-------
- 42 -
CO
OC
o
S3
co 5 Q-
*0
O
(O
O
if)
o
CO
o
CM
O _,
E CO
£ UJ
z O
o X
f= UJ
"2u.
§H
s§
o
UJ
CO
UJ
QC
z>
g
u.
CO
2
UJ
CO
CO
H
UJ
5
LJJ
O
Q.
UJ
OC
•H
J-l
a
rt
OJ
0)
1?
jj
a)
pa
•H
H
(|B6 OOCH/D) 1SOO W31SAS
-------
-43 -
Influence of Empty Bed Contact Time
The tradeoff between bed depth (empty bed contact time) and reactivation
frequency for a given quality of water is an important relationship. To
make this comparison the ratio between reactivation frequency in months and
the EBCT in minutes was calculated. For example, for one of the systems in
Table XIV with an EBCT of 10 minutes and a reactivation frequency of 0.5
months, the ratio, "R", is as follows:
R- °'5 = 0.05 (1)
10 minutes
The R values were calculated for the sand replacement systems in Table XIV
XIII with contact times of 10 and 20 minutes (lines 4 and 5) . The system
with 20 minutes contact time has twice the bed depth of the system with 10
minutes contact time.
The cost in cents per thousand gallons is plotted versus R in Figure 14
for these two systems. These data show that the longer contact time bed is
always more expensive for the same value of R. When the R value remains the
same, this means that a direct proportionally exists between empty bed
contact time and reactivation frequency, which may or may not be true in a
given situation. The data collected with the pilot columns described in the
previous sub-section will provide the necessary information for this analysis.
For a given value of R and a given EBCT (P^ the straight line drawn
horizontally to the curve representing the longer contact time bed (?„) ,
represents the increase in the period between reactivations that is required
for the longer contact time to be economically equivalent to the shallower
bed. In Figure 14 at P , R = 0.01, the total unit cost for the 10 minute
empty bed contact time bed is 45.1 cents/1000 gal. Drawing a horizontal
line to the 20 minute contact time curve yields an R at P~ of 0.0124. The
necessary reactivation frequency for cost equivalancy can be calculated as
follows:
P2 = 0.0124 (2)
20
P
2 = 0.25 months
-------
-44-
(|B6 000170) 1SOO IN31SAS
g
LL
-------
- 45 -
Therefore, the period between reactivations would have to increase by 150
percent to be able to use a iQO percent longer EBCT system at the same cost.
This non-proportional relationship between EBCT and reactivation frequency
is caused by the increased activated carbon inventory that must be maintained
for deeper beds.
At larger values for R (longer periods between reactivation), however,
the difference between systems becomes small so that little economic penalty
occurs when choosing a system with a longer EBCT and a longer period between
reactivations. Performance data from the recommended pilot column test (see
previous sub-section), to determine whether or not a more than proportional
lengthening in reactivation frequency will occur with an increase in contact
time, can be used to make this type of analysis in a given location.
Influence of Type of System Chosen
In the following analysis when post-filter adsorbers and sand replacement
systems are compared, the comparison will be made between the last two systems
in Table XIV because the design empty bed contact times are equal. Figure 15
shows the cost in cents per thousand gallons for a sand replacement system and
the post-filter adsorber having equal empty bed contact times (10 minutes). For
very short reactivation periods (less than 2 weeks) post-filter adsorbers are
always less expensive than sand replacement. When the reactivation period is
greater than 0.5 months, however, sand replacement becomes less expensive because
of less capital expense.
Figure 16 shows the relationship between a sand replacement system with a
10 minute EBCT and a post-filter adsorber with a 20 minute EBCT. As might be
expected from the previous analysis, with smaller R values the longer EBCT post-
filter adsorber is less costly than the shorter EBCT -sand replacement system,
but as the period between reactivations increases, the sand-replacement system
becomes relatively less expensive. Therefore, under this set of
-------
-46-
UJ
N m
o co
< 2
_l UJ
Q. H
LU CO
DC >•
f Q CO
« o Z Z
^ s < O
z CO d
= Q t
2 O
£H
Q. co
in
6
3z=!
S O "r
co H
E co
< O
OL Q-
o
U)
o
co
(|B6 OOOI./0) 1SOO W31SAS
Q
o z
o <
in
UJ
oc
O
u.
w
c
o
•H
4J
Cfl
U
cfl
0)
Pi
c
OJ
0)
M
01
•H
H
-------
-47-
g
H
> Q
H Z
O <
m
q
d
LJJ
CC
Z
UJ
uj Q 2
i Z LU
H < O
o
<
o
o
Q.
LU
cc
o
(O
o
CM
(|B6 0001/3)) 1SOO IN31SAS
> < z
o H <
£ ZCO
=3 O M-
Q ii ^^
H
2 DO (/) tt
< > O O
g > e o to
UJ
tr
u.
O
O
U. LU CO
U. DC O
LU U. Q.
CO
HI
CC
D
g
LL.
-------
- 48 -
assumptions, the increased cost of building post-filter adsorbers cannot be
overcome by obtaining longer periods between reactivations through the use
of longer EBCTs, if the reactivation frequency is greater than 0.4 months.
Because the cost differential is not great, however, other considerations
may dictate final designs. Note: the cost comparisons of other combinations
than those presented in Figures 13-16 can be made using the data in Table
XIV.
Influence of Granular Activated Carbon Cost
To minimize costs, a water purveyor might consider the use of the least
expensive granular activated carbon available. A lower cost activated
carbon, however, may also require shorter periods between reactivation, if
its performance is reduced. For example, using the data shown in Figure 17,
for a 10 mgd post-filter adsorption plant, if a pound of activated carbon
costs $0.70 and the reactivation frequency is three months, then the system
cost would be 17.5 cents/1000 gal. With a less expensive activated carbon,
perhaps one costing $0.30/pound, the reactivation frequency would have to be
2 months or greater to achieve a favorable economic tradeoff under these
assumed conditions.
Influence of Inflation
Table XIII shows that under static economic conditions, sand replacement
systems are slightly less expensive than post-filter adsorption systems.
Because post-filter adsorption is less labor intensive than sand replacement,
it might become less expensive some time in the future because of inflation.
Figure 18 illustrates the impact of inflation on the two 100 mgd systems
(one of each type) assuming an inflation rate of 5 percent per year and
shows that sand replacement system does become more expensive than post-
filter adsorption in year 18.
-------
- 49 _
.0.0.0
o
0
t-
CO
CM m
Q
O
QC
UJ
Q.
q
in
CJ
q
d
q
in
q
d
q
in
q
6
(|B6 00017$)
IN31SAS dO 1SOO
H o
O (0
< o
O LLI
CO H
O CO
8o.
> cc =
i s§
5 t "•
UJ < .
£ >fc
UJ CO
O
UJ
oc
D
O
LL
-------
- 50 -
Q. CO
O
CM
in
q
in
q
6
q
in
q
6
(|e6 00(H/3>)
IAI31SAS dO 1SOD
o>
LU
z
g
i-
o
D
CC
(/)
Z
o
o
£ z
t 2
DC
<
LU
£2
°H
h- 0-
s§
og
z <
UJ Z
I- UJ
25
CC ui
u_
00
LiJ
CC
3
O
-------
-51 -
Figure 19 shows the impact of inflation on the two granular activated
carbon configurations, over a 20-year period assuming inflation at 7 percent
per year. Post-filter adsorption becomes less expensive than sand replacement
in year 12 under this set of assumptions for a 100 mgd plant.
Figures 18 and 19 illustrate that over the life of the two types of
systems, because of the labor intensive nature of the sand replacement type,
it eventually becomes more expensive than a post-filter adsorber on a yearly
expenditure basis. This phenomenon occurs, of course, because the capital
expenditure remains fixed over the life of the investment, while operating
costs, particularly labor costs, are subject to inflation.
Figures 18 and 19 also illustrate that the total expenditure over time
is less for the sand replacement system than for post-filter adsorbers. To
account for total expenditures, a "present value" analysis was made for the
systems listed in Table XIII. Two discount rates (6% and 8%) and three
inflation rates (5%, 7% and 9%) were used in the analyses. The results are
summarized in Table XV. As can be seen from the Table, for the larger
plant, at the highest inflation rate, for both discount rates, the difference
in present value for the two systems is small. In no case, however, is the
present value of the expenditure for post-filter adsorbers less than those
for sand replacement.
The unit costs in Table XIII show that small treatment systems in
general are more expensive,on a per unit of product basis, than larger
systems. These costs can be reduced significantly, however, by the use of
truck transport and regional reactivation systems. This effect is illustrated
3 4
in the Trihalomethane Interim Treatment Guide. '
-------
-52 -
CC
LU
I-
o
o_
•o
O)
E
o i
o
UJ
CO
oz
og
z H
uj o
I-CO
-------
- 53 -
TABLE XV
SUMMARY OF PRESENT VALUE ANALYSIS
FOR POST-FILTER ADSORPTION AND SAND REPLACEMENT SYSTEMS
System
Sand Replacement
Inflation at
Inflation at
Inflation at
10 mgd 100 mgd
Discount Discount Discount Discount
at 6% at 8% at 6% at 8%
5% 225.67 190.96 119.43
7% 262.37 219.50 140.86
9% 305.30 252.58 166.12
101.
117.
137.
67
64
13
Post-Filter Adsorber
Inflation at
Inflation at
Inflation at
5% 259.96 221.20 134.69
7% 289.80 244.72 150.63
9% 323.35 270.50 169.12
114.
127.
141.
76
15
40
-------
- 54 -
Summary
This sub-section has discussed some of the important factors that a
water purveyor must consider when making decisions regarding the use of
granular activated carbon systems for removal of organic contaminants.
Because the economics of choosing and designing a granular carbon system
are complicated, individual utilities or their consultants, or both may need
assistance in developing cost and economic design criteria. The Water
Supply Research Division is prepared to assist those utilites affected
by treatment regulations. Computer programs currently being utilized by
WSRD and instructions in the use of these programs will be made available
to interested parties. The Water Supply Research Division is prepared
to provide a limited economic analysis for individual utilities affected
by the treatment regulations. This assistance will aid the utility in
making general decisions regarding overall implementation strategies,
but will not be sufficient for specific designs.
SUMMARY
In summary, the Interim Treatment Guide provides information that demonstrates
that granular activated carbon adsorption is the best available treatment technology,
Fall 1977, for treating water to remove organic contaminants, thereby improving
finished water quality and providing the American consumer with a more healthful
and esthetically pleasing drinking water. For a more detailed summary the reader
is referred tc the Executive Summary at the front of this document. Three
Appendices with more detailed information on various aspects of this subject
follow this Guide.
-------
- 55 -
ACKNOWLEDGMENTS
The Authors wish to extend their special appreciation to Ms. Maura
M. Lilly without whose dedication, secretarial skill, diligence and
patience, this Guide could not have been completed in a timely manner.
They also wish to thank the Organic Contaminants Research Staff, J.K.
Carswell, J. DeMarco, P. Dorsey, W.C. Elbert, W.A. Feige, D.L. Guttman,
D. D. Hinderberger, K.L. Kropp, O.T. Love, Jr., B.W. Lykins, L.A. Moore,
D.R. Seeger, C.J. Slocum, B.L. Smith, A. A. Stevens and R. Stevie, all
of whom contributed to this document.
REFERENCES
1. Dostal, K.A., Pierson, R.C., Hager, D.G. and Robeck, G.G., "Carbon
Bed Design Criteria Study at Nitro, West Virginia," JAWWA, 57, No. 5,
663-674 (May 1965).
2. Stevens, A.A. and Symons, J.M^, "Measurement of Trihalomethane and
Precursor Concentration Changes," JAWWA, 69, No. 10, 546-554
(October 1977).
3. Symons, J.M., et al., "Interim Treatment Guide for the Control of
Chloroform and Other Trihalomethanes," Water Supply Research Division,
U.S. Environmental Protection Agency, Cincinnati, Ohio, June 1976,
mimeo, 48 pp. plus 4 Appendices, unpublished.
4. Clark, R.M., Guttman, D.L., Crawford, J.L. and Machisko, J.A.,
"The Cost of Removing Chloroform and other Trihalomethanes from
Drinking Water Supplies," Municipal Environmental Research Laboratory,
Office of Research and Development, U.S. Environmental Protection
Agency, Cincinnati, Ohio 45268, EPA-600/1-77-008, March 1977.
See Also:
"Activated Carbon in Water Treatment" Proceedings of a Water Research
Association Conference at the University of Reading, April 3-5, 1973,
Water Research Centre, Medmenham, United Kingdom, Available from the
Water Research Centre, Henley Road, P.O. Box 16, Medmenham, Marlow,
United Kingdom, SL7 2HD.
Translation of Reports on Special Problems of Water Technology -
Volume 9 - Adsorption" Proceedings of a Conference in Karlsruhe,
Federal Republic of Germany, 1975, EPA-600/9-76-030, December 1976.
-------
APPENDIX A
PERFORMANCE OF GRANULAR
ACTIVATED CARBON FOR THE
REMOVAL OF ORGANIC COMPOUNDS
Written By:
J.K. Carswell
R.M. Clark
J. DeMarco
P. Dorsey
W. A. Feige
D.L. Guttman
O.T. Love, Jr.
B. W. Lykins
A. A. Stevens
R. Stevie
J.M. Symons
Reviewed By
Gordon G. Robeck
-------
APPENDIX A
TABLE OF CONTENTS
Literature Review Al
Taste and Odor Control Al
Removal of Organic Compounds as Measured by General
Organic Parameters A5
Removal of Specific Organic Compounds A9
Reactivation of Granular Activated Carbon A22
Current Water Supply Research Division Findings A34
Class I Compounds (Taste and Odor Producing) A34
Class II Compounds (Synthetic Source Water Contaminants) A36
Naphthalene A36
Other Polynuclear Aromatic Hydrocarbons A37
Carbon Tetrachloride A37
Trichloroethylene, Tetrachloroethylene and Similar
Chlorinated Organics A40
Prediction of Granular Activated Carbon Bed Performance,
Studies with Humic Acid and Carbon Tetrachloride,
Dieldrin, PCB, Benzene, and _p_-Dichlorobenzene A41
Mixtures from Actual Waters A42
Unidentified Compounds A52
Summary A56
Class III Compounds (Disinfection By-Product Precursors) A56
Class IV Compounds (Disinfection By-Products) A63
General Organic Parameters A66
Influence of Empty Bed Contact Time A69
Biological Growth and Endotoxin Production A76
Standard Plate Count A76
Endotoxins A80
Ozone Enhanced Granular Activated Carbon Adsorption (Biological
Activated Carbon) A81
Literature Review A81
WSRD Pilot Plant Results A86
Economic Analysis Information A91
Basic Costs A91
Influence of Empty Bed Contact Time A92
Influence of Granular Activated Carbon Cost A99
Influence of Inflation A99
Cost of Reactivation A108
Summary Alll
Acknowledgments A112
References A113
-------
PERFORMANCE OF GRANULAR ACTIVATED CARBON FOR THE REMOVAL OF
ORGANIC COMPOUNDS
The purpose of Appendix A is to summarize the current state of
knowledge concerning the performance of granular activated carbon
adsorption as a treatment unit process. Appendix A will be in two major
parts, one, a review of the literature, and two, a summary of the current
status of the Water Supply Research Division's projects on this subject.
The literature review will not be exhaustive, but will be a summary of
the subject. Because many of the research projects discussed in the
second portion of Appendix A are on-going, the research findings to date
(Fall 1977)* will be presented as progress reports. Detailed papers on
these studies will be published in the technical literature as they are
completed.
Literature Review
Taste and Odor Control
Many water treatment plants in the United States are practicing taste
and odor control by the addition of powdered activated carbon, which also
removes some organic matter. Currently (Fall 1977) however, about 35 plants
are using granular activated carbon, either alone or on top of some sand, as
both a filter media for particulate control and an adsorption media for
organic contaminant control. This type of system is hereafter called a sand
replacement system and is used primarily for taste and odor control. A few
water purveyors installed these granular activated carbon beds because of
the organic pollution in their raw water as well as taste and odor problems.
In their 1964 paper, Woodward, Dostal, and Robeck reported on five
installations of granular activated carbon beds during the 1930's. A
2
more recent survey of the known granular activated carbon installations in
this country, see Table I, shows that approximately 80 percent of the water
*Some data collected beyond that contained in the November 1977 draft of
Appendix A has been included in the report.
-------
- A 2 -
TABLE I
DATE GEANULAR ACTIVATED CARBON FIRST INSTALLED2
Experimental Use (9 Plants) Routine Use (38 Plants)
1935
1961
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
0
0
0
0
0
1
3
1
1
1
1
0
1
1
1
2
1
4
6
7
7
5
3
1
-------
-A3 -
treatment plants routinely using granular activated carbon beds have
been installed since 1970.
A discussion of some of the early installations is of interest. In
the fall of 1963 the management of the Nitro, West Virginia, water utility
installed two beds of granular activated carbon to investigate treatment
for removal of taste and odor. During the 3-month testing period the
Threshold Odor Number (TON) of the raw water varied between 500 and
1000. During this same time period the TON of the influent water to the
experimental granular activated carbon beds varied between 100 and 200,
with one excursion to 400. These beds, with empty bed contact times (EBCT)
of about eight minutes, were able to produce an odor-free water for as
3
long as 26 days. The success of these tests encouraged the water purveyor
to convert the entire plant to granular activated carbon beds in 1965.
Another interesting feature of this installation was the construction of an
on-site 10,000 Ib/day multiple-hearth reactivation furnace. From that time,
until the plant was closed several years later, because the municipality
obtained a different source of water, the granular activated carbon was
reactivated approximately every six months.
The success of granular activated carbon beds for controlling taste and
odor demonstrated at Nitro, West Virginia, has been duplicated at many other
water utilities. For example, at Piqua, Ohio; Mt. Clemens, Michigan; Lawrence,
4
Massachusetts ; and Davenport, Iowa, the water purveyors were having difficulty
providing their consumers with an acceptable drinking water in spite of the
use of large doses of powdered activated carbon. After conversion to granular
activated carbon beds, these purveyors reported successful control of their
taste and odor problems. Figure 1 shows that at Lawrence, Massachusetts,
during a time period when the TON in the settled water was in the 6-10
range, the granular activated carbon beds controlled the effluent TON to
acceptable levels for over a year.
-------
-A4-
(CI1OHS3}IHiMO139)
O
=>
O
LU
00
fid
LU
00
Z
O
CO
0<
Q>
Ofi
_ u
LU CO
Z<
O ..
U_ LU
-z z
LU fi£ UJ
fid UJ Qd
o<5
U.O. _i
to
LU
a:
X
aOI«3d >I33M CL iiOd NO1 39V33AV
-------
-A5-
When conversion was first contemplated, the Massachusetts State Health
Department insisted that some sand be left in the bottom of the filters to
avoid any problems with particulate (turbidity) breakthrough. A six-month
comparison of the turbidity in the effluent of a conventional sand filter
and an all granular activated carbon system, see Figure 2, showed that their
performance was equal. Note, pilot plant scale comparisons by the Water
Supply Research Division Laboratory has confirmed this finding. When the
entire plant was converted to granular activated carbon, no sand was left in
the filter boxes.
2
In the survey cited above the water purveyors were asked how long
their current charge of granular activated carbon had been in service and
whether or not the beds were still effective for taste and odor removal. The
shortest effective life reported was 23 months and some beds had been in
service for four years and were still effective. Reports from the United
Kingdom substantiate these findings from United States practice.
Removal of Organic Compounds as Measured by General Organic Parameters
The previous sub-section detailed the performance of granular activated
carbon beds for the removal of taste and odor causing compounds, but investigators
have also been interested in the performance of the granular activated
carbon adsorption process with respect to the removal of analytic parameters
that would be reflective of the "total" organic content of water. As pointed
out by Stevens and Symons, no analytic test currently available measures
the "total concentration of organic compounds" in water. Several tests have
been proposed to approximate or be proportional to this parameter. A few of
these are: 1) the organics-carbon adsorbable test producing a carbon chloroform
extract (CCE), 2) chemical oxygen demand (COD), and 3) total organic carbon (TOG).
-------
-A6-
u
z
LU
ae
z
z
LU
s
oc
LU
Q_
X
u
z
LU
ot
>-
o
LU
X
—
z
<
o
Z
O
co
V
V
<
s
Q£
Q.
<
z
<
a:
O
O
Z
<
CO
u.
si
CO Q_
" S^
89
. OQ
a: »-
=>c*
OO
simn
-------
-A7-
These tests have been used to monitor various operating granular activated
carbon adsorption beds.
o
In 1965, Dostal, Pierson, Hager and Robeck reported that the percent
removal of chemical oxygen demand declined from 72 percent at the start of
the test to 56 percent after 31 days at the Nitro, West Virginia water
treatment plant, mentioned previously, while 97.5 percent of the odor was
still being removed after 31 days. More recently, monitoring of operating
granular activated carbon beds has been carried out at five field locations:
Nitro, West Virginia; Piqua, Ohio; Mt. Clemens, Michigan; Lawrence, Massachusetts;
and Davenport, Iowa. This monitoring began in September 1968 and spanned
several years. Because analytic capabilities have changed and improved
gradually, the same monitoring techniques were not used at all of these
installations. In general, however, the purpose of the monitoring was to
compare the time of breakthrough (first detectable increase) of organic
compounds as measured by a general organic parameter with the time of odor
breakthrough.
In every case, the breakthrough of odor occurred much later than the
breakthrough of the parameter measuring the removal of general organic
compounds. Except for the Nitro, West Virginia situation, the control of odor
was successful for years, whereas control of organic compounds as measured by
a general organic parameter was successful only for weeks. The following
are some typical data: Nitro, West Virginia, CCE-hf* removal** lasted 5 weeks;
Piqua, Ohio, CCE-lf removal lasted 10 weeks; Mt. Clemens, Michigan, CCE-m
removal lasted less than 28 weeks; Lawrence, Massachusetts, CCE-m removal
lasted 16 weeks, and Davenport, Iowa, CCE-m removal lasted 6 weeks. Figure 3
from Lawrence, Massachuetts, is a typical data plot of the breakthrough
curve for a general organic parameter.
*The lower case letters refer to the method of operation of the activated
carbon adsorption collection and the chloroform extraction system and are
defined in Reference 8.
**Effluent concentration approximates influent concentration.
-------
-A8-
Q31VAI10V M3N'
CO
CN
« 2
UJ
UJ
o z
z
UJ
Z
UJ
S
UJ
a. ^
UJ
DC z
o 2
< o
2 ffl
CC
ui <
o °
Z Q
UJ UJ
DC |-
^ 5
< >
-I H
-
53
o 2
UJ
es
CO
cs
00
o
•o
6
CN
O
UJ
O
O
CO
UJ
CC
O
UJ
I/Bui ' (uj-300) 1QVU1X3 IflaOJOUOIHO N08UVO
O
-------
-A9-
6 *
Figure 4 from Ford indicates that experiences in the United Kingdom
are similar to those in the United States. The experience in Germany is
9*
similar to that in the United Kingdom as shown by Heymann of Duisburg who
investigated the breakthrough pattern of the parameter dissolved organic
carbon (DOC) through a granular activated carbon bed with sampling points at
various depths. Figure 5 shows that the maximum depth, corresponding to an
empty bed contact time of 4 minutes, showed some increase in DOC after about
4 to 5 days. These reports lead to the generalized conclusion that if
controlling the general organic content of drinking water is desirable, the
useful life of a given charge of granular activated carbon in an adsorption
bed will be much shorter than might be anticipated from performance based on
taste and odor compound control.
Removal of Specific Organic Compounds
Early demonstrations of granular activated carbon's ability to remove
specific organic compounds from water came from its use as an analytic
procedure in which organics were adsorbed on granular activated carbon and
desorbed (extracted) with a solvent, chloroform or ethyl alcohol. Analysis
of these extracts revealed what organic compounds that were in the original
sample were able to be adsorbed onto granular activated carbon under prescribed
conditions, and then be desorbed with a solvent. For .example, in 1956,
Middleton and Rosen found the following organic compounds or classes of
compounds in a carbon chloroform extract (CCE) from a surface water:
*Note, the entire document "Activated Carbon in Water Treatment," a Water
Research Association conference held at the University of Reading, April 3-
5, 1973 and available from the Water Research Centre, P.O. Box 16, Henley Rd.,
Medmenham, U.K., SL7 2HD, is an outstanding volume well worth reviewing. A
second document "Translation of Reports on Special Problems of Water Technology
Volume 9 - Adsorption" a conference held in Karlsruhe, Federal Republic of
Germany, 1975, EPA-600/9-76-030, December 1976 is an excellent companion
document to the one just previously cited and summarizes the experiences of
water purveyors in continental Western Europe with activated carbon.
-------
-A10-
o- 6
e* ?«/T
< C LU
UZ*^
IP
*z £ 4
^^ ^J ill
\J H-
-------
-All-
o
CM
in
O
Z <
1«
.. < o
o u-
Q S O
O
CO
Z
< s
S i
LU
O
cs
cs
oo
•O
6
cs
o
|/6ui 'NOIiVUiN3DNOD NO8HVD DINVOHO QHAlOSSia
-------
-A12-
Substituted benzene compounds Kerosene
Polycyclic hydrocarbons Phenyl ether
Acrylonitrile Alcohols
Aldehydes Ketones
Organic acids Esters
DDT
In the same year, Rosen, Middleton and Taylor isolated alkyl
benzene sulfonate from a carbon-alcohol (ethyl) extract (CAE). In
12
1963 Rosen, Skeel and Ettinger isolated the following organic compounds
from a river water CCE.
Naphthalene Tetralin
Styrene Acetophenone
Ethyl benzene Bis- (2-chloroisopropyl) ether
2-Ethylhexanol Bis- (2-chloroethyl) ether
Di-isobutyl carbinol Phenyl methyl carbinol
2-Methyl-5-ethylpyridine
13
In 1970, Rosen, Mashni and Safferman reported on finding both geosmin
and 2-methyl-isoborneol in a CCE from an Ohio Lake. In 1972 a report was
14
published concerning studies on organic contamination of drinking
water conducted in the lower Mississippi River area. Organic compounds
were collected on granular activated carbon and desorbed by both heat
and solvents. One sample contained the following compounds:
acetylene dichloride ethyl benzene
benzene methyl chloride
carbon tetrachloride propyl benzene
chloroform toluene
1,2-dichloroethane vinyl benzene
dimethyl sulfoxide
-------
A second series of samples
acetone
acetophenone
benzene
b r omob en z ene
bromochlorobenzene
bromoform
bromophenyl phenyl ether
(positional isomer?)
butyl benzene
a-camphanone
chlorobenzene
chloroethyl ether
chloromethyl ether*
chloroform
chloronitrobenzene
chloropyridine
dibromobenzene
dichlorobenzene
(positional isomer?)
1,2-dichloroethane
-A13-
contained the following organic materials:
dichloroethyl ether
dimethoxy benzene
2,6-dinitrotoluene
endo-2-camphanol
ethyl benzene
exo-2-camphanol
hexachlorobenzene
l-isobromobenyl-4-isopropyl
benzene (1,2 isomer)
isocyanic acid
methyl biphenyl
methyl chloride
nitrobenzene
o-methoxyphenol
p-menth-en-l-8-ol
tetrachloroethylene
toluene
1,1,2-trichloroethane
vinyl benzene
*The identity of this compound is very questionable.
-------
-A14-
Also in 1972 Kleopfer and Fairless found the following compounds in CCE's
from drinking water taken from the Ohio River at Evansville, Indiana,
using gas chromatographic - mass spectrometric techniques.
Bromodichloromethane Toluene
Dibromochloromethane Tetrachloroethylene
Bromoform Xylene
Ethylbenzene Bis-(2-chloroisopropyl) ether
Styrene Bis-(2-chloroethyl) ether
Hexachloroethane Hexachlorobenzene
Chlorohydroxybenzophenone
Finally, Symons and Stevens reported on organic compound identification
from an ethyl alcohol extract of granular activated carbon (CAE) that included
the following:
Trichlorobiphenyl Di (2-ethylhexyl) phthalate**
Tetrachlorobiphenyl Methyl Ester of Benzoic Acid*
Pentachlorobiphenyl Methyl Ester of Benzene Sulfonic Acid
Hexadecane Methyl Ester of Palmitic Acid*
Octadecane Methyl Ester of Stearic Acid*
Eicosane Methyl Ester of Lignoceric Acid*
Docosane
Diethylphthalate**
* - Probably esterified during extraction.
** - Plasticizers
These reports show some of the types of organic compounds that can
be adsorbed from water by granular activated carbon as evaluated by
this technique.
-------
- A15-
Below are typical references reporting on the removal of specific
organic compounds in conjunction with other organics by granular activated
carbon beds in use at operating water treatment plants. In 1965, Robeck,
Dostal, Cohen and Kreissl demonstrated that coal-base granular activated
carbon, partially exhausted for CCE-hf removal, could reduce the concentration
of endrin, dieldrin, lindane, 2,4,5-T ester, DDT, and parathion dosed into
3
river water. In the same year, Dostal, Pierson, Hager and Robeck showed that
the seven compounds listed below that were present in the Kanawha River water
after aeration could be reduced to below detectable concentrations by fresh (2-
day old) granular activated carbon beds. These compounds were bis-(2-chloroethyl)
ether, 2-ethylhexanol, bis-(2-chloroisopropyl) ether, a-methylbenzyl alcohol,
acetophenone, isophorone and tetralin.
Forty days later, however, all of these compounds with the exception of
acetophenone were detected at a bed depth equal to an empty bed contact
time of about 8 minutes. Providing an additional 2 minutes of empty bed contact
time did remove these seven compounds at this time (40 days), although another
organic compound, ethyl benzene, was detected at a bed depth equal to 15 minutes
of empty bed contact time.
Stieglitz, et al. reported on removals by 2 month-old granular activated
carbon beds of 61 organic compounds amenable to analyses by the Grob closed loop
18
stripping and gas chromatographic procedures . The granular activated carbon
appeared to have lost its effectiveness for adsorbing compounds eluting from the
chromatograph early, such as chloroform and trichloroethylene, but was still
quite effective for later eluting chlorinated aromatics. Some questions remain
as to documentation of analytical recovery efficiencies and the operation of the
adsorption beds, however. The authors claim that some higher concentrations of
-------
-A16-
a few aliphatic and aromatic hydrocarbons in the adsorber effluent can be explained
by chromatographic or biological effects on the adsorber. Dissolved organic
chlorine (DOC1) was still being reduced from 85 to 2 yg/£ through the adsorber at
this time. The sum, as organic chlorine, of the concentration of chloroform,
carbon tetrachloride, trichloroethylene, and tetrachloroethylene in the treated
water accounted for all the DOC1. To summarize, after two months of operation,
all of the typical low molecular weight chlorinated compounds were breaking
through the bed although DOC1 was significantly removed from a chlorinated influent.
19
In 1965 Kolle, Sontheimer, and Stieglitz reported on studies of pilot
granular activated carbon adsorbers receiving ozonated pre-filtered Rhine River
water. Two or three adsorbers in series were used, each one meter in depth.
Each meter of depth represented four minutes of empty bed contact time. After
six months of operation, granular activated carbon samples were taken from the
top of each section and from the bottom of the last section. These activated
carbon samples were extracted with dioxane and the extracts analyzed for specific
organics. The following organics were found in various concentrations: chloroform,
1,2-dichloroethane, 1,2-dichloropropane, tetrachloroethylene, trichloroethylene,
bis-(2-chloroisopropyl) ether, o-dichlorobenzene, hexachlorobutadiene, hexa-
chlorocyclohexane, and tris-(2-chloroethyl)-phosphate.
According to the authors, the substances identified in these extracts can
be classified into three groups, the aromatic chlorohydrocarbons, (o-dichlorobenzene)
that were completely adsorbed by the uppermost layer of the adsorber, the aliphatic
chlorohydrocarbons (hexachlorobutadiene and hexachlorocyclohexane) that were much
less strongly adsorbed, but still are adsorbed well enough so that the lowest
filter layers do not contain these substances, and the oxygen-containing organic
-------
-A17-
chlorine compounds, bis-(2-chloroisopropyl) ether and tris-(2-chloroethyl)
phosphate, that were breaking through the activated carbon beds and were identified
in the bottom layer of the filter.
In addition to the above information, chloroform, 1,2-dichloroethane, and
1,2-dichloropropane were present at various levels throughout the adsorbers,
although tri- and tetrachloroethylene were confined to the upper and middle
layers of the three-layered adsorber. The presence of these first three compounds
throughout the depth of the adsorber would tend to weaken the authors' gross
classification of the adsorption of aliphatic hydrocarbons (see above).
20
In 1977 Suffet, et al. reported on the performance of granular activated
carbon and some adsorbent resins for the removal of trace organics from Philadelphia
drinking water. Suffet used computer-reconstructed gas chromatographic profiles
plus mass spectrometric identifications to assess the ability of the adsorbents
to remove twenty-seven identified organic compounds. For the activated carbon
column (Calgon F-400*, 9.7 minute empty bed contact time) in one experiment, the
adsorbent was shown to be quite effective for removal of most of the compounds
identified although exhaustion was noted for the organic compounds with lower
boiling points by the 18th week of the run. Gas chromatographic profile analysis
of the F-400 column effluent indicates chloroform and trichloroacetone first
broke through after 3 weeks and dibromochloromethane and tetrachloroethane first
broke through after 4 weeks.
* Mention of commercial products does not imply endorsement by the U.S. Environmental
Protection Agency.
-------
-A18-
In a second experiment, breakthrough patterns were presented for 1,2,7,9 and
15 weeks of the run plus the respective organic profiles for the influent water.
Twenty-nine compounds were identified from the gas chromatographic profiles for
this experiment. Empty bed contact time in this experiment was 7.3 - 7.5 minutes.
Again, detection in the effluent of organic species with lower boiling points occurred
sooner than organic compounds with higher boiling points. Suffet cautions the
reader, however, that the data are largely qualitative and interpretation of
results is complicated by the highly variable nature of the organic content of
the influent to the adsorbent column.
21
In 1977 McCarty, et al. reported on the performance of "Water Factory 21"
for removal of organic materials. Water Factory 21 is an advanced waste treatment
facility designed to reclaim wastewater to provide injection water needed for a
sea water barrier system to protect ground waters in Orange County, California.
Part of the treatment train includes packed-bed, upflow pressure adsorbers filled
with Calgon Filtrasorb 300. The empty bed contact time is 30 minutes.
During a period when the plant was operating on a continuous basis, single
activated carbon adsorber influent and effluent samples were taken and subjected
1 ft
to a rather rigorous organic analysis by closed loop stripping. Relative
influent and effluent concentrations were reported for sixteen compounds and
absolute concentrations for twelve of these. The general trend was toward removal
of these compounds to widely varying degrees. An examination of the gas chromatograms
verifies this trend for a large number of unidentified compounds (see section on
"Unidentified Compounds" , page A52, for results of a similar comparison). The
meaning of the results presented by McCarty, et al., however, are difficult to
interpret in the context of predicting activated carbon adsorber life for
removal of the specific compounds identified, because few data are presented on
-------
-A19-
the condition of the activated carbon itself (time in-place, reactivation frequency,
and so forth). Measurement of influent/effluent organic concentration profiles
vs. time is part of the planned future Water Factory 21 work, however.
Most of the information on reducing various concentrations of trace organics
has been gathered through laboratory studies and pilot-scale experiments. The
22
National Interim Primary Drinking Water Regulations established maximum contaminant
levels for six organic chemicals: endrin, lindane, methoxychlor, toxaphene, 2,4-
D and 2,4,5-TP (Silvex). These six specific organic contaminants can be grouped
under the general term "pesticides." The "Manual of Treatment Techniques for
23
Meeting the Interim Primary Drinking Water Regulations" reported that adsorption
on granular activated carbon is the most effective treatment process for reducing
the concentrations of these contaminants.
The U.S. EPA library in Cincinnati made a computer search of the literature
on the subject of adsorption of organic contaminants on granular activated carbon.
Listed below in Table II are 50 organic compounds in addition to those reported
in the text that have been reported to be reduced in concentration through granular
activated carbon treatment. Only those studies where the concentration of the
specific organics before treatment were below the one mg/Jl level were included,
thus eliminating studies on industrial wastes where the concentrations are usually
much higher. Even these concentrations are higher than usually found in source
waters, making direct extrapolation or prediction of adsorption behavior speculative
when low concentrations are present. These data are based on both isotherm and
column type studies.
-------
-A20-
TABLE II
ADDITIONAL ORGANIC COMPOUNDS THAT HAVE BEEN REPORTED IN THE LITERATURE TO BE
ADSORBABLE ON GRANULAR ACTIVATED CARBON
acetophenone
39
i
36
29
,, . 39,43,49
aldrin
baygorT
a-BHC
43,49,50,54
benzocaine
32
29
benzole acid
29
butyric acid
dibrom
,. ,, . 25,39,43,49
dieldrin
di (n-butyl) phthalate
48
di (2-ethylhexyl) phthalate
48
diuon
diquat
50
30,51
dimenthoate
49,50
m-dinitrobenzene
P27,39,43,46,49
38
DDT
endosulfan
),5
26
49,52
, . 50,53
endrin
gasoline
heptachlor
49
heptachlor epoxide
49
hexachlorobenzene
32
49
juglone~
lindane
27,31,49,50
49
linuron
MS-22232
malathion
43,49
49,95
methyoxychlor
, 38
nitrobenzene
oil (fuel)26>35,37,45
30,51,52
paraquat"
parathion
50
,29
phenylacetic acid
,33,34,40,42,45
phenols"
p-nitrophenol
42
propionic acid
29
pyridine
.39^41,49
29
PCB"
rotenone
. 36
sevin
simazine
24,32
28,31,50
53
strychnine
3-trif luoromethyl-4-nitrophenol (TFM)
I.A-D4"'50
2 , 4-dinitrophenol
2,4,5-T (ester)43'50
, 24,50
toxaphene
32
tetrachlorobenzene
49
telodrin
triazine
28
-------
-A21-
A report by Giusti, et al. includes a list of 12 alcohols, 8 aldehydes, 11
amines, 4 pyridines (and morpholines), 8 aromatics (benzene derivatives),
11 esters, 3 ethers, 14 glycols and glycol ethers, 2 halogenated hydrocarbon
solvents, 10 ketones, 8 acids, and 2 oxides that were studied. Single dose
studies (isotherms not determined) were carried out on each of the compounds
with 100 ml of a 1000 mg/£ solution being dosed with 0.5 g of activated
carbon (equivalent dose 5000 mg/£). Concentrations of solute dosed were
less where solubilities so dictated. These data were used to draw conclusions
and test hypotheses about effects of pH, polarity, functional groups,
molecular weight, and other differences in physical and chemical characteristics,
in-so-far as amenability to adsorption is concerned. These data might
relate to relative adsorbability of the respective compounds at these very
high concentrations, but do not evaluate competitive effects of mixtures or
the relative effects of 5 to 6 order of magnitude lower concentrations of
organic matter and adsorbent that are experienced in the drinking water
treatment situation.
Isotherm tests were run for only 5 compounds at varying pH and although
linear, the isotherms were determined only at high concentrations. Parallel
column studies and multi-solute isotherm studies were conducted on these
compounds and the authors claim a fairly high level of predictability of
the column capacities from the isotherm data. Whether this translates well
to lower concentrations (where isotherms may become non-linear) or to real
systems where the solute species number in the hundreds, cannot be determined
from these studies.
In summary, although the individual literature citations are often
vague on critical details of the study, and data have been collected under
a variety of circumstances, many atypical, in total they do demonstrate
that granular activated carbon is correctly described as a "broad-spectrum"
adsorbent.
-------
-A22-
Reactivation of Granular Activated Carbon
Although the internal pore structure providing a large surface area
per unit weight is developed during initial manufacture (thermal activation)
of granular activated carbon, the surface area is finite and eventually
becomes covered with adsorbate, and adsorption ceases. To continue with
effective adsorption the granular activated carbon is processed to remove
these adsorbed materials (reactivation). The most common technique for
processing granular activated carbon to renew the adsorption capacity of
its surfaces is to drive off and oxidize the adsorbed organic compounds in
environment containing steam and little oxygen at high temperature (approaching
1000°C). The problem is to design a facility and choose reactivation conditions,
such that the maximum amount of adsorbed materials is removed with a minimum
of change in the properties of the granular activated carbon.
The four basic types of furnaces currently (Fall 1977) in use for the
reactivation of granular activated carbon are: the multiple-hearth, the
rotary-kiln, the infra-red-tunnel, and the fluidized-bed furnace. Of the
five currently operating on-site reactivation facilities in Europe, two are
multiple-hearth furnaces and three are various designs of the fluidized bed
furnace. Figure 6 is a schematic diagram of a multiple-hearth furnace,
Figure 7 is a cross-section of a one-bed fluidized bed furnace, Figure 8 is
a diagram of the infra-red-tunnel furnace, and Figure 9 shows a rotary-kiln
furnace. The Water Supply Research Division currently has underway three
projects for the evaluation of reactivation factilites.
Proper design and operation of a thermal reactivation facility is
necessary to avoid any unwanted change in the properties of granular activated
carbon. For example, Juhola reported a change in pore size distribution
upon several cycles of reactivation as shown in Table III.
-------
-A23-
FEED IN
GAS OUT
BURNER, STEAM
ENTRY OR AIR
INJECTION •
GAS FLOW
PRODUCT OUT
BRICK HEARTHS
SUPPORTED AT
WALL ONLY
RABBLE ARM (NOT
SHOWN ON OTHER
HEARTHS)
GRANULAR
ACTIVATED
CARBON
CENTER SHAFT
(ROTATES)
SHOWING SOLIDS IN
PERSPECTIVE (TYPICAL
FOR ALL HEARTHS)
FIGURE 6 CROSS SECTION MULTIHEARTH FURNACE
-------
-A24-
GAS INLET
GRANULAR
ACTIVATED
CARBON
INLET
GAS
OUTLET
OUTLET
GAS CHAMBER
INCOMING
AIR CHAMBER
REACTIVATED
GRANULAR
ACTIVATED
CARBON
FIGURE 7 CROSS SECTION OF FLUIDIZED BED FURNACE
NOTE: THIS DIAGRAM DEPICTS A ONE-BED FURNACE AS AN EXAMPLE.
OTHER DESIGNS OF FLUIDIZED BED FURNACES ARE COMMERCIALLY
AVAILABLE.
-------
-A25-
UJ
OC
OC
UJ
N
l_ 5
UJ < -I ^
o oc o z
< UJ OC UJ
z o. j- o
OC 2 Z >•
3 UJ O .
u. h- O O
UJ
o
UJ
o
UJ
cc
o
z
o
o
UJ
CO
(0
o
oc
o
00
UJ
oc
o
-------
-A26-
LU
O
QC
D
U.
O
cc
U.
O
o
LU
CO
(/)
O
CC
O
O)
LU
CC
D
g
u_
-------
-A27-
TABLE III
CHANGES IN GRANULAR ACTIVATED CARBON PROPERTIES ON SUCCESSIVE REACTIONS
Cycle
Ash Content - %
Iodine Number*
Measure of
Pores Between
1 nm and 2.5 nm
Molasses Number*
Measure of
Pores Greater
than 2.8 nm Bulk Density
Initial
1
2
3
5.7
7.6
8.6
9.5
1090
1040
935
940
250
310
290
350
0.469
0.468
0.469
0.473
*These numbers are based on the adsorption of the respective material under
a standard set of conditions.
This change in pore structure was then reflected in a decrease in
performance as measured by a decrease in initial percent chemical oxygen demand
reduction from 55 percent to 40 percent over eight reactivation cycles.
Smith agreed pointing out that based on Figure 10, practically no change
occurred in pore size distribution in the 1 to 1.9 nm diameter range, although a
marked change occurred in the pores between 1.9 and 2.5 nm in diameter. He also
noted that from 2.5 nm diameter and above, again, very little change occurred in
58
the pore structure. Finally, DeJohn and Hutchins also reported that the
effect of thermal reactivation may be to reduce the surface area in the
micropores of granular activated carbon. They further state that the properties
of granular activated carbon derived from lignite are not changed as much as
the properties of granular activated carbon derived from another source,
bituminous-coal. Their data showed a decline in specific surface area of
2
bituminous-coal based activated carbon from 1100 m /gm for virgin material to 700
2
m /gm (36 percent) after four reactivations, although lignite-based activated
i-\
carbon only declined from 650 m^/gm for virgin material (initially lower) to 500
2
m /gm (23 percent) after five reactivations.
-------
-A28-
O
O
O
uT
O
UJ
cc
O
Q.
O
OC
O
UJ
O
.22
.20
.18
.16
.14
.12
.10
.08
.06
.04
.02
ORIGINAL
_TJT^--3rd REACT.
Z-~- 1st REACT.
5th REACT.
L 5TH
ORIGINAL
1.0 1.5 2.0 3.0 4.0 5.0 6.0 8.0 10.0 15.0 20.0
MICROPORE DIAMETER , NANOMETERS
FIGURE 10. CUMULATIVE MICROPORE VOLUMES AS A FUNCTION
OF MICROPORE DIAMETER
REFERENCE 57
-------
-A29-
59
In contrast to these reports, Juntgen noted that, based on Figure 11, an
appropriate time-temperature relationship (800°C - 20 min.) during reactivation
can be developed that will not alter activated carbon significantly from its
virgin state. In support of this, Weissenhorn collected data on the decline of
granular activated carbon performance as measured by ultraviolet absorbance of
the effluent with relation to the volume of water treated per unit volume of
granular activated carbon and showed, according to Figure 12, that virgin activated
and reactivated carbon were nearly equivalent. Finally, the Shirco Company of
Dallas, Texas, claims to have collected data as shown in Figure 13 and Table IV.
(unpublished promotional literature).
TABLE IV
TYPICAL RESULTS OF CARBON REACTIVATION IN THE SHIRCO ELECTRIC INFRARED FURNACE
Apparent
Activated Density Iodine Molasses
Carbon (g/cc) Number Number
ICI HYDRODARCO 3000
Virgin
Spent
Reactivated
CALGON FILTRASORB 400
Virgin
Spent
Reactivated
WESTVACO NUCHAR WV-L
Virgin
Spent
Reactivated
CARBORUNDUM
Virgin
Spent
Reactivated
0.352
0.458
0.397
0.48
0.62
0.48
0.594
0.678
0.593
0.525
0.667
0.523
550
413
596
1167
480
1122
834
588
868
950
320
1132
333
302
380
355
250
376
371
178
380
220
137
230
-------
-A30-
Z 30
UJ
o
oc
ui 20
UJ
2-10
01
z g
5 u
12
< H -20
Q UJ
"g
gS-40
t=20 min_
t=15 min
2 4 6 8 10 12
REACTIVATION CYCLE
14
T=800 °C
t=15 (min)
t=20 (min)
fe30
in)
2 4 6 8 10 12 14
REACTIVATION CYCLE
FIGURE 11. EFFECT OF RESIDENCE TIME
ON ACTIVITY AND CHANGE OF
WEIGHT OF ACTIVATED CARBON
REFERENCE 59
-------
-A31-
100
50
a
cc
o
(A
O
3 20
ACTIVATED
CARBON "F"
REACTIVATED
VIRGIN
I
10
024 6 8 10 12
THROUGHPUT, CUBIC METERS /LITER
OF ACTIVATED CARBON
100
A
3? 50
a.
cc
O
(A
O
3 20
ACTIVATED
CARBON "L"
VIRGIN
REACTIVATED
10
024 6 8 10 12
THROUGHPUT, CUBIC METERS/LITER
OF ACTIVATED CARBON
FIGURE 12 INFLUENCE OF REACTIVATION ON
GRANULAR ACTIVATED CARBON
PERFORMANCE
REFERENCE 60
-------
-A3 2-
DC
UJ
100 -I
± 70^
O
O
EFFECT OF REPEATED REACTIVATION BY
THE SHIRCO PROCESS
VIRGIN ACTIVATED CARBON (MINIMUM)
S: SPENT ACTIVATED CARBON O
R: REACTIVATED CARBON •
VIRGIN S-1 R-1 S-2 R-2 S-3 R-3
FIGURE 13
INFLUENCE OF REACTIVATION ON
IODINE NUMBER
FROM SHIRCO, INC. PROMOTIONAL
LITERATURE
-------
-A3 3-
In summary, some changes may occur in granular activated carbon properties
during thermal reactivation, and any reactivation system should be carefully
designed to minimize the changes. Because some granular activated carbon losses
occur through burn-off and mechanical attrition, fresh granular activated carbon
will be added during each operating cycle. The addition of this fresh material
will help overcome any losses in performance of the reactivated granular carbon,
as compared to virgin material.
A final design consideration for any reactivation system must include proper
consideration of handling the off-gases. Both dust collecting devices and gas
after-burners may have to be considered in certain circumstances. Reactivation
systems, however, can be designed in order to avoid any possible problems of air
pollution during the reactivation process.
Because, as discused above, most water utilities using granular activated
carbon as part of their water treatment process are doing so to control tastes
and odors, and because tastes and odors are removed by granular activated carbon
adsorption beds for periods of several years, reactivation of granular activated
carbon for drinking water purposes is not widespread. In most cases, when a
water purveyor changes the granular activated carbon charge in its treatment
plant, the material that is replaced in the beds is virgin rather than
reactivated granular carbon. The only known on-site reactivation facility
at a water treatment plant in the Unted States was at Nitro, West Virginia,
a plant that has now been closed for several years because the utility
built a new plant on a cleaner source of water. At this writing (Fall
1977) the five known on-site reactivation facilities at water treatment
plants are all in Europe, one in Switzerland, two in the
-------
-A3 4-
Federal Republic of Germany, one in Sweden, and one in the United Kingdom.
If granular activated carbon adsorption systems are to be used to control
other organic contaminants beyond taste and odor producing compounds, as noted in
the literature review above and as will be discussed in the sub-sections below,
the period between reactivations will be much shorter than currently practiced.
Current Water Supply Research Division Research Findings
In an attempt to provide additional information not currently in the literature
on the performance of granular activated carbon as an adsorption medium, the
Water Supply Research Division has an active in-house and extramural research
program on this subject The results of these experiments, many still on-going,
are summarized below. Detailed papers on many of these projects will eventually
be published in the technical press.
Class I Compounds (Taste and Odor Producing)
Snoeyink at the University of Illinois*working under EPA sponsorship found
that the odorous compounds 2-methylisoborneol (MIB) and geosmin are both strongly
f\~\ f\ 9
adsorbed by activated carbon. ' When present, humic substances significantly
reduce the capacity of activated carbon for adsorption of these compounds, more
so before equilbrium is achieved than at equilibrium. Commmercial humic acid
(HA) and the humic substances from well water each had differing competitive
effects on MIB. The capacity of activated carbon for geosmin adsorption was
reduced by commercial HA to a greater extent than was observed for MIB. The
performance of laboratory columns was consistent with the isotherm results.
Application of distilled water to a partially saturated activated carbon bed
resulted in almost no elution of MIB indicating that it was strongly adsorbed.
-------
-A3 5-
Using the data collected in this study and assuming, 1) complete saturation
of the activated carbon, 2) no desorption and 3) no biological activity, Snoeyink
predicted the bed life for the reduction of MIB or geosmin from 10 yg/& to its
threshold odor level of 0.1 yg/£ in a 7-8 minute empty bed contact time bed to
be much greater (several months to years) than the predicted life for the reduction
of humic substances from 5 to 1 mg/Jl (1 to 2 months). When both MIB, or geosmin,
and humic substances must be removed, humic substance removal will control the
life of the bed.
Chlorophenols are adsorbed very strongly by activated carbon at the yg/£
level, which is near the threshold odor limit for these compounds. The extent of
adsorption of 2,4-dichlorophenol (DCP) and 2,4,6-trichlorophenol (TCP) is a
function of pH. The neutral species of these compounds predominate at pHs below
the pK (pH at which the concentrations of the free acid and the acid anion are
3.
equal) values (7.85 and 6.00, respectively, at 25°C) and are adsorbed more strongly
than the anionic species. As the number of chlorine atoms substituted on the
phenol increases, the solubility of the neutral species decreases and the adsorbability
increases. As substitution increases, the pK of the species is lowered, however.
a
When water containing phenol is chlorinated with low levels of chlorine, a
mixture of chlorophenols will form and thus the extent of adsorption of one
chlorophenol in the presence of another chlorophenol is an important consideration.
Significant reductions in adsorption capacity (up to 50 percent) of one chlorophenol
was caused by the presence of a second chlorophenol. Evaluation of the competitive
effects of commercial HA, soil fulvic acid (FA) and leaf FA showed that the
presence of these materials decreased the capacity of activated carbon for chlorophenol
adsorption and that each of the materials competed somewhat differently.
-------
-A3 6-
Even in the presence of humic substances and another chlorophenol species,
however, the adsorption capacity is even greater for chlorophenol than it is for
MIB and that bed life for chlorophenol adsorption will be greater than for MIB
and much greater than for humic substances.
In another study, a joint effort between the AWWA Research Foundation, 14
water utilities, the EPA Water Supply Research Division, and the Universities of
Illinois, Iowa State and Missouri-Columbia, is underway to determine the removal
of trace organics (particularly taste and odor compounds and haloforms) on granular
activated carbon and polymeric adsorbents.
In this study, pilot scale columns containing seven different adsorbents are
located at the Kansas City Water Treatment Plant, Kansas City, Missouri. The
adsorbents include granular activated carbon made from bituminous coal, lignite,
petroleum, and peat, and carbonaceous anion exchange resins. The adsorbers are
presently arranged so that both a coal-base-and a lignite-base granular activated
carbon system can provide up to 33 minutes empty bed contact time (EBCT). The
applied water is unstable because of precipitative lime softening and a lack of
recarbonation, resulting in calcium carbonate deposition on the adsorbents. The
effect of this will be evaluated. Work will continue on the pilot column studies
until the spring of 1978 and the final report will be available by early summer.
Taste and odor removal results are not available at this time (Fall 1977).
Class II Compounds (Synthetic Source Water Contaminants)
Naphthalene
A long-term (started July 7, 1974) experiment comparing the adsorption of
naphthalene with that of background organic content of Cincinnati tap water on
granular activated carbon beds was terminated during June 1975.
-------
-A3 7-
330-day span of the experiment, 18,500 liters of Cincinnati tap water spiked with
an average concentration of 30 yg/£ of naphthalene were passed through the granular
activated carbon column (16.9 min EBCT). Although the non-purgeable organic
carbon (NPOC) 50 percent removal front had penetrated the entire 27-inch length
of the column by May 1975, the maximum penetration of the naphthalene 50 percent
removal front was only about 6.5 inches (this was quite variable throughout the
experiment). Therefore granular activated carbon columns can be expected to
remove compounds of low polarity and solubility such as naphthalene for a much
longer period of time than they can remove the diverse organic group represented
by NPOC.
The variability of the depth of the naphthalene penetration into the adsorbent
(1-6.5 inches) is something of a mystery. Suspected causes are (1) variable
influent concentrations, (2) variable constitution of NPOC (competitive adsorption),
(3) biological activity, (4) variable flow, (5) variable temperature, or (6)
influence of backwashing.
Other Polynuclear Aromatic Hydrocarbons
r n
Snoeyink, at the University of Illinois, reported that limited
experimentation with the polynuclear aromatic hydrocarbon (PAH) anthracene led
to the conclusion that no significant association between its adsorption and that
of humic substances occurred. Thus the possibility of PAH passage through
granular activated carbon beds because of its association with the less adsorbable
humic substances is not a cause for concern.
Carbon Tetrachloride
The adsorption of this contaminant on granular activated carbon beds was
observed in the U.S. EPA Cincinnati Laboratory during the fall and winter of
1976-77 when carbon tetrachloride was in the tap water. At that time, granular
activated carbon columns that had been in service for two months were being
monitored for trihalomethane reduction. Table V shows the monthly average carbon
-------
-A38-
tetrachloride concentrations in the tap water and corresponding monthly average
concentrations in the effluent. In spite of being in service for two months, the
granular activated carbon was effective for 3 to 5 months, probably because this
contaminant was "new" to the adsorbent when it occurred during the third month.
Low level desorption of the carbon tetrachloride continued for several months
after the contaminant disappeared from the influent.
TABLE V
ADSORPTION OF CARBON TETRACHLORIDE ON GRANULAR ACTIVATED CARBON
Time in Service, months
2 3 4 56789 10 11
Monthly Average
Influent,
Monthly Average
Effluent, yg/£
NF
NF
NF 5
13 15 44 NF NF NF NF
NF NF NF NF <6* 14 16 13
Filtrasorb 400 granular
activated carbon
10 min. Empty bed contact time
NF = none found
* - Single Value during month
During the first week of January 1977 a 6-inch diameter by 30-inch
depth bed (11 minutes EBCT) of virgin Calgon Filtrasorb 400 granular
activated carbon was placed in upflow operation treating Cincinnati tap water,
Coincidently with the start of operation, the concentration of carbon
tetrachloride in the tap water began to increase.
Figure 14 shows the performance of the adsorbent bed during the first
43 weeks of 1977. During the first 7 weeks, the granular activated carbon
effectively removed carbon tetrachloride during the period of extremely high
influent concentrations. As the influent concentration returned to the
limit of detection (0.1 yg/£), the adsorbent, acting under the influence
of the adsorption-desorption equilibrium phenomenon, began to desorb carbon
-------
-A39-
O)
oc
LU
O
O
o
oc
O
oc
LU
O
m
oc
<
o
60
50
40
30
20
10
UPFLOW OPERATION
11 MINUTES EBCT
INFLUENT (CINCINNATI TAP WATER)
^ GRANULAR ACTIVATED
CARBON COLUMN EFFLUENT
0 5 10 15 20 25 30 35 40 45 50
TIME IN SERVICE, WEEKS
FIGURE 14 PERFORMANCE OF GRANULAR ACTIVATED
CARBON BED FOR CARBON TETRACHLORIDE
REMOVAL
-------
-A40-
tetrachloride. This desorption, which was monitored approximately every 6 weeks,
has continued for about 9 months.
Figure 14 is important because it illustrates the protection a granular
activated carbon adsorption barrier can afford the consumer during periods of
sudden contamination of the raw water supply. These data also demonstrate that
granular activated carbon reactivation will probably be necessary fairly frequently
to afford continuing consumer protection and guard against desorption of unwanted
contaminants.
Trichloroethylene, Tetrachloroethylene and Similar Chlorinated
Organics
Because several incidences of tri- and tetrachlorethylene contamination of
drinking water supplies have been reported, the Water Supply Research Division is
currently (Fall 1977) studying the effectiveness of granular activated carbon
beds for removing these pollutants. In most instances the water source is ground
water and the contamination can be linked to some industrial activity involving
the present or past use of cleaning solvents in the aquifer recharge area.
Columns containing granular activated carbon (10 minute empty bed contact time)
have been installed in three locations in New England and the first few weeks of
data indicate that tri- and tetrachloroethylene are very well removed when the
adsorbent is fresh, see Table VI. As indicated in the footnote to Table VI, the
contaminant, 1,1-dichloroethane was not as well removed. At another location
(not shown in Table VI) the tetrachloroethylene contamination has exceeded 2500
Pg/&> Yet the granular activated carbon bed produced an effluent with less than
0.5 yg/£ tetrachloroethylene, for 15 weeks. These tests are continuing.
-------
-A41-
TABLE VI
USE OF GRANULAR ACTIVATED CARBON BEDS TO REMOVE ORGANIC CONTAMINANTS FROM
A GROUND WATER
Contaminant
Influent
Concentration
Range, yg/£
Average Effluent Con., yg/£
0-4 weeks 4-8 weeks 8-12 weeks
1,1,1-trichloroethane 10.9-36.7
trichloroethylene 1.9-7.7
tetrachloroethylene 0.1-16.9
NF = None Found
NF
NF
NF
NF
0.4
Note: After six weeks of testing, two new contaminants, 1,1-dichloroethane
and cis-l,2-dichloroethylene began appearing in the influent to the granular
activated carbon column. No cis-1 , 2-dichloroethylene has yet been detected
in the treated water, but the average concentration of 1,1-dichloroethane
in the effluent for the 4 to 8 week period after first appearing was 0.7
Prediction of Granular Activated Carbon Bed Performance, Studies with
Humic Acid and Carbon Tetrachloride, Dieldrin, PCB, Benzene, and
p-Dichlorobenzene
Weber, at the University of Michigan, is studying the adsorption of
several specific organic compounds on granular activated carbon. The main thrust of
this WSRD grant activity is to investigate the possibility of predicting activated
carbon column adsorber performance on the basis of mathematical parameters
obtained from simple batch kinetic and equilibrium (isotherm) studies. Although
the studies include only the use of model systems (humic acids as the
competitive species) , certain general conclusions about relative activated
carbon adsorber operating life can be drawn on the basis of some of the data
obtained to date from the modeling studies.
The results of the work with carbon tetrachloride support the in-house work
reported above in that an adsorber life of several weeks to months can be expected.
The work with dieldrin indicates that this compound is very strongly adsorbed,
supporting evidence to that effect previously reported in the literature. ^3
PCB mixtures resemble pesticides in properties and are also strongly adsorbed.
-------
-A42-
Adsorber operating life for PCB's are likely to be on the order of months to one
year. Adsorber operating periods for treatment of benzene and jD-dichlorobenzene
are likely to range somewhere between carbon tetrachloride and the PCS mixtures.
Mixtures from Actual Waters
In an attempt to further evaluate the capability of granular activated
carbon beds to remove actual mixtures of raw water organic contaminants in competition
with other organics, several studies are being conducted on actual waters at
locations other than Cincinnati, Ohio. One of these research projects, using
bench scale pilot granular activated carbon columns is being performed with a
ground water in Southern Florida. The granular activated carbon columns are 1-
inch in diameter and contain 30-inches of activated carbon. The bed was loaded
2
at 3 gpm/ft , which resulted in a 6.2 minute empty bed contact time. The influent
to the granular activated carbon column was untreated ground water for the first
phase of the study and finished water for the second and third phase. The source
has a color of approximately 50 color units, a TOC concentration of about 10
mg/£, and a pH of about 7.1.
The project includes routine analysis for the concentration of 19 organic
compounds, total organic carbon, and trihalomethane formation potential, both
before and after the granular activated carbon column. The average influent
concentration to the granular activated carbon bed for the duration of the study
is shown in Tables VII, VIII, and IX. Tables X, XI, and XII show the corresponding
effluent concentrations for specific substances with time over the duration of
each study. These data show that at these short contact times, some organic
compounds did break through fairly early in the test, but that the chlorinated
aromatic compounds were well removed. Longer empty bed contact time tests are
now underway, see pages A69 to A78.
-------
-A43-
TABLE VII
AVERAGE CONCENTRATIONS OF TENTATIVELY IDENTIFIED ORGANICS INFLUENT TO
GRANULAR ACTIVATED CARBON BED
(Jan. 18 - May 20, 1977)
Compound yg/£
Vinyl chloride 0.8
Methylene chloride 0.08
Methyl iodide nil
trans-1,2-Dichloroethylene 1.3
1,1-Dichloroethane 0.3
cis-1,2-Dichloroethylene 29.0
Chloroform (Tr. to 2.1) 0.16
1,1,1-Trichloroethane
1,2-Dichloroethane ^(Sum) 0.11
Carbon tetrachloride
Trichloroethylene 0.13
Bromodichloromethane (Tr to 0.9) 0.11
Tetrachloroethylene 0.06
Dibromochloromethane (Tr. to 0.4) 0.04
Chlorobenzene 0.19
Bromoform (Tr. to 0.19) 0.02
p-Chlorotoluene 0.11
o,m,p-Dichlorobenzene (sum) 1.1
-------
-A44-
TABLE VIII
AVERAGE CONCENTRATION OF TENTATIVELY IDENTIFIED ORGANIC COMPOUNDS
IN INFLUENT TO GRANULAR ACTIVATED CARBON BED (Aug. 26 to Oct. 18, 1977)
EBCT =6.2 minutes
Chemical Name
Vinyl chloride 5.4
Methylene chloride not determined
trans-1,2-Dichloroethylene 1.0
1,1-Dichloroethane 0.3
cis-1,2-Dichloroethylene 19
Chloroform 57
1,1,1-Trichloroethane
1,2-Dichloroethane -(Sum) 5.3
Carbon tetrachloride
Trichloroethylene 0.1
Bromodichloromethane 39
Tetrachloroethylene nil
Dibromochloromethane 27
Chlorobenzene 0.8
Bromoform 2.5
£-Chlorotoluene 0.2
m-Dichlorobenzene
£-Dichlorobenzene—— (Sum) 0.3
o-Dichlorobenzene
-------
-A45-
TABLE IX
AVERAGE CONCENTRATIONS OF TENTATIVELY IDENTIFIED ORGANIC COMPOUNDS IN INFLUENT
TO GRANULAR ACTIVATED CARBON BED, EBCT =6.2 min.
(Run No. 2, Nov. I, 1977 - Jan. 3, 1978)
Compound yg/£
Vinyl chloride 7.4
Methylene chloride
trans-1,2-Dichloroethylene 0.63
1,1-Dichloroethane 0.11
cis-l,2-Dichloroethylene 24.5
Chloroform 76.6
1,1,1-Trichloroethane
1,2-Dichloroe thane ^(sum) 7.9
Carbon tetrachloride
Trichloroethylene 0.52
Bromodichloromethane 48.4
Tetrachloroethylene nil
Dibromochloromethane 29.0
Chlorobenzene 0.56
Bromoform 1.95
£-Chlorotoluene nil
m-Dichlorobenzene nil
p-Dichlorobenzene 0.19
o-Dichlorobenzene 0.09
-------
-A46-
TABLE X
CONCENTRATIONS OF TENTATIVELY IDENTIFIED ORGANICS PRESENT IN EFFFLUENT
FROM ADSORBER, EBCT = 6.2 min. (Jan. 18 - May 20, 1977)
All data in yg/£ - Compare to Table VII
Time,
Weeks
0
1
2
3
4
5
6
7
8
9
10
11
12
13+
14
15
16
17
0
0
0
0
0
0
0
0
0
0
0
0
0
A
N
N
N
.002
N
.002
N
.002
.04
.02
.05
.07
.07
.66
.45
.90
.76
.60
B
N
N
N
0.02
0.08
0.007
0.013
0.015
0.02
0.03
0.04
0.1
0.02
0.32
0.17
0.38
0.34
0.23
C
N
0.18
0.2
2.7
4.2
7.8
5.3
12.0
18.0
17.3
21.6
10.2
18.7
25.7
30.4
23.3
21.9
21.3
D
N
N
N
0.002
N
N
N
N
N
N
N
0.01
0.006
0.06
0.07
0.05
0.06
0.028
E
0.007
0.006
0.004
0.001
0.002
N
N
N
N
N
N
N
N
N
N
N
N
N
F
N
N
N
0.01
N
N
N
N
0.001
N
N
N
N
N
N
N
N
N
G
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
H
0.02
N
0.02
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
I J
0.
0.
N
0.
N
0.
N
0.
0.
0.
0.
N
0.
0.79 0.
0.70 0.
0.81 0.
0.75 0.
0.78
07
003
003
006
002
003
003
002
03
01
10
06
05
N
- = Not determined
N =
A =
B =
C =
None found
trans-1 , 2-Dichloroethylene
1 , 1-Dichloroethane
cis-1,
2-Dichloroethylene
D = Trichloroethylene
E = Tetrachloroethylene
F = Chlorobenzene
G = p-Chlorotoluene
H = m,p,o-dichlorobenzene
I = Vinyl chloride
J = fl,l,l-Trichloroethane
|l,2-Dichloroethane
^Carbon tetrachloride
-------
-A47-
TABLE XI
CONCENTRATIONS OF TENTATIVELY IDENTIFIED ORGANICS PRESENT IN EFFLUENT FROM ADSORBER
EBCT = 6.2 min., (Run No. 1, Aug. 26 - October 18, 1977)
All Data in \ig/i - Compare to Table VIII
Time
Weeks A
0
1
2
3
4
5
6
7
N =
A =
B =
C =
N
N
0.
N
N
0.
N
N
B C D E
N N N N
N N N N
40 N 10.7 0.59 N
0.55 2.19 0.03 N
N 2.87 N N
12 0.62 8.9 0.003 N
N 8.77 N N
N 9.91 N N
F G H I J
N N N - N
N N N 0.165 N
N N N 0.64 N
N N N 0.33 N
N N N N 0.31
N N N 4.7 0.34
N N N 3.25 0.09
N N N 1.85 N
None found
trans-1,
2-Dichloroethylene
1 , 1-Dichloroethane
cis-1 , 2-Dichloroethylene
D = Trichloroethylene
E = Tetrachloroethylene
F = Chlorobenzene
G = _p_-Chlorotoluene
H = m,p,o-Dichlorobenzene
I = Vinyl chloride
J =p.,l,l-Trichloroethane
<1,2-Dichloroethane
(Carbon tetrachloride
-------
-A48-
TABLE XII
CONCENTRATIONS OF TENTATIVELY IDENTIFIED ORGANICS IN EFFLUENT FROM ADSORBER
EBCT = 6.2 min. (Run No. 2, Nov. 1, 1977 - Jan. 3, 1978)
All data in yg/£, Compare to Table IX
Time
Weeks A
0
1
2
3
4
5
6
7
8
9
N =
A =
B =
C =
N
N
N
N
N
N
N
N
0.58
0.97
None found
B C
N N
N N
N N
1.05 1.28
0.21 3.8
N 11.1
0.30 5.5
0.8 12.0
0.66 9.1
N 9.1
D
N
N
N
N
N
N
0.22
0.29
0.57
0.04
E
N
N
N
N
N
N
N
N
N
N
F
N
N
N
N
N
N
N
N
N
N
G
N
N
N
N
N
N
N
N
N
N
H
N
N
N
N
N
N
N
N
N
N
I
0.30
N
N
N
N
N
N
N
N
N
J
N
N
N
N
N
N
N
N
N
N
K L
N N
3.2 1.2*
8.2 N
N N
5.2 N
N N
1.7 l.f
1.8 4.9
2.5 4.C
4.4 4.7
trans-1 , 2-Dichloroethylene
1 , 1-Dichloroethane
cis-1, 2-Dichloroethylene
D = Trichloroethylene
E = Tetrachloroethylene
F = Chlorobenzene
G = £-Chlorotoluene ?•
H = m-dichlorobenzene
I = _p_-Dichlorobenzene
J = jo-Dichlorobenzene
K = Vinyl chloride
L = fl,l,l-Trichloroethane
t1,2-Dichloroethane
(Carbon tetrachloride
-------
-A49-
Another actual plant site research project is being performed using full-
scale granular activated carbon beds located at a water treatment plant in
the lower Mississippi Valley. In this project one bed is used as a post-filter
adsorber and another as a sand replacement system in parallel operation. The
post-filter adsorber is an existing conventional rapid sand filter that had
the 30 inches of sand replaced by 30 inches of 12x40 mesh granular activated
carbon. The post-filter adsorber is in series after a rapid sand filter,
so that the granular activated carbon received coagulated, settled, softened, and
filtered water. The empty bed contact time was 27 minutes at the start of
the test and 20 minutes at the end because of granular activated carbon
loss, caused by inadvertent excessive backwashing.
The sand replacement system is an existing rapid sand filter that had
the top 24 inches of the 30 inches of sand replaced by 24 inches of 12x40 granular
activated carbon. This unit receives coagulated, settled and softened water
directly from the precipitator of the full-scale plant. The empty bed contact
time was 24 minutes at the start of the test and 18 minutes at the end because
of granular activated carbon loss caused by inadvertent excessive backwashing.
The project includes routine analysis for the concentration of at least
35 organic substances and total organic carbon and trihalomethane formation
potential both before and after the granular activated carbon beds. Table XIII
shows the effluent concentration for each substance found, as contrasted to
the influent concentrations listed at the bottom of the Table for the duration
of the study, 25+ weeks. For this report only the post-filter adsorber data
will be shown as an example. Note, for this project, too, all data is
preliminary in nature as the gas chromatograph-mass spectrometry confirmations
have not yet been evaluated for consistency of identification (Fall 1977).
-------
-A50-
TABLE XIII
REMOVALS OF TENTATIVELY IDENTIFIED ORGANICS BY POST-FILTER
GRANULAR ACTIVATED CARBON ADSORBER
Empty Bed Contact Time, 27 min. Start, 20 min. End
Time,
Weeks
0
1
2+
3
4
5
6
7
8
9
10+
11+
12+
13+
15+
20+
25+
Benzene
Effl. Cone.
ygM
ND
0.1
ND
ND
ND
0.1
ND
ND
ND
ND
ND
0.1
ND
ND
ND
ND
ND
1 , 2-Dichloroethane
Effl. Cone.
ygM
0.4
0.2
ND
0.1
ND
0.3
1.1
0.7
0.8
1.3
1.4
1.0
2.0
3.5
4.5
10.1
9.4
Trichloroethylene
Effl. Cone.
vg/A
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.4
ND
ND
ND
ND
ND
ND
ND
Toluene
Effl. Cone.
yg/£
ND
0.1
ND
ND
ND
0.1
ND
ND
ND
ND
ND
0.2
ND
ND
ND
ND
ND
Avg. Inf.
Cone. yg/£ 0.3 8.0
Range Inf.
Cone. yg/£ ND-2.1 1.2-23.7
0.3
ND-0.9
0.2
ND-0.9
ND
Not detected
-------
-A51-
Within the precautions stated, Table XIII can be used to show the
trends for removal by the granular activated carbon bed. The adsorber
removed the ambient concentrations of toluene, benzene and trichloroethylene
more efficiently than other substances consistently present during the
first 10+ weeks of operation. The 1,2-dichloroethane was removed at an
average of greater than 80 percent for about 8 weeks, but the effluent
concentration consistently exceeded 0.5 yg/£ after six weeks. Desorption
occurred from about the 16th week to the end of the test. Preliminary
data show that trace concentrations (ng/£) of chlorinated hydrocarbon insecticides
were generally reduced to below detectable concentrations throughout the 25+
week study. Finally, the sand replacement system performed similarly to the
post-filter adsorber, but the removals were not as long-lasting.
A third study on the performance of granular activated carbon beds is
being conducted in full scale at a water treatment plant in the upper Ohio
Valley. As this project is just starting, the data are preliminary, mainly
because the identification of specific organics thus far is based only on
gas chromatographic retention times without mass spectrometry confirmation.
The following generalizations are made after 52 days of virgin granular
activated carbon being on-line. Compounds for which granular activated
carbon adsorption may serve as a control mechanism and their influent concentration
ranges are: 1,1,1-trichloroethane or carbon tetrachloride or both (0.17 -
1.14 yg/£, not resolved); 1,2-dichloropropane or trans-1,3-dichloropropylene
or both (0.10-0.64 yg/£) (not resolved); trichloroethylene (0.16-0.96 yg/£);
cis-1,3-dichloropropylene or 1,1,2-trichloroethane or both (0.15-24.0 yg/£)(not
resolved); dichloroiodomethane (0.10-1.22 yg/£) and chlorobenzene (0.24-3.36
yg/£). Although present occasionally, these compounds were not consistently
present in the influent to the adsorber, therefore removals could not be calculated.
-------
-A52-
Unidentifled Compounds
After concentration, gas chromatography can be used to separate many organic
compounds, producing a gas chromatogram in which the separated organic compounds
are represented by "peaks" on a chart. Although unidentified, the absence of
certain peaks after a given type of treatment gives an indication of the success
Q
of the treatment. In 1972, influent and effluent carbon chloroform extracts
from an operating granular activated carbon sand replacement system in Lawrence,
Massachusetts were compared gas chromatographically. Figure 15 shows the
reduction of many of the organic peaks when the granular activated carbon was
fresh, but after 16 weeks of operation the influent and effluent gas chromatograms
were similar, Figure 16. Note, Figure 3, page A8, shows the breakthrough pattern
of the general organic parameter CCE-m from this same treatment unit.
Recently, in the WSRD laboratories this approach was repeated using
improved analytic procedures. Weekly one gallon samples of influent and effluent
from a virgin granular activated carbon bed (about 9 minute EBCT), receiving
coagulated, settled, and dual media filtered Ohio River water, were collected
and extracted with one 250 ml and two 100 ml portions of redistilled methylene
chloride. After concentration of the extract 2 yg of anthracene, in methylene
chloride was added as an internal standard.
In-house analyses are being conducted on a gas chromatograph employing a
30 meter SP-2100 wall coated glass capillary column. Injections are made at
20°C, and after 5 minutes the oven temperature is programmed at 2°/min to
240°C. Detection limits for anthracene (internal standard) using a flame ionization
detector is approximately 0.4 ng or approximately 20 ng/£ for similarly responding
compounds from the original 4£ sample. Preliminary results, Figure 17, indicate
that after one week almost all of the compounds present in the influent were
not detectable in the effluent. Further, no major new peaks occurred, indicating
that organics detectable by this procedure were not leaching off the granular
-------
-A53-
-------
-A54-
* I
< O
LU _
Q. LU
u
10 >
-------
-A55-
UJ
>
O
CO
o
fiUj UJ
Si
si
U2
o o
iu o
ii
-------
-A56-
activated carbon. Analysis of later samples showed; 1) that the influent
quality was very variable and 2) that although some "peaks" began occurring
in the effluent it always contained fewer "peaks" than the influent, even
after 3 months. These studies are continuing. Figure 18 shows the increase
in effluent NPOC concentration during the time of the collection of the
9
above samples. Similar data have been collected by Heymann in Duisburg,
Federal Republic of Germany.
Note, because the influent to the pilot plant (Ohio River water) is
stored in open tanks for several days prior to treatment, relatively volatile
chlorinated organics were not present in the influent to the granular activated
carbon bed.
Summary
Taken in total, these results, which were collected under conditions
representative of those found in drinking water, support the concept of
granular activated carbon adsorption being a "broad-spectrum" organic removal
unit process. These data do, however, indicate the variation adsorbability
of different organic compounds and show the importance of site specific
evaluations of granular activated carbon performance as outlined in the
Interim Treatment Guide (see pages 23 to 26).
Class III Compounds (Disinfection By-Product Precursors)
Because no direct measurement exists for Class III organics (for example,
trihalomethane precursors), the degree of precursor removal can be judged by
comparing trihalomethane concentrations upon chlorination of an untreated
/TO
control water (called the terminal trihalomethane concentration) to similar
data collected on a treated water after similar chlorination and storage.
For example, if the effluent from a sand filter that was chlorinated and
stored for two days yielded 50 yg/£ chloroform and the same effluent passed
through an adsorbent, then chlorinated and stored under similar conditions
produced 25 yg/£ chloroform, the adsorbent would be 50 percent effective in
removing chloroform formation potential. This example assumes that no
-------
-A57-
o
z
<
o
DC
o
LU
_J
OQ
<
LU
O
cc
O
Z
u_
o
z
o
o
DC
< CD
HI DC
CC <
00
111
DC
O
I/Bui 'NOIlVdlN30NOO NOSUV3 3INV9UO 319VOand-NON
-------
-A58-
trihalomethanes were present in the filter effluents, that is the terminal
trihalomethane concentration and trihalomethane formation potential are
equal.
64
Using Ohio River water as source, Love et al. observed that the
relative effectiveness of granular activated carbon adsorbers to prevent the
formation of trihalomethanes was highest for chloroform and lowest for
dibromochlormethane. A granular activated carbon bed receiving coagulated,
settled, but undisinfected water was initially effective for preventing
trihalomethane formation upon subsequent chlorination (see Figure 19).
Similar pilot plant studies have been carried out by Sylvia at the
Lawrence Experiment station in Lawrence, Massachusetts under a WSRD research
Contract. The Merrimac River receives considerable industrial contamination
upstream from the Lawrence study site, however, relative to the Ohio River,
the Merrimac has a very low turbidity (2-4 NTU vs 10-75+ NTU, for the Ohio
River at Cincinnati), and experimentation has shown that trihalomethane
precursors in the Merrimac are only slightly reduced in concentration through
coagulation and settling, yet are removed for long periods by granular
activated carbon adsorption. For example, three different types of granular
activated carbons exposed to treated yet undisinfected Merrimac River water
were found to reduce the trihalomethane formation potential (2 day) of 40-60
yg/£ to less than 1 yg/& after 6 months of operation. Details on this will
be contained in the final project report, which is due in early 1978.
f O
Work by Snoeyink at the University of Illinois on the adsorption of
humic and fulvic acids, major trihalomethane precursors, relates to this phase
of the overall problem of organic contamination. He found that activated carbon
adsorbed humic substances in all cases that were studied, but the adsorption
properties of the substances from different sources varied widely as did the extent
-------
-A59-
100
- 75
O)
=5.
w 50
oc
LU
25
100
o> 75
3.
50
cc
LJ
H
25
4 DAY
25°C
INFLUENT
SAND REPLACEMENT ADSORBER EFFLUENT
(F-200 EBCT=10 MIN.)
10 15 20
TIME IN OPERATION, WEEKS
25
30
_
o-a~—
/ SAND REPLACEMENT
ADSORBER EFFLUENT
(HD 10x30 EBCT=20 MIN.)
i _ i _ i _
25
30
0 5 10 15 20
TIME IN OPERATION, WEEKS
FIGURE 19 USE OF GRANULAR ACTIVATED ACTIVATED CARBON
FOR REMOVING TRIHALOMETHANE FORMATION
POTENTIAL. REFERENCE 64
-------
-A60-
to which they competed with selected trace organics for adsorption sites on
activated carbon.
Humic substances from leaf- and soil extract, a well water, and a commercial
source were examined in detail. Extent of adsorption depended upon solubility,
with the less soluble humic acid (HA) fraction being more adsorbable than the
fulvic acid (FA) fraction from the same source. - The lower molecular weight
species from a given FA or HA fraction are more adsorbable than the high molecular
weight species, presumably because more surface area is accessible to them.
The adsorption characteristics of the humic substances are also dependent
on the method of analysis used to quantify them. The species that fluoresce the
most were found to be the lower molecular weight species and these adsorb best.
Ultra-violet absorbing species did not adsorb as well as those that fluoresce.
Solution pH and phosphate concentration also had a marked effect on adsorbability
of the humic materials, with adsorption generally improving with decreasing pH
and increasing phosphate concentration. The trihalomethane formation potential
of the humic substances varied widely from source to source, with only one
exception, but no dependence on molecular weight was found for fractions of FA
or HA. This work reemphasizes the need for on-site pilot studies to determine
adsorbability for that particular location.
The removal of disinfection (chlorination) by-product precursors was also
studied at the three projects operating at the actual water treatment plants
cited above. The data in Table XIV and Figure 20 show the same variability in
the treatability of trihalomethane precursors as noted in the previous studies.
Note: The data in Table XIV show that although breakthrough did occur, complete
exhaustion for the removal of trihalomethane formation potential did not, at least
during the time of the study. In summary, removal of trihalomethane precursors
by adsorption on granular activated carbon beds is variable and site specific.
-------
-A61-
TABLE XIV
REMOVAL OF SUMMATION** TRIHALOMETHANE FORMATION POTENTIAL (THMFP)
BY GRANULAR ACTIVATED CARBON BEDS
Southern Florida Ground Water
6.2 min. EBCT
Inf. Eff. Percent
Time THMFP THMFP Reduction
Weeks yg/& yg/&
Lower Mississippi River Water
EBCT-28 min. start, 20 min. end
Inf. Eff. Percent
Time THMFP* THMFP* Reduction
Weeks yg/£ yg/£
0
1
2
3
4
5
6
7
8
9
10
11
12
17+
384
485
758
878
807
733
710
936
617
639
616
348
655
575
11
335
558
560
579
557
524
437
342
424
445
397
452
394
97
34
28
37
30
27
29
53
45
34
39
-17
33
31
0
1
2+
3
4
5
6
7
8
9
10+
11+
12+
15+
20+
25+
59
56
70
113
104
203
239
151
159
217
184
240
236
205
297
218
2.3
1.3
6.7
22
18
25
44
49
92
86
73
100
89
93
148
157
96
98
91
80
82
87
81
68
42
60
60
58
62
54
50
28
*Note: These data are shown in Figure 11, page 20, of the Interim Treatment Guide.
**The arithmetic sum of the individual trihalomethane species determined.
This parameter is called "Total" Trihalomethane Formation Potential in the
Regulations.
-------
-A62-
o>
5.
<
CC
LU
o
z
o
o
HI
h-
o
Q.
cc
o
LL.
UJ
UJ
2
O
Z
O
CO
260
240
220
200
180
160
140
120
100
80
60
40
20
I-
A
WVW 14x40 GRANULAR
ACTIVATED CARBON
Upper Ohio Valley
Water Treatment
Plant
SAND REPLACEMENT
ADSORBER
INFLUENT
SAND REPLACEMENT
ADSORBER EFFLUENT
-" APPROX. EMPTY BED CONTACT TIME = 6.5 MINUTES
i i i i
i i i i
11 13 \S 18 20 22 25 27 29 1 3 5 8 10 12 15 17 19 22 24 26 29 31
JULY AUGUST
FIGURE 20 REMOVAL OF TRIHALOMETHANE PRECURSORS BY
GRANULAR ACTIVATED CARBON BEDS
-------
-A63-
Class IV Compounds (Disinfection By-Products)
To determine the effectiveness of granular activated carbon to remove
trihalomethanes, Cincinnati tap water, which contains these compounds, was
exposed to granular activated carbon columns at various hydraulic loadings
(different contact times). Chloroform was reduced 90 percent or more for about
three weeks, (10 min. EBCT), then breakthrough was steady until the adsorber
was exhausted at about the ninth or tenth week. The trihalomethanes containing
bromine were effectively reduced by the granular activated carbon for 30 or
more weeks. Figure 21 shows the difference in effectiveness of six different
types of granular activated carbon for removing all of the trihalomethanes
summed together. In earlier field studies where the applied water and the
effluent from granular activated carbon beds were sampled, the findings regarding
trihalomethane reductions were very similar to the laboratory results.
The second phase of the Florida study, which studied adsorption of organics
from chlorinated water,also showed that chloroform was the trihalomethane least
effectively adsorbed, while bromoform was adsorbed the best of the trihalomethanes.
At the lower Mississippi River water treatment plant,ammonia is added after the
addition of chlorine so the concentrations of trihalomethanes reaching the
granular activated carbon bed were low, chloroform (1.8-46 yg/£), bromodichloromethane
(NF-6.2 yg/£), dibromochloromethane (NF-12.8 yg/£), bromoform (NF). Under
these conditions the bromine-containing trihalomethanes were well removed,
although chloroform began to appear consistently in the adsorber effluent at
low concentrations after the fifth week. Adsorption-desorption cycles
began after 13 weeks.
Figure 22, data from the water treatment plant in the upper Ohio Valley,
shows a breakthrough pattern for the trihalomethanes very similar to the pilot
plant data described above and shown in Figure 21. Here too, the bromine-
contadning trihalomethanes were removed better than chloroform.
-------
-A64-
-------
-A65-
Upper Ohio Valley Water
Treatment Plant
DICHLOROBROMOMETHANE
ADSORBER INFLUENT
CHLOROFORM
ADSORBER INFLUENT
SAND REPLACEMENT
ADSORBER EFFLUENT
SAND REPLACEMENT
EFFLUENT
SUMMATION TRIHALOMETHANES
DIBROMOCHLORO METHANE
ADSORBER INFLUENT
ADSORBER INFLUENT
SAND REPLACEMENT
ADSORBER
EFFLUENT
SAND REPLACEMENT
ADSORBER EFFLUENT
FIGURE 22. REMOVAL OF TRIHALOMETHANES BY GRANULAR ACTIVATED CARBON BEDS-
6.5 MINUTE EMPTY BED CONTACT TIME-VIRGIN WVW 14x40 GRANULAR
ACTIVATED CARBON
-------
-A66-
In summary, if the aromatics, taste and odor compounds, and certain pesticides
can be categorized as strongly adsorbed onto granular activated carbon, then
chloroform is located near the other end of the adsorption spectrum.
General Organic Parameters
Although not directly related to any single class of organic compounds discussed
above, the use of an easily measured general parameter to monitor the performance
of an adsorption treatment system is appealing from the standpoint of convenience
and cost. The Water Supply Research Division pilot plant studies routinely
include the following general organic measurements: non-purgeable organic
carbon (NPOC) , ultra-violet absorbance (UV) at 254 nm , and fluorescence —
both the emission scan (EMS)" and the fixed wavelength procedure described by
Silvia as the rapid fluorometric method (RFM) .
Figure 23 shows the pattern of each general organic parameter for the
first 15 weeks after a fresh granular activated carbon adsorber was put into
service. In an attempt to develop a simple test to predict the organic carbon
content of effluents, UV, RFM and MS data were correlated to NPOC. Table XV
summarizes the regression analysis. The most promising relationship thusfar is
between RFM and NPOC. A clearer understanding of these relationships and
others, such as between some general parameter and trihalomethane (Class IV)
or by-product precursor (Class III) concentrations is being sought through
additional statistical analyses and should be available in the near future.
TABLE XV
RELATIONSHIP BETWEEN NPOC AND OTHER GENERAL ORGANIC PARAMETERS
FOR MONITORING A GRANULAR ACTIVATED CARBON ADSORBER
UV RFM EMS
r\
(R ) Coefficient of Determination
(R) Correlation Coefficient
Significant at 95% by F and t test
95% Confidence, interval, mg/£
0.22
0.47
Yes
±0.73
0.67
0.82
Yes
±0 . 48'"
0.52
0.72
Yes
+0.58
-------
-A67-
'30NV9dosav A n
S1INH NtfOS NOISSIW3 3DN30SdOmd
.
-------
-A68-
Thusfar, the best predictor of NPOC, RFM, has such a high 95 percent
confidence interval (+0.48 mg/Jl) that its potential use is discouraging.
64
Sylvia , however, obtained better results correlating RFM measurements with
granular activated carbon bed performance as measured by CCE-m, so some
possibilities for this approach may exist in certain waters .
As noted above, in addition to correlating a more easily measured general
organic parameter against NPOC, the use of a general parameter to predict
performance of a granular activated activated carbon bed with respect to
breakthrough of a specific organic compound or group of specific organic compounds
is also an attractive concept. In the lower Mississippi River water treatment
plant study, four general parameters have been correlated to total trihalomethanes,
trihalomethane formation potential, and the sum of all of the non-trihalomethane
chlorinated organic compounds in the adsorber effluent. Table XVI shows that,
except for UV adsorbance, the correlation coefficients have good statistical
significance. The confidence limits on these data are, however, not available
at this time (Fall 1977). This effort will continue at this field site, as
well as in the Water Supply Research Division pilot plant, and at other water
treatment plant project locations.
-------
-A69-
TABLE XVI
RELATIONSHIP BETWEEN VARIOUS GENERAL ORGANIC PARAMETERS AND THREE CLASSES
OF ORGANIC CONTAMINANTS
NPOC
(50
a
r
OBS)
Signif.
EMS
(42 OBS)
r Signif.
RFM
(47 OBS)
r Signif.
UV
(50 OBS)
r Signif
Class II
Total Non-
Trihalo-
me thane
Organics 0.585 .00001 0.622 .00001 0.740 .00001 0.282 .047
Class III
XTHMFP 0.723 .00001 0.658 .0001 0.770 .00001 0.366 .009
Class IV
ETHM 0.631 .00001 0.745 .00001 0.747 .00001 0.306 .031
a - "r" is the correlation coefficient
b - "Signif" is the significance of the correlation coefficient, r (the
smaller the value for "Signif" the greater the significance.)
OBS - Observations
Influence of Empty Bed Contact Time
One phase of the project studying the use of adsorbents to remove
organics from a southern Florida groundwater includes the effect of empty
bed contact time on organic removal using the techniques similar to those
outlined in the "Procedure for Collection of Site Specific Design Data" in
the Interim Treatment Guide (pages 23 to 26) and Appendix C. Finished water
from the treatment plant was diverted to four pilot granular activated
carbon columns connected in series. Each column contained 30 inches of
12x40 mesh activated carbon and was operated at a hydraulic loading of 3
2
gpm/ft . The nominal empty bed contact time for each column was 6.2 min.
and thus the contact times were 6.2 min., 12.4 min., 18.6 min., and 24.8
min., respectively for the four columns. Twenty purgeable halogenated
organic substances were monitored at the inlet and outlet of each column.
Also TOC and Terminal THM concentration data were collected.
-------
-A70-
The ability to achieve a given quality of effluent for longer time periods
(larger water volumes) is related to longer empty bed contact times (larger
activated carbon volumes), Figures 24, 25, 26 and 27. Figure 26 for vinyl
chloride is an exception as the relationship is not as definite for the data
collected to date. Prior experience also showed that vinyl chloride removal by
granular activated carbon adsorption was the most sporadic of all substances
tested at this site. So far this has been the only substance to show such an
erratic pattern. Figures 24, 25, and 27 show a "broad-wave" front for the
substances tested, thus providing a more gradual approach toward equilibrium
concentration than if a more "narrow-wave" front had occurred.
In order to assess the period of time between activated carbon reactivations
for a given empty bed contact time, the three performance criteria discussed in
the Interim Guide were applied to the data in Figures 24-27. Whenever a
performance criterion was consistently exceeded, the first time at which the
criterion was exceeded (see arrow on Figures) was used and the corresponding
operating time and water volume treated calculated. The concentrations between
data points were taken as linear.
Table XVII presenting the maximum duration of time in days of operation
and liters treated when complying with each criterion shows that the TOG
criterion is the most limiting of the three at all contact times for the
conditions tested. TOG would have limited the operating time to 1.5 days, 4.5
days, 11.2 days and 18.6 days for the respective empty bed contact times.
Table XVIII is a different display of the prior data in which the activated
carbon volumes used and water volumes treated are recalculated using the 6.2
min EBCT data as unity and presenting all data as ratios. Thus for the most
stringent criterion, TOG, if two times the activated carbon volume is used,
-------
-A71-
o
LU
O
o
o
o
HI
CD
O.
5
LU
LL
O
LU
O -J
z <
m £
D O
l
£ ± cc
LU z
I O
i- m
LU
<
X ^
LU O
CM
LU
DC
D
O
I/Bui 'NOIlVaiN30NOD NO8dVD OINVOUO 1V1O1
-------
-A72-
at
ui
Z
iu
^
>-
O
DC
O
z
o
u
<
I—
z
8
u
z
O
LU
_J
Q_
10
CM
O
|/Bn 'NOIlVaiN3DNOD
-------
-A73-
-------
-A74-
0.80
15 20 25
TIME IN SERVICE, DAYS
30
35
FIGURE 27 EXAMPLE OF THE INFLUENCE OF EMPTY BED CONTACT
ON TERMINAL SUMMATION TRIHALOMETHANE REMOVA
-------
Empty Bed
Contact Time
Minutes
6.2
12.4
18.6
24.8
-A75-
TABLE XVII
DURATION OF RUN ALLOWED BY EACH CRITERION
Criterion No. 1 Criterion No. 1 Criterion No, 2
(Vinyl Chloride) (cis-l,2-Dichloro- (TOG)
ethylene)
Days Liters
MCL
Days Liters
3.5 313.8
10.8 966.4
35.7 3196.1
35.3 3096.5
_ Days Liters Days Liters
17.5 1573.3 1.5 132.5 2.3 204.8
60.5 5437.4 4.5 402.3 9.3 832.7
11.2 999.6 14.5 1297.1
18.6 1653.7 19.5 1733.2
- = Not yet exceeded criterion.
TABLE XVIII
RATIO OF VOLUME OF WATER TREATED IN COMPLIANCE WITH EACH CRITERION
EBCT
Min.
6.2
12.4
18.6
24.8
V.C. =
cis =
Ratio of Act.
Carbon
Required
1
2
3
4
Vinyl chloride
Criterion No. 1
Specific Organics
(V.C.) (cis)
1 1
3.1 3.5
10.2
9.9
Criterion No. 2
TOG
1
3.0
7.5
12.5
MCL
1
4.1
6.3
1.5
cis-1, 2-Dichloroethylene
- = Not yet exceeded criterion
-------
-A76-
three times the amount of water can be treated before the criterion is exceeded.
Further, again doubling the empty bed contact time increases the length of time
of operation over six more times, a favorable improvement. Similar beneficial
results occurred relative to criterion one and the MCL requirements with a
doubling of the "base" empty bed contact time, but redoubling produced only a
proportional increase in operating time according to the MCL requirement.
This evaluation of empty bed contact time shows that increased quantities
of activated carbon, at fixed flow conditions, resulted in the ability to treat
more water before exceeding a given criterion. In each case the first doubling
of empty bed contact time resulted in adsorber operating times that were proportionately
greater than the increase increse in empty bed contact times for all criteria.
This has an important impact on the economics of treatment, see pages 43-45 in
the Interim Guide.
Two in-house studies on the influence of empty bed contact time on adsorber
operating times, Figures 28 and 29, showed that for the two organic compounds,
chloroform and carbon tetrachloride, the time to reach a given breakthrough
point was approximately proportional to the empty bed contact time. Because of
these different results in two different locations, this type of information
should be collected in the location under study, as indicated on pages 23-26
of the Interim Treatment Guide.
Biological Growth and Endotoxin Production
Standard Plate Count
Controlling bacterial populations (and particularly killing or inactivating
pathogenic microorganisms) is a primary goal of water treatment. Some concern,
therefore, has been expressed about the possibility of bacteria proliferating
within granular activated carbon beds. To investigate this, the bacterial
quality of the untreated and treated water from the WSRD pilot plant was
-------
-A77-
100
AVERAGE CHLOROFORM CONCENTRATION
IN APPLIED CINCINNATI, OHIO TAP WATER=46A»g/L
GAC DEPTH: 90 cm (36 INCH)
GAC TYPE: FILTRASORB 400
50% EFFECTIVE
CHLOROFORM = 23 ftg/L
234567
TIME IN SERVICE WEEKS
FIGURE 28. EFFECT OF EMPTY BED CONTACT TIME ON CHLOROFORM
REMOVAL FROM TAP WATER.
8
-------
-A78-
I- >
*- o
.
o Z
" 8e
at
Q O
>• %
f <
£ *
«/> 5 i—
^ iz "j
UJ UJ |_
Z O
CM
UJ
fid
3
O
o
10
O
CO
O
CN
\/Brf 'NOIlVillN3DNOD
-------
-A79-
routinely monitored over a four-month period by the staff of the WSRD
Microbiological Treatment Branch using the Standard Plate Count (SPG) test.
Samples were also collected intermittently and analyzed for total and fecal
coliforms, however, these indicator organisms seldom survived the coagulation
and settling processes and were never detected in the filter or adsorber effluents.
The monthly average SPC (expressed as the geometric mean) for the pilot plant
studies (see Table XIX), in general, show a two log (99 percent) reduction in
the bacterial count through the treatment plant. Note that no disinfectant was
added anywhere in the treatment process. Although an increase in bacterial
populations was expected a. priori, the SPC of the effluent from the granular
activated carbon adsorber was slightly lower than the SPC in the companion
sample taken from the dual media filter effluent. No attempt was made to
measure attached growths. An attempt was made, however, to isolate and identify
the predominant populations in the pilot plant. From the adsorber effluent,
five or six different types of colonies could be recognized and two genera,
Flavobacterium and Xanthomonas were identified from smear plates.
TABLE XIX
MONTHLY MEAN (G ) STANDARD PLATE COUNT
m
(Pilot Plant Studies - Ohio River Water)
all bacterial counts are No./I ml
Time in
operation,
Months ,
1
2
3
4
Raw
19,600
12,000
7,170
6,680
Settled
1650
1000
790
700
Effluent from
Dual Media
Filter
137
270
80
100
Effluent from
Granular Activated
Carbon Adsorber
63
72
29
37
-------
-A80-
Endotoxins
Endotoxins are lipopolysaccharide-protein complexes produced in the cell
walls of Gram-negative bacteria. The lipopolysaccharide portion of the complex
r Q
is pyrogenic and as little as 1 yg can produce fever in a 700 kg horse.
Concern, therefore, has been expressed regarding the possible formation of
endotoxins in granular activated carbon adsorbers because of bacteriological
activity. For a 6-month period in 1977, the U.S. EPA Health Effects Research
Laboratory (HERL) monitored bacterial endotoxin concentrations in untreated and
treated water from the WSRD organics removal pilot plant. These were companion
samples with those collected for Standard Plate Count analyses.
Using the Limulus lysate bioassay , HERL scientists observed a marked
reduction in pyrogenic activity as a result of chemical coagulation (and settling)
and a slight additional decrease through filtration by either dual media or
granular activated carbon. The encouraging finding was that no increase in
pyrogenic activity occurred in the effluent from the granular activated carbon
bed. These data are shown in Table XX. An extramural project entitled "Pyrogenic
Activity of Carbon-Filtered Waters" (EPA Grant No. R-804420010) is underway at
Texas A&M University. Samples for endotoxin concentrations are being collected
from full-scale water treatment plants utilizing granular activated carbon
adsorption and the results from this definitive study should provide further evidence
as to whether or not a problem exists.
TABLE XX
MONTHLY MEAN ENDOTOXIN CONCENTRATIONS IN OHIO RIVER WATER
(Pilot Plant Studies)
Endotoxin Concentration, yg/£
Time Dual Granular Activated
in Operation, Media Carbon Adsorbent
Months Raw Coagulated/Settled Effl. Effluent
2
3
4
5
6
7
158
236
205
500
45
35
16
63
36
66
20
11
16
7
41
16
5
11
9
6
11
15
4
11
-------
-A81-
Ozone Enhanced Granular Activated Carbon Adsorption (Biological Activated Carbon)
One organic removal unit process being used in some locations has not yet
been discussed in this report. It is currently (Fall 1977) receiving much attention,
therefore, although much still needs to be known about the process, a discussion
of a variation of granular activated carbon adsorption for organic contaminant
control in which ozonation precedes the granular activated carbon treatment processes
is included here. These two processes in combination are frequently called biological
activated carbon.
Literature Review
As part of a research grant with the WSRD, Dr. Rip Rice, in collaboration
with Public Technology, Inc., Washington, D. C. prepared a chapter of the final
report on the "Status of Ozonation and Chlorine Dioxide Technologies for Treatment
of Municipal Water Supplies" project summarizing what is known about biological
activated carbon. The first few pages of that chapter, with some editing, are
presented below.
A REVIEW OF THE STATUS OF OZONATION PRIOR TO GRANULAR
ACTIVATED CARBON FOR REMOVAL OF DISSOLVED ORGANICS AND AMMONIA
FROM WATER AND WASTEWATER
1. Introduction
In a recent article that discusses the use of granular activated carbon in
water treatment, McCreary & Snoeyink state that "beds of granular activated
carbon (GAG) are a convenient place for microorganisms to grow because bacteria
attach themselves to the irregular external surfaces of the activated carbon
particles and are very difficult to dislodge via backwashing procedures." In the
presence of soluble carbonaceous matter, which serves as food for these organisms,
and in the absence of oxygen, anaerobic bacteria can develop. There are numerous
instances in which sulfidic odors have been reported emanating from granular
activated carbon columns used for the removal of dissolved organic materials
contained in sewage treatment plant effluents' and drinking water supplies.^
-------
-A82-
On the other hand, with sufficient dissolved oxygen and carbonaceous matter, the
bacteria that develop in activated carbon beds will be aerobic. These do not
produce sulfidic odors.
Many of the advantages of biological granular activated carbon (BAG) were
first recognized by German water treatment scientists in the 1960's in drinking
water plants along the Rhine River in the Dusseldorf area. Subsequently, BAG
processes also have been installed in Swiss and French drinking water treatment
plants, and are subjects of active pilot studies in Holland and Belgium. In the
United States, the U.S. Environmental Protection Agency's Water Supply Research
Division in Cincinnati, Ohio has been testing a pilot BAG column since late in
1976 (see below.)
2. Fundamental Principles
Granular activated carbon is made biologically active by the deliberate
introduction of sufficient dissolved oxygen (DO) to aqueous streams just before
they are passed through GAG columns. As long as the water contains sufficient DO
to maintain aerobicity of the bacteria and sufficient dissolved carbon to provide
74
food, the aerobic bacteria will thrive in this environment. Eberhardt has
likened bacterial activity in such an ideal environment to a "herd of cows grazing
in a luscious meadow." Pre-ozonation can convert larger, less biodegradable
organic molecules into smaller, more biodegradable organics, for example, into
acetic and oxalic acids. Sontheimer has summarized the German findings to date
which have led to the current theories of operation of BAG.
-------
- A83-
Although aerobic bacteria are necessary to obtain the benefits from BAG, so
also is the adsorptive capacity of the GAG for the dissolved organic materials
that will serve as food for these bacteria. This means that the surface area and
pore volume of the activated carbon should be high. Stated another way, the
organic materials present in solution should be adsorbable onto the activated
carbon column, because the contact times of solutions with the activated carbon
particles in the columns or beds are normally short (15-30 minutes). This does
not necessarily give the bacteria sufficient time to degrade larger organic
carbon molecules, ideally to carbon dioxide and water. Therefore, retaining the
dissolved organic molecules in the activated carbon columns so that the bacteria
then will have sufficient time to degrade them is important, even though the
actual contact times involved are rather short.
Many organic materials are readily adsorbed onto GAG, but many others are
not. For example, high molecular weight natural humic acids, so prevalent in
drinking water supplies, are not readily adsorbed by activated carbon. If
solutions of these non-sorbable organic materials are ozonized before passage
through the GAG columns, they are converted to more readily biodegradable organic
materials. ' ' At the same time, ozonation introduces a large quantity of
oxygen into the water, which promotes aerobic bacterial growth.
-------
-A84-
3. Advantages of Biological Activated Carbon
In European pilot studies and in drinking water treatment plants by many
workers ' ' have shown that ozonation followed by granular activated
carbon adsorption results in:
- More effective removal of dissolved organics from solution by the BAG,
- Increased operating life of the activated carbon columns before
having to be reactivated especially if the GAG can be kept free of halogenated
organics.
- Biological conversion of ammonia in the GAG columns,
- Use of less ozone for removing a given amount of organics than using
ozonation alone. (BAG is more cost-effective over ozonation in
removing Dissolved Organic Carbon - DOC),
- Filtrates from BAG columns in drinking water plants can be treated
with small quantities (0.1-0.5 mg/£) of chlorine or chlorine dioxide,
which produces drinking water of acceptable bacterial quality
(zero coliforms) and provides a residual disinfectant for distribution
systems.
Independent studies on physical-chemical treated sewage at the Cleveland
7 ft Rf*
Regional Sewer District, and in Israel have confirmed these advantages with
respect to removing organic materials.
4. European Background
Introduction of granular activated carbon into European drinking water treatment
practices occurred after World War II. Its initial application was for dechlorination,
Q 7
then for tastes and odor removal. Many surface waters containing ammonia undergo
break-point chlorination at the beginning of the treatment process. This technique
effectively removes ammonia, but produces considerable amounts of residual chlorine
Q C
and chlorinated products in the water. German water treatment objectives are
to process surface waters to the same quality as that of natural groundwater
(which does not have to be treated in many cases). Therefore, waters
-------
-A85-
treated by break-point chlorination have to be dechlorinated before they are
Q Q QQ
treated further or distributed. Schalekamp points out that in Switzerland a
residual chlorine concentration of only 0.05 mg/H is permitted in finished drinking
water. Therefore, the raw water chlorine dose is removed by granular activated
carbon prior to the addition of 0.05 mg/£ of chlorine dioxide for final disinfection.
Combinations of ozone and granular activated carbon first were installed in
87
the Dusseldorf area in the early 1960's but nearly ten years passed until the
biological activity in the activated carbon columns was recognized as being beneficial.
By the early 1960's, the lower Rhine River water quality had declined, and advantage
was taken of filtration of organic materials from the river water through the sand
banks of the Rhine. Wells were dug into the river banks and water is drawn from
these wells as the treatment plant raw water.
In the intervening years since the introduction of ozone/activated carbon at
Dusseldorf, the beneficial effects of biological activity in the activated carbon
columns have been recognized, characterized and optimized. After ozonation, the
water is allowed to stand for 20-30 minutes to allow the more refractory organic
compounds sufficient time to react with residual ozone. This retention time also
allows residual ozone to be utilized, rather than simply to be destroyed when
passed through the activated carbon column.
On the other hand, the Rhine River also contains considerable amounts of
chlorinated organic materials that are not removed during river bank filtration.
These halogenated materials also are more resistant to oxidation by ozone than are
non-halogenated organics and thus are less likely to be converted into easily
biodegradable materials. In addition, halogenated organics are variably
1 9 89
adsorbed onto activated carbon. '
-------
-A86-
Combining the variable adsorptivity of halogenated organics compounds on
granular activated carbon with their lesser reactivity upon ozonation and
their lower biodegradability means that breakthrough of halogenated organic
compounds occurs sooner than does breakthrough of non-halogenated organic
compounds, even if the granular activated carbon columns are biologically
activated. Thus German water works along the Rhine in the Dusseldorf area
monitor their activated carbon column capacities for Total Organic Chlorine
90—92 93
(TOC1) as well as by DOC and/or UV absorption. Activated carbon
columns at three Dusseldorf plants along the Rhein (Flehe, Am Staad, Holthausen)
are backwashed every 4-6 weeks and reactivated every 6 months.
When Dusseldorf activated carbons are thermally reactived, however,
only some 80 percent of the activated carbon charge is taken out of the
columns. This leaves a portion of biologically active activated carbon in
the column so that the level of bioactivity will not drop significantly when
fresh or reactivated activated carbon is added. With fresh activated carbon
columns, about 15 days of operation usually are required for biological
activity to build up to an effective "steady state", particularly for ammonia
i 94
removal.
WSRD Pilot Plant Results
In late 1976, WSRD began a pilot plant study on the use of ozone to
extend the life of a granular activated carbon adsorber used as a sand
replacement system as measured by Total Organic Carbon - and Trihalomethane
Formation Potential (2 day, 25°C) concentrations. This study was prompted
by the successful results from pilot- and full-scale BAG tests conducted in
Europe (principally West Germany) cited above.
-------
-A87-
In this test a granular activated carbon adsorber receiving settled ozonated
water and a granular activated carbon adsorber receiving settled unozonated water
were operated for a 10 month study period. Both units had about a 9 minute empty
bed contact time. Ozone dosages applied to the settled water (20 minute plug
flow contact time) and dissolved ozone concentrations prior to filtration or
adsorption are shown in Figure 30. Because of an increase in ozone demand
within the system and the resultant loss of ozone residual, the applied ozone dose
was increased following ozone contactor cleaning early in the 9th month of operation.
Figure 31 compares the TOG concentrations in both effluents. Figure 31
illustrates the extension in bed life afforded by the combination of ozone followed
by granular activated carbon. Operating time was extended from about 3-3/4 months
to about 7-1/2 months according to Criterion 2, increase of 0.5 mg/£ of TOC concen-
tration.
Figure 32 shows a similar trend for the removal of trihalomethane formation
potential, although the MCL requirement was never exceeded in either effluent
during this study. The results of this study are encouraging enough to prompt
additional in-house and extramural research on biological activated carbon during
FY78, particularly to determine whether or not the use of ozonation prior to
adsorption improves granular activated carbon performance with respect to Criterion
1, the control of low molecular weight halogenated organic compounds.
-------
-ASS-
CO
UJ
o
<
cc
UJ
>
>- CO
-1 Z
fs
li
«5
Ujg
ttg
:
<=l
X
.0
_i
-iu_
<0
1
1
N
t^
"V.
00
•s^
N
Q
UJ
CO
<
UJ
cc
o
z
UJ
CO
o
Q
UJ
Z
0
N
O
o
UJ
_l
CL
Q.
<
I i
O
(N
I
DC
UJ
h-
_J
>>.
CO
0
5
£E
UJ
h-
<
^
UJ
_J
1-
1-
UJ
CO
o
h-
o
UJ
_l
Q.
Q.
<
UJ
CO
o
Q t
tt 1
UJ
OQ
K
0
CO
Q f
1
1-
<
0
UJ 1
OC T
^M I
j
CO
<
ml
1
1
I
•>>
O)
£
UJ 1.
ZT
01
N
O
' Q <-
_l
0
CO
CO
' Q
-------
-A89-
l/Bui 'NOIlVdlN30NOO OINVOdO "IV1O1
-------
-A90-
1/BrT 'NOI1VU1N30NOO
'Oo93 'AVQ Z
-------
-A91-
Economic Analysis Data
This text supplements the economic analysis information presented in
the Interim Guide on pages 36 to 54, and is divided into five sections. The
first section contains basic cost information that can be used to present a
generalized framework for considering the cost impact of using granular
activated carbon adsorption. In the next two sections, specific operational
configurations and their influence on system costs are considered. For
example, the tradeoffs that exist between Empty Bed Contact Time (EBCT) and
Reactivation Frequency (RF) are examined in one section, and the effect of
activated carbon cost and reactivation frequency is considered in the other.
In the fourth section the impact of inflation on the choice of systems is
evaluated. Because reactivation is a significant portion of the cost of a
granular activated carbon treatment system, it is considered separately in
the final section. Another source of supplementary economic analysis information
is the Interim Treatment Guide for the Control of Chloroform and Other
Trihalomethanes.
Basic Costs
The data utilized in this section are the same as used to develop costs
in the Interim Guide. Tables XI, XII, and XIII, and Figure 13 in the
Interim Guide pages 37, 38, 40, and 42, respectively provide examples of the
use of this baseline information.
-------
-A92-
Influence of Empty Bed Contact Time
As discussed in the Interim Guide examining the cost tradeoffs that exist
between bed depth (EBCT) and reactivation frequency (RF) for a given quality of
water is important. This relationship is examined using an "R" ratio. The "R"
ratio allows the changing of two variables at one time (EBCT and RF). For example,
in a sand replacement system, if the removal efficiency of granular activated
carbon is proportional to EBCT and RF, then a system with an EBCT of 10 minutes
and an RP of one month (R = 0.1) might be assumed to be equivalent in cost to one
with an EBCT of 20 minutes and an RF of two months (R = 0.1). As shown on page 44
in the Interim Guide, however, although the R's are the same, the costs are not.
Table XXI duplicates the data in Table XIV (page 41 of the Interim Guide) with the
exception that values for "R" have been added.
Another way of examining the tradeoff between cost, RF, and EBCT is shown in
Figures 33, 34, 35, and 36. In these figures the relationship that exists between
the cost, EBCT, and RF for 10 and 100 mgd sand replacement systems and for 10 and
100 mgd postfilter adsorbers can be seen. The problems of non-proportionality
between cost and performance can also be seen in these data. For example, in
Figure 35 if the required EBCT was 10 minutes with a one month reactivation
frequency, the unit cost would be 6.6 cents/1000 gal. If the EBCT were increased
to 18 minutes, the period between reactivations would have to be increased to
approximately 2.5 months or greater to achieve a favorable economic tradeoff.
Therefore, to merit an 80 percent increase in empty bed contact time, the period
between reactivations would have to increase by at least 150 percent. This important
concept is examined more fully both in the Interim Guide (see pages 27, 34, and 35)
and earlier in Appendix A, pages A69 to A76.
-------
-A93-
C/l
w
H
0
O
O
o
rH
- — .
O-
CU
e
T3 1H
CQ en
>* CJ U
p J l r^
S C 1-1
W O S
u •-'
^
4-J
•H ,—,
.*: cj TJ
rO CO OC
a) a. E
a. co ^
u
C 3 T3
00 0 00
01
a
a
a
en a
:>N >-
en H
o
r-.
O
O
.
a.
CN
in
0
C3
in
_J
rH
O
O
00
in
rH
rH
O
O
a*
r^
CN
O
O
O
00
o
m
o
o
r^
rH*
rH
m
CN
O
o
in
rH
m
o
o
rH
rH
CM
CN
rH
O
O
r>-
r*
m
o
o
o
rH
in
00
0
CM
O
O
0
o
^r
r^
rH
in
CM
0
d
r^
rH
CN
rH
O
O
rH
O
00
o
o
o
CO
^
in
co
O
o
o
rH
^
CM
rH
0
O
O
a
o
rH
d
r^
00
in
r^
0
d
en
<*
O
in
o
o
00
rH
rH
m
CM
0
o
^.
r^
rH
O
O
0
r^
CN
O
rH
O
O
u
CU
rH IH
•H CD
tt4 ^
1 U
en in
0- <
o
in
O
o
o
CO
rH
OO
CO
o
o
e*
^r
r-t
m
CM
o
o
>o
OO
rH
CN
O
o
ON
OO
CN
in
o
o
o
00
,v|
in
O
CN
O
o
CO
O
o
0>
vO
rH
in
CM
O
o
10
ON
rH
rH
O
O
ON
<^
CM
CO
o
o
o
OO
ON
O
O
d
o
CN
CO
O
0
o
-------
-A94-
,-v
•o
4J
C
o
CJ
CO
£_4
CO
CO
s
o
03
Oi
CJ
a
u
H
^
rH
H
CJ
as
3
1
o
u
o
^
z
o
CO
03
p-l
s
o
CJ
H
CO
o
CJ
M
y
rH
CO
H
en
4J
c
0
s
c
'H
o
c
41
3
cr
CO
d
o
•H
4J
CO
4J
O
(0
01
at
o
e
o
c
S
o
in
•
0
S
o
0
B
o
en
o
B
O
CN
•o
0)
03
^
Q
g
-a:
re
01
d
0
•H
d
a
OS
,_(
CD
cm
o
o
0
rH
O
OS
CO
00
o
o
o
I— f
o-
OS
CTJ
00
o
o
o
rH
-— ^
a!
rH
10
00
o
o
o
rH
o
*
OS
rH
CO
00
o
o
o
o-
01
_B
H ^-~
01
4-1 CO
CJ 4-1
CO 3
C 'r
0 g
CJ -^
4_(
u t
CO O
a. e
CO ^
o
3 ?
1 O o
^ s
Ct, N^
^ 0)
c/i a
>. >
o o o
CN r- 1 •
in r-« ej^
O o en
O O en
-3" CN rH
O O O
O*i rH rH
m ao o
rH
O O O
o m o
en i-H rH
o o d
^O CO O
1-1
O O r^
CN rH O
O O O
cj\ o^ in
\O CT\ CN
O O O
rH CN en
O O C
O O O
S
rH V4
•H cy
i M
U 0
CO -3
•H 4-1
4J
O 4J
CO U
G
G O
OJ O
0)
3 'O
D .£3
•^
QJ 4-1
•S a
11
cd
-K
-------
-A95-
Hl-Ht- H
OOOU Q
OQ OQ OQ OQ 00
UJ LLlUJUJ LU
(/> h-
o|
< 5
Z Q.
< UJ
03 <
"iS>
UJ T3
5 o>
c E
o w
S8
0.0)
lj(0
u-z
00
o§
SE
LL UJ
LLlQu
CO
CO
UJ
CC
O
W31SAS JO 1SOD
-------
-A96-
I- I- h- -
O O OOO
00 OQ 00 CQfiQ
UJ LU UJ WQJ
UJ UJ UJ
h- I- l-
D 3 ZD
(D •* CMOro
eg CM CMCM£:
X
I-
CNI
^ (/)
z
g
s I
I-
o
00
(O
UJ
cc
UJ
UJ
UJ
CQ
O
g
cc
UJ
Q.
CM
1
O
10
CNJ
1
O
d
CM
1
O
in
1
o
d
^
1
0
in
0
d
O H
m c/)
LU O
s_x (L
ILI T3
z u.
O i-
O (/>
UJ O
> UJ
I- I-
Q. (/)
UJ (/>
U. Z
O O
I- Q
U^
t LJ
UJ CL
CO
UJ
CC
IAI31SAS dO 1SOO
-------
-A97-
LU
00000
CQCQ CQQQOQ
LULU LULULU
(/)(/>(/> COW
LULU LULULU
CO
CM
(0
X
h-
o
(0
z
g
<
o
<
LU
CC
Z
LU
LU
LU
CQ
Q
O
QC
LU
Q.
n'S
S o
CQS
> LU
Q- CO
S >•
LU CO
LL Z
00
LL
LL
CC
LU
q
in
q
o
q
in
q
d
LU Q.
LO
CO
LU
CC
3
O
JO 1SOO
-------
-A98-
CQ CQ OQOQCQ
LU UJ LULU HI
W W (/)
UJ LU LULU HI
55556
(f)
5 5555
CM
00
CM
q
in
q
O
o
in
o
d
g cc
i- u
>§
o
CC DC
II
z b £
o m O
5 HI Q.
* u7?>
s 1E
P po
< I- O
> o ^
I- < CC
It*
•" / \ ^^
z ^ CO
LU Q O
LU UJ O
^ ffl5
lii ^" "J
OQ h- H
Q ^ CO
o S >-
£ UJ CO
s feg
«s
UJ ^
LL DC
U. UJ
UJ Q.
(O
CO
UJ
CC
D
g
u.
(|BB OOOI-/0)
W31SAS dO ISOO
-------
-A99-
Influence of Granular Activated Carbon Cost
Because of the wide variations in the types of activated carbon available and
their prices and performances, a water purveyor must have some understanding of
the relationship between activated carbon cost and performance. If a lower cost
activated carbon also has lower performance, requiring more frequent reactivation,
then its use may result in higher total system cost as opposed to using a higher
cost carbon with better performance. Figures 37 through 39 show the effect of
activated carbon cost and RF on the cost of 10 and 100 mgd sand replacement systems
and for a 100 mgd post-filter adsorber system, see also page 48 of the Interim
Guide.
Influence of Inflation
In the Interim Guide, Table XV (page 53) contains a summary of the "present
value" of the expenditures for 10 and 100 mgd sand replacement and 10 and 100 mgd
post-filter adsorber systems. These data have been calculated for two discount
rates (6 percent and 8 percent) and three inflation rates (5 percent, 7 percent,
and 9 percent). Capital expenditures were assumed to be amortized over 20 years
at 7 percent interest, and reflect a continuous, constant expenditure pattern over
the life of the facility.
Tables XXII, XXIII and XXIV show the costs over time for 10 and 100 mgd sand
replacement systems and 10 and 100 mgd post-filter adsorption systems at assumed
inflation rates of 5 percent, 7 percent, and 9 percent. Table XV (page 53) in the
Interim Guide summarizes present value calculations for the given investment
streams for each combination of system configuration and inflation rate.
Figures 40 and 41 supplement Figures 18 and 19 in the Interim Guide
(pages 50 and 52) by showing the effect of 5 percent and 7 percent inflation
rates for two 10 mgd systems (one of each type.) In contrast to the LOO mgd
systems, the post-filter adsorbers never become less expensive than the sand
-------
-A100-
O
Q
O
oc
111
Q.
en
H
Z
O
ot-
o
01-
u->-
UJCO
CO
UJ
oc
D
O
EE
(|B6
INH1SAS JO 1SOO
-------
-A101-
UJ
I-
,_CL
000
N into
606
<*
-------
-A102-
DC
UJ
CO
CC
OQ
H<
IE
CM
o
Z
g
o
<
m
cc
Z
LLJ
UJ
H
UJ
CQ
Q
g
cc
UJ
0.
DO
ZO-
I- O)
(/) E
8§
CO Q
CC LL
UJ
I-
UJ
l_l — mm
q Oo
1
0
10
M
1
O
0
CM
1
0
in
v—
1
0
6
1
o
in
o
d
LL DC
LL UJ
UJ Q.
0>
CO
UJ
CC
O
(|B6 OOOi-/U>)
dO 1SOO
-------
-A103-
Table XXII. INFLATIONARY IMPACT FOR INVESTMENTS IN POST FILTER ADSORPTION
AND SAND REPLACEMENT (5%)
Costs in C/1000 gal
,,£ars 10 mgd 100 mgd
After Post-Filter Adsorption Sand Replacement Post-Filter Adsorption Sand Replacement
Construction Capital O&M Total Capital O&M Total Capital O&M Total Capital O&M Total
CO >3- rH
OO rH
CN r~
0 0
rH rH
ON ON
in ON
CM CM
rH rH
ON CO
r- oo
rH CO
CM rH
CN CO
CM CM
CO O
CM CM
OO ON
rH rH
rH VD
CJ ON
CM CM
r~ CM
O CO
in m
rH rH
sr in
rH rH
rH P^
o m
CO CO
rH r~-
CO CO
rH rH
rH rH
O -3-
3-
•^ -* m
vO ON ^~
-------
-A104-
fr"
r-
H
!3
W
s
w
o
•<
rJ
PH
W
Pi
P
5
C/3
53
o
H
H
PH
O
tt)
Q
g ?"
H
O
4-1
i*
W
s
H
en
W
>
H
O
K^ '
|
O
t-l
H
X
X
43
<0
OO mgd
T3
M
S
O
C. H
M
O
W
T3
fn
a
Q)
P.
M
O
CO
-H
P-i
O
•H
a.
Cap
ti
O
•H
4-1
CO O
rfl Q) (-1
C 4J 4-1
p-i MH co
< e
O
CJ
oo
CO
CXI
O
r-
vD
0\
in
CM
CTi
O
00
CMCM
*3" CO ^
m o*> co
in vr>
CM
vO *«^* *^ vO
vo o\ csi m
oo oo
rH O^ C^* rH
O CM m o^
rHrH
CM
o
-------
-A105-
H
TD REPLACEMI
Q
ORPTICM
*$ rH
cd
H O
i-J O
it 3
H ~<>
CO
0 d
Pi -H
Z, CO
M 4J
CO O
w
CO
55
M
O
my IMPACT
fNOIIVIi
M
M
CU
rH
,0
icement
P.
cu
&
$
CO
•o
0 g
0 -H
O.
)-i
O
CO
TJ
-Filtei
CO
o
Pi
uacement
o.
cu
w
TJ
00
1(
: Adsorptic
:-Filtei
CO
o
CO
n) cu
0) 4J
rH
rd
H
2
rH
«0
•H
a
rd
u
"ca
4-1
H
rH
cd
4-1
P.
cd
u
rH
cd
o
H
a
i— (
•H
O.
CJ
rH
cd
0
H
1— 1
cd
4-1
Construction Cap:
oo sr
CO en
oo -*
rH VO
m m
o
rH
VO rH
vO O
00 ON
rH VO
O en
in
-*
rH ON
en cn
rH rH
"•3" **3"
CM O
0> O
rH
en
r^ i~i
vo r^
Q\ p*-.
m CM
rx. co
ON
O rH
•vf ON
CO en
p- CO
-3- ON
rH vo
vO vo
ON rH
cn co
ON ON
n cn
P^- CO ON
ON rH r —
O r-. 3"
in OO O
CN CM en
rH ON VO
CM cn P~
sr vo co
CM CM CM
-3-00
rH CO VO
cn
-------
-A106-
I
00
o
13
oc
t-
(0
z
o
O
OC
UJ
(0
oc
<
UJ
U_ UJ
O 5
h- u
5)0
o <
UJ O
I- Z
< <
CC 00
Z 0
oz
Zh-
— a.
SCC
u.
UJ U.
UJ
tr
=)
o
\L
(IBB
JO 1SOO
-------
-A107-
Z
o
i-
u
D
QC
I-
o
o
QC
LU
V)
CC
LU
05
Q.UJ
O H
i- z
LL LU
OS
fcS
OS
7UJ
CCO)
ZQ
OZ
ii
~Qu
u^O
LLJ u_
LLJ
QC
O
LL
IN31SAS dO 1SOD
-------
-Ales-
replacement systems. For small systems, post-filter adsorbers are less desirable
from an economic point of view than for larger systems. A possible alternative
that minimizes the economic impact of using granular activated carbon adsorption in
small systems would be to consider the use of truck transport and regional
reactivation systems.
Cost of Reactivation
The most common (but not the only) method of reactivating carbon is by the
use of multi-hearth furnaces, and a significant portion of the cost of using granular
activated adsorption treatment is associated with on-site reactivation. Figures 42 and
43 show the annual capital (amortized over 20 years at 7 percent interest) and
annual operating and maintenance cost for reactivation based on reactivation rate
in Ibs/day and can be used to calculate this cost.
Using the data from Table X (page 34) in the Interim Guide, an example will
be constructed to demonstrate a convenient method for estimating a unit cost based
on the activated carbon use rate for reactivation. Assume a 100 mgd treatment plant
with an activated carbon use rate as given in line 3, column 4 in Table X (page 34)
(530 mg/£). The use rate converted to pounds per day as follows:
Use Rate = 530 mg/£ x 8.34 Ibs/gal =
4,420 Ibs/mil gal
Multiplying by 100 mgd yields a reactivation rate of =
442,000 Ibs/day
Entering Figures 42 and 43 at 442,000 Ib/day yields an approximate annual
capital cost of 430,000 $/yr and annual operating cost of 6,500,000 $/yr. The
total annual cost, therefore, is 6,930,000 $/yr for this rather high adsorbent
use rate. The total yearly cost is divided by the annual flow (100 mgd x 365 days/yr =
36,500 mgy) to yield:
-------
-A109-
O
m
cc
<
o
o
UJ
I-
o
<
cc
cc
UJ
Q.
O
Z
o
Q.
LU
h-
CC
Ul
(/>
^
CC
0
z
0
p
<
_
V-
o
UJ
cc
z
0
CO
cc
o
Q
UJ
t—
^-
o
^
cc
o
X
cc
UJ
I
H~
3
5
cc
o
LL.
1-
co
0
o
UJ
0
z
z
UJ
H
2
^
5
Q (0
55
1»
H CO
< _
O
3 O
Z <
z ui
< cc
oi
UJ
cc
O
aad savmoa 'isoo 30NVN3iNivw QNV Nouvuado IVHNNV
-------
-A110-
uv3A
savmoa 'isoo iviidvo IVHNNV
-------
-Alll-
Unit Cost = $6-93 x 10 /vr = $l90/mg or 19 cents/1000 gal.
36,500 mgy
The above unit cost is for buying and operating the reactivation furnace only.
Summary
This material is intended to supplement the material on pages 36 to 54 of
the "Interim Treatment Guide for Controlling Organic Contaminants in Drinking
Water Using Granular Activated Carbon." It presents data that can be used to
provide additional understanding of the cost of using granular activated carbon
adsorption treatment for controlling organic contaminants in drinking water.
-------
-A112
ACKNOWLEDGMENTS
Because of the size and complexity of this report, its importance, and
the short time that was available for its completion, the authors judge that this
is one of the more difficult assignments given to them. It could not have been
completed without help. The authors wish to acknowledge the contribution of
many members of the Physical and Chemical Contaminant Removal Branch (listed
at the end of the Interim Treatment Guide) who helped gather the data contained
in Appendix A, as well as Dale W. Dietrich and his staff of artists,
Nancy J. Quilhot, Stephen E. Wilson and Newell J. Maston who prepared all of
the Figures in the Interim Treatment Guide and Appendix A. They also wish to
acknowledge the efforts of the Principal Investigators of the extramural projects
from which additional data were obtained. Finally, this difficult project
could not have been successfully completed without the dedicated and skillful
efforts of Ms. Maura M. Lilly who typed all of the drafts and the final
manuscript so willingly and rapidly.
-------
-A113-
REFERENCES
1. Woodward, R.L. , Dostal, K.A. and Robeck, G.G., "Granular-Activated
Gabon Beds for Odor Removal," JAWWA, 56_, No. 3, 287-295 (May 1964).
2. Symons, J.M, "Summary of Granular Activated Carbon Practice Data,"
Water Supply Research Division, U.S. Environmental Protection Agency,
Cincinnati, Ohio, February 17, 1976, mimeo, 10 pp.
3. Dostal, K.A., Pierson, R.C., Hager, D.G. and Robeck, G.G., "Carbon
Bed Design Criteria Study at Nitro, West Virginia," JAWWA, 57, No. 5,
663-674 (May 1965).
4. Love, O.T., Jr., Robeck, G.G., Symons, J.M. and Buelow, R.W.,
"Experience with Activated Carbon in the U.S.A.," In: Papers
and Proceedings of a Water Research Association Conference at the
University of Reading, United Kingdom, Paper 13, 279-312 (February 1974)
398 pp.
5. Love, O.T., Jr. and Symons, J.M., "Evaluating the Need for a Sand
Layer in a Granular Activated Carbon Filter," Water Supply Research
Division, U.S. Environmental Protection Agency, Cincinnati, Ohio,
April 22, 1977, Mimeo, 10 pp.
6. Ford, D.B., "The Use of Granular Activated Carbon Filtration for
Taste and Odour Control," In: Papers and Proceedings of a Water
Research Association Conference at the University of Reading,
United Kingdom, Paper 12, 263-278 (Feb. 1974) 398 pp.
7. Stevens, A.A. and Symons, J.M., "Measurement of Organics in Drinking
Water," In: Proceedings of the AWWA Water Quality Technology Conference,
pp. XXIII-1 to XXIII-25 (Dec. 1973), American Water Works Association,
Denver, CO (1974).
8. Buelow, R.W., Carswell, J.K., and Symons, J.M., "An Improved Method
for Determining Organics by Activated Carbon Adsorption and Solvent
Extraction," JAWWA, 65_, No. 1, 57-72 and No. 3, 195-199 (Jan., March, 1973)
9. Heymann, E., "Practical Experience in the Use of F.locculation Filtration
Connected in Series to Granular Activated Carbon Filters," In:
Translation of Reports on Special Problems of Water Technology -
Volume 9 - Adsorption, Conference held in Karlsruhe, Federal
Republic of Germany, 1975, pp. 112-127, EPA-600/9-76-030, U.S.
Environmental Protection Agency, Cincinnati, Ohio, (Dec. 1976), 454 pp.
10. Middleton, P.M. and Rosen, A.A., "Organic Contaminants Affecting the
Quality of Water," Public Health Reports, 7J., 1125-1133 (Nov. 1956).
11. Rosen, A.A., Middleton, P.M. and Taylor, N.W., "Identification of
Anionic Synthetic Detergents in Foams and Surface Waters," JAWWA,
48, 1321-1330 (October 1956).
-------
-A114-
12. Rosen, A. A., Skeel, R.T. and Ettinger, M.B., "Relationship of
River Water Odor to Specific Organic Contaminants," JWPCF,35,
No. 6, 777-782 (January 1963).
13. Rosen, A.A., Mashni, C.I., and Safferman, R.S., "Recent Developments
in the Chemistry of Odour in Water: The Cause of Earthy/Musty
Odour," Water Treatment and Examination, 19, No. 1, 106-119 (1970).
14. Anon., "Industrial Pollution of the Lower Mississippi River in
Louisiana," Surveillance and Analysis Division, U.S. Environmental
Protection Agency, Region VI, Dallas, Texas (April 1972), Mimeo, 146 pp.
15. Kloepfer, R.D. and Fairless, B.J., "Characterization of Organic
Components in a Municipal Water Supply," Environmental Science and
Technology, jS, 1036-1037 (Nov. 1972).
16. Robeck, G.G., Dostal, K.A., Cohen, J. and Kreissl, J.F., "Effectiveness
of Water Treatment Processes in Pesticide Removal." JAWWA, 57, No. 2,
181-199 (Feb. 1965).
17. Stieglitz, L. , Roth, W., Kuhn, W. and Leger, W., "The Behavior
of Organohalides in the Treatment of Drinking Water," Vom Wasser, 47,
347-377 (1976).
18. Grob, K. and Grob, C., "Organic Substances in Potable Water and in
Its Precursor. Part II. Applications in the Area of Zurich,"
Journal of Chromatography, 90, 303-313 (1974), Grob, K., Grob, K. and
Grob, C., "Part III. The Closed-Loop Stripping Procedure Compared
with Rapid Liquid Extraction," Journal of Chromatography, 106,
299-315 (1975).
19. Kolle, W., Sontheimer, H. and Stieglitz, L., "Quantification Test
of Activated Carbon for Waterworks Based on Its Adsorption Properties
for Organic Chlorine," Vom Wasser, 44_, 203-217 (1975).
20. Suffet, I.H., Radziul, J.V., Cairo, P.R., and Coyle, J.T., "Evaluation
of the Capability of Granular Activated Carbon and Resin to Remove
Trace Organics from Treated Drinking Water," In: Proceedings of
Second Conference on Water Chlorination - Environmental Impact and
Health Effects, Gatlinburg, Tennessee, October 31 - November 4, 1977,
In Press.
21. McCarty, P.L., Reinhard, M. and Argo, P.G., "Organic Removal by
Advanced Wastewater Treatment," Presented at 97th Annual AWWA
Conference, Anaheim, California, May 8-13, 1977.
22. Anon., "National Interim Primary Drinking Water Regulations," Office
of Water Supply, U.S. Environmental Protection Agency, EPA-570/9-76-003,
Washington, D.C., 23 pp + 2 Appendices.
23. Sorg, T.J., Logsdon, G.S. and Love, O.T., Jr., "Manual of Treatment
Techniques for Meeting the Interim Primary Drinking Water Regulations,"
Water Supply Research Division, U.S. Environmental Protection Agency,
EPA-600/8-77-005, Cincinnati, Ohio (May 1977), 73 pp.
-------
-A115-
24. Cohen, J.M., Kamphake, L.J., Lemke, A.E., Henderson C. and Woodward,
R.L., "Effects of Fish Poison on Water Supplies, Part 1. Removal of Toxic
Materials," J. Amer. Water Works Assoc., 52, 1551-1566 (1960).
25. Goodrich, P.R. and Monke, E.J., "Insecticide Adsorption on
Activated Carbon," Trans. Amer. Soc. Agr. Engr. Gen. Ed., 13,
No. 1, 56-60 (Jan.-Feb. 1970).
26. Hyndshaw, A.Y., "Activated Carbon to Remove Organic Contaminants
from Water," J. Am. Water Works Assoc., 64, No. 5, 309-311 (1972).
27. Moergeli, B., "Removal of Pesticides from Drinking Water," Sulzer
Tech. Rev., 54, No. 2, 91-96 (1972). Also see Gas, Wasser, Abwasser,
_52, No. 5, 142-146 (1972).
28. Shevchenko, M.A., Marchenko, P.V., Taran, P.N., and Kravets, E.V.,
"Removal of Triazine-Series Herbicides from Natural Waters,"
Vodopodgotovka Ochistka Prom. Stokov (USSR), No. 9, 103-112, (1972).
29. Spahn, H., Brauch, V., Schluender, E., and Sontheimer, H., "Active
Carbon Filters for Water Purification, I. Adsorption on a Single
Particle," Verfahrenstechnik (Mainz), j?, No. 8, 224-231 (1974).
30. Parkash, S., "Adsorption of Cationic Pesticides (Diquat and
Paraquat) from Aqueous Solution by Activated Carbion," Carbon, 12,
No. 5, 483-491 (1974).
31. Shevehenko, M.A., Marchenko, P.V., Taran, P.N., and Kravets, E.V.,
"Removal of Some Pesticides from Drinking Water by Adsorption Methods,"
Vodosnabzh. Sanit. Tekh. (USSR), 10, 31-32 (1974).
32. Dawson, V.K., Marking, L.L. and Bills, T.D., "Removal of Toxic
Chemicals from Water with Activated Carbon," Trans. Am. Fish Soc.,
105, No. 1, 119-123 (1976). ~~
33. Thielemann, H., "Identification and Removal of Odor and Taste-Imparting
Substances in Filtered Groundwater," Mikrochim Acta (E. Ger.) No. 1,
79-80 (1972).
34. Willey, B.F., "Surveillance and Control of Phenolic Tastes and Odors
in Water to Prevent Their Effects on Taste and Flavor of Foods,"
Amer. Chem. Soc. Symp. Series 26: Phenolic, Sulfur, Nitrogen Compd.
Food Flavors Symp., 71-84 (1975).
35. Engels, C. "The Removal of Extra Light Fuel Oil from Polluted
Water," Vom Wasser, 45, 45-51 (1975).
36. El-Dib, M.A., Ramadan, F.M. and Ismail, M., "Adsorption of Sevin
and Baygon on Granular Activated Carbon," Water Res., 9, No. 9,
795-798 (1975).
37. Allen, D.H., "Evalution of Activated Carbon for Fuel Oil Adsorption
from a Potable Water Supply," Natl. Tech. Info. Ser. (NTIS), AD Rep.,
AD-A037478 (1976).
-------
-A116-
38. Koganovskii, A.M. and Leuchenko, T.M., "Calculation of Isotherms for
Adsorption of Organic Nonelectrolytes from Aqueous Solutions,"
Colloid J.(USSR). 2i» N°- 3> 305-307 (May-June 1969).
39. Bernardin, F.E., Jr., "Selecting and Specifying Activated Carbon
Adsorption Systems," Chem. Engr. (NY), 83, No. 22, 77-82 (Dec. 18, 1976).
40. Zogorski, J.S., Faust, S.D. and Haas, J.H., Jr., "The Kinetcs of
Adsorption of Phenols by Granular Activated Carbon," J. Coloid.
and Interface Science, 55, No. 2, 329-341 (1976).
41. Kondo, M., Kishi, T., Shimamoto, T., Nohihara, T., and Watabe, K.,
"Studies on the Adsorption Effect of Activated Charcoal on Polychlorinated
Biphenyls in Water," Memoirs of the Faculty of Pharmaceutical Sciences,
(Japan), 18, 135-138 (1974).
42. Snoeyink, V.L., Lai, H.T., Johnson J.H. and Young, J.F. "Activated
Carbon: Dechlorination and the Adsorption of Organic Compounds."
Amer. Chem. Soc. Symp. on the Chemistry of Water Supply, Treatment
and Distribution, Dallas, Texas (April 1973).
43. Whitehouse, J.D., "A Study of the Removal of Pesticides from Water,"
Univ. of Kentucky Water Resources Institute, Research Report No. 8,
Lexington, Ky. (1967). (Also available from NTIS No. PB-242-971).
44. Zogorski, J.S. and Faust, S.D., "Removal of Phenols from Polluted
Water," Natl. Tech. Info. Ser., No. PB-237-859 (June 1974).
45. Gauntlett, R.B. and Packham, R.F., "The Removal of Organic Compounds
in the Production of Potable Water," Chemistry and Industry, No. 17,
812-817 (1973).
46. Rambow, C.A., "Effect of Flow Rate on Carbon Filter Performance,"
J. Amer. Water Works Assoc., 55, No. 8, 1037-1043 (Aug. 1963).
47. Aly, O.M. and Faust, S.D., "Removal of 2,4-D Derivatives from Natural
Waters," J. Am. Water Works Assoc., 57, 221-230 (1965).
48. Morita, M., Nakamura, H. and Mimura, S., "Phthalic Acid Esters in
Water," Water Res., 8, No. 10, 781-788 (1974).
49. Schmidt, K., "Effectiveness of Standard Drinking Water Preparation
for Eliminating Pesticdes and Other Pollutants," Gas. Wasserfach Wasser-
Abwasser,, 115, No. 2, 72-76 (1974).
50. Vrochinskiy, K.K., "Removal of Pesticides from Water by Various
Water Preparation Methods," Gig. Sanit. (USSR), W, No. 1, 76-78
(1973).
51. Calderbank, A., "Environmental Considerations in the Development of
Diquat and Paraquat on Aquatic Herbicides," Outlook Agr., ]_, No. 2,
51-54 (1972).
-------
-A117-
52. Greve, P.A. and Wit., S.L., "Endosulfan in the Rhine River," J. Water
Poll. Contr. Federation, 43, No. 12, 2338-2348 (1971).
53. Stoltenberg, D.H., "How to Reclaim a Poisoned Lake," Pub. Works,
103. No. 3, 59-61 (1972).
54. Hansen, R.E., "Organics Removal of Mount Clements, Michigan," Presented
to Michigan Section AWWA, Ann Arbor, Michigan (Jan. 1977).
55. Giusti, D.M., Conway, R.A., and Lawson, C.T., "Activated Carbon
Adsorption of Petrochemicals, JWPCF, 45, No. 5, 947-965 (May 1974).
56. Juhola, A.J., "Regeneration of Activated Carbon, In: Papers and
Proceedings of a Water Research Association Conference at the
University of Reading, United Kingdom, Paper 9, 177-202, (Feb. 1974),
398 pp.
57. Smith, S.B., "Techniques of Activated Carbon Regeneration," In:
Proceedings of the Physical-Chemical Treatment, Activated Carbon
Adsorption in Pollution Control Technology Transfer Seminar,
Sponsored by Environment Canada and Pollution Control Association of
Canada, 1975 Report M 1000.01, Chemical Division, Westvaco Corporation,
Covington, Virginia, (1975), Mimeo, 50 pp.
58. DeJohn, P.B. and Hutchins, R.A., "Treatment of Dye Wastes with Granular
Activated Carbon," Textile Chemist and Colorist, 8^ No. 4, 34-38,
(April 1976).
59. Juntgen, H., "Phenomena of Activated-Carbon Regeneration," In:
Translation of Reports on Special Problems of Water Technology -
Volume 9 - Adsorption," Conference held in Karsruhe, Federal
Republic of Germany, 1975, p. 269-283, EPA-600/9-76-030, U.S.
Environmental Protection Agency, Cincinnati, Ohio (Dec. 1976), 454 pp.
60. Weissenhorn, F.J., "Testing of Activated Carbon Filters in Waterworks,"
In: Translation of Reports on Special Problems of Water Technology -
Volume 9 - Adsorption," Conference held in Karlsruhe, Federal Republic
of Germany, 1975, pp. 238-250, EPA 600/9-76-030, U.S. Environmental
Protection Agency