand
column effluent residuals 11
5. Sampling and analysis schedule ..... ..... . ...... 13
6. Gas chromatography detection limits
for VOC analysis .-... .................... . ...... i6
7. Gas chromatography detection limits
for HAA analysis .... ........................ fff 17
8. Gas chromatography detection limits
for DBF analysis .... .................... ....... 18
9. Confirmatory GC/MS standard concentrations and
qualitative detection limits .... ............ „ . . 19
10. Analytical precision as measured by
duplicate sample analysi-s ................... . . „ 23
11. Analytical accuracy as measured by internal
quality control sample analysis ......... . ..... .. 25
12- USEPA performance evaluation DMR009,
April 1989 ..................................... 25
13. USEPA performance evaluation WS024,
May 1989 ................................ . ...... 26
14. USEPA performance evaluation WS025,
November 1989 ........................... . ...... 26
xvi
-------
TABLES
Number ^ page
15. L'SEPA performance evaluation WS026,
May 1990 -,-,
Z. /
16. L'SLPA performance evaluation WS027,
November 1990 28
17. Percent GC/MS confirmation and average mass
spectral fit for the influent of each
process stream 30
13- Percent GC/MS confirmation and average
mass spectral fit for terminal distribution
system simulation 31
19. Average sand filter effluent DBP
concentrations . 152
2y. Average chlorinated distribution system DBP
concentrations for sand filtration 163
21- Average chlorinated distribution system DBP
concentrations for GAG filtration 164
22. Average chloraminated distribution system DBP
concentrations for sand filtration 166
23. Average chloraminated distribution system DBP
concentrations for GAG filtration 167
24. Comparison of HPC across each process stream . 178
25. Comparison of total coliform across each
process stream ; 179
26. MS2 coliphage reductions by ozone and
cnlorine dioxide 182
27. MS2 coliphage reductions by chlorine
and cnloramine 133
xvi a.
-------
LIST OF ABBREVIATIONS
AMU — atomic mass units
---OC — assimilable organic carbon
BAA — bromoacetic acid
5CAA — bromochloroacetic acid
— bromochloroacetonitrile
f
— degree centigrade
CAA — chloroacetic acid
C-eq/L — carbon equivalents per liter
cfu — colony forming units
3H — chloral hydrate
C? — chloropicrin
C12 — free chlorine
~^2 • — terminal samples 'treated with free chlorine and
• incubated at-30 C for 5 days
C102 — chlorine dioxide
£T10 — residual concentration in mg/L x T,g in minutes
DBAA — dibromoacetic acid
D3AN — dibromacetontrile
33P — disinfection by-product
3CAA — dichloroacetic acid
DCAN — dichloroacetoni'tr i le
XVIXI
-------
LIST OF ABBREVIATIONS
DCP
DFTPP
DO
DPD
EBCT
EPA
FAS
FC43
ft
GAG
gal
gpm/ft'
GC/MS
gpm
HAA
HPC
ID
in
infl
m
mgd
mg/L
mg C/L
1,1-dichloropropanone
decafluorotriphenylphosphine
dissolved oxygen
n,n-diethyl-p-phenylenediamine
empty bed contact time
U. S. Environmental Protection Agency
ferrous ammonium sulfate
perfluorotributylamine
foot
granular activated carbon
gallon
gallons per minute per square foot
gas chromatograph/rnass spectrometer
gallons per minute
haloacetic acid
heterotrophic plate count
inner diameter
inch
influent
meter
million gallons per day
milligram per liter
v
milligrams of carbon per liter
xix
-------
LIST OF ABBREVIATIONS
mm
mm
msec
NH2C1
ntu
°3
PRO
SWTR
T10
TC
TCAA
TCAN
TCP
THM
TOG
TOX
ug/L
ug Cl/L
ul
VGA
minute
millimeter
milliseconds
samples treated with ammonia followed by
chlorine and incubated at 30°C for 5 days
chloramine
nephelometric turbidity unit
ozone
percent relative deviation
surface water treatment rule
the time (minutes) in which 90 percent of the water
passing through the column is retained within the
column.
total coliform
trichloroacetic acid
trichloroacetonitrile
1,1,1-trichloropropanone
total trihalomethane
total organic carbon
total organic halogen
microgram per liter
micrograms of chloride per liter
microliter
volatile organics analysis
xx
-------
METRIC CONVERSION TABLE
ENGLISH UNIT
1
1
1
1
1
1
ft3
ft
gal
gpm
gal/min
per ft
in
mgd
1
1
1
Ib
psi
sq ft
METRIC EQUIVALENT
0.02834 m3
30.49 centimeter
3.788 liter
3.788 liters/min
40.78 liters/min
per m
2.54 cm
3788 m3/day
453.5 g
0.1785 kg/cm2
. 0.0929 m2
XXX
-------
ACKNOWLEDGMENTS
The administrative assistance and technical advice of the
Project Officer, Ben Lykins, was most valuable in meeting the
objectives of this study.
The technical staff of the Jefferson Parish Water Quality
Laboratory was instrumental to the success of this project
through their long hours of dedicated work. These included
Gulser Wood, Robert Grant, Mark Flynn, Leonardo Tapia, Salvador
Maffei, Cleofe' Beltran, Douglas Meadowcroft, Michelle
Lutenbacher, James Massony, Daniel Acosta, Paul O'Rourke, Sushma
Pahwa, and Anita Turner. The secretarial services of Marilyn
Harding in preparing this report were well performed and
appreciated.
Appreciation must also be expressed to Julius Belteky and
William Gwinn for their assistance in maintaining the analytical
equipment required to complete this study.
The assistance and cooperation of the plant superintendent,
Elaine Elstrott throughout the course of this study was also
appreicated.
xxi i
-------
SECTION 1
INTRODUCTION
Chlorinated disinfectants have historically been widely used
throughout the United States in the disinfection of drinking
water. During the disinfection process, chlorinated
disinfectants have been found to react with naturally occurring
organic matter to form a number of halogenated disinfection
by-products. Because recent ammendments to the Safe Drinking
Water Act will require that these by-products be regulated in the
near future, this research project was developed to evaluate the
formation of disinfection by-products by the alternative
disinfectants' and their removal by granular activated carbon
filtration as well as to evaluate the microbiological quality of
tne treated water. More specifically, the objectives of this
project were:
1. To measure the effects of the alternative disinfectants
on the formation of halogenated disinfection by-products
including the trihalomethanes, the haloacetic acids, the
haloacetonitriles, the haloketones, chloral hydrate, and
chloropicrin.
2. To measure the effectiveness of granular activated
carbon filtration following sand- filtration in removing
halogenated disinfection ' by-products arid their
precursors.
3. To measure the general microbiological quality of water
treated with the alternative disinfectants.
4. To assess the levels of assimilable organic carbon
formed during disinfection with ozone and chlorine.
5. To evaluate the effectiveness . of the alternative
disinfectants in the,inactivation of MS2 coliphage.
To meet these objectives, a pilot column system consisting of
four disinfected process streams ' (ozone, chlorine dioxide,
cnlorine, and chloramine) and a nondisinfected process stream was
continually operated for one year beginning in July 1989. This
-------
report contains a complete description of the pilot column system
employed, the analytical methods used to monitor the parameters
of interest, and the quality assurance associated with the
analytical results obtained. A thorough discussion of each
parameter is presented relative to the project objectives
indicated.
-------
SECTION 2
CONCLUSION
i. Wnen high levels of disinfection by-product (DBF)
precursers are present, the level of halogenated DBFs
reaching the consumer is primarily dependent upon the
type of postdisinfectant (chlorine or chloramine) used
in the distribution system as opposed to the
predisinfectant or filtration process employed.
2. The combination of pre-- and postdisinfectants which
resulted in the lowest levels of halogenated DBFs
reaching the consumer was preozonation and
postchloramination wnich produced annual' distribution
system averages of 27 ug. Cl/L for TOX and 12 ug/L for
total halogenated DBFs when sand filtration was
employed. With subsequent GAC filtration having a 20
min empty bed contact time, these annual averages were
reduced to 13 ug Cl/L TOX and 7 ug/L total halogenated
DBFs indicating that, even with annual GAC replacement
or reactivation, GAC filtration would not be very
beneficial for DBF removal in a chloraminated
distribution system.
3. The use of postchlorination following sand- filtration
and preozonation completely negated any beneficial
effects of ozonation due to the high level of DBF
precursors present, producing annual distribution system
averages of 330 ug Cl/L for TOX and 309 ug/L for total
halogenated DBFs, 50% of which were THMs. Even with GAC
filtration following preozonation and sand filtration,
postchlorination produced annual distribution system
averages for TOX and total halogenated DBFs of 127 and
138 ug/L, respectively. In order to maintain THM levels
below 50 ug/L using this process, GAC replacement or
reactivation would be required approximately every 100
days.
4. Pretreatment with chlorine dioxide followed by sand
filtration and postchloramination produced only slightly
higher DBF levels than pretreatment with ozone which had
-------
an annual halogenated DBF level of 19 ug/L. However,
cnlorine dioxide pretreatment also produced an annual
average TOX level of 80 ug/L which was three times that
observed for ozone pretreatment, as well as a chlorite
residual of 0.5 ug/L (chlorate was not measured).
5. Relatively high levels of assimilable organic carbon
(AOC) were observed following ozone pretreatment ranging
from 50-270 ug acetate C-eq/L with an annual average of
170 ug acetate C-eq/L. These AOC levels correlated with
cnanges in water temperature, with the lowest AOC levels
being produced at the lowest water temperature. The
biologically active ozone sand column reduced AOC levels
by 77 percent to an average of 30 ug acetate C-eq/L with
subsequent GAC filtration reducing the AOC level to 4 ug
acetate C-eq/L. Chlorination produced only a slight
reduciton in the nondisinfected influent AOC which
averaged 10 ug acetate C-eq/L, while subsequent.GAC
filtration reduced the AOC level to 3 ug acetate C-eq/L.
6. Reductions of MS2 'coliphage of greater than 6 logs
occurred across the ozone contact chamber with CT,fl
values ranging from 3-5 mg/L-min while those across trie
chlorine dioxide and chlorine contact chambers were
equal to or greater than 5 logs .with respective CT,~
values ranging from 7-10 .mg/L-min and 14-20 mg/L-min:
Relatively low •coliphage removals ..of 0.2-1.8 logs
occurred across the chloramine contact chamber with CT-, n
values ranging from 22-31 mg/L-min. -
7. High heterotrophic plate counts (HPC) were observed in
the ozone sand column effluent averaging 28,000 cfu/mL
and in the GAC effluents of all process streams
averaging 2,000-16,000 cfu/mL. This indicated that free
chlorine contact of a relatively short duration would be
required to reduce the HPC to an acceptable level before
the addition of ammonia to form chloramine- and thereby
minimize DBP formation in the distribution system.
8. A relatively good correlation with a coefficient of 0.88
and a standard deviation of 26 ug Cl/L, was observed
between TOX' and the sum of 18 DBPs for the terminally
chlorinated distribution simulation. Correlations -
within tne treatment train were highly variable due to
the short disinfectant contact times employed and the
variations in reaction rates observed for TOX .-and the
DBPs. The use of TOX as a DBP surrogate in a
chloraminated distribution system -was determined to be
impractical since.the correlation coefficient would vary
dramatically with the free chlorine contact time.
employed in the treatment process.
-------
SECTION 3
RECOMMENDATIONS
1. Further studies of ozone disinfection should be
conducted with regard to the formation of AOC and
specific nonhalogenated ozone by-producrs at various
ozone dosages and their reductions across various
filtration medias over a range of hydraulic loadings.
2. The effects of AOC on distribution system regrowth
should be thoroughly investigated particularly with
regard to the presence of excess ammonia from chloramine
generation and phosphate from corrosion inhibitors.
3. The free chlorine contact time required to produce an
acceptable heterotrophic plate count level following
ozonation and GAC filtration should be assessed for a
chloraminated distribution system along with the
subsequent levels of halogenated disinfection
by-products formed.
4. Further studies of virus inactivation in natural waters
should be conducted for the alternative disinfectants to
confirm tne higher log removals observed in this study.
-------
SECTION 4
PILOT COLUMN CONFIGURATION AND OPERATION
Lower Mississippi River water entering the 34 mgd Permutit
treatment plant was dosed with 1-6 mg/L diallyldimethylammonium
cnloride and/or dimethylamine polyelectrolyte polymers for
clarification, 0.1-0.3 mg/L fluosilicic acid (as fluoride) for
fluoridation, and 2 mg/L powdered activated carbon for spill
prevention. After clarification via Permutit upflow
precipitators, a small portion of clarified water was diverted to
the pilot column system and was filtered through one of two
pressure sand filters at a hydraulic loading of 1.7 gpm/ft .
Eacn sand filter contained 30 in of 0.45 mm filter sand and
provided an average nondisinfected sand filtered water flow of
8.5 gpm to the rest of the pilot column system as indicated in
Figure 1. The. filtered water was then split into five process
streams, one for each of the four disinfectants and a
nondisinfected process stream which was used as a control. Each
disinfected process stream consisted of a 30 min disinfectant
contact chamber followed by series filtration through a sand
column_ and a granular activated carbon (GAC) column, while the
nondisinfected process stream consisted of only a sand column in
series with the GAC column.
Eacn disinfection contact chamber was constructed using a 12
in diameter stainless steel pipe and was 10. ft in height, except
for the ozone contact chamber which was 11 ft high. The sand and
GAC columns were constructed from 10 ft sections of 6 in diameter
glass pipe. All pilot column components were constructed from
stainless steel, glass, and teflon. Further details of pilot
column construction have been presented in a previous report.
The sand columns were charged with 30 in of 0.45 mm filter sand
while the GAC columns were charged with 6.8 ft of 12 x 40 mesh
GAC to achieve a 20 min empty bed contact time (EBCT) at a flow
of 0.5 gpm as indicated in Table 1. Each column was backwashed
only when necessary to achieve the desired flow rate. No media
loss was observed during backwashing. The GAC used in this study
was Cecarbon GAC 40 and was selected after a thorough evaluation
of various types of GAC as previously reported.
The average water quality of the nondisinfected water
-------
. c
o
•H
-p
Cn
•H
m
G
O
O
g
Q)
-P
CO
O
U
-P
O
no
0)
M
-------
TABLE 1: Physical Pilot Column Operating Conditions
Average
Average Hydraulic Average Total
Flow
Rate
(gpm)
non-
disinfected 0.50
Loading GAC
(gal/min EBCT
/ft ) (min)
Volume
Filtered.
Number of
Backwashes
(gal x!0^) Sand GAC
ozone
chlorine
0.50
2.5
2.5
20
20
251
239
5
3
1
1
dioxide
chloramine
chlorine
0.50
0.50
0.50
2.5
2.5
2.5
20
20
20 •
252
252
251
2
2
6
0
0
0
entering the five pilot column process streams after sand
filtration is indicated in Table 2. During the course of the
operational period, water temperature fluctuations were observed
from 3-29 C as indicated in Figure 2. After the addition of the
various disinfectants, slight variations in pH were observed due
to the acids and bases contained in the disinfectant solutions.
On the average, the pH of the chlorine dioxide contact chamber
effluent decreased 0.6 units to pH 7.0 while that for the
chlorine and chloramine contact chambers increased 0.1 &
units to pH 7.7 & 7.8, respectively. No change in pH
observed for the ozone process stream. Chlorine dioxide
generated with a 96% yield by the in-line mixing
nypochlorite/chlorite and sulfuric
injection into the process stream.
0.2
was
was
of
acid solutions prior to
Chloramines were formed
within the process stream with the injection of hypochlorite
followed within a few seconds by that of ammonia hydroxide. The
average 30 min demands for each disinfectant determined during
the operational period are compared in Table 3 along with their
respective average disinfectant contact chamber effluent residual
concentrations. While all residual concentrations were measured
as chlorine using the DPD titrimetric method, all residual
concentrations are reported as the specific disinfectant
indicated, and not as free chlorine. With the exception of the
ozone demand, these demands were determined by the difference in
the 30 min residual readings of the various disinfectants applied
at the same dosages to deionized water and nondisinfected
influent water. The ozone demand was determined by measuring the
-------
c
a
3
C
M
s
a
4-1
w
>1
en
4J
O
r-i
•H
a*
M-t
O
o
o
a.
E
O
O
Q)
S-i
Q. LU o: < f- ID a:
a iu o o uj
-------
TABLE 2: Average Nondisinfected Influent Water Quality
calcium
carbonates
(mg/L as CaC03) 0
bicarbonates
(mg/L as CaCO.,) 88
nardness
(mg/L as CaC03) 131
pH
7.6
(mg/L as CaCO_) 39
chlorides (mg/L) 28
turbidity (ntu) 0.3
temperature (°C) 19.4
TABLE 3: Average Disinfectant Contact Chamber Demands and
Residuals
Average
30 Minute Process Stream Process Stream
Disinfectant Average Average
Demand Disinfectant Disinfectant
Determinations Contact Time Residuals
(mg/L) (min) (mg/L;
ozone
chlorine dioxide
chlorine
monochloramine
2.5
0.7
1.8
0
30
30
30
3'0
0.5
0.5
1.0
2.2
difference in the concentration of ozone in the gas entering and
exiting the ozone contact chamber and subtracting the ozone
residual after making the appropriate corrections for gas and
water flows through the ozone contact chamber. The highest
demand was observed for ozone at 2.5 mg/L followed by chlorine at
1.8 mg/L and chlorine dioxide at 0.7 mg/L. Essentially no 30 min
demand was observed for chloramine, however significant demands
of 1-1.5 mg/L were observed after several days of storage at a
temperature of 30 C. Tne average concentrations of residual
disinfectant species observed in the contact chamber and sand
column effluents are presented in Table 4. While the ozone
residual dissipated completely across the sand column, the other
tnree disinfectant residuals were only slightly reduced. Ko
residual species of any disinfectant were observed in the
10
-------
TABLE 4: Average Disinfectant Contact Chamber and Sand
Column Effluent Residuals
Process
Stream
ozone ,
chlorine
dioxide
Disinfectant
Species
ozone
chlorine
dioxide
Contact
Effluent
0.5
0.5
Chamber
(mg/L)
Sand Columns
Effluent (mg/L)
0.0
0.3
chloramine
chlorine
chlorite 0.3
mono-
cnloramine 0.2
dichloramine 0.2
chlorine 0.0
mono-
chloramine 2.2
dichloramine 0.1
chlorine 0.0
chlorine 1.0
dichloramine 0.2
mono-
chloramine 0.1
0.4
0.0
0.1
0.0
1.7
0.3
0.0
0.9
0.2
0.1
effluent of the GAG columns during the one year 'operational
period.
-------
SECTION 5
SAMPLING, ANALYSIS, AND QUALITY ASSURANCE
Sampling, Preservation, and Storage
The sampling and analysis schedule for the pilot column
system during the one year operational period is indicated in
Table 5 with the numbered sampling point locations corresponding
to those in Figure 1. Those analyses performed included total
organic carbon (TOC), total organic halide (TOX), volatile
organics (VOC), haloacetic acids (HAA), chlorinated disinfection
by-products (CDBP), heterotrophic plate count (HPC), total
coliform (TC) , assimilable organic carbon (AOC), MS2 coliphage,
dissolved oxygen (DO), disinfectant residuals, pH, and
temperature. The CDBP analysis included such chlorinated
by-products as the haloacetonitriles, haloketones, chloral
hydrate, and chloropicrin. Qualitative confirmatory GC/MS
analysis was performed for the THMs, HAAs, and the CDBPs on the
influent to each process stream (the nondisinfected influent and
the disinfectant contact chamber effluents), as well as on the
cnlorinated terminal formation potential samples collected at
these locations. Those sampling locations in Figure 1 designated
17-36 were terminal formation potential samples which were dosed
in thg laboratory with chlorine or chloramine and held for 5 days
at 30 C in order to simulate the maximum formation of by-products
in the distribution system as well as- to obtain a measure of the
level of disinfection by-product precursors present. The
chlorinated terminal formation potential samples (-Cl~) were
spiked with varying amounts of chlorine dependent upon•the level
of demand in the sample. Over the course of the project, the
precipitator and nondisinfected influent and sand filter effluent
were spiked with 8-12 mg/L chlorine, all other sand filter
effluents were spiked with 6-7 mg/L, and all GAC effluents were
spiked with 4-4.5 mg/L. The chloraminated terminal formation
potential samples (-NH2C1), having a considerably lower demand,
were spiked with 5 mg/L ammonia-nitrogen followed by 4 mg/L
chlorine at all locations. After,the five day storage period at
30 C had elasped, all terminal samples were quenched with sodium
thipsulfate (VOC), sodium sulfite (TOX), or ammonium chloride
(HAA & CDBP) and stored at 4 C until analyzed. This same
quenching and storage procedure was also used for all
12
-------
c
o
4-1
IT
U
O
•-1
r
c
f^l
c.
F
ID
tr.
0)
I-H
•c
0'
.c
0
u.
tn
•r-4
in
£•
S.
•o
c
1C
CP
c
•••4
rH
C.
E
1C
V.
•
IT
(U
rH
.C
n:
r-
U"
n
r
n1
r
n
r
c
*"
c
r
o
0
r
o
V
^
ir
r*
^
r
^
o
1—4
C
c:
fV
C
o:
-*
r-
r~J
~*
IT
1-1
-
^
cv
^
3
c
o
Cr.
r-
V
IT!
n
CV
i-H
j,.
• ij 0)
O Q* O1
7. 0. 2-
m
•T-l
,
rH
TO
C
IS7ESIIIIIIIIII)
1 F S I E. 1 1 | | | | | | | | |
I^ESEIIIIIIIIIII
I555SIIIIIIIIIII
I55S5IIIIIIIIIII
ISESEIIIIlllllli
'ISEEE 1 | | | | | | | | | |
-
1 S £ E E 1 1 p 1 | | | | | | |
•ssysiiiiiiiiiii
lEsysiiiiiiiiiii
ISSSSSSFIIIIIIl'l
ISSSSIIIIIIIIiii
IEEEES E' £-111 || III
1 F 5 £ E 1 1 1 1 1 1 1 1 1 1 1
ISSSSSS-yilliliii
I££E£IIII|||||||
IEEE.ESESIIIIIIII
ISFEEIIIIIIIIlll
ISSS-SESSItllllll
lrs?£ll|l|||||||
S 2 2- 2 2- 1 1 1 t- E- 1 1 X 2 1 1
22222-1 1 IE-E-1 1X21 1
22-2-22 1 1 t E- E- 1 1X2-1 1
22223 1 1 IE-E-E 1X2 1 1
2-222-2 1 1 1 . t- E- 1 1 X 1 1 |
22222- 1 IIE-E-I 1X2 1 1
2- 233SIIIE-E-E" IX2II
3 2323 1 I.IE-E- 1 1X2 1 1
222221 1 IE-E-S 1X2 1 1
2"- 23321 1 IE-E-I IXI 1 l
22S:22222E-E-ET IXfcu2 |
2 " S '£ 2- 3 2 2 2 E^ E- 1 E X fc, 3 I
22322322.E"E-IFXb.S |
222-223S2- E-E-EEX 12- 1
23232 1 1 IE-E-I 1 1 1 12-
«• rH —i —i ,-H IT i1" vr o- or cv ^t c a. in ~t
r-H CV C^: CV CV -c: TJ p-i (V,
T;
in ^
E: < ~z at a> cj
E «5" rf PI CE O
E- X 3T 10 ^
III £
tr en en a c •
0. ET S, E" -M -rt Q.
CJXCJ<;cQ\vv\Cj CJ r-( W F
OCO E CJ C3 IT O E t- rf CJ a C.' D. E-
^
0-*
a1
^
ir;
m
c.
t
rH
^
Q
s
JJ
0'
2
^;
£
^
0)
2
CM
al
X
^y-
*Q)
•;,
^"
t— <"
3-
13
-------
instantaneous samples. NO preservative was used for TOG other
than immediate storage at 4°C. All organics samples were
collected in duplicate with the duplicate being analyzed when the
original result was suspected of being in error based on the
existing data trend, as determined by the section analyst. HPC
and TC samples were both collected and analyzed in duplicate and
the average value reported.
Analytical Procedures
Total Organic Carbon (TOC)—
EMSL Method 415.2 was used for TOC analysis with the
exception that the samples were stored at 4°C without
preservative and a linearity check was performed on a weekly
basis. J
Total Organic Halide (TOX)—
_EMSL Method 450.1 was used for TOX analysis without
modification.
Disinfection By-Products--
Those volatile organics and disinfection by-products
monitored during this project are listed in Tables 6, 7, and 8
along with their respective method detection limits and, where
applicable, their percent recoveries. For the HAA and CDBP
methods, which were still under development at the time of this
study, the detection limit was defined as the lowest
concentration at wnich the method response was still linear.
This process is exemplified in Figure 3 which compares the true
and reported concentrations of trichloro- and dichloro-
acetonitrile with respective detection limits of 0.2016 and
0.0032 ug/L.
EMSL Method 502.2 was used for VOC analysis with the
following exceptions:
The analytical column was a 30 m x 0.53 mm I.D. DB-624
column which was held initially at 33 C for 4 min,
ramped at 2.4 C/min to 40°C, held 1 min, ramped at
6 C/min to 160UC, and held for 10 min.
The 25 cm x 2.7 mm I.D. trap was comprised of equal
volumes of Tenax, silica gel, and charcoal.
The internal standard used was 1,2-dichloropropane.
14
-------
CM
0)
o .3
4-J
•r-l
G
O
CD
O
z
o
(0
10 J
ui -g
M
O
CO
0)
•H
8
4J
0)
CM
OO
S
o
Cn
•H
15
-------
Table 6. Gas Chromatography Detection Limits for VOC "Analysis
Standard Cone, Method Detection
Compound ug/L Limit, ug/L
dichloromethane 2.1 .003
1,2-dichloroethylene 5.0 0.5*
chloroform 2.4 .001
tetrachloromethane 2.6 .005
1,2-dichloroethane 3.0 0.01
1,1,2-trichloroethylene 3.5 0.005
bromodichloroir.ethane 3.2 0.02
1,1,2-trichloroethane 2.3 0.008
chlorodibromomethane 5.9 0.017
bromoform 11.6 0.04
*Interference from methyl-t-butylether used as the extraction
solvent in methods 551 & 552.
EMSL Draft Method 352 was used for the anlaysis of the
haloacetic acids with the following exceptions:
- The DB-210 column was used as the analytical column and
wasoheld at 32 C for 10 min and ramped at 10 C/min to
200 C with a final Qhold of 5 min. The injection port
temperature was 160 C and a splitless injection with a
20 sec delay was employed.
EMSL Draft Method 551 was used for the analysis of
chlorinated disinfection by-products which included the
haloacetonitriles, naloketones, chloral hydrate, and
chloropicrin, with the following exceptions:
A 3fim x 0.325 mm I.D. DB-210 column with a 0.5 urn film
tnickness was used as the analytical column and was
16
-------
Table 7. Gas Chromatography Detection Limits for HAA Analysis
Method
Compound
monochloro-
acetic acid
dichloro-
acetic acid
monobromo-
acetic acid
trichloro-
acetic acid
bromochloro-
acetic acid
dibromo-
acetic acid
2,4,6-trichloro-
phenol
2,4-dichloro-
phenol
2-chloro-
phenol
*Method contaminant
Standard
Cone, ug/1
3.0
3.5
3.0
1.0
2.0
0.25
1.5
25
100
interference
Dectection
Limit, ug/L
0.1
0.012
0.01
0.01
0.02
0.0025
0.02
4.0'*
0.81
Method
Recovery %
46.2
83.1
63.0
71.6
43.9
62.4
87.8
70.0
101
programmed as previously indicated for Method 552.
Confirmatory GC/MS Analysis—
The standard concentrations and method detection limits for
those volatile organics and disinfection by-products that were
qualitatively confirmed by GC/MS analysis are indicated in Table
9.
EMSL method 624 was used for VOC confirmatory analysis by
17
-------
Table 8. Gas Chromatography Detection Limits for CDBP Analysis
Compound
dichloroaceto-
nitrile
trichloroaceto-
nitrile
bromochloroaceto-
nitrile
dibromoaceto-
nitrile
chloral hydrate
chloropicrin
1,1-dichloro-
propanone
1,1,1-trichloro-
propanone
Standard
Cone, ug/1
.20
.10
.15
.20
.40
.15
.20
.5
Method
Detection
Limit, ug/1
.008
.004
Method
Recovery %
.0032
.0016
.0024
.0032
.0256
.0024
97.3
89.1
99.3
1.19
55.2
96.3
62.5
91.6
GC/MS with the following exceptions:
standards
- 20mL aliquots were spiked with internal
(2-broflio-l-chloropropane, fluorobenzene, and
bromof luorobenzene) and purged for 15 min at a helium
flow of 40 mL/min on to a Tenex and silica gel trap.
The trap was rapidly heated to 180°C, desorbed for 4 min at
30 mL/min, and then vented to atomsphere and purged for an
additional 110 min.
A 60 m x 0.25 mm ID DB-5 capillary column with a 25
micron coating thickness was used for analysis. A 20:1
split ratio was employed during injection.
Spectra were acquired on a Finnigan 4023C GC/MS in the
full scan mode from 35-260 amu each second. A compound
was considered confirmed if the signal-to-noise ratio
was greater than 5 and the spectrum obtained met an
18
-------
Table 9. Confirmatory GC/MS Standard Concentrations and
Qualitative Detection Limits
Standard Method Detection
Compound Concentration, ug/L LJT.it, ug/L
chloroform
bromodichlorome thane
cnlorodibromomethane
bromof orm
monochloroacetic acid
dichlor'oacetic acid
monobromoacetic acid
trichloroacetic acid
bromochloroacetic acid
dibromoacetic acid
2,4,6-trichlorophenol
2 , 4-dichlorophenol
2-chlorophenol
dichloroacetonitr ile
trichloroacetonitr ile
bromochloroacetonitr ile
dibromoacetonitr ile
chloral hydrate
chloropicrin
1, 1-dichloropropanone
1,1, 1-trichloropropanone
8.0
8.0
8.0
8.0
30
3.5
3.0
1.0
1.0
0.25
25
1.5
100
0.20
0.10
0.15
0.20
0.40
0.15
0.50
0.25
0.05
0.10
- 0.10
0.30
0.10
0.05
0.10
0.10
0.05
0.10
0.60
0.04
0.25
0.05
0.03
0.10
0.10
0.10
0.07
0.20
0.15
acceptable fit criteria (usually >850) when compared to
external standard spectra. The GC/MS was calibrated
each,day with FC43 and verified with the bromofluoro-
benzene in each sample. The other two internal
standards were used to assess compound recovery.
EMSL Draft Method 551 & 552 were used 'for the initial
extraction and concentration of the chlorinated disinfection
by-products and haloacetic acids with further processing for
GC/MS confirmation as follows:
After GC analysis, the remaining extracts were sealed
and stored glong with their corresponding standards and
blanks at 4 C until GC/MS analysis was performed. A 200
uL aliquot of the haloacetic acid and chlorinated
disinfection by-product extracts of each sample were
combined and blown down by gentle air stream to 5 uL. A
2 uL aliquot of this concentrate was injected splitless
onto a 60 m x 0.25 mm diameter DB-5 capillary column
with a 25 micron film thickness, held for 5 min at 40 C,
19
-------
andoprogrammed to 70 C at 3 C/min and then to 200°C at
6.5 C/min. Spectra were generated using multiple ion
detection (MID) with eight MID descriptors which were
sequentially changed during the chromatographic run.
Each discriptor scanned up to 24 different masses in
one second scans. Each mass _+ 0.1 amu was scanned in 12
msec to maximize signal-to-noise. The instrument was
calibrated daily with FC43 and periodically checked for
agreement with DFTPP tuning specifications. The sample
spectra obtained were then compared to those of external
standards and blanks and were confirmed if the sample
response was significantly greater than that of the
blank, the signal-to-noise ratio was greater than 5, and
an acceptable fit, usually >850, was obtained.
Heterotrophic Plate Count (HPC), Total Coliform (TC), and
Dissolved Oxygen (DO)—
Analysis for HPC, TC, and DO were performed using Parts 907C,
909A, and 421F of the 16th Edition of Standard Methods2.
Assimilable Organic Carbon (AOC)—
Samples were collected in 45 mL vials which were prewashed
with detergent, rinsed 5 times with hot water, acid rinsed with
0.1 N HC1, rinsed 3 times with carbon-free deionized water,
dryed, capped with foil, and heated to 550 C for 6 hours. The
teflon-lined silicone seota were soaked in a 10% potassium
persulfate solution at 60 C for one hour and then rinsed with
carbon-free deionized, distilled water.
After collection, samples were returned to the laboratory
wereQthey were quenched with' 10% sodium thiosulfate and placed in
a J0 C waterbath for 30 min. The samples were then cooled to
15 C and innoculated with a culture of Pseudomonas P-17 to
achieve an -initial plate count of 50-500 cfu/mL. After
subsequent incubation at 15 C for 7, 8, and 9 days, three vials
from each ^ocation were assayed using the standard spread plate
technique . Each vial was assayed using two dilutions in
triplicate. The viable counts for the three days were then
averaged and the AOC concentration determined using the Vander
Kooij yield factor .
MS2 Coliphage—
The initial MS2 coliphage culture was received from EPA in
Cincinnati in a tryptone yeast e.xtract broth (10 g bacto
tryptone, 8 g sodium chloride, 1 g glucose, & 1 g yeast extract/L
20
-------
distilled water) with a titer of approximately 10 plaque
forming units (pfu) per mL. The stock culture was prepared from
the_6original culture by plating four 20 mL agar plates with a
10 dilution as indicated below. These plates were then
harvested by mixing the contents of the plates with 200 mL
tryptone yeast extract broth, adding 0.28 g of disodium EDTA and
0.026 g of lysozyme for each 100 mL of agar-broth mixture,
stirring for 30 min,. centrifuging at 2200 rpm for 30 min, and
collecting the supernatant. Each month, the stock culture was
prepared in this manner from the previous month's stock culture
and consistently contained 3-6 x 10 pfu/mL. The stock culture
was subsequently diluted 1:10 in sodium chloride-calcium chloride
diluent (8.5 g sodium chloride and 0.22 g calcium chloride/L
distilled water) to produce 2 L of phage seed solution containing
10 pfu/mL. This seed • solution was metered into the
nondisinfected water entering the p-ilot column system such that
the final concentration was 10 -10 pfu/mL. After feeding the
seed solution a minimum of 2-3 hours to ensure saturation of the
contact chambers, 200 mL samples were collected dropwise from
each location over a period of 10 min. This was accomplished by
inserting a small capillary tube into the side of a section of
latex tubing attached, to the sample spigot and slightly
restricting the 0.8 gpm flow in the latex tubing such that a
dropwise flow was obtained at the capillary.
Prior to sample collection, petri dishes were prepared with
15-20 mL of bottom agar (10 g bacto tryptone, 10 g bacto agar,
2.5 g sodium chloride, 2.5 g potassium chloride, and 1 mL 1 M
calcium chloride/L distilled water) and refrigerated. A host E.
coli solution (ATCC 15597) was prepared by innoculation of 10-2
ml of tryptone yeast extract broth in a 20 mL test tube and
incubating at 35 C. Top agar (10 g bacto tryptone, 8 g sodium
chloride, 8 g bacto agar, 1 g glucose, 1 g yeast extract/L
distilled water) was also prepared and refrigerated.
After sample collection, 3 ml of melted top agar was
dispensed into a test tube and held at 44 C in a water bath.
Sample dilutions were made with sodium chloride-calcium chloride
diluent and one mL of the diluted sample was added to the test
tube along with 4 drops of. host E. coli solution. After thorough
mixing, the test tube solution was added to the petri dish on top
of the bottom agar. The petri dish was then swirled, allowed to
s.tand for 15 min, inverted, and incubated at 35°C for 18-24
hours. The plates were then read and the number of plaque
forming units recorded.
Disinfectant Residuals—
All disinfectant residuals were measured via the
LaMotte-Palin DPD-FAS Method using a LaMotte Model DT Laboratory
21
-------
titration kit available from LaMottte Chemical Products,
Chestertown, MD. The stock standard solution of ferrous ammonium
sulfate (FAS) was freshly prepared each month and standardized
against a primary standard dichromate solution as indicated in
Part 408D of the 16th Edition of Standards Methods .
Temperature and pH—
Temperature and pH measurements were made in accordance with
Parts 212 and 423 of the 16th Edition of Standard Methods .
/
Quality Assurance
Precision was assessed for all organic analyses by comparing
the results of duplicate samples. One duplicate sample was
analyzed each week for TOG, TOX, VOC, HAA, and CDBP analyses.
The sample which was analyzed in duplicate was the first and the
last sample to be analyzed within each weekly sample set. In
this manner, the precision determined was indicative of the
entire sample set over the time period in which the sample set
was analyzed. The average precision observed for each organic
parameter monitored during the operational period is indicated in
Table 10 along with the concentration range and the number of
duplicate samples analyzed. The number of observations for each
parameter varied with sampling frequency and detectability. The
percent relative deviation (PRO) of each duplicate was calculated
by determining the percent difference between the higher of the
two data points (H) and their mean (X), i.e. PRO •= 100 (H-X)/X.
The average PRO represents the average of all PRD's determined
during the one year operational period. The PRO for all
parameters monitored were within 10% except for monochloroacetic
acid, monobromoacetic acid, dibromoacetic acid,
trichloroacetonitrile, and dibromoacetonitrile which were within
20%. The higher PRD's observed for these substances resulted
from the very low concentrations observed, which were in most
cases, below the method detection limit.
Accuracy was assessed through the evaluation of the relative
error determined for the internal quality control samples
obtained from EPA and through performance audits conducted by
EPA. Unfortunately, quality control samples for the HAA and CDBP
methods were not available during the project and they were not
included in the EPA performance audits until the very end of the
operational period in May, 1990. Hence, the subsequent
perofrmance evaluation in November, 1990 has been included to
further qualify the data presented in this report. Internal
quality control samples were introduced approximately once each
month while EPA performance evaluations were conducted every six
montns. The average percent relative error observed for the
internal quality control samples, which were only available for
22
-------
Table 10. Analytical
Analysis
Parameter
TOC, mg/L
TOX, ug/L
chloroform, ug/L
dichlorobromo-
me thane, ug/L
1, 2-dichloro-
ethane, ug/L
carbon tetra-
chloride, ug/L
1,1,1-tricnloro-
etnene, ug/L
monochloro-
acetic acid, ug/L
dichloro-
acetic acid, ug/L
trichloro-
acetic acid, ug/L
monobromo-
acetic acid, ug/L
dibromo-
acetic acid, ug/L
bromochloro-
acetic acid, ug/L
chloral
hydrate, ug/L
dicnloro-
acetonitrile, ug/L
tricnloro-
acetoni trile, ug/L
chloropicrin, ug/L
1, 1-dichloro-
propanone, ug/L
1 , 1, 1- trichloro-
propanone, ug/L
bromochloro-
acetoni trile, ug/L
dioro.Tio-
acetoni trile, ug/L
Precision as
Cone. Range
2.497-4.405
68.8-129.5
0.967-7.830
0.131-1.808
0.009-0.207
0.007-0.681
0.004-0.314
0.018-0.938
0.870-6.111
0.264-2.119
0.006-0.658
0.005-0.064
0.193-1.723
0.027-2.911
0.078-3.557
0.001-0.125
0.007-1.182
0.256-1.502
0.037-0.179
0.008-0.566
0.0003-0.029
Measured by
No. of
Observ.
52
50
50
50
12
6
7
53
53
53
29
53
43
53
53
'49
53
53
53
53
53
•Duplicate
Sample
Average Percent
Relative
Deviation
4.6
4.4
5.2
5.1
5.8
3.6
3.1
17.2
8.6
9.4
13.0
11.5
7.1
8.9
4.3
17.7
6.2
4.9
4.7
6.1
12.5
23
-------
TOG and VOC analyses, were at or below 10% (Table 11). In the
EPA performance evaluations, the acceptance range was -18% to 39%
for TOG and +_ 20% for the VOC' s as indicated in Tables -12-16.
The performance evaluations for these substances were acceptable
for the most part except that chloroform and bromoform were
occasionally just outside the upper acceptance limit. The
average relative error determined for chloroform in the
performance evaluations averaged 16% while that for bromoform
averaged 7%. The acceptance ranges for those substances analyzed
by the HAA and CDBP methods varied by compound and were quite
broad, ranging from the detection limit (DL) to over 100% (Tables
15 & 16). The performance evaluations for all HAA and CDBP
compounds were acceptable except for one trichloroacetic acid
result which slightly exceeded the upper acceptance range. Those
unacceptable results indicated for dibromoacetic acid,
dichloroacetic acid, monochloroacetic acid, and monobromoacetic
acid were considered invalid because sample No. 1 did not contain
these compounds. Sample No. 2 was accurately analyzed at the
same time and with the same method as sample No. 1. However,
sample No. 1 contained no measurable chromatographic peaks even
though the concentration was only one-sixth to one-third that of
Sample No. 2 and was 80-500 times that of the method detection
limits (Table 7).
GC/MS Qualitative Confirmation
Qualitative confirmation by GC/MS analysis was performed for
the trihalomethanes, haloacetic acids, haloacetonitriles,
haloketones, chloral hydrate, and chloropicrin. Those locations
confirmed by GC/MS were the influent of each process stream after
30 min of disinfectant contact time and the terminally
chlorinated sand filter effluents of the nondisinfected, ozone,
cnlorine dioxide, and chlorine process streams as well as the
terminally chloraminated sand filter effluent of the chloramine
process stream. The percentage of possible GC/MS confirmation
and the average mass spectral fit for each compound are presented
in Table 17 for the process stream influents and in Table 18 for
the terminal distribution system simulation. A value of 1000 for
tne mass spectral fit would indicate that the compound spectra of
the sample matched perfectly with that of the computer based mass
spectral library derived from external standards.
In general, higher levels of confirmation were observed in
the terminal distribution system simulation than in the process
stream influents due to the significantly higher concentrations
observed in the terminal samples. Those low confirmation levels
observed for the process stream influents resulted from the low
concentrations observed and the significantly lower detection
limits of the gas chromatography methods (Tables 6, 7, & 8) as
compared to those of the GC/MS confirmatory procedure (Table 9).
24
-------
Table 11. Analytical Accuracy as Measured by Internal Quality
Con'trol Sample Analysis
Concentration No. of Average %
Parameter Range ug/L Observ. Relative Error
TOG
chloroform
bromodichloro-
methane
chlorodibromo-
methane
bromoform
carbon tetra-
chlor ide
1,1,1-trichloro-
ethylene
23.0 - 61.4
10 - 45
20
20
20
10
10
13
6
2
2
2
1
1
8.36
7.0
2.45
6.2
10.3
6.0
10.0
Table 12. USEPA Performance Evaluation DiMR009, April 1989
True Reported Acceptance
Value Value Limit Performance
Parameter mg/L mg/L ' mg/L Evaluation
TOG - 11.5 13.1 9.45-16.0 acceptable
Those low confirmation levels observed for the terminal
distribution system simulation resulted, for the most part, from
the low levels observed with average concentrations at or below
the GC/MS -detection limits.
25
-------
Table 13. USEPA
Parameter
bromodichloro-
methane
bromof orm
chlorodibromo-
methane
chloroform
total
trihalomethane
Table 14. USEPA
Parameter
bromodichloro-
methane
bromoform
-
chlorodibromo-
methane
chloroform
total
trihalomethane
r
Performance
Sample
No.
1
2
1
1
2
1
2
1
2
True
Evaluation
Reported
Value Value
ug/L
22.5
57.8
12.3
66.9
7.66
80.5
10.6
63.8
53.1
269
Performance
Sample
No.
1
2
1
2 -
1
2
1
2
1
2
True
Value
ug/L
71.4
8.40
48.1
9.12
47.8
15.6
77.8
24.4
245.1
57.5
ug/L
22.2
49.8
12.8
69.2
8.40
69.5
12.3
50.8
55.7
239
Evaluation
Reported
Value
ug/L
75.0 .
9.28
47.9
11.3
47.8
17.3
82.5
30.2
253.2
68.1
WS024, May
Acceptance
Limit
ug/L
18.0-27.0
46.2-69.4
9.84-14.8
53.5-80.3
6.13-9.19
64.4-96.6
8.48-12.7
51.0-76.6
42.5-63.7
215-323
19,89
Performance
'Evaluation
acceptable
acceptable
acceptable
acceptable
acceptable
acceptable
acceptable
unacceptable
acceptable
acceptable
WS025, November 1989
Acceptance
Limit
ug/L
,57.1-85.7
6.72-10.1
38.5-57.7
7.30-110.9
38.2-57.4
12.5-18.7
62.2-93.4
19.5-29.3
196-294
46.0-69.0
Performance
Evaluation
acceptable
acceptable
acceptable
unacceptable
acceptable
acceptable
acceptable
unacceptable
acceptable
acceptable
26
-------
Table
Parameter
USEPA Performance Evaluation WS026, May 1998
True Reported
Sample Value Value
No. ug/L uq/L
Acceptance
Limit
ug/L
Performance
Evaluation
bromodichloro-
methane
bromoform
chlorodibromo-
methane
chloroform
total
tribal ome thane
dibromoacetic
acid
dichloroacetic
acid
monobromoacetic
acid
monochloroacetic
acid
tr ichloroacetic
acid
2,4,6-trichloro-
phenol
bromochloro-
acotoni trile
dibromoaceto-
ni trile
1
2 .
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
3
4
3
4
• I I WT ..mm
25.8
60.5
15.4
77.4
22.1
72.9
7.53
59.1
70.8
270
1.27
6.77
2.58
12.0
4.10
13.9
8.33
16.7
1.10
5.87
5.04
12.6
0.671
4.58
0.762
2.37
i — • ,.—
26.46
65.12
17.43
74.38
24.43
71.54
10.10
63.92
78.42
27.5.0
____
3.81
10.4
____
11.8
____
18.5
0.767
5.38 '
3.85
11.6
0.571
4.38
0.642
2.34
20.6-31.0
48.4-72.6
12.3-18.5
61.9-92.9
17.7-26.5
58.3-87.5
6.02-9.04
47.3-70.9
46.6-85.0
216-324
D.L.-4.98
D.L.-21.2
.0212-4.12
D.L.-22.5
1.30-4.01
D..L.-18.4
D.L.-8.00
D.L.-20.6
D.L.-1.58
D.L.-0.01
D.L.-11.2
D.L.-28.1
0.201-1.37
3.55-6.27
0.309-1.03
1.01-3.17
acceptable
acceptable
acceptable
acceptable
acceptable
acceptable
unacceptable
acceptable
acceptable
acceptable
unacceptable*
acceptable
unacceptable*
acceptable
unacceptable*
acceptable
unacceptable*
acceptable
acceptable
acceptable
acceptable
acceptable
acceptable
acceptable
acceptable
acceptable
27
-------
Table 15. USEPA Performance Evaluation WS026, May 1990 (Cont.)
Parameter
1,1-dichloro-
propanone
trichloro-
acetonitrile
True Reported Acceptance
Sample Value Value Limit
So. ug/L ug/L ug/L
3
4
3
A
1,1,1-trichloro- 3
propanone 4
0.692 0.587
4.04 3.47
0.393 0.295
2.84 2.05
0.452 0.372
7.68 5.41
Performance
Evaluation
0.313-1.09 acceptable
2.22-7.23 acceptable
0.075-0.72 acceptable
0.271-5.13 acceptable
0.136-0.753 acceptable
D.L.-14.5 acceptable
'suspect performance evaluation samples,
Table 16. USEPA Performance Evaluation WS027, November 1990
Parameter
bromodichloro-
methane
bromoform
chlorodibromo
methane
chloroform
total
trihaolmethane
dibromoacetic
acid
dichloroacetic
acid
True Reported Acceptance
Sample Value Value Limit Performance
No- ug/L ug/L ug/L Evaluation
1
1
1
1
32.4 29.5
22.7 23.28
30.8 30.35
87.0 82.85
172.9 166.0
4.84 2.45
6.96 5.57
25.9-38.9 . acceptable
18.2-27.2
24.6-37.0
69.6-104
138-207
acceptable
acceptable
acceptable
acceptable
D.L.-16.4 acceptable
D.L.-13.5 acceptable
28
-------
Tabla 16. USEPA performance Evaluation
(Cont.)
Parameter
True Reported
Sample Value Value
No.
monobromoacetic 1
acid
monochloroacetic 1
acid
trichloroacetic 1
acid
2,4,6-trichloro- 1
phenol
bromochloro- 2
acetonitrile
dibromo- 2
acetonitrile
dichloro- 2
acetonitrile
1,1-dichloro- 2
propanone
trichloro- 2
acetonitrile
1,1,1-trichloro- 2
propanone
ug/L ug/L
8.40 4.22
10.5 10.97
8.33 16.09
7.32 6.80
1.98 1.72
5.33 5.38
0.075
1.33 1.29
1.76 1.81
VvS027, November 1990
Acceptance
Limit Performance
ug/L Evaluation
D.L.-10.7 acceptable
D.L.-15.9 acceptable
D.L.-14.6 unacceptable
D.L.-16.6 acceptable
D.L.-4.33 acceptable
0.0846-6.90 acceptable
D.L.-D.L. unacceptable
0.67-2.07 acceptable
0.442-3.14 acceptable
0.989 0.93 0.118-1.35 acceptable
29
-------
Table 17. Percent GC/MS confirmation and Average Mass Spectral
Fit for the Influent of each Process Stream.
Compound
cnloroform
bromodicnloro-
methane
chlorodibromo-
methane
monochloro-
acetic acid
dichloroacetic
acid
monobromoacetic
acid
trichloroacetic
acid
bromochloro-
acetic acid
dibromoacetic
acid
dichloroaceto-
nitrile
trichloroaceto-
nitrile
bromochloro-
acetonitrile
dibromoaceto-
nitrile
chloral hydrate
chloropicrin
1,1-dichloropro-
panone
1,1,1-trichloro-
propanone
Nondis
Ozone
CIO.
NH2C1
Chlorine
%/Fit
92/962
83/889
57/931
100/982
96/979
30/932
98/990
95/986
39/952
78/996
23/979
13/971
70/980
28/907
32/958
47/969
16/964
%/Fit
98/956
68/913
56/925
100/986
98/981
93/978
98/989
96/986
100/986
83/995
24/922
15/954
62/975
24/853
20/944
28/953
21/966
%/Fit
96/967
83/901
56/890
97/987
98/980
77/943
98/987
98/994
98/994
89/995
29/960
27/949
58/974
31/906
40/936
51/962
36/969
%/Fit
100/983
96/956
58/897
94/990
98/980
47/923
100/989
98/996
60/982
96/998
30/959
56/975
62/969"
83/909
60/959
85/971
69/975
%/Fit
96/996
96/999
98/984
94/986
96/982
100/969
98/990
98/998
98/997
98/998
70/966
96/996
98/991
96/952
96/985
96/972
94/985
30
-------
Table 18. Percent GC/MS Confirmation and Average mass Spectral
Fit for Terminal Sand Filtered Distribution System
Similation.
Compound
chloroform
bromodichloro-
methane
chlorodibromo-
methane
monochloro-
acetic acid
dichloroacetic
acid
monobromoacetic
acid
trichloroacetic
acid
bromochloro-
acetic acid
dibromoacetic
acid
dichloroaceto-
ni tr ile
trichloroaceto-
nitrile
bromochloro-
acetonitrile
dibromoaceto-
nitrile
chloral hydrate
chloropicrin
1, 1-dichloro-
propane
1,1,1-trichloro-
propane
Nondis-Cl
%/Fit
100/996
100/998
100/978
98/994
100/974
100/999
"
100/963
100/995
100/999
100/989
63/953
92/973
92/973
92/946
100/991
42/962
89/961
., Ozone-Cl0 C109-C1_
£+ £.* £ £*
%/Fit
100/994
77/998
92/994
99/992
100/968
100/999
100/982
100/995
100/999
92/989
45/967
92/994
92/994
92/947
100/988
50/981
92/967
%/Fit
100/993
100/998
100/999
100/971
100/977
100/999
100/986
100/997
100/999
100/997
•
72/973
92/993
92/993
92/944
100/988
58/955
92/982
ci2-ci2
%/Fit
100/992
91/999
100/998
100/995
100/969
100/999
100/976
100/997
100/999
100/995
91/972
91/988
91/988
100/946
100/984
36/967
100/978
NH2C1
%/Fit
100/998
100/999
100/999
100/972
100/972
100/999
100/970
100/996
100/999
92/995
54/953
92/997
92/997
92/951
92/987
42/959
92/982
31
-------
SECTION 6
RESULTS AND DISCUSSION
DISINFECTION BY-PRODUCTS
Total Organic Carbon (TOG)—
While TOC itself is not a disinfection by-product (DBF) , it
is a general measure of the amount of precursor material
available for DBF formation. Seasonal variations of total
organic carbon levels in the influent of each process stream are
indicated in Figure 4. Tne average TOC levels in the contact
chamber effluents across the one year operational period were
3.12, 2.86, 3.18, 3.23, and 3.15 mg C/L for the nondisinfected,
ozone, chlorine dioxide, chloramine, and chlorine process
streams, respectively. As expected, sand filtration had no
effect on TOC in the precipitator water entering the pilot column
system_ which averaged 3.16 mg C/L. Relative to the
nondisinfected influent, ozonation produced an average TOC
reduction of 0.26 mg C/L (8%) while an additional removal
averaging 0.56 mg C/L (18%) occurred across the ozone sand column
(Figure 5) for a total removal of 0.82 mg C/L (26%) prior to GAC
filtration. Based on the levels of heterotrophic bacteria in the
effluents of the'ozone contactor 'chamber and ozone sand column
which are discussed later in this report, the reduction of TOC
across the ozone contact chamber appears to have resulted
primarily from oxidation while that across the ozone sand column
can be attributed to biodegradation. Essentially no reductions
in TOC attributable to predisinfection or sand filtration were
observed for the other process streams.
TOC removals across the GAC columns were very similar for the
nondisinfected, chlorine dioxide, chloramine, and chlorine
process streams (Figure 6) with each column reaching steady state
on day 200. Average removals relative to the respective sand
column effluents of 22-23% were observed after steady-state was
reached with average GAC effluent concentrations of 2.5-2.6 mg
C/L. The ozone GAC column also reached steady-state on day 200
with a subsequent average removal of 26% relative to the ozone
sand column effluent and an average GAC effluent concentration of
1.8 mg C/L. The overall TOC removal observed for ozonation
32
-------
M WUJ
OQZCMO o
UZOOCM.J
0:0 N_IX x
Q.ZOOZO
o
in
to
o
o
o
in
CM
013
CM
Lu
O
-in Q
o
o
o
in
-o
o
o
m
o
o
o
O
o
in
to
O
o
o
to
o
o
m
CM
I- O O
o
g
m
Cl)
i-i
4J
CO
to
CO
0)
o
O
M
O
CO
0)
o
•P
4-1
£
H
Q)
JC
4J
c
•H
0)
Q)
r-1
U
O
EH
o
CO
•H
M
«J
I
U
Q)
tn
•H
En
33
-------
O
to
fO
O
O
to
o
in
CM
CM
U.
O
O
O
o
in
- o
T
O
o
o
T~
O
O
in
ro
H- O O
o
o
o
ro
O
o
to
CM
—I-
o
o
o
CM
in
0)
J-l
3
Cn
•H
34
-------
o
o o
-------
followed by sand filtration and subsequent GAG filtration after
steady-state was reached on day 200 was 47%.
Total Organic Halide (TOX)—
With an average nondisinfected influent concentration of 25
ug Cl/L, TOX levels increased significantly after 30 min of
disinfectant contact time across the chlorine dioxide,
chloramine, and cnlorine contact chambers to 86, 99, & 246 ug
Cl/L, respectively (Figure 7). A reduction in TOX averaging 33%
occurred across the ozone contact chamber with an average
effluent concentration of 16 ug Cl/L. Further reduction to an
average concentration of 11 ug Cl/L was observed across the ozone
sand column for a combined reduction of approximately 50% for
ozonation followed by sand filtration (Figure 8) . While some
variation was observed due to experimental error, no significant
change in TOX was observed across the sand columns of the
chlorine dioxide, chloramine, and chlorine process streams.
Treatment of the sand filtered effluents with free chlorine
followed by 5-day storage (TOX-C12) significantly increased TOX
levels for all process streams as indicated in Figure 9. Similar
TOX-C12 levels were observed for the precipitator and the
nondisinfected and chlorine sand column effluents with respective
averages of 585, 557, and 540 ug Cl/L. Predisinfection with
ozone and chlorine dioxide reduced the levels of TOX-C1 by 39 &
32 percent, respectively, relative to the nondisinfected sand
column effluent with average TOX-C10 levels of 339 and 379 uq
Cl/L, respectively. ^
Treatment of the sand filtered effluents with chloramine
followed by 5-day storage (TOX-NH-C1) resulted in an increase to
an average of 27 ug Cl/L for the ozone process stream, while that
for the chlorine dioxide process stream remained relatively
constant at 89 ug Cl/L (Figure 10). Similar treatment in the
chloramine process stream resulted in an reduction of 32% to 59
ug Cl/L relative to the sand column effluent. Tnis unusual
reduction in TOX for the chloramine process stream has been
observed from terminally stored samples in previous studies and
is, as yet, unexplained. The TOX-NH Cl levels observed for the
precipitator and nondisinfected sand column effluent were
essentially the same as that observed for the chloramine sand
column effluent with respective averages of 54 and 44 ug Cl/L.
GAG filtration with a 20 min empty bed contact time resulted
in TOX levels of less than 30 ug Cl/L over the one year
operational period for all process streams except for the
chlorine process stream (Figure 11). The nondisinfected and
ozone GAG columns reached apparent steady-state conditions on day
and day 130, respectively, achieving average percent removals
36
-------
o
h m
o
o
ro
o
in
CM
Q)
x;
u
u
m
4J
G
O
U
C
(C
-p
u
CD
U-t
CM
o
"- m
o
- o
_ o
m
M
•H
P
CD
tf* «
•P g
(0
C 0)
•H M
-P
w en
iH
OJ CO
> w
Q) (I)
X
O
Q) O
X! (0
-P 0)
4-1 IM
O O
C -P
O fl
to
0)
i- O x
ID O O
37
-------
e-^-
T"
O
in
T"
O
O
K)
CM
T~
O
O
CM
h- O X Z) O
O
in
o
T~
o
o
o
ro
O
O
ro
o
in
CM
CM
Lt.
O
O
-o
_o
in
-o
o
m
x:
o
to
Q)
c
0)
w
I
rH
O
u
G
ns
en
c
•H
tn
Q)
O
EH
Q)
Q)
m M
O -P
w
c
O W
.tn to
•rH Q)
M 0
(0 O
04 M
S Pn
O
U
CO
Q)
S-l
3
tn
•H
38
-------
CM
CM CM
i_l_l
vv
Oil CM
I aacM_i
QL-z.-z.-m
o « o l-
I- W W I Q
< az
H-2UJ2<
H W 2
-------
o
m
ro
o
o
ro
O
in
013
• 1
SSI
u
CN
o
o
o
in
- o
o
-------
I- O X
13 O O
tn
•H
En
41
-------
after steady-state of 41% for both columns with respective
average effluent TOX concentrations after steady-state of 13 and
6 ug Cl/L. Steady-state was reached about day 150 for the
cnlorine dioxide GAG column with an average percent removal and
average effluent concentration after steady-state of 77% and 19
ug Cl/L. The chloramine GAG column reached steady-state about
aay 100 while that for the chlorine GAG column occurred on day
120. Tneir respective average removals after steady-state were
74 and 65 percent with corresponding average effluent
concentrations of 24 and 82 ug Cl/L. Thus after one year of
operation, GAG filtration continued to produce sianificant TOX
removals in all process streams.
As with the sand filtered effluents, treatment of the GAG
column effluents with free chlorine followed by 5-day storaae
(TOX-C12) significantly increased TOX levels in all process
streams as indicated in Figure 12. All GAG columns reached
steaay-state on about day 150 with average TOX-C10 steady-state
erfluent concentrations of 290, 280, and 310 ug^Cl/L for the
nondisinfected, chlorine dioxide, and chlorine GAG columns
respectively. The average TOX-C1- level of 310 ug Cl/L in the
cnlorine GAG effluent was comprised of approximately 27 percent
(84 ug Cl/L) TOX breakthrough and 73 percent (226 ug Cl/L)
unreacted TOX precursors. The ozonated sand filtered water
entering the GAG column contained a lower level of TOX-C1
(Figure 8) and resulted in subsequently lower GAG effluent!
concentrations averaging 167 ug Cl/L TOX-C1, after reachinq
steaay-state on day 150. While TOX-C1, levels2 in the. chlorine
dioxide sand filter effluent were similar to those of the ozone
sand filter effluent, significantly higher levels of TOX-C1 were
observed i'n the chlorine dioxide GAG effluent as compared fo the
ozone GAC effluent. Despite the nigher levels of TOX-C1
observea in the GAC effluents, significant TOX-C10 reduction!
were still evident after steady-state was reached wi«i respective
average removals of 49, 51, 26 & 43% for the nondisinfected,
ozone, chlorine dioxide, and chlorine GAC columns. The TOX-C1
levels in the chloramine GAC column effluent were similar to
those of the nondisinfected ,and chlorine process streams
indicating similar levels of TOX-C12 removal.
Treatment of the GAC column effluents with chloramine
followed by storage for 5 days (TOX-NH,C1) resulted in minimal
increases in TOX as indicated in Figure^ 13. Average increases
over tne TOX levels in the GAC column effluents (Figure 11) of
17, 7, 11, & 12 ug Cl/L were observed for the nondisinfected,
ozone, chlorine dioxide, and chloramine process streams with
'respective average TOX-NH Cl levels of 29, 13, 26, & 32 ug Cl/L
Essentially no increase was observed in the chlorine process
stream witn an average TOX-NH^Cl concentration of 69 ug Cl/L, all
of wnich was derived from the^TOX breakthrough on the GAC column.
Comparison of these average TOX-NH2C1 GAC effluent levels to
42
-------
o
f- o
O
O
o
in
ro
o
o
ro
O
in
CN
o
o
CM
o
in
o
o
o
m
—r
o
O
o
(N
o
o
o
in
4J
c
0)
U-l
w
u
u
rtj
CN
iH
U
X
o
CD •
,q g
-P 03
(1)
O -P
CO
O en
en IQ
•H 0)
M u
nd O
DJ ^)
g 0,
O
U
OJ
H- O X
O O
Cn
•H
PM
43
-------
o
0)
u
(M
X
o
J! g
•P
-------
those of tne sand columns (Figure 10) across the project oeriod
indicates that GAC filtration resulted in average reductions of
34, 52, 71, and 46% for the nondisinfected, ozone, chlorine
dioxide, and chloramine process streams.
Trihalomethanes (THM)—
Chloroform was the most abundant THM formed accounting for
approximately _S3% of the total followed by bromodichloromethane
with 15s ana dioromocnloromethane with the remaining 2°-
Bromoform was rarely observed, even in the free chlorine
formation potential samples.
No significant levels of THMs were observed in the
disinfectant contact cnamber and sand column effluents of the
nondisinfected, ozone, and chlorine dioxide process streams with
concentrations averaging 1 ug/L. An average THM level of 3 ug/L
occurred in the cnloramine disinfectant contact chamber and sand
column effluents while that in the chlorine contact chamber
effluent averaging 39 ug/L, increased 10 ug/L (25%) to 49 ug/L
across the chlorine sand column due to the additional contact
time of approximately 30 min across the sand column (Figure 14)
.Treatment of the sand column effluents with chloramine followed
by storage for 5 days at 30UC resulted in slightly elevated
terminal THM levels (THM-NH-Cl) averaging 8.5, 3.2, 4.2, and 9.4
ug/L for tne nondisinfected, ozone, chlorine dioxide, and
chloramine process streams (Figure 15). Similar treatment and
storage of the sand column effluents with free chlorine produced
relatively high terminal THM levels '(THM-C1-) with average
concentrations of 236 and 225 ug/L for the nondisinfected and
chlorine process streams (Figure 16). Average reductions of 35
and 41 percent relative to the nondisinfected sand column
effluent were observed for pretreatment with ozone and chlorine
dioxide resulting in average THM-C19 levels of 154 and 138 ug/L,
respectively. ^ y/ '
Tne THM concentrations in tire GAC column effluents, compared
in Figure 17, all reached saturation in 60-80 days. In each
case, saturation was preceeded by a rise and a subsequent fall in
the sand column effluent concentration (Figure 14) exemplifying a
chromatographic effect. The cnlorine, chloramine, and chlorine
dioxide GAC columns all reached saturation on day 60 with the
nondisinfected and ozone GAC columns reaching saturation on days
70 and 80, respectively. The average GAC effluent concentrations
observed after saturation was reached were 1.2, 1.0, 1.4, 4.1,
and 43 ug/L for the nondisinfected, ozone, chlorine dioxide'
chloramine and chlorine process streams. while the GAC effluent
levels after saturation equaled their respective influent or sand
column effluent levels in the chlorine process stream, those in
the other process streams exceeded their respective influent
45
-------
t!
___. e-
o
(tj
(U
>H M
O -P
W
c
o tn
w tn
•H Q)
M u
to o
Q> ^
g P<
O
U
MJ>
s
tn
•H
I— X
46
-------
_JO
O
-------
is '1
O
O
fO
O
IO
CM
0=)
oo:
CM
fe
4J
C
0)
UH
4-1
W
C
O
u
C
tO
CO
c
•rH
in
o < "3
,2° 5
o
o
O
m
o
in
•<*•
o
o
o
in
o
o
o
m
CM
o
o
CM
o o
m o
o
m
- o
o
CN
rH
U
I
g
K
4-1 rd
0)
m M
o -P
en
c
O ra
en to
•rH 0)
r-l U
(0 O
cu M
S ^
o
u
3
CD
•r-l
48
-------
e-—».
o
<
a o
-------
concentrations by 41-73 percent exhibiting a significant level of
desorption. A typical breakthrough profile' followed by
aesorption is presented in Figure 18 for the chlorine dioxide
process stream.
Treatment of tne GAG column effluents with free chlorine and
subsequent storage for 5 days at 30°C resulted in THM-C1 levels
which were significantly lower than those of similarly treated
sand column effluents (Figure 19) . Steady-state was reached on
day 150 for the chlorine GAG column with an average steady-state
THM-C12 concentration of 161 ug/L which was comprised of
approximately 27 percent (43 ug/L) THM breakthrough and 73
percent (118 ug/L) unreacted THM precursors, which coincidently,
were tne same percentages observed for TOX-C1-. Both the
nondisinfected and chlorine dioxide GAG columns reached
steady-state on day 130 with steady-state THM-C1- levels of 118
and 105 ug/L resulting in steady-state THM-C12 removals of 48 and
24 percent, respectively. The ozone GAG column did not reach
steady-state until day 180 with an average steady-state effluent
concentration of 95 ug/L which^was indicative of an average
THM-C12 removal of 37 percent. ""The THM-C12 levels observed in
the cnloramine GAG effluent were similar to those of the
nondisinfectant GAG effluent indicating a similar THM-C1,
removal. ^
Treatment of the GAG column effluents of each process stream
with chloramine and subsequent storage for 5 days produced only
slight increases in THMs relative to the GAG column e.ffluents,
which were also slight decreases in THM-NH^Cl when compared to
similarly treated sand column effluents. Tne average THM-NH^Cl
levels observed across the operational period in the
nondisinfected, ozone, chlorine dioxide, and chloramine GAG
column effluents were 5.2, 2.2, 2.2, and 5.5 ug/L, respectively
(Figure 20). A significantly higher THM-NH2C1 concentration was
observed for the chlorine GAG effluent whicTi reached 60 ug/L by
the end of the operational period. However, this was due solely
to tne GAG breakthrough of THMs formed during predisinfection
with chlorine.
Haloacetic Acids (HAA)--
The highest levels of the haloacetic acids were formed using
free chlorine with primary constituents of dichloroacetic acid
(DCAA), trichloroacetic acid (TCCA), and bromochloroacetic acid
(BCAA). Chloroacetic acid (CAA), bromoacetic acid (BAA), and
dibromoacetic acid (DBAA) were also formed to some extent, but
were very minor constituents. Halophenols were also analyzed via
the haloaceitc acid method, but were not found. Despite the
additional disinfectant contact time observed across the sand
columns, no increase was observed for ^any of the haloacetic acids
50
-------
a)
S-l
en
en
0)
o
O
M
Oi
OJ
'a
•H
x
o
•H
a
QJ
c
-H
S-l
O
OJ
•H
4-1
O
S-l
Oi
D1
p
o
S-l
0)
S-l
s;
s
EH
oo
rH
0)
S-i
3
Cn
•H
51
-------
CM
O
CM
I CM
O_J CM
rH
tn
o< ^
- m Q CD
CN
rH
U
O
l~ o
_ o
m
EC
-P n3
CD
4-1 J-)
O -P
en
C
O tn
tn tn
•H Q)
M u
(0 O
04 M
e CM
o
- o
o
o
o
in
o
o
o
m
QJ
52
-------
CM OO
O CM _J
-IXX
ozo
o
CD
o
m
T"
O
T-
o
-]—
o
—r~
o
—r
o
o
in
o
o
ro
o
in
CN
- in
o
o
o
m
o
(0
QJ
4-1
O
4J
c
w
c
o
c_>
o
o
Z M
CN -H
U_
O
cu
CO
a)
<-t
r\i
S
•P n3
0)
M-f M
O -P
CO
c
O CO
CO CO
•H 0)
H o
-------
in tne chlorine, chloramine, and chlorine dioxide process
streams.' vvhile some indication of biodegradation was observed
across the nondisinfected and ozone sand columns, significantly
nigher removals occurred across the GAG columns at steady-state
wnicn may be attributed to an adsorption/biodegradation process.
One of the more prevalent haloacetic acids observed following
cne use of a chlorinated disinfectant was dichloroacetic acid
(DCAA). Contact chamber and sand column effluent concentrations
averaged 1.7, 3.7, and 13 ug/L for the chlorine dioxide,
cnloramine, and chlorine process streams, respectively (Figure
21). Minor reductions in DCAA attributable to biodegradation
were observed across the nondisinfected and ozone sand columns
with average respective influent concentrations of 0.9 and 1.1
ug/L and a corresponding effluent concentration of 0.6 ug/L for
ootn columns (Figure 22). Treatment of the sand column effluents
with free cnlorine followed by 5-day storage indicated that the
DCAA-C1- levels entering the pilot column system averaged
approximately 65 ug/L (Figure 23) with little change occurring
across the nondisinfected and chlorine sand columns. Reductions
in DCAA-C12 of 271 (to 44 ug/L) and 26% (to 38 ug/L) were
observed for chlorine dioxide and ozone pretreatment followed by
sand filtration. Similar treatment of the sand filter effluents
with cnloramine and a 5-day storage period resulted in only
slightly elevated DCAA levels (Figure 24) with DCAA-NH-Cl
concentrations averaging 5.6, 7.3, 8.6, and 9.2 ug/L for the
ozone, nondisinfected, chlorine dioxide, and chloramine process
streams, respectively. Although these DCAA-NH2C1 levels were,
respectively, 9, 13, 5, and 3 times that of their average sand
column effluent levels, they were very similar to the levels of
THMs formed during similar treatment with chloramine suggesting
tnat both were formed by reaction with free chlorine during the
in situ formation of chloramine.
GAG filtration resulted in significant DCAA removal
efficiencies throughout the operational period (Figure 25) with
continued removals of 80% or greater after steady-state was
reached, which occurred about day 150 for all process streams.
Steady-state effluent concentrations were proportionate to their
respective influent concentrations. With steady-state influent
concentrations of 13, 3.7, and 0.6 ug/L and effluent
concentrations of 2, 0.55, and 0.07 ug/L, respectively, the
cnlorine, chloramine, and ozone GAG columns achieved 86% removal
after steady-state was reached. A steady-state removal of 80%
was observed for both the chlorine dioxide and nondisinfected GAC
columns witn respective influent concentrations of 1.7 and 0.6
ug/L and corresponding effluent concentrations averaging 0.3 and
rj. 1 ug/L. . Similar GAC breakthrough profiles were observed for
tne DCAA-C12 levels in all process streams except for that of
ozone (Figure 26). Steady-state was reached around day 250 for
all process streams at a DCAA-C12 effluent concentration of 40
54
-------
T-
in
-------
I-
in
•
ro
o
to
IO
CM
T"
O
CM
in
T-
o
T~
in
•
o
o
m
ro
o
o
ro
o
m
CM
z
013
CM
Lu
O
o
o
o
m
- o
o
•
o
0)
4J
O
0)
4-1
fi
•H
CO
•H
na
c
o
CD
4-)
S-l
o
m
en
rH
Q)
cu
i-}
u
Q
4J
C,
s
•H
,— w
c
4-) e
c 3
0) -H
P O
^1 U
m
C« T3
H C
to
o
(U
c c
o o
CO N
•H O
tO T!
a. c
g (0
o
u
fM
CN
(D
Q O « ID
tn
•H
56
-------
x-.
JO
CM CM
CM_J _J
-JO O
QT 7 CM
£KZ_)ZO
O
Q.O1CO M
MM Q1UJ
OQCM O Z
UJZO _J O
O:O_JXN
Q.ZOOO
PEl^f O
I • I
!!
! j I
3Eli*i
i i i i i i i i i i
oooooooooo
o>ooi^eoin^j-K)cN'-
o
m
ro
o
o
o
in
CN
o
(T3
u-i
O
4-)
c
CD
W
i
3
,— I
O
O
•a*
o ^ &
« rv v
o 5
c
> -H
0< co
in Q -H
o
o
o
m
- o
CM
rH
O
I
o s
a rt
0)
,M-t S-l
O 4J
02
c
o to
en en
•H a)
M o
(0 O
a M
6 cu
o
o
ro
CN
Q)
a o «
Dl
•H
57
-------
o
an
K)
o
o
ro
o
an
0<
in Q
o
o
o
m
- o
o
(N
in
m
CN
0)
Q O
58
-------
& —
o
<
o
0
o
<
°
M<
cno
i_i (£ j_j
QUJ _jo:
TZCSJOO
Z O O CM _j
ON-JXI
zo o zo
T
m
CM
—r-
o
CM
—r~
in
—r~
m
*
o
"1-
O
o
o
o
in
CM
o<
m Q
O
O
O
«n
u
13
O
C
O
c
3
i-H
O
CJ
u
<
u
o
£1
C
•H
M
o
0)
O -P
w
c
o en
en en
M o
(3 O
e cu
o
LO
CN
o
i-l
a
tn
•H
Q O
59
-------
\
o
m
-------
ug/L except for the ozone process stream which had an average
steady-state effluent concentratior, of 20 ug/L. Average GAG
DCAA-C12 removals of 63, 79, 67, 61, and 62 percent relative to
the precipitator influent were observed for DCAA over ' the one
year operational period for the nondisinfected, ozone, chlorine
dioxide, cnloramine, and chlorine process streams. Respective
removals' of 48, 73, 53, 46, and 51 percent continued. after
steady-state was reached. Treatment of the GAG column effluents
witn chloramine and subsequent storage for 5 days resulted in
average DCAA-NH Gl reductions of 53-56% (Figure 27) relative to
the similarly treated sand filter effluents (Figure 24). Average
DCAA-NH2C1 levels ranged from' 2.5-4.8 ug/L over the project
period with steady-state concentrations ranging from 4-7 uq/L
after day 250.
Trichloroacetic acid (TCAA) was an equally prominent
haloacetic acid with equivalent contact chamber and sand column
effluent concentrations averaging 0.6, 0.9, and 10 ug/L for the
cnlorine dioxide, chloramine, and chlorine process streams
(Figure 28). Unlike DCAA, almost negligible reductions of TCAA
were observed across the nondisinfected and ozone sand columns
with respective influent levels averaging 0.5 and 0.6 ug/L, and
average effluent concentrations of 0.4 and 0.5 ug/L. The levels
of TCAA-C12 in the precipitator effluent entering the pilot
column system ranged from 30-120 ug/L with an average of 74 ug/L.
Little change in TCAA-C12 levels were observed across the
nondisinfected and chlorine sand columns while average reductions
of 32% and 58% occurred for the chlorine dioxide and ozone sand
columns relative to the nondisinfected influent (Figure 29).
Treatment of the sand column effluents with chloramine followed
by storage for 5 days produced similarly elevated TCAA levels for
all process streams (Figure 30) with operational period averages
of 2 ug/L or less.
As with DCAA, excellent removals across the GAC columns were.
observed for TCAA throughout the operational period.
Steady-state occurred on or about day 150 in all process streams
(Figure 31) with subsequent removals of 85, 87, 73, 77, and 94
percent through the remainder of the operational period for the
nondisinfected, ozone, chlorine dioxide, chloramine, and chlorine
GAC columns with average effluent concentrations of 0.07, 0.06,
0.15, 0.23, and 0.68 ug/L. Similar TCAA-C1- breakthrough
profiles were observed for all process streams except that of
ozone (Figure 32). As with DCAA-C1-, steady-state was reached
around day 250 at a TCAA-C12 level of about 60 ug/L for all
process streams except that or ozone which was 20 ug/L. Average
percent removals of TCAA-C12 relative to the precipitator
influent of 57, 83, 60, 54, and 58 percent were observed across
the operational period for the nondisinfected, ozone, chlorine
dioxide, chloramine, and chlorine process streams with continued
removals averaging 43, 79, 47, 39, and 48 percent after
61
-------
o
in
ro
o
o
K)
o
O
ho
-O
o
r«
0)
-U
c
0)
w
c
o
u
u
OJ
c
•H
to
rH
0)
u
CM
ffi
s
I
U 6
Q (0
CU
M-l 5-1
O -P
CO
C
O W
w en
•H OJ
M u
n3 O
Q, H
e CM
o
CJ
*- o
Q O «
CN
CD
M
62
-------
Q
za
«n
o
<
Q (/)
Z < Ul
cn < (n z
C/) (/) M
M Ul _l K
QZCMOO
ZOO •& ©
in
CM
T-
O
CM
O
I" IO
ro
O
O
10
O
m
CM
013
-T-
m
"T"
O
CM
u_
o
,o<
; in Q
o
- o
.o
in
- o
• (C
a)
c
0)
m
w
i
o
u
c
en
-------
<"
\
?$
'>
<
i V
CM CM
CM_J _J
df?
TOO
cgzz
o«
i-en en
fr
**-**.
T"
O
00
"T"
O
CO
-r-
o
O
m
to
o
o
o
m
CM
O
O
o
m
- o
O
CN
u
(0
Q)
c
G)
g
O
u
C
(0
CO
CM o>
fe 5
•H
Cfl
rH
gJ
cu
^
CN
iH
U
X §
E-(
-------
-JO
OCM
CMX-J «J
XZO _JO
Tax CMX
HwaTa
< zaz
HZ < Z <
HI M (ft < V)
Q.CO (ft
HJ HI _I UJ
OQOCMZ
UIZCM O O
(TOX _JN
Q.ZZOO
'TTT
I I
CM
T"
O
oo
CO
T"
CN
O
m
ro
o
o
K)
o
in
CM
u.
o
0$
m
O
o
o
m
- o
o
n3
0)
-u
c
cu
W
c
B
O
U
C
nS
CO
0)
C/3
iH
0)
I
O
CN
u g
cu
O -P
CO
c
o en
CO CO
•H 0)
M O
rti O
a, M
5 P-J
o
u
o
ro
H- O «
3
tn
•H
65
-------
o
o o
-(
a -p
e en
o
CJ
n
0)
SH
a
tn
•H
i-o «
66
-------
CM
YN
0-J
MO
cn
M UJ
QZ
zo
ON
zo
CM
CM.JO
-JO <
o T o
To
O
-------
steady-state was reached. Post disinfection using chloramine
with a 5-day hold produced consistently low TCAA-NH-C1 levels
averaging from 0.3-0.8 ug/L across the project period (Figure
33}. Steady-state concentrations after day 250 averaged only
sightly higher at 0.4-0.9 ug/L.
Bromochloroacetic acid (BCAA) was also formed to a
significant extent during the disinfection process. Contact
chamber and sand column effluent concentrations were essentially
identical for the chlorine dioxide, chloramine, and chlorine
process streams with respective averages of 1, 0.8, and 6 ug/L
(Figure 34). Low level reductions attributable to biodegradation
were observed across the nondisinfected and ozone sand columns
with respective average influent concentrations of 0.21 and 0.26
ug/L and average effluent levels of 0.13 and 0.12 ug/L.
Treatment of aliquots of process stream water with free chlorine
followed by 5-day incubation indicated that the BCAA-C1? levels
in the precipitator effluent entering the pilot column system
ranged from 10-30 ug/L with an average of 17 ug/L (Figure 35) .
Essentially indentical BCAA-Clg levels were observed for the
nondisinfected, ozone, chlorine dioxide, and chlorine sand
columns with overall averages of 17, 14, 18, & 16 ug/L,
respectively. Similar treatment of sand column aliquots with
chloramine and subsequent 5-day storage again resulted in
increases which were 3-13 times that of the sand column effluents
witn average BCAA-NH-Cl levels of 1.7, 1.0, 3.1, and 2.0 ug/L for
the nondisinfected, ozone, chlorine dioxide, and chloramine sand
column effluents (Figure 36).
As with the previous haloacetic acids, relatively good
removals of BCAA were observed across the GAC columns throughout
the operational period. The average BCAA removals observed
across tne project period for the nondisinfected, ozone, chlorine
dioxide, chloramine, and chlorine GAC columns were, respectively,
74, 79, 90, 91, and 93 percent with average effluent
concentrations of 0.02, 0.01, 0.07, 0.06, and 0.34 ug/L (Figure
37). Steady-state was not readily observable for the
nondisinfected and ozone GAC columns due to their relatively low
influent concentrations which averaged 0.13 ug/L. As with the
previous haloacetic acids, steady-state was again observed on day
150 for the chlorine dioxide, chloramine, and chlorine GAC
columns. The percent removals observed for these three GAC
columns after steady-state averaged, repsectively, 87, 88, and 89
percent with average effluent concentrations of 0.10, 0.07, and
0.46 ug/L BCAA. Free chlorine treatment of the GAC effluents
produced similar BCAA-C12 breakthrough profiles across each GAC
column (Figure 38) . A relatively rapid breakthrough occurred to
a Nquasi' steady-state on or about day 50 in all process streams
at a removal of 35-45% which slowly decreased to 0% by the end of
the project period. The average BCAA-C1? removals observed
across the project period were similar for all process streams
68
-------
O_J
CM O
X CM
3?
-------
0
(0
0)
4-)
c
-------
CM
0
o:z
o
H-Z
IH M
CLCfl
MM
OO
LJ
cs
O CM_J
Z-JO
< o T
to T o
oz
UJ Z <
Z <(/>
WC/)
(£ LU
O CMZ
-JOO
X-JN
ooo
TTt
I !
i
UA
o
in
to
o
K)
in
T~
o
10
"T"
O
-T-
m
-r
o
o
in
o
o
o
w
CM
OZ)
hoo:
CN
u.
o
- m
o
o
o
in
x:
o
as
OJ
4-1
O
4J
C
CD
W
i
O
u
C
-------
O CM
CM
^O-lO
T CM OCM
OX CMX
ZZXZ
< TzT
(/) Q TO
zoz
Q. w
MM _J UJ
O OO CMZ
LUZ CM O O
QCOX ~JM
Q.ZZOO
hi
T?T
! !
I ! 1
bE3
O
in
o
o
ro
o
CM
03
CM
U.
O
- mo
o
o
o
m
- o
OO
(O
CM
u
re
0)
4J
Cl
CD
3
r-I
4-1
4-1
W
O
U
c
(0
en
Q)
X
-p
c
•H
en
rH
0)
>
(1)
u
(N
ffi
3
I
O
CQ
0)
(P
4-1 M
O 4J
01
C
O W
w o:
•H 0)
M U
ft) O
PJ M
g ^
O
U
n
0)
m o
72
-------
o
-------
CM
O-JCM_lO
O I O 1 O
o T o
Z< O4H
o
+J
c
0)
3
4-1
w
c
o
u
u
o
(U
-P
c
•H
to
CN
rH
U
I
u
CD •
^ e
4-) (0
0)
m s-i
O 4J
cn
c
o w
w LO
•H QJ
!H u
m o
o
u
CM
to
CM
o
CM
in
m o
0
S-l
3
tn
•r-i
CM
74
-------
ranging rrom 2J-25% with average precursor effluent
concentrations ranging from 9-12 ug/L. Chloramine treatment of
tne GAG column effluents and subsequent storage for 5 days
resulted in similar average BCAA-NH2C1 effluent levels of 0.6,
0.8, 1.7, 1.6, and 1.2 ug/L across the project period for the
nondisinfected, ozone, chlorine dioxide, chloramine, and chlorine
process streams (Figure 39). Comparison to similarly treated
sand filtered effluents indicates that GAG filtration produced
BCAA-NH Cl removals of 64, 24, 45, and 20 percent in the
nonaisinfected, ozone, cnlorine dioxide, and chloramine oroc-ss
streams. *• '~~oa
Chloroacetic acid (CAA) was a minor constituent with an
observed maximum average of approximately 0.5 ug/L in the process
stream effluents and 1.7 ug/L in the free chlorine terminal
CAA-C12 determinations. The CAA data obtained was somewhat
irratic oecause tne CAA cnromatographic peak was usually a small
peak which was not completely separable from that of a rather
.Larger impurity requiring a somewhat subjective integration The
contact chamber and sand column effluent concentrations were
essentially equivalent for the chlorine dioxide, chloramine, and
chlorine process streams with respective averages of 0.12, 0 20
ana U.59 ug/L (Figure 40). m the effluents of "the
nondisinfected and ozone process streams, CAA was detected in
less tnan 50 percent of tne samples analyzed at levels averaging
.0.07 and 0.09 ug/L, respectively. Terminal CAA-C10 levels were
similar in the sand column effluents of all process streams
ranging from 2-6 ug/L (Figure 41) and averaging '3-4 ug/L.
Similar CAA-NH^Cl levels were also observed for the sand column
aliquots treated with chloramine and stored for 5 days with
averages ranging from 0.9-1.2 ug/L (Figure 42). As with several
of the naloacetic acids previously discussed, these levels were
5-15 times those of the sand column effluents suggesting that
some reaction between free chlorine and the CAA precursors had
occurred during the in situ formation of chloramine.
The only GAG column effluent in which CAA was consistently
detected was that of the chlorirre process stream (Figure 43). By
the end of the project period (days 300-365) effluent CAA levels
were averaging 0.07 ug/L resulting in CAA removals averaging 92%
witn respect to the sand column effluent. While influent
concentration spikes caused temporary breakthroughs in the
process streams (day 127) , no consistant CAA breakthrough was
observed in the GAG effluents of the other process streams. in
contrast, significant CAA precursor breakthrough occurred for all
GAG columns as indicated by the CAA-C19 levels depicted in Figure
44. Tne average CAA-C1 levels in "the GAG effluents of the
nondisinfected, ozone, chrorine dioxide, chloramine, and chlorine
process streams were essentially identically at 1.5, 1.1, 1.3,
1.4 and 1.6 ug/L. The GAG removal of CAA precursor diminished
across tne operational period with continued removals of 31, 53,
75
-------
o
_J 0
1 1 1
ai&A
^^
CD in TJ- ro
CM
I
o
o
in
ro
o
o
ro
O
in
CM
CM
u.
O
O
(0
0)
^j
c
0)
a
W
C
o
u
u
o
01
4-)
c
•H
to
O< a)
IOQ £
O
ho
Lo
in
- o
O
(N
u
CQ
0)
0)
O 4-J
C/3
C
O tO
10 tO
•H a)
M O
(0 O
S ft
o
u
(Tl
m o < < ID
0)
3
Cn
•H
76
-------
-------
MMZ
Q.WM
MMK
OQO
010X
o
in
o
o
o
in
CM
§i
0)
CM
r-H
u
I
u
a;
(U
m !-(
O -P
cn
C
O W
tn en
•H cu
u o
(0 O
& H
E P4
O
oo
CO
CM
O «
Cn
•H
-------
Q.CO
OOOCMZ
UJ Z C4 O O
CCOX -JN
Q.ZZ OO
m
!!
I ! I
«n
o
ro
in
CM
o
T"
O
-------
o
<
o o
O M
M UJ _! OL
Q Z
-------
CM CM
d H
T CM CM T
O-JCM_JO
< O _l O <
oTp To
oio
O n
IHUJ _jo:
QZCMOO
ZO O CM_j
ONI-JXX
zo ozo
o
K>
r~
in
CM
—r~
o
B--'--^
-------
27, and 33 percent on days 320-360 for the nondisinfected, ozone,
chlorine dioxide, and chlorine process streams as compared to
similarly treated sand .column effluents. Since the cnloramine
GAG column exhibited similar CAA-C12 effluent concentrations,
similar CAA precursor removal was assumed. 'Chloramine treatment
of the GAG effluents and subsequent 5-day storage resulted in
similar CAA-MH2C1 concentrations for all process streams with
project period averages ranging from 0.3-0.5 ug/L (Figure 45).
Comparison to similarly treated sand column effluents indicated
that average removals of 60-70% were produced by GAG filtration
over the one year operational period with removals diminishing to
27-53% by tne end of tne year (days 320-360).
Dibromoacetic acid (DBAA) was another minor constituent with
a maximum average process stream concentration of 0.3 ug/L and an
maximum average terminal DBAA-Cl- concentration of 1.1 ug/L.
UnliKe CAA, reasonably good precision was obtained for DBAA in
the 0.02-0.1 ug/L range as evidenced by the comparison of the
chlorine dioxide influent and sand effluent in Figure 46. As
exemplified for the chlorine dioxj.de process stream, the contact
chamber and sand column effluent concentrations were essentially
identical for all process streams except that of ozone.
Ozonation generated additional DBAA apparently resulting from the
presence of trace amounts of bromide in the raw water, increasing
the average nondisinfected concentration from 0.007 to 0.088 ug/L
(Figure 47). After passage through the ozone sand column, the
average DBAA concentration was reduced to 0.037 ug/L, presumably
due to biodegrada t i on. The average DBAA levels in'-the sand
column effluents of the nondisinfected, ozone, chlorine dioxide,
chloramin.e, and chlorine process streams were, respectively,
0.004, 0.0*37, 0.096, 0.018,"and 0.33 ug/L (Figure 48). The level
of terminal DBAA-C1- entering the pilot column system averaged
0.6 ug/L. Increases of approximately 65% were observed for both
chlorine dioxide and ozone resulting in DBAA-C12 levels averaging
1.0 ug/L (Figure 49). The addition of chloramine and storage for
5 days produced moderate increases in DBAA levels of
approximately 2-3 times that of the sand column effluent for the
ozone, chlorine dioxide, and ctiloramine process streams with
respective averages of 0.10, 0.22, and 0.06 ug/L (Figure 53).
Similar treatment of nondisinfected sand filtered water resulted
in an average DBAA-NH-C1 concentration of 0.08,ug/L which was 20
times that of the initial average nondisinfected sand column
concentration of 0.004 ug/L.
Significant levels of DBAA breakthrough occurred across the
GAC columns of the chlorine and chlorine dioxide process streams
wnile relatively minor intermittent breakthrough was observed for
tne nondisinfected, ozone, and chloramine GAC columns at the
0.S02 ug/L level (Figure 51). The average DBAA removals observed
for the cnlorine and cnlorine dioxide GAC columns were both 94%
relative to their respective sand column influents. Both of
82
-------
QZCM
ZOO
ON_J
ZOO
00
•
o
CO
•
O
I
-------
o
in
o
o
n
o
in
CM
cs
o
o
o
o
in
- o
Q QQ < <
OJ
84
-------
U __ I Q
z u. z
MZ<
M UJ UJ
O Z Z
zoo
O N N
ZOO
CM
O
o
CN
CD
I
lO
I
o
o
If)
ro
o
o
o
IT)
(N
CM
U.
O
0<
m Q
o
o
o
in
- o
in
o
o
o
4-)
O
(0
4-1
c
o
u
CD
C
O
N
O
Q)
4J
U
0)
M-J
C
•H
C
O
OJ •
^ tn
4-) 4J
c
C OJ
•H 3
M
en IH
i-i IH
0) w
>
0) C
^ e
3
rf r— I
-------
©•——""-""""—™
O
ro
0
n
O
m
CM
o
CM
o
h in
o
l~ o
50
0
o>
o
00
CD
i
CO
I
in
o
ro
•
o
I
CM
O
*
O
o
(0
0)
o
c
iH
M t
n i
pa
i
o
o
to
Q)
Q) •
x; e
a)
o -P
w
c
O en
w en
•H Q)
n o
-------
dY¥
T Q a
CKZ Z
o «
H- en en
I-ZUJ
M M Z
OQO
UZ_J
CEO X
CLZO
CM
CM_J
_JO
Q I
T Q
QZ
Z <
M U
ft O
Oi S-l
£ fij
o
0)
tn
•H
Q CD < <
87
-------
—I U
O
-------
Q m < <
89
-------
tnese GAG columns reached apparent steady state about day 200
with respective average steady-state removals of 92 and 93% and
average effluent concentrations of 0.017 and 0.005 ug/L. The
average GAG effluent concentrations observed across the
operational period for the nondisinfected, ozone, and chloramine
process streams were 0.001, 0.001, and 0.002 ug/L. Similar GAG
breakthrough profiles were observed for DBAA-C12 in all process
strea.ns with a very rapid breakthrough to an apparent steady
state prior to day 50 (Figure 52) . DBAA-Cl, steady-estate
effluent concentrations averaging 0.97, 1.25, I.IT, 1.03 and 0.96
ug/L were observed for the nondisinfected, ozone, chlorine
dioxide, chloramine, and chlorine GAG columns, respectively.
These DBAA-C1- levels were 59, 30, 19, and 55 percent greater
tnan those or similarly treated sand filter effluents in the
nondisinfected, ozone, cnlorine dioxide, and chlorine process
streams. No explanation for this increase was apparent.
Chloramine treatment of GAG filtered water- followed by 5-day
storage resulted in average DBAA-NH-Cl concentrations of 0.02,
0.11, 0.17, 3.12, and 0.08 ug/L for the nondisinf ected; ozone,
chlorine dioxide, chloramine, and chlorine process streams,
respectively (Figure 53). The cause of the relatively 'irratic
data obtained for the chloramine 'treated samples could not be
determined and may have resulted from slight variations In free
chlorine contact.
Bromoacetic acid was also a minor component of the total
haloacetic acid concentration ranging from nondetectable in many
samples to a maximum average . concentration of 0.25 ug/L in the
cnlorine sand column effluent and 1.5 ug/L in the terminal free
chlorine formational potential determinations (BAA-C12) . Like
chloroacetic acid, the. low-level BAA data obtained was consider-
ably irratic due to interference from other chromatographic
peaks, between which BAA eluted, with quantitation dependent on
somewhat subjective intergration. In many instances, BAA could
not be quantitated due to variations in chromatographic condi-
tions. BAA was detected at a frequency of only 60% in the
influent of the pilot column • system at an approximate iconcen-
tration of 0.01 ug/L. The frequency of detection and approximate
average concentration in the contact chamber and sand column
effluents were very similar for all process streams (Figure 54)
except that of ozone, with 95% and 0.10 ug/L for the chlorine
dioxide pro.eess stream, 55% and 0.02 ug/L for the chloramine
process stream, and 85% and 0.25 ug/L for the chlorine process
stream. BAA was detected in the ozone contact chamber effluent
at a frequency of 96% and an average concentration of 0.;22 ug/L
(Figure 55) , while that for the ozone sand column was 88% and
0.07 ug/L suggesting that some biodegradation had occurred. BAA
was detected in less than 50% of the GAG effluent samples with
breakthrough to an apparent steady-state occurring on
approximately day 120 for all locations (Figure 56) . The BAA
levels after steady-state were below 0.02 ug/L on all GAG
90
-------
OS
—I
O
i CM
0_J
g?
O
2<
WO
CO
M LU
02
20
ON
2O
CM
CM_lO
-J0<
o T o
To
o
-------
o
_J CM
CMO_IOZ
ZCMOCM T
OTz T o
-------
o
(0
0)
c,
0)
CQ < <
o
u
C
(0
CO
cu
x;
C
•H
W
tH
.
0)
5 a
-P tO
0
o -P
CO
c
O CD
CQ CO
•H Q)
!-) O
nS O
S f^4
O
u
in
(1)
M
S
CD
•H
93
-------
ItlUI
5§
NN
OO
o
to
o
o
ro
O
JO
CM
CM
C
(0
CO
13
G
£3
U
-P
U
tfl
-P
c
o
u
CD
O
ISJ
O
Lu
O
- in Q
o
f-o
.o
m
-o
in
•
o
I
CM
c
•H
r-l
a)
o
•
o
to
0) -P
4J (U
-------
OQ < <
95
-------
effluents except that of the chlorine process stream which
averaged approximately 0.05 ug/L. A greater number of inter-
ferences with SAA quantitation occurred in all formation poten-
tial samples sucn that only 20-60% of the data was quantifiable
at all locations. From the data obtained, it would appear that
all process streams exhibited similar BAA-C12 and BAA-NH2C1
levels. For 3AA-C17, quantifiable data ranged from 0.5-3 ug/L in
the sand effluents while that for the GAG effluents ranged from
0.4-1.5 ug/L. Comparable quantifiable data for BAA-is!H2Cl ranged
from it).02-0.5 ug/L both before and after GAG filtration.
Chloral Hydrate (CH)—
Chloral hydrate (CH) was formed predominantly in the chlorine
process stream with an average contact chamber effluent
concentration of 2.9 ug/L which increased 55% to 4.5 ug/L across
tne sand column due to an additional .30 min of chlorine contact
time (Figure 57) . The CH levels in the effluent of the chlor-
amine contact chamber and sand column were identical averaging
0.25 ug/L. CH 'was detected intermittently in the contactor
chamber and sand column effluents of the chlorine dioxide, ozone,
and nondisinfected process streams at respective frequencies of
56, 32, and 26 percent with similar average concentrations
ranging from 0.01-0.07 ug/L. Treatment of the sand column
effluents with free chlorine and storage for 5 days produced
CH-C1- levels averaging 79, 55, 45, and 75 ug/L (Figure 58) for
the nondisinfected, ozone, chlorine dioxide, and chlorine process
streams, respectively. Treated precipitator water had essential-
ly an identical CH-C1? content as the nondisinfected sand column
effluent. Similar treatment with chloramine resulted in average
CH-NH-Cl levels of 0.03 ug/L for the nondisinfected and ozonated
sand column effluents, and, 0.08 & 0.3 ug/L for the chloramine
and cnlorine dioxide sand column effluents (Figure 59).
Granular activated carbon filtration resulted in 100 percent
removal of CH throughout the one year project period in all
process streams. Essentially no chloral hydrate was detected in
any GAC effluent greater than 0.001 ug/L. Similar breakthrough
profiles for terminal CH-C1- were observed across the GAC columns
in each process stream (Figure 60). Steady-state was reached
about day 150 for CH-C19' in all process streams. Steady-state
removals relative to the sand column effluents ranged from 42%
and 50% for the chlorine-dioxide and chlorine GAC columns, to 60%
for the nondisinfected and ozone GAC columns with respective
average steady state CH-C1? concentrations of 35, 41, 36, and 28
ug/L. Treatment of the CTAC column effluents with chloramine
followed by 5-day storage produced essentially the -same CH-NH2C1
levels in each process st,ream ranging from 0.1-0.3 ug/L (Figure
61) . These levels were essentially the same as those obtained
for the similarly treated sand column effluents.
96
-------
©—
o
a <
zo a co
|_i
zoozo
ill
00
T"
to
T~
CM
00
tO
T-
CM
O
IO
ro
O
O
IO
O
IO
CM
0
o
—*
_ O
in
-o
o
m
c
a)
3
w
oo: 'g
CM •§
Lu tn
O 0)
to
tH
0)
Q)
g
cu
CO
O to
to to
•H Q)
H 0
(0 O
g, s-i
S P-I
o
u
IT)
0)
O
tn
•H
97
-------
W (N
CM
Oil CM
T QQCM_J
C£ZZ_JO
0«0 1
h-tn
-------
(NX
XZ
? '
TQ
aiz:
o<
I-OT
<
1-Z
MM
Q.U)
MM
00
LU2:
(TO
O-Z
O_IO
CMO C«l
XCMX
Q I Q
ZQZ
-------
CM
I CM C. .
O-JCM-JO
o To To
oT o
z < o < uj
tn o M
MLU ,-JCC
QZCMOO
ZOO04J
ON_JXX
zoozo
TTT
1i
o
CO
o
m
o
•<*•
o
ro
o
CM
I
o
o
in
o
o.
to
o
in
CM
CM
Lu
O
o
o
o
m
-o
,
u
cO
CD
•4-1
c
0)
O
u
o
0)
,G
.4-)
G
•H
03
iH
-------
O_J
- Q o: < i- uj
-T-
CM
o
—T-
o
o
o
in
ro
o
o
o
in
CM
CM
u.
O
o<
^>Q
o
o
o
m
- o
xi
o
a
CD
4-1
c
CD
w
c
I
o
CJ
u
o
10
r-i
0)
CD
u
CM
w
I
as
u
o
0
O -P
CO
c
O CO
CO CO
-H 0)
M O
(0 O
Oi M
S At
O
vo
Q)
3
tn
•H
CM
101
-------
Haloacetonitriles--
Relatively low levels of the haloacetonitrlies were formed
across each process stream with the chlorine process stream
producing the highest levels averaging 3.1 ug/L total
haloacetonitriles. Levels of haloacetonitriles averaging less
than 1 ug/L were observed across" the other process streams. The
predominant haloacetonitrile (HAN) was dichloroacetonitrile
(DCAN) followed by bromochloroacetonitrile (BCAN),
dibromoacetonitrile (D3AN), and trichloroacetonitrile (TCAN). No
consistent breakthrougn of the HANs was observed through any GAG
column except that of the chlorine process stream, whach was
still removing over 95% of the influent HANs by the end of the
one year operational period.
Dichloroacetonitrile (DCAN) was primarily formed in the
chlorine process stream with an average influent concentration of
1.5 ug/L which increased 30% to 1.9 ug/L across the chlorine sand
column. The contact chamber and sand column effluents of the
chloramine, chlorine dioxide, and nondisinfected process streams
were essentially identical with respective sand column effluent
concentrations averaging 0.2, 0.1, and 0.05 ug/L (Figure (52).
Some evidence of biodegradation across the ozone sand column was
observed with respective average influent and effluent
concentrations of 0.07 & 0.02 ug/L. Treatment of the chlorine
sand column effluent with additional free chlorine and subsequent
5-day storage produced an average DCAN-C1- concentration of 1.9
ug/L (Figure 63) which was the same as that of DCAN in the sand
column effluent indicating that all DCAN precursors had reacted
across tne chlorine contact chamber and sand column., 'Similar
treatment of the nondisinfected, ozone, and chlorine dioxide sand
column effluents produced respective DCAN-Cl- levels averaging
1.3, 1.6, & 4.1 ug/L suggesting that significantly higher levels
of DCAN precursors were formed by pretreatment with chlorine
dioxide. Similar treatment of the sand column effluents with
chloramine followed ,by 5-day storage produced respective average
DCAN-NH^Cl levels of 0.03, 0.02, 0.21, and 0.04 ug/L for the
nondisinrected, ozone, chlorine dioxide, and chloramine process
streams (Figure 64). Comparison of these DCAN-NH2C1 levels to
corresponding DCAN levels present in the sand column effluent
indicated that the addition of chloramine had essentially no
effect on the DCAN levels of the nondisinf ected and ozone sand
column effluents while it increased the_DCAN in the chlorine
dioxide sand column effluent by a factor of three to an average
of 0.21 ug/L, again suggesting the formation of DCAN precursors
by cnlorine dioxide pretreatment. The further addition of
cnloramine to the chloramine sand column effluent resulted in an
average DCAN reduction of approximately 80 percent after 5 days
of storage. This reduction was similar to that observed for TOX
after 5-day chloramine treatment. '
102
-------
-IT"0
e=n
et——
o <
ZQ QW
-------
^^ ^ "
* **^^
CM CM
C>|_J _J
-JO O
ol T CM
T Q04Q_1
QCZ-JZO
o
0<
o
o
o
to
o
I
CO
I
If)
CM
£',
O
n)
a)
o
4->
c
OJ
p
^-^
m
tw
w
I
rH
O
O
QO
in
.
'0)
a
I
o
a
Q)
'•H H
O -P
U3
c
O to
w to
-H Q)
M O
(0 O
ft H
g to
O
u
10
Q)
cn
•H
104
-------
OCM
z T CM OCM
Taxtxx
H- waTa
< ZQZ
OQ OCMZ
UIZCMOO
Ct OX_lN
CLZZOO
ft?
i ! 1
0
o
ro
o
m
CM
-o
o$
m Q
o
o
o
in
U
m
Q)
G
S
rH
M-t
m
w
c
g
3
rH
O
U
Q)
G
-H
w
'•H
Q)
U
CM
ffi
U
P
0)
Q)
IH M
O -P
w
G
O w
w 01
•H Q)
M O
(0 O
O
U
CO
•
o
ro
•
O
CM
•
O
o
•
o
QO <
105
-------
No consistent breakthrough of DCAN was observed across the
GAG column of any process stream other than chlorine (Figure 65).
breakthrough on the cnlorine GAG column occurred on day 70 with
an increasing effluent concentration until steady-state was
reached after 140 days of operation. The average effluent
concentration after reaching steady-state was 0.05 ug/L for an
average steady-state removal of 98 percent., Treatment of the GAG
column effluents with free chlorine and storage for 5 days
resulted in average DCAN-C12 of 2.4, 1.1, 2.9, 2.3, and 2.0 ug/L
for the nondisinfected, ozone, chlorine dioxide, chloramine, and
chlorine process streams, respectively (Figure 66). Comparison
of these data with those of the similarly treated sand column
effluents indicated that in the ozone and chlorine ;dioxide
process streams, saturation was reached on approximately days 140
& 150 with subsequent effluent levels essentially equal to their
respective influent levels. In contrast, after a relatively
rapid precursor breakthrough to saturation on day 80, DCAN-CLp
levels in the nondisinf ected GAG effluent were cons i:s tantly
greater by an average of approximately 80% (Figure 67). A
similar anomaly occurred in the chlorine process stream as
indicated in Figure 68. The DCAN and terminal DCAN-C1- levels in
the sand column effluent were essentially identical indicating
that all DCAN precursors present had reacted to form DCAN; Since
the DCAN levels in the GAG effluent were essentially zero, the
DCAN-C12 levels in the GAG effluent should also have been
nonexistant. However, concentrations from 0.5 - 5 ug/L DCAN-C12
were observed in the chlorine GAG effluent. Saturation was
reacned on day 180 with the DCAN-C12 levels in the chlorine GAG
effluent averaging 42% higher than those in the sand' column
effluent, similar to that observed for the nondisinfected GAG
effluent. These increases in .precursor concentration were also
observed for bromochloroacetonitrile and dibromoacetonitrile, as
indicated in subsequent sections, suggesting that, these
precursors are being formed on or released from the GAG columns.
Treatment of the GAG column effluents with chloramine
followed by 5-day storage resulted in average DCAN-NH2C1 levels
of 0.02, 0.03, 0.2, 0.03, and 0.02 ug/L for the nondisinf ected,
ozone, chlorine dioxide, chloramine, and chlorine process
streams* respectively (Figure 69). No significant difference
between the DCAN-NH.,C1 levels of the sand -and GAG. column
effluents was observed in any process stream, with the DCAN-NH Cl
levels in the chlorine GAG column effluent being essentially
identical to those of the chloramine GAG column.
Bromochloroacetonitrile (BCAN) was the second most abundant
haloacetonitrile formed primarily in the chlorine process stream
with an ave'rage contact chamber effluent concentration ;of 0.70
ug/L which increased 25% to 0.88 ug/L across the chlorine sand
column. A similar increase of 35% was observed across the
cnloramine sand column raising the average concentration from
106
-------
o
o
<
o
ZO O
M< O < UJ
o
MOM
?UJ _J OL
ZCM QO
Z O O CM _J
O N _J X X
ZO O ZO
o
1
o
•
o
Q O < -Z.
—r~
o
o
—r~
o
d
—T—
o
•
o
~T~
o
•
o
o
in
fO
o
o
ro
o
m
(N
F—
o
(0
cu
-p
G
-------
CM
O
CM CM I
_J
-------
,
A [71
s^^ • •>"*"""^^
xx , ~--'~~~'~
X ••••"""""
Mf CTV ^ ^
*^*"x. *•"**» ^
•«• *«. 4
-^ **..„ ^
^^& . '""EL *
-
»— — .» 13
^^ ^* ••». ^
^^ ^^ ^^ ^
*^* ^» ^^ v
^^ ^P* LJJ
^ . ^
CM ^-^ t
-I CM ^,' \_
O _j "**-,^ EJ
V ^-^ \
2 2 O O "*"*^^^ \ ^
CO CO O O ^^— -""^P !
to to to co ., --•"'"" /
l-l MM M •! --^ -*,
o a o o "^^-^ 9
2222 ^--^ »
O O O O ^^ »
2222 **>. \
i Ty \ ; ^
• i \ »
! i ^ &
1 1 / *
' 1 f \
\ \
v\ \ ^
,JE1 ^
c
•H
05
'H
Q)
>
Q)
iJ
CM
rH
U
1
Q
13
C
(0
U
o
IH
O
c
CQ S
•H to
M 0
(d s-i
G.-P
S co
O
U
•
VD
o o < z:
0)
S-i
3
tn
•H
109
-------
QQ
I
I
4k
V
>- u
0<
• m Q
CM
U
I
o
Q
•n
U
Q
C
o
to
-H
M
o
u
CO
CD
M
0
Cn
•H
110
-------
d_
CNIO
o
CM
_ O Z
XCMOCMT
ZXCMXO
T ZXZ<
oTz T o
-------
0.031 to 0.042 uy/L. The sand column effluent concentrations of
BCAN compared in Figure 70, averaged 0.006, 0.002, 0.019, 0.042,
and 'J.88 ug/L, respectively, for the nond is infected, .ozone,
cnlorine dioxide, chloramine, and chlorine process streams.
Evidence of BCAN biodegradation was observed across the
nondisinfected and ozone sand columns. An average removal of 80%
was observed across the ozone sand column v/ith average influent
and effluent concentrations of 0.014 and 0.002 ug/L. A lower
removal of 20% occurred across the nondisinfected sand column
with average influent and effluent concentrations of 0.012 and
3.306 ug/L. No change in BCAN concentration was observed across
the chlorine dioxide sand column. Treatment of the sand column
effluents with free chlorine followed by storage for ,5 days
produced respective average BCAN-C12 levels of 0.44, 0.66, 1.39,
and kJ.61 ug/L for the nondisinfected, ozone, chlorine dioxide,
and chlorine process streams (Figure 71), again indicating
elevated levels in the chlorine dioxide process stream.
Treatment of the sand column effluents with chloramine followed
oy 5-day storage also produced slightly elevated levels in the
cnlorine dioxide process stream with average BCAN-NH-Cl levels of
U.J06, £.007, 0.05, and 0-.00S ug/L for the nondisinfected, ozone,
cnlorine dioxide, and chloramine process streams (Figure 72) . In
comparison to the BCAN levels in the sand column effluents,
slight increases were observed for BCAN-NH2C1 in the chlorine
dioxide and ozone process streams with no change :in ' the
nondisinfected process stream. The chloramine process stream
again exnioited lower (80%) BCAN^NH,01 levels than the BCAN-NH2C1
in the sand filter effluent after Tzhe additional chloramination
and 5-day storage.
Filtration througn GAC with a 20 minute EBCT removed
essentially 100% BCAN in every process stream except that of
chlorine (Figure 73). Breakthrough above 0.001 ug/L was observed
on day 75 with steady-state occurring on day 150 at the 0.015
ug/L level. Relative to' the sand filtered effluent, an iaverage
removal, of 98% was observed after steady-state was reached.
Treatment of the GAC filter effluent with free chlorine and
suosequent storage for 5 days resulted in respective average
BCAN-C1- levels of 1.1, 0.8, 1.2, 1.1, .and 0.9 ug/L for the
nondisiriiected, ozone, chlorine dioxide, chloramine, and chlorine
process streams. These levels were equivalent or slightly
greater tnan those BCAN-C1- levels in the similarly treated sand
column effluents indicating that precursor breakthrough to
saturation occurred fairly rapi-dly, within the first 50 days
(Figure 74). As indicated previously for DCAN, increased
precursor levels were observed across the nondisinfected and
chlorine GAC columns. The average- BCAN-C12 concentration after
saturation in the nondisinfected GAC column effluent at 1.3 ug/L
was more than 3 times that of tne sand column effluent; at 0.4
ug/L (Figure 75). Witn equivalent chlorine sand column effluent
BCAN and BCAN-C1,, levels indicating that essentially no
112
-------
o
ZO 0
z
(O £/) M
WUJ _J £C
QZCMOO
ONIJXX
zoozo
!M
!!
i ! 1
in
CM
°T-
o
CM
in
"T"
m
o
o
in
ro
o
o
to
o
in
CM
CM
b,
O
o<
o
o
o
in
o
CD
O
m
0)
4J
a
0)
3
o
u
tJ
c
c
•H
CO
0)
O g
PQ <0
0)
m M
O -P
en
a
o to
to w
•H Q)
M o
(fl O
o
o
o
Q)
mo <
tn
•H
CH
113
-------
CM
CM CM
CM
0< f €>
i I I
i I
o
(N
T~
03
n 1 1 1——r-
CD ^ CN O 00
• • • • •
T- T- T- T- O
mo
a)
P
O
o
in
- o
CM
rH
U
I
o e
(Q (C
-------
OCM
CMX_J _J
XZO_|Q
TCMOCM
I
o- -H
O <
IOQ ^3
Q)
O
CN
ffi
3
§ -
O g
a to
a)
4-1 J-|
o -P
w
c
O CD
W M
•H 0)
M O
e
o
CN
I~.
Q)
QQ O <
tn
•H
115
-------
€)
Si^*
3
So o"
O < O < UJ
M O M
Mill .JQZ
Q 2 CM O O
ZOO
-------
CM
CM_lO
-------
4
Ik 0 <
^ / <
<
^^ / 4
9 S ^" '' \
To ^ ,' J
Q Q 1 -,-" ' ^
z z o o -ft EL
(/) U) CD CD \ %X
\
to to c/) c/> fc. in
WI_J • — t • _* J^^^ — , tvJ
I™! (M| ^n| ^i*^ ^^»
Q Q Q Q "^^.^ \x
oood ^"*^"">B xkl
.f^ \
f?t^ s' \
i {SL u i
1^*^ 1
^ •
1 X «
1 \ /
^jmiA ^^ "*"* i
—••*"*"*" i
#*""**" gh
N \
X \
N. I
X l
X l
X 1
^& p]
\
\ /
\ a
^ \
\ \
\ *
"Bs. __p1 4
«-->'
i
r
>
J
f
T
D
O
-m
ro
o
- o
r^J
o
-in
CM
™"
OP
- O Q^
CM
U_
Q
^_
0<
- in Q
y™r
0
-o
t~
_ O
m ;
- o
•a
a)
u
Q)
c
•H
0)
•rH
C
O
is
0)
c
•H
w
a;
Q)
O»3
rH
U
1
u
«
T3
C
§
U g
PQ (0
Q)
0 -P
CO
O CQ
CO CO
•H CD
M O
to o
DJ M
S 0-*
O
U
.
CM
Q)
m o <
ZD e>\
tTI
•r)
118
-------
precursors were entering the GAG column, and with all SCAN in the
chlorine sand column effluent being removed by GAG filtration, it
would be expected that the BCAN-C1., levels in the chlorine GAG
effluent would be essentially nonexistent (Figure 76). However,
the actual BCAN-C12 levels in the chlorine GAG effluent were
equal to or greater than those of the sand column effluent
implying that the GAG column was the source of the precursors,
possibly through some form of microbiological transformation.
The_ addition of chloramines to the GAG filter effluent followed"
oy s-day storage produced respective average BCA.N-NH.~C1 levels of
0.005, 0.008, 0.11, 0.009, and 0.006 ug/L for the nondisinfected,
ozone, chlorine dioxide, chloramine, and chlorine process streams
whicn were essentially equivalent to those of the similarly
treated sand column effluents (Figure 77).
Contact cnamber effluent concentrations of -dibromoaceto-
nitrile (DBAN) were equivalent for the chlorine and ozone orocess
streams averaging 0.26 ug/L (Figure 78). • Lower levels averaging
0.007, 0.027, and 0.03 ug/L were observed for the nondisinfected
influent and the chlorine dioxide and chloramine contact chamber
effluents. While no change in DBAiSl was observed across the
chlorine sand column (Figure 79) , an average reduction of 90%
occurred across the ozone sand column resulting in an average
effluent concentration of 0.027 ug/L (Figure 80). The average
sand column effluent concentrations for the nondisinfected,
chlorine dioxide, and chloramine process streams at 0.006, 0.033,
and 0.07 ug/L were essentially equivalent to their respective
contact chamber effluent concentrations. Chlorination of the
sand column effluents and storage for 5 days resulted in similar
average DBAN-C12 levels of 0.23, 0.39, 0.45, and 0.26 .ug/L for
the nondisinfected, ozone, chlorine dioxide, and chlorine process
streams, respectively (Figure 81). Similar treatment of the sand
column effluents with chloramine resulted in average DBAN-NH_C1
levels of 0.005, 0.011, 0.026, and 0.007 ug/L for Ihe
nondisinfected, ozone, chlorine dioxide, and chloramine process
streams (Figure 82) . Other than in a few instances during the
operational period, these DBAN-NH2C1 levels were essentially the
same as the DBAN levels in the respective sand column effluents,
except for the chloramine process • stream which exhibited a
reduction of 87%, similar, to that observed for TOX (Figure 83) .
GAG filtration reduced the average DBAN levels in all process
streams to .003-.005 ug/L with intermittant breakthrough observed
at or below the .05 ug/L level (Figure 84). The breakthrough of
DBAN across the cnlorine GAG column was more or less continuous
after day 40 with an average steady state concentration of 0.006
ug/L resulting in a 98% removal. Chlorination of the GAG
effluent followed by 5-day storage resulted in essentially the
same DBAN-C12 levels in all process streams with respective
average concentrations of 0.45, 0.53, 0.50, 0.45, and 0.41 for
tne nondisinfected, ozone, chlorine dioxide, chloramine, and
119
-------
c
•H
CO
^H
CD
>
•CU
h^
tN
rH
U
I
«
T3
(0
'§
U
CQ
m
o
G
O
CO
•H
M
CO
O
U
0)
QQ O <
•H
Fn
120
-------
\ /^/^
M
^ *""""*"*"" ^
^fr-**" \
^^* 1
^^~^^ \
*"""-- — 4
*"•—-*—• 1
o "^ ^
-J CM \ «
O _J _l X \
CM 0 _l 0 Z \
XCM OCM T <& ^
ZXCMXO r '
Tzxz< i /;
o T z T o / /A
< o T o A .Atfi
O < O < UJ ^*N. »f¥
0< 0 Z "\ VA/
V) O M N^jj
QZCMOO ^aS
ZOOCM_I . *" \V
ON-JXX .-; —~ V
zoozo —- — I
•jJQ^^Y "^*-*»^.^ j
Jij • "^""H
til •' i
j j / \
! \
— 4 \
T i i i r i 1
anoinomomc
** rO CNJ CN| T— r— Q ^
• • ... , . .
O ' O . O O O OO c
3
i
i
0
a
3
3
r
1
&•
rf
3
L
r
D
D
•
-i
O
-in
ro
O
-0
ro
O
-in
(N
"Z.
** f~) (y
csl
i
O
^^"
o<
-mo
r—
0
-o
o
in
o
U
(fl
Q)
4-)
C
4-1
o
u
o
w
c:
o to
W Kl
•H Q)
M O
nj o
Q, M
e dt
o
u
Q)
m o <
Cn
•H
121
-------
T
00
tO
•
o
to
•
o
oca •
ro
•
o
CM «-
• •
O O
i
O
O
m
ro
O
O
O
IT)
CM
CN
U.
O
O
O
O
to
,
o
to
CD
4J
C
cu
'•M
CU
m
u
O
C
O
u
CD
C
•H
CO
0)
| ^
O
-------
o
(0
0)
c
s
•H
M-l
o
u
C
rt
CO
0)
J3
4J
a
•H
in
« 6
Q m
0)
o -P
U>
c
o w
to en
•H Q)
M O
rd o
O. M
K fit
o
u
-------
124
-------
CM CM
V
o
S55
o <
H- (ft
H- Z
M Hi
Q_ CO
M M
OQ
UJ Z
ECO
^Y^ ^^
CM
Q CM _1
3d?
(O T Q
oz
UJZ<
z < t/>
M tfl
a: uj
o CM z
-JOO
X _JN
ooo
ii
>s *
<
. , ^^?
00
*
O
co
•
O
in
•
o
—\ 1-
CN «-
• •
O O
O
in
o
o
o
«n
U.
O
-in
o
o
o
in
- o
o
•
o
o
(0
4J
G
0)
3
W
c
§
rH
O
U
G
a
to
0)
X!
4J
c
•H
to
CN
^H
U
I
« s
Q (0
O -P
O W
w to
-H 0)
M O
(0 O
£4 M
g eu
o
u
H
oo
0)
M
3
01
•H
Pn
m
125
-------
^
r—
*
o
I
CM
I
o
o
in
to
o
o
o
- in
CM
CM
u.
o
0<
- in Q
o
o
o
in
r o
00
O
(O
o
o
•
o
CM
O
O
o
CN
oo
0)
m <
Cr>
••H
126
-------
m < 2 ID e>\
127
-------
O
m
o
o
K)
CM
U_
O
in
o
o
o
ID
-o
00
O
T"
O
•
O
to
o
o
•
o
T-
o
•
o
T
O
•
o
-------
chlorine process streams (Figure 85).
Precursor breakthrough to steady-state for all process
streams occurred very rapidly, prior to day 40. As indicated for
tne previous acetonitriles, the DBAN-C1- levels in th~
nondisinfected GAG column effluent were more than twice those of
tne nondisinfected sand column effluent (Figure 86). Also, the
D3AN and DBAN-C12 levels in the chlorine sand column effluent
-were equivalent indicating the absence of DBAN precursors in the
chlorine GAG influent. Yet the levels of DBAN-C1 were
significantly higher than the DBAN levels in the chlorine GAG
eftluent indicating the presence of additional precursors (Figure
87). Similar treatment of the GAG effluents with chloramine and
subsequent 5-day storage produced significantly lower levels with
respective average DBAN-NH-Cl concentrations of 0.010, 0 012
0.034, 0.012, and 0.005 u^/L for the nondisinfected, ozone'
_cnlorine dioxide, chloramine, and chlorine process streams
(Figure 88). These DBAN-NH Cl levels were essentially the same
as those of the respective sand column effluents.
The average concentrations of trichloroacetonitrile (TCAN) in
the sand column effluents were equivalent to those of the
respective disinfectant contact chambers in all process stream's
except for that of chlorine which increased 86% across the sand
column to i».06S ug/L due to increased contact time. The TCAN
sand column_ effluent concentrations for the nondisinfected,
ozone, cnlorine dioxide, and chloramine process streams were very
similar averaging 0.001, 0.001, 0.002, and 0.004 uq/L,
respectively (Figure 89). Chlorination of the sand c.olumn
effluents ana subsequent 5-day storage produced similar TCAN-C1-
levels for all process streams with respective averages of 0.062,
0.058, 0.104, and 0.054 for the nondisinfected, ozone, chlorine
dioxide, and chlorine process streams (Figure 90). The higher
average level for the chlorine dioxide process stream resulted
from two somewhat irratic data points at the beginning of the
operational period. Treatment of the sand column effluents with
chloramine followed by 5 days of storage produced TCAN-NH Cl
levels less than 0.001 Ug/L in the nondisinfected-, ozone,
chlorine dioxide, and chloramine process streams.
No consistent breakthrough of TCAN was observed across the
GAG columns of the nondisinfected, ozone, chlorine dioxide, and
chloramine process streams with average concentrations of 0.001
ug/L or less. A relatively consistent breakthrough of TCAN was
observed across the chlorine GAG column after day 50 with an
average concentration of 0.003 ug/L (Figure 91). Chlorination of
tne GAG column effluents followed by 5-day storage produced
TCAN-C12 levels slightly lower than the similarly treated sand
column effluents. TCAN-C1- levels averaged 0.1243, 0.034, 0 070
0.tf3fa, and 0.046 in the nondisinfected, ozone, chlorine dioxide,'
chloramine, and chlorine process streams (Figure 92). Similar
129
-------
CM
I CM C
O _J CM _J O
U_ x:
• O +'
a
v_ -H
in
o
o
- O
- o
o
CD
D)
0)
CNJ
t-H
u
I
ca S
a (0
Q)
<4-4 S-)
o -P
en
c
O w
U) CO
•H 0)
M O
(0 O
ft M
13 P.!
o
CJ
in
CO
(U
tr>
•H
Q CD <
130
-------
to
0
o
o
•a
-------
_1 CM
t V
QQ T
ZZOO
CO (/) O O
UJ UJ 111 UJ
zzzz
HI HI HIM
C£ Q£ Q£ Q£
OOOO
_J_J _l_l
XXII
OOOO
—r~
oo
T"
(0
•
O
10 •** ^^ CM
• • • • •
OOOO
T
o
o
O
in
ro
o
o
s
CM
CM
' Ll.
o
tOQ
o
o
o
in
-o
tc
CD
w
CD
O
O
M
CM
QJ
G
•H
S-l
o
U
-------
CM O
X CM
<0
o<
O
in
Hi LU
Q2
2O
ON
ZO
<1
o
CM
-IX
-JO2
OCMT
CMX O
X2<
2 I O
I O
O
-------
Q <
ZQ ato
C/)
-------
CM CM
CM_J_I
_J O O
Oil CM
I QQCM_J
(£ZZ_JO
O«O I
»- V) W T Q
t- ZUJZ <
MMZ
Q)
rH
a
u s
EH tO
0)
O -P
02
O W
w to
•rH 0)
(0 O
u
0)
M
3
tn
•H
H- O <
135
-------
o
o ,o
-------
(£>
*•—
•
O
T~
CM
o
*•—
*
o
00
O
(O
O
o
•
o
—r-
CM
o
—r-
o
o
o
to
10
o
o
o
in
CM
It-
CD
- tn Q
O
O
o
in
u
m
Q)
o
jj
c;
M-l
w
o
u
•r-i
05
CM
r-1
O
I
tH tS
(U
4H M
O -P
CO
c
O 05
CO CO
•H (I)
M 0
(0 O
a M
u
0)
I- O <
CD
•H
137
-------
treatment of GAG effluents with' chloramine again produced
TCAi\;-iSlH2Cl concentrations less than 0.001 ug/L in all process
streams.
Haloketones--
Only two haloketones, 1,1,1-trichloropropanone (TCP) and
1,1-dichloropropanone (DCP) were detected with the highest levels
(1-2 ug/L) being observed in the chlorine process ;stream.
Although consistent breakthroughs of these haloketones were
observed across the GAC columns, removals remained above 85%
throughout the one year operational period.
The average TCP concentrations in the effluent of the
disinfectant contact chambers were essentially identical to those
of the sand column effluent in all process streams except for
tnat of ozone which exhibited some evidence of biodegra'dation.
Average sand column effluent concentrations ranged from 0.034
ug/L for the nondisinfected process stream to 1.64 ug/L for the
chlorine process stream with 0.036, 0.062, and 0.15 ug/L .for the
ozone, chlorine dioxide, and chloramine process streams,
respectively (Figure 93). With an average influent concentration
of 0.75 ug/L, a 52% reduction attributable to biodegradation was
observed across the ozone sand column (Figure 94). Chlorination
of the sand column effluents followed by 5-day storage produced
similar TCP-C1- levels in each process stream with respective
average concentrations of 2.14, 2.45, 4.23, and 2.45 ug/L for the
nondisinfected, ozone, chlorine dioxide, and chlorine ^process
streams indicating the presence of 2-4 ug/L of TCP precursors
(Figure 95) . Treatment of the sand column effluents with
chloramine and subsequent 5-day storage produced significantly
lower levels with average TCP-NH-Cl levels of 0.003-0.007 ug/L
(Figure 96). Similar to the reductions observed for TOX: in the
terminal chloramine samples, these levels were lower tha;n those
of the sand column effluent by a factor of 10 or greater as
indicated in Figure 97 for the chlorine- dioxide and chloramine
process streams. •
The breakthrough of TCP across the GAC columns was observed
on or about day 100 for all process streams (Figure 98).
Steady-state was reached shortly thereafter with average
concentrations of 0.002, 0.001, 0.003, 0.004, and 0.057 ug/L for
tiie nondisinfected, ozone, chlorine dioxide, chloramine, and
chlorine GAC columns resulting in respective removals of 93, 98,
95, 97, and 96 percent throughout the remainder ;of the
operational period. Chlorination of the GAC effluents and
subsequent storage for 5 days resulted in respective average
TCt>-Cl2 levels of 1.6, 2.1, 2.7, 1.3, and 2.3 ug/L for the
nondisinfected, ozone, chlorine dioxide, chloramine and chlorine
process streams (Figure 99) . These levels were comparable to
138
-------
o
Q <
ZQ Qtn
0)
u e
EH (0
(U
c:
o m
CQ W
•H CD
M O
tO O
2-4 n
S A<
o
U
Q)
O Q.
Cn
•H
139
-------
o
u
c
(0
en
fd
cu
x!
U
O
td
4->
G
O
U
0)
O
N
O
Q)
X
-P
C
•H
CD
i-l
0)
>
CU
U
O •
03
C -P
o c
U Q)
•H 3
H rH
(0 '4-1
U
0)
J- O Q.
tr>
•H
140
-------
'^ft-,.
CJCM
CM
OQCv)_j
KZZ_IO
o«pT
h-C/Jtnla
< QZ
I-ZUJZ<
M M Z < CO
Q.C/) MC/)
WW ££ UJ
OQOCMZ
UJZ.JO O
ceox_j N
CLZOOO
I I I
*—-<*
(O
in
T-
CM
O
-in
o
ro
ro
.
O
in
CM
0=)
-oo:
04
L.
O
O
-o
-O
to
o
m
OJ
G
cu
M-t
w
G
e
3
rH
O
u
G
(d
CO
Q)
J5
4J
G
•H
CO
rH
SJ
Q)
CN
i-t
U
I
a. •
u g
^ ns
cu
M-t M
O -P
c
o
en
T
o
\
en en
•H Q)
M o
(0 O
P4M
e 04
o
o
in
i-oo.
d)
s-i
3
tn
•H
141
-------
-JO
OCN
CMX
o:z
o<
O-JO
04O04
X04X
Q i O
ZQZ
I- Z < Z <
M M CO
-------
O
CM
O
CM
X
oo
into
_i«j
CMCMOO
OO CM CM
-I-IXX
oozz
,TT
T
O
Tf
O
O
O
ro
IO
CM
O
CM
T~
10
T"
o
"T"
IO
O
-O
f-io
ro
o
FO
ro
O
[-10
CM
o
h o
CM
u.
o
0<
O
ho
. O
IO
~T
O
O
c
m
Q)
t)
•H
X
o
•r)
Q
0)
G
•H
U
0)
X!
-P
•H
W
rH
Q)
> •
Q) W
iJ -P
i-i 0)
O 3
CMrH
tC M-l
S M-t
I H
PU
U C
C O
(XS U
u c
to
u-i
O 0)
fi -H
o e
w td
,
e u
o
r-
o\
-------
o
o o
f
!'!!
! ! '
IT-
CM
T
O
—IT-
CO
O
T~
CO
O
-y-
O
•
o
T-
CM
O
O
in
n
o
o
o
JO
CM
013
ott:
CM
u.
o
0<
IOQ
o
o
-O
-o
o
o
u
ftj
0)
G
0)
4-1
w
G
o
o
u
o
Q)
-P
G
•H
CO
0)
Cn •
U g
EH n3
CD
m VH
O 4J
CO
C
o m
U} CO
0)
n o
(d o
& M
g &
o
CJ
i
CO
<7l
1- O CL
•H
144
-------
e*
QZCMOO
ZOOCM.J
ON-JIX
ZOOZO
i I
(O
HOQ.
-------
e:f fluent
chlorine
those of the sand column effluents indicating essentially no
removal of TCP precursors by GAG filtration. Except for a few
outlying data points, the TCP-NH2C1 levels in the GAG effluents
after similar treatment with chloramine were essentially the same
•as those of the chloramine treated sand column effluents -(Figure
100). [
The levels of 1,1-dichloropropanone (DCP) in the effluents of
the disinfection'contact chambers and the sand columns were
essentially equivalent. Average sand column
concentrations were very similar in the chloramine and
process streams at 0.43 and 0.47 ug/L, respectively, while those
of the nondisinfected, ozone, and chlorine dioxide process
streams were somewhat lower at 0.07, 0.04, and 0.10 ug/L !(Figure
101) . Chlorination of the sand column effluents followed by 5
days of storage produced similar DCP-C1, levels in all process
streams with averages ranging from 0.11-^.15 ug/L (Figure 102).
Comparison of these data to the average DCP sand column effluent
concentrations indicated a 0.36 ug/L reduction (75%) for the
chlorine process stream and a 0.11 ug/L increase (400%) for the
ozone process stream, while those of the nondisinfected and
chlorine dioxide process streams were essentially unchanged.
Cnloramination of the sand column effluents and subsequent 5-day
storage resulted in DCP-NH-Cl levels in the nondisinfected,
ozone, and chloramine sand column effluents which were
essentially identical averaging 0:.04-0.05 ug/L (Figure 103). As
seen earlier for a number of other parameters, a significant
decrease (9U%) in the DCP levels in the chloramine sand column
effluent was observed after additional chloramination and storage
for 5 days (Figure 104). Also, the average DCP-NH2C1 level for
the cnlorine dioxide sand column effluent of 0.30 ug/L was 0.16
ug/L (114%) greater than the DCP level in the sand' column
effluent.
The breakthrough of DCP across the GAG columns at the 0 „ 002
ug/L level occurred on or about day 60 for the chlorine ; process
stream, while breakthrough was observed about day 100 for the
other process streams (Figure 105). Steady-state was reached in
all process streams on day 130 with respective average
steady-state concentrations of 0.003, 0.003, 0.006, 0.008, and
0.047 ug/L for the nondisinfected, ozone, chlorine dioxide,
chloramine, and chlorine GAG columns corresponding to removals of
95, 93, 94, 98, and 90 percent. Chlorination of the GAG column
effluents and subsequent storage for 5 days resulted in average
DCP-C1 levels of 0.06-0.09 ug/L for all process streams^ (Figure
106) w&ich were comparable to the similarly treated sand column
effluents. The same was true for the chloramine treated GAG
effluents with respective average' DCP-NH2C1 levels of 0.03, 0.06,
0.2d, 0.04, and 0.02 ug/L for the nondisinfected, ozone, chlorine
dioxide*, chloramine, and chlorine process streams (Figure' 107) .
146
-------
0_J
CMO
XOJ
II
O<
CD
If)
MUI
QZ
§O
1*4
ZO
o
CM
_IX
-IOZ
OCMT
CMIO
zz<
zTo
O
-------
Q <
ZQ OCO
•
o
00
•
o
•» to
3 CD
QO Q.
in
*
o
CM
•
O
I
o
o
o
ro
o
in
CM
CM
U.
O
" mo
o
-o
-O
in
-o
D
(C
QJ
-P
C
0)
w
o
u
Q)
c
•H
CO
0)
u e
Q td
cu
m M
O 4J
CO
fi
O en
CO 10
-H Q)
>-t U
(0 O
& M
e PI
o
u
i-H
O
H
0)
3
tn
148
-------
O
fO
*
O
IO
CM
T~
O
CM
o
V
•
o
T-
«o
o
o
in
10
o
o
to
o
in
CM
• IOQ
o
t"0
IO
-o
O
O
o
td
0)
+J
c
(1)
4-1
G
E
rH
2 5
oo:
CM
u.
o js
C
(d
CO
01
iH
Q)
Q)
CM"
iH
U
I
O. •
U g
Q (0
o w
tfl W
•H a)
^ o
a o
£ £M
o
u
(N
o
Q)
M
Q O Q.
149
-------
.—»
"^\
^*- .^
CMX-J _J
X Z O _ J O
Z T W C) CM
T Q X CM X
IK Z Z X Z
O < T z T
I— (/) Q I Q
auto co
UJ Z CM O O
a: o x - j M
O.ZZOO
0
in
fO
o
o
ro
o
10
CM
o^
oo:
CM
u.
o
o<
IOQ
o
ho
.o
to
-o
I
00
I
r-
CO
o
to
o
QOQ.
ro
d
CM «~
d o
I
O
.c
u
0)
M-t
O
JJ
c
s
rH
M-t
«R
H
C
3
O
U
•73
C
Cfl
0)
0)
Q)
CM
m
u e
a a
Q)
^W M
C
O
CO
CO
CO CO
•H Q)
n o
(S O
Cli 5-1
e cu
o
u
ro
O
rH
0)
3
150
-------
O 04
CM I
TOO
OOZZ
wen
CMCMOO
OO oj o-i
ddll
T"
03
"T"
CO
•
o
T~
10
•
o
O
t~m
o
o
ro
o
in
CM
"T r~
CM T-
• •
O O
T
O
•
O
CN
Lu
O
o
o
o
m
-o
0)
•o
O
•H
Q
Q)
C
•H
M
O
rH
U
0)
c
•H
to
CD
Q) to
i__n .jj
•H Q)
U 3
CNrH
M I J |
Sl.|_|
HH
I W
U C
°l
C O
U C
Q (0
0)
,C 0)
•P C
•H
^ti Cj
b m
G O
O rH
to &
•H O
e 10
o
O
rH
0)
Q O Q.
CD
•H
151
-------
o
o o
-------
CM
z
^ o
fd o
o
U
vo
o
rH
s
P i
Q O Q.
•H
Pn
153
-------
O_J
ojo
X01
Sf
*-
I
1^
O
in
ro
O
O
ro
in
CM
U
(0
Q)
4-1
O
-P
G
s
O
U
U
<
o
0)
CM
CD
o
Ll_ -H
o n
rH
U
Cv
SB
a
Q
m
CO
•
o
in
•
o
ro
•
o
-------
Cnloropicrin (CP)—
The concentrations of chloropicrin (CP) in the effluents of
tne nondisinfected, ozone, chlorine dioxide, and chloramine
contact cnambers were essentially equivalent averaging 0.004,
0.UU4, 0.015, and 0.038 ug/L (Figure 108), respectively, while
that for tne cnlorine process stream was somewhat higher
averaging 0.43 ug/L. Chlorination and subsequent storage of the
sana column effluents produced elevated CP-C19 levels averaging
1.3, 7.7, 1.4, and 1.3 ug/L for the nondisinfected, ozone,
chlorine dioxide, and chlorine process streams indicating that
ozonation resulted in a significant (6.4 ug/L) increase in CP
precursor levels (Figure 109). Only slight elevations in CP
concentrations were observed after similar chloramine treatment
and storage of the sand column, effluents•with respective average
CP-NH2C1 levels of 0.03, 0.04, 0.11, and 0.09 ug/L for the
nondisinfected, ozone, chlorine 'dioxide, and chloramine sand
columns (Figure 110") .
No consistent breakthrough of CP above the 0.003 ug/L level
was observed across the GAC column of any process throughout the
operational period (Figure 111). This resulted in average
removals of 61, 54, 84, 95, and 99 percent for the
nondisinfected, ozone, chlorine dioxide, chloramine, and chlorine
process streams, respectively, witn the indicated higher removals
being _associated with higher sand column effluent concentrations.
Chlorination of the GAC column effluents followed by 5-day
sfc°*a9e resulted in respective average CP-C19 levels of 0.51,
0.2b,. 0.48, and 0.37 ug/L for the nondisinfected, ozone, chlorine
dioxide, and chlorine process streams indicating relative
removals of 67, 96, 69, and 74 percent (Figure 112).' No
steaay-state condition was evident for any process stream. The
average CP-C12 concentration in the chloramine GAC column
effluent was similar at 0.49 ug/L. While ozonation produced the
highest level of CP precursors, ozonation followed by GAC
filtration resulted in the lowest levels of CP precursors
Similar treatment of the GAC effluents with chloramine produced'
respective average CP-NH.Cl levels'of 0.012, 0.011, 0.050, 0.022
ana 0.006 ug/L in the nondisinfected, ozone, chlorine dioxide,
cnloramine, and cnlorine process streams (Figure 113) which were
slightly lower than those of the similarly treated sand column
effluents.
Summary of Disinfection By-Products—
Tne annual average sand filter effluent concentration of
total organic carbon (TOC), total organic halide (TOX)
trihalomethane (THM), haloacetic acid (HAA), chloral hydrate
(CH), naloacetonitrile (HAN), chloropicrin (CP) , and haloketone
(HK) for each process stream are compared in Table 19. The total
155
-------
___.o
o
ZQ
*i
T
CN
o
in
ro
o
Q
JO
O
to
CN
-001
T"
O
00
*
O
T~
CO
•
o;
CM
*
O
—r~
o
•
o
-in
o
o
o
in
-o
U
tfl
0)
o
-P
c
.s
rH
U-l
m
w
o
u
m
c
-H
CO
u
^c g
4J (0
0)
m M
O 4J
o in
co u
•H Q)
M o
ns o
8
oo
o
H
0)
OX-JOQIOQ-WO
156
-------
0--:
CM 04
04 _J .J
. J O O
O I I
I Q OC
ITZZ_
O«C
i~tnto i Q
< QZ
I- ZUJ Z <
Ml W Z
w
O W
tn in
M o
m o
o
o
a)
M
3
157
-------
-JO
OCM
Si
I-W
<
HZ
MM
D. CO
W HI
00
bJZ
o:o
O.Z
d
<
-i _j
0_IO
CM OCM
XCMX
zxz
TzT
Ql Q
ZQZ
o
o
to
o
in
CM
002
CM
u.
o
4J
C
Q)
w
c
O
u
m
CO
Q),
-P
c
•H
O
o
o
m
u
CM
o
x: g
-P
-------
o
o o
m M
O 4J
o w
tQ IQ
•H 0)
M O
(0 O
G. M
u
(U
V)
id
tn
•rH
PL,
159
-------
CM
CM
T CM CM I
O _5CM JO
<
O . _ .
O i O
Z*A
O
in
o
o
ro
o
in
CNI
O
O
CM
- in
o
o
u
(8
c
0)
H
C
rH
O
U
o
in
0)
0)
CM
rH
U
I
to
O
0)
O 4-)
Cfl
c
o w
ID 05
•H Q)
(0 O
O
CM
00
•
O
(O
•
o
CM
•
o
o
*
o
0)
S-l
3
tn
•H
160
-------
0
CMO_JOZ
XCMOCMT
ZXC^XO
O TzTO
-------
TOG, mg/L
TOX, ug Cl/L
THM, ug/L
HAA, ug/L
CH, ug/L
HAN, ug/L
CP, ug/L
HK, ug/L
3.1
23
1.1
1.2
0.03
0.06
0.004
0.1
2.3
11
0.9
1.3
0.01
0.05
0.004
0.07
Table 19. Annual Average Sand Filter Effluent Concentration.
:Chlorine
Nondisinfectea Ozone Dioxide Chlorine Chloramine
,3.1 3.2 3.'l
87 233 39
1.2 49 3.3
3.7 30 5.6
0.34 4.5 • 0.3
'0.1 '3.1 0.3
0.U2 3.4 0.04
0.2 2.1 0.:7
Total ;
DSPs, ug/L 2.5 2.3 5.3 39 10 :
level of DBFs produced was estimated as the sum of the individual
DBFs indicated (THM, HAA, CH, HAN, CP, and HK). The highest
level of DBFs in the sand filter effluent were produced by free
chlorine at 89 ug/L followed by chloramine at 10 ug/L. Treatment
with chlorine dioxide produced a, total of 5.3 ug/L DBFs while
ozonatibn resulted in the lowest level of DBFs at 2.3 ug/L which
was just slightly below that of the nondisinfected control! at 2.5
ug/L. These levels of DBFs formed during disinfection were
evaluated relative to the CT,0 requirements of the Surface Water
Treatment Rule using the average disinfectant residuals, water
temperature, and pH, as well as the T,Q contact times determined
for the ozone contact chamber at 10 min and that for the ;contact
chamber _in series with the sand column of the other disinfectants
at 17 min. With an average water temperature of 16°C, a pH of
7.7, and a sand column effluent chlorine residual of 0.9 mg/L,
the chlorine contact chamber in series with the sand column
produced a DBF level of 89 ug/L achieving a CT,0 of 15.3 mg/L-min
whicn corresponds to a Giardia lamblia removal, of approximately
0.5 log and a virus removal of greater than .4 logs. The chlorine
dioxide contact^ chamber and sand column. achieved a CT1M;of 5.1
mg/L-min at 16 C for a Giardia removal of 0.8 log ana a virus
removal of 2.4 logs while producing only 5.3 ug/L of DBFs. The
ozone process stream which contained the lowest level of DBFs in
162 ;
-------
the sand column effluent at 2.3 ug/L, achieved a CTn of 5
mg/L-min in the ozone contact cnamber which exceeded a T.0 log
removal for Giardia and viruses. Chloramine, while producing
only 13 ug/L of DBFs, achieved a CT, of 34 mg/L-min resulting in
only a 0.07 log removal of Giardia and a 0.17 log removal of
viruses.
_ The maximum average level of DBFs in a chlorinated
Distribution system after ' 5 days at 30°C following initial
predisinfection and sand filtration have been estimated by the
terminal-C!2 data obtained for each process stream. Relatively
high levels of DBFs were formed in the chlorinated distribution
simulation (terminal-Cl ) samples, regardless of the type of
preaisinfectant employed, as indicated in Table'20. Pretreatment
with ozone and chlorine dioxide -followed by sand filtration
resulted in the lowest maximum levels of DBFs averaging 309 and
308 ug/L, respectively, while maximum average DBF levels of 474
and 446 ug/L were observed for' the nondisinfected and chlorine
process streams. From these data, it is evident that extended
cnlorine contact nas almost completely negated any beneficial
effects derived from ozonation which were apparent in the sand
column effluent, and, that the use of chlorine in the
distribution system is prohibitive for those waters containing
nign levels of DBF precursors in systems using only sand
Table 20. Maximum Annual Average Chlorinated Distribution
. System Concentrations for Sand Filtration.
Chlorine
Nondisinfected Ozone Dioxide Chlorine
TOX, ug Cl/L 557 339 379 540
THM, ug/L 236 154 138 225
HAA, ug/L 153 87 ' 113 139
CK, ug/L 79 55 , 45 75
HAN, ug/L 2.0 2.7 6.0 .2.9
cp' U9A 1.3 7.7 ±.4 1.3
HK, ug/L 2.2 2.6 4.3 2.5
Total
DBFs, ug/L 474 309 308 446
163
-------
filtration. ;
Similar maximum average DBF levels for a chlorinated
aistrioution system following initial preclisinf ection, sand
filtration', and GAG filtration witn a 20 min empty bed contact
time are presented in Table 21. The respective maximum average
annual DBF levels after GAG filtration and 5 days in a
cnlorinated distribution system were 138, 183, 232, and 223 ug/L
for pretreatment witn ozone, chlorine dioxide, chlorinje, and
cnioramine with the nondisinfected process stream averaging 213
ug/L. i-vhile these levels are also relatively high, it is
possible to produce water with • lower levels of DBFs by
reactivating the GAG within the filters at a frequency of less
tnan 12 months. For example, by assuming that the DBF levels in
the effluents of all the treatment plant filters are evenly
distributed across the DBF breakthrough profiles in Figure 114,
tnese profiles can be used to estimate the GAG reactivation
frequency for maintaining the running average of DBFs below a
desired level, such as 100 ug/L. Using these criteria, the
reactivation frequencies required to maintain a running average
of total DBFs of less than 100 ug/L after 5 days in a chlorinated
distribution system were estimated at 240, 170, 115, 100, .'and 90
days for pretreatment with ozone, chlorine dioxide, chloramine,
chlorine, and no pretreatment, respectively. If the criteria for
Table 21. Maximum Average Chlorinated Distribution
System Concentrations for GAC Filtration. '
.Chlorine ;
Nondisinfected Ozone Dioxide Chlorine Chloramine
TOX, ug Gl/L 231 127 ,216 246 235
THM, ug/L 107 74 ' 84 125 112 ;
HrtA, ug/L .72 39 i 65 70 ' 75 '
CH, ug/L 28 20 26 31 . 30 ;
HAN, ug/L 4.1 2.54.7 3.3 3.9
CP, ug/L 0.5 0.3 ^ 0.5 0.4 -0.5
HK, ug/L 1.6 2.2 2.8 2.3 1.3
Total
DBFs, ug/L 213 138 183 232 223
164
-------
CM
CM
I CM CM
O-J CM_lO
ra
0)
o
ho
m
-o
4J
o
EH
m
o
o
CO
•H
!H
rci
i1
O
u
Q)
OU.
QCDO.C/)
165
-------
reactivation was a THM level of. less than 5fc! ug/L, the
reactivation frequency would be reduced to approximately 130 days
for the ozone and chlorine dioxide process streams and
approximately 50 days for the remaining process streams :(Figure
19). ;
The maximum average DBF levels present in a chloraminated
distrioution system after 5 days at 30 C were estimated from the
terininal-NH-Cl samples from each process stream in which ammonia
was added pri-or to chlorine. The average DBP levels in a
chloraminated distribution system for predisinfection followed by
sand filtration, presented in Table 22, indicate that the; use of
cnloramine- in the distribution system is particular effective in
minimizing DBFs with respective maximum annual averages of 12,
19, 24, and 21 ug/L for predisinfection with ozone, chlorine
dioxide, cnloramine and no predisinfection. While ozone and
cnlorine dioxide can produce effective removals of Giardia and
viruses as previously indicated,' chlorine dioxide pretreatment
also produces a significantly hig;her level of unknown DBFs than
does ozone, as measured by TOX, with respective averages of 89
and 27 ug Cl/L. Chlorine dioxide also produces the inorganic
by-products of chlorite (est. at 0.5 mg/L) and chlorate (not
measured) which will soon be regulated. Ozone, on the oth'er hand
Table 22. Maximum Average Chloraminated Distribution
System Concentrations for Sand Filtration.
Chlorine
Nondisinfected Ozone Dioxide Chloramine
TOX, ug Cl/L 44 21 89 59
THM, ug/L 8.5 ; 3.2 4-2, 9'4
HAA, ug/L 12 8.8 14 14 ;
CH, ug/L 0.03 0.03 0.3 0.08
HAN, ug/L 0.04 : 0.04 0.3 0.05
CP, ug/L 0.03 0.04 0.1 0.09
HK, ug/L 0.04 0.06 0.3 0.05
Total
D6Ps, ug/L 21 12 19 24
166
-------
in
produces aldehyde and ketone by-products and results exr^m^y
hign heterotrophic plate counts (HPCs) in the sand column
effluent as Described in a subsequent section. In order to
reduce the HPC to an acceptable level, a short chlorine contact
period on the order of several minutes would be required prior to
tne ado.it ion of ammonia for the formation of chloramines. This
would, of course, slightly increase the level of DBFs in the
distribution system.
Slightly lower DBF levels can be achieved in a chloraminated
distribution system by using GAG as a filtration media as
indicated in Table 23. Maximum annual DBF levels averaging 7,
IvJ, 13, and 45 ug/L were observed for ozone, chlorine dioxide
cnloramine, and chlorine predisinf ection with the nondisinf ected
process stream averaging 12 ug/L. The relatively high level of
UBPS m tne chlorine process stream was due primarily to the
oreaKthrough of instantaneous THMs with' some contribution from
the naloacetic acids. The use of GAG filtration in conjunction
with ozone or chlorine dioxide predisinf ection to minimize DBF
levels in a chloraminated distribution system would appear to be
of little benefit and would be cost prohibitive.
Table 23. Maximum Average Chloraminated Distribution
System Concentrations for GAG Filtration.
Nondis-
- infected
29
5.
6.
0.
0.
0.
0.
2
4
04
03
01
04
Ozone
13
2.
4.
0.
0.
0.
0.
2
5
08
05
01
07
Chlorine
Dioxide Chlorine
26
2
6
0
0
0
0
.2
.6
.3
.3
.05
.2
69
37
7.
0.
0.
0.
0.
5 .
06
03
006
03
Chloramine
32
5.
6.
0.
0.
0.
0.
5
7
2
05
02
05
TOX, ug Cl/L
THM, ug/L
HAA, ug/L
CH, ug/L
HAN, ug/L
CP, ug/L
HK, ug/L
Total
DBFs, ug/L 12 7 10 45 13
167
-------
TOX as a Surrogate for DBFs—
The TOX data ootained on a weekly basis from the sampling
locations in each process stream were compared with the
corresponding sum of the- eighteen disinfection by-products
measured during this study which include the THMs (4), haLoacetic
acids (6), haloacetonitriles (|4) , haloketones (2), :chloral
hydrate, and chloropicrin. Because the response of,one mole of
oromide is equivalent to that of one mole of chloride'by TOX
analysis, the difference in weighlt: of the brominated species was
compensated for by determining the percent halide as chloride for
each compound and applying this factor to tne concentration
found. This resulted in ug chloride/L which was directly
comparable to the ug chloride/L determined by TOX measurement.
For example, while 10 ug bromofbrm/L is equivalent to 9.48 ug
bromide/L, the concentration as chloride is only 4.21 ug
cnloride/L. This difference in response was empirically verified
by the TOX analysis of bromobenzene and chlorobenzene.
The TOX and DBF data from each sampling point in each -process
stream are compared in Figur.e 115 by process stream, and in
Figure 116 by the sampling location within each process ;stream.
The center line in each figure represents the regression line
with the outside lines representing two standard deviations.
Wnile a relative good correlation was observed for these 1094
data points with a correlation coefficient of 0.90 and a standard
deviation of 17, it was evident from the figures that those data
points in tne area of 0-50 ug Cl/L for the DBFs and 150:-350 ug
Cl/L for TOX appeared to skew the correlation. Examination of
both figures revealed- that these data originated from the
chlorine contact chamber and sand column effluents. -This
suggests that the reactions occurring at these locations were not
yet near equilibrium due to their relatively short contact times,
and that the reaction rate of TOX formation was apparently
greater than that of the DPBs. A closer examination of the data
in the 0-150 ug Cl/L range for TOX in Figures 117 & 118 indicated
that the contact chamber and sand column effluent data : of the
chlorine dioxide and chloramine process streams also appeared to
skew the correlation for similar reasons. In addition,
examination of the data for the chloramine process stream in
Figure 119 suggested that some TOX species may be precursors of
some of the DBFs, such as the THMs, haloacetic acids, and
cnloropicrin. While the average terminal chloramine DBF |and TOX
levels in the chloramine sand column effluent were 9 and 59 ug
Cl/L, respectively, the average of the corresponding
instantaneous DBF and TOX levels from the same location: were 4
and 94 ug Cl/L. This indicated that a 35 ug Cl/L (37%) reduction
in TOX had occurred with a corresponding increase in DBFs ;of 5 ug
Cl/L (125%) following treatment with additional chloramine and
storage for 5 days. ;
168
-------
1
o
in
CN
1
o
o
CM
1
O
in
T—
i
O
0
1
o
ID
3P-
o
o
CO
O
CO QOQQ_(/)
. 169
Q)
M
0
Cn
•H
EH
-------
o
in
CM
o
o
CNJ
O
in
o
o
o
in
Q)
a,
>i
E-)
Q)
iH
!•
(0
c
o
•H
•P
m
o
o
(0
to
ffl
P
0)
•P
G
m
o
en
•H
CO
(0
G
O
•H
CO
CO
0)
tn
CD.
CD
Cn
-H
CO
O
-------
171
-------
172 '
-------
Q CD Q_ (7) Z) CO
00 Z)
-------
Because of the apparent difference in the reaction rates of
TOX and DBF formation, and the relatively short disinfectant
contact times observed for the instantaneous samples, all
instantaneous data was removed from the data pool. This included
tne removal of the instantaneous GAG effluent data because, while
at equilibrium due to the lack of a disinfectant residual,
additional disinfection would be required before being
distributed to consumers. with only 254 termi-nal chlorine and
cnloramine data points remaining, a better correlation was
observed with a higher correlation coefficient of ui.95 along with
a higher standard deviation of ;20.3. From Figure 120, it is
apparent that two seperate data sets were present, one for
chlorine and one for cnloramine. While separation, of these two
data sets in Figures 121 and 122 produced poorer correlations
with lower correlation coefficients of 0.88 and 0.71 for
terminal-Cl^ and terminal-NH-Cl, respectively, these data sets
were the most normally distributed and were considered more
representative of the distribution system. A better correlation
may oe obtained by using actual distribution samples since
laboratory distribution system simulation (terminal) samples are
generally less precise than instantaneous pilot column or
distribution system samples. The correlation of terminal
chloramine TOX and DBFs may not have much practical application
since these terminal values were generated by adding .ammonia
prior to cnlorine and, thus, are indicative of the minimum levels
wnich may be observed in the distribution system. In most
treatment plant applications, the order of addition is reversed
with tne free chlorine contact time varying with plant, design
which may soon require modification depending upon the
requirements of the Surface Water Treatment and Disinfection
By-Product Rules under the Safe Drinking Water Actv With
increased free chlorine contact time, the DBFs and TOX produced
will fall somewhere in between the terminal chloramine and
terminal chlorine regression lines suggesting that TOX may only
be of value as a DBF surrogate for those systems which employ
free chlorine in the distribution system.
MICROBIOLOGICAL OBSERVATIONS :
Heterotrophic Plate Count (HFC)—
The geometric means of the heterotrophic plate count (HFC)
observed across each process stream, presented in Table 24,
indicated that ozone exhibited the greatest level of disinfection
after 3W minutes of disinfectant contact, followed closely by
chlorine and chlorine dioxide. Chloramine was. considerably less
effective. While a 58% reduction in HFC was observed across the
nondisinfected sand column, the HFC in the ozone sand column
effluent increased dramatically to 28,000 cfu/mL due to the
174
-------
CO Q DQ Q_ 00 Z> O
00 =)
O b_
-------
GO ID
O
-------
C\J O CO CD
U~i ID
-------
Table 24. Comparison of HPC Across each Process Stream.
Geometric Mean of HPC, cfu/mL
Nondis infected
Ozone
Chlorine Dioxide
Chlorine
Chloramine
Contact
Chamber
Effluent
10,900
10
21
.19
230
Sand
Column
Effluent
4,600
28,000
17
! 13
195 '
GAC
Column
Effluent
1,800
2,300 i
16,300
12,800 ;
1,900
dissipation of the ozone residual,, the relatively high levels of
assimilable organic carbon available from ozone oxidation, and
the abundance of dissolved oxygen which averaged 9 mg/L. 'The HPC
levels in the sand column effluents of the chlorine, chlorine
dioxide, and chloramine process streams were similar to those of
their respective contact chamber effluents. Even though ithe HPC
levels in the ozone sand column effluent were exceedingly high,
those of the ozone GAC column were lower by a factor of ,ten and
were similar to those of the nondisinfected and chloramine GAC
columns. The HPC levels of the chlorine and chlorine dioxide GAC
columns were 6-8 times higher than those of the other iprocess
streams and may have resulted from a rapid regrowth response due
to the dissipation of the disinfectant residual in the upper
position of each GAC column. As indicated in a forthcoming
section, elevated levels of assimilable organic carbon were not
evident in the chlorine process stream. ',
Total Coliform—
The percentage of positive coliform samples which occurred
across eacn process stream are presented in Table 25 along with
the average coliform count of the positive samples observed at
each location. As expected, the nondisinfected process stream
contained the highest levels of coliforms. A reduction in
coliform density of 54% was observed across the nondisinfected
sand column with an additional 42% occurring across the
nondisinfected GAC column reducing the coliform density to an
average of 4 cfu/100 mL. While ozone completely inactivated all
coliform bacteria in the disinfectant contact chamber, low level
i
178
-------
Table 2b. Comparison of Total Coliform Across each Process
Stream.
Total Coliform
(% Positive/Average, cfu/190 tnL)
Contact Sand GAC
Chamber Column Column
Effluent Effluent Effluent
Nondisinfacted 100/95 100/44 , - 83/4
Ozone 0/0 14/2 3/4
Chlorine Dioxide 0/0 0/0 28/1
Chlorine 0/0 0/0 ' 28/2
Chloramine 19/1 1/1 7/1
coliforms were frequently observed in the effluent of the
biologically active ozone sand column. Similar low levels
occurred in the ozone GAC column effluent but at a reduced
frequency. While essentially no coliforms were observed in the
contact chamber and sand column effluents of the chlorine dioxide
and chlorine process streams, low coliform levels occurred with a
relatively high frequency in both GAC column effluents. A
relatively high frequency of low level coliforms was also
observed in the chloramine contact chamber effluent and • is
believed to have resulted from the use of a pulsating metering
pump for the hypochlorite feed. In a previous one year study
which had a somewhat higher influent coliform density, a similar
ammonia solution feed pump followed by a continuous chlorine gas
feeder produced the same chloramine residual .in the contact
chamber effluent with essentially no coliforms. ; This suggests
that the higher frequency of detection observed in this study may
have resulted from the pulsations of the metering pump even
though tne hypochlorite solution was fed ahead of -the ammonia
solution. While additional cnloramine contact across the sand
column removed essentially all of the remaining coliform
bacteria, low level coliforms were occasionally observed in the
effluent of the chloramine GAC column.
Assimilable Organic Carbon (AOC)—
179
-------
Due to the work intensive: natura of the AOC analysis
procedure and limited manpower, AOC measurements were limited to
_tne strongest oxidants, ozone and chlorine, and were conducted on
a monchly basis using pseudomohas fluorescens strain; P17 to
assess seasonal variations (Figure 123). The AOC levels in the
nondisinfected influent ranged from 5-25 ug acetate C-eq/L with
an average of 10 ug acetate C-eq/L. The highest levels of AOC
were observed in the effluent of the ozone contact Ichamber
averaging 166 ug acetate C-eq/L and ranging from 51-268 ug
acetate C-eq/L. The seasonal variations of AOC observed^ in the
ozone contact chamber effluent correlated quite well with water
temperature (Figure 2) tnroughout the operational period with
lower AOC levels being formed at lower water temperatures. With
tne exception of one outlying data point on -day 300, the
biologically active ozone sand column reduced the AOC to an
average level of 39 ug acetate C-eq/L for a reduction of 77
percent. The further reduction of AOC down to the 4 ug'. acetate
C-eq/L level occurred across the: ozone GAC column. Exdept for
one outlying data point which also occurred on day 300, 'the AOC
levels in the chlorine contact chamber effluent were relatively
low averaging 5 ug acetate C-eq/L. Similar AOC levels averaging
3 ug acetate C-eq/L were observed; in the effluent of the chlorine
GAC column. ;
HS2 Coliphage—
g Each month, an MS2 coliphage seed containing approximately
IB pfu/mL was continuously pumped into the nondisinfected
influent following pressure sand filtration 5 to achieve a
nomogeneous concentration of approximately 10 pfu/mL in the
influent of each disinfected process stream. The log removals
observed for MS2 coliphage across the disinfection contact
cnamber of each process stream are presented in Tables 26 and 27
along with the corresponding water temperature, disinfectant
residual, and calculated CT. value relative to the Surface
Water Treatment Rule (SWTR) promulgated under the Safe Drinking
Water Act. While the plug-flow contact time calculated for each
disinfectant contact chamber was 30 min, the T, ~ contact >:time as
specified by SWTR guidelines, was measured at 10 min for the
ozone contact chamber and 17 min for the other disinfectant
contact chambers. • The shorter contact time in the counterQurrent
ozone contact chamber evidently resulted from the additional
turbulence created by the rising gas bubbles. I
Wifn CT-,0 values , ranging from 2.7-5.1 mg/L-min acrbss the
ozone contact chamber, all coliphage was inactivated resulting in
removals greater than 4-6 logs. This agrees with the SWTR
guidance manual in that the manual requires a CT,~ value: of 1.2
mg/L-min at 5 C and 0.3 mg/L-min at 25 C for a 4 log virus
removal wnich are approximately 4-10 times lower than those CT,(,
values observed across the ozone contact chamber. Cokiphage
180
-------
o: *
o
tK O
O <
O ZO
-J< O<
U-f- O O O
ZZZO
MO «UJ UJ
ocnozz
-------
Taole
26. HS2 Coli-phage
Dioxide.
Reductions
Ozone
oT'- "
25.5
26.1
24.7
13.9
15.3
3.4
3.3
3.3
10.4
11.9
17. 6
18.9
24.4
25.0
Residual ,
mg/L
'J. 34
0.45
3.27
J.51
0.41
id. 37
0.44
0.44
S.51
0.47
v> . 44
•3.37
0.44
CT,,.,1
3.4
4.5
2.7
5.1
4.1
3.7
4.4
4.4
5.1
4.7
4.4
3.7
4.4
Log
Removal
>4.1 '.
>4.7
>5.6 1
1
>5.9
>5.7
>5.9
>5.6
>5.8 !
>5.8 !
>5.6 '
>5.2 ,
>5.3 i
>4.9
by Ozone and Chlorine ]
Chlorine Dioxide
Residual
mg/L
0.38
0.48
0.48
0.38
0.86
0.48
0.55
0.40
0.51
0.49
0.57
0.48
0.53
2"
_
6.5
8.2
8.2
6.5
14.6
8.2
9.4
6.8
8.7
8.3
9.7
8.2
9.0
Log :
Removal
>4.7
5.6
>5.9
>5.7
>5.9 '
.5.5
5.3
5.4
>5.4
>5.6
>5.2
>5-3
>4.9
T, .. for Ozone = 10 min
JL t)
'T, ., for; C102 = 17 min
levels ranging from 0-3 pfu/mL jwere observed in the chlorine
dioxide chamber effluent resulting in log removals of 5 or more
corresponding to CT^ values ranging from 7-10 mg/L-min.
182
-------
Table
Temp.
28.
26.
24.
18.
15.
8.
3.
8.
10.
11.
17.
18.
24.
25.
1T]_
5
1
7
9
3
4
3
3
4
9
0
9
4
0
27. MS 2
Residual
mg/L
0.7
1.
1.0
0.8
1.0
1.2
0.8
0.9
0.8
3Q
* iP
1.2
1.8
1.2
1.1
17 min
Colipnage
Chlorine
,
11.
18.
17.
13.
17.
20.
13.
15.
13.
15.
20.
30.
20.
18.
Reductions by
Chlorine and Chloramine.
Chloramine
-^ Log Residual
,, Removal mg/L CT1 ,,
9
7
0
6
0
4
6
3
6
3
4
6
4
7
>4.1
>4.7
>5.6
>5.9
5.7
5.2
4.8
5.7
>5.8
>5.4
>5.6
>5.2
>5.3
>4.9
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
5
4
3
8
5
5
6
7
6
6
4
4
6
6
25.
23.
22.
33.
25.
25.
27.
28.
27.
27.
23.
23.
27.
27.
5
8
1
6
5
5
2
9
2
2
8
8
2
2
Log
Removal
1.
1.
0.
0.
0.
0.
0.
0.
0.
0.
0..
0.
1.
0.
4
8
9
8
5
5
4
4
6
2
5
6
3
6
Comparison of these data to the requirements in the SWTR
guidance manual indicated that the values in the guidance manual
are 5-8 times greater than those observed across the chlorine
dioxide contact chamber. For example, the required CT value
for a 4 log virus removal at 3 C in the manual is 38 mg/L-min
with a 5 log removal in the 60-70 mg/L-min range, while a 5.5 log
coliphage removal was observed across the chlorine dioxide
183
-------
contact chamber at 3°C with a CT-j^ of 8.2 mg/L-min. Similar to
chlorine dioxide, chlorine inactivated essentially all .of the
coliphage applied witn contact qhamber effluent levels .ranging
from 0-7 pfu/mL resulting in log removals of approximately 5 or
greater witn CT,,, values of 14-20 mg/L-min. As was the case for
ozone, the CT.. ,. ** va lues observed for chlorine were 2-10 times
greater tnan tnose in the guidance manual of 6 mg/L-min at 5 C
and 2 mg/L-min at 25 C for a 4 'log virus removal. Coliphage
levels in the, effluent of the ^chloramine process stream; ranged
from 4.8 x 10 pfu/mL to 4.0x12 pfu/mL resulting in log removals
of only:j.2-1.8 with CT,^ values of 22-31 mg/L-min. Relative to
the SwTR guidance manual" which requires a CT,0 of 857 mg/L-min at
4°C and 214 mg/L-min at 25 C for a 2 log virus removal, the CT, „
values observed across the chloramine contact chamber were 2-8
times lower than those required i;n the manual for equivalent log
removals. Because of the dramatic difference between the CT,~
requirements of free chlorine and chloramine, e.g. 2 mg/L-min
versus 500 mg/L-min for a 4 log virus removal at 25 C, this
greater '. inactivation rate was obviously the result of the amount
of free chlorine contact achieved prior to reaction with ;ammonia
to form chloramine. while the in-line free chlorine contact time
prior to entering the chloramine contact chamber was estimated at
only a few seconds with an average free chlorine residual of 1.5
mg/L, a similar free chlorine residual for a period of only one
minute would result in a 3 log virus removal according to "the
SwTR guidance manual. Thus, a-4 log virus removal could be
achieved with a 2 mg/L free chlorine residual and a T,g contact
time of only 4 min at 5 C and 1 min at 25 C before quencning with
ammonia to form chloramine so as to minimize the level of
disinfection by-products formed.. ;
-------
REFERENCES
1. Koffskey, Wayne E. and Brodtmann, Noel V., Organic
Contaminant Removal in Lower Mississippi River Drinking
Water by Granular Activated Carbon' Adsorption,
EPA-600/S2-83-032, June 1S83.
2. Koffskey, toayne E., Alternative Disinfectants and Granular
Activated Carbon Effects on Trace Organic Contaminants,
EPA-600/S2-87-006, April 1987.
3. , Standard Methods for the Examination of water and
/•/astewater, 16th Edition, 1985.
4. riowes, J. E. et al, Determination of Dioxin Levels in
Carbon Reactivation Process Effluent Streams,
EPA-600/2-84-013.
5. Process Design Manual for Carbon Adsorption, U. S. EPA,
EPA-625/1-71-002A, October 1973.
6. Neukrug, Howard M. et al, Effect of a Spill Event on an
Ozone - Granular Activated Carbon Treatment Plant, Water
Chlorination, Volume 5, 1985.
185
-------
------- |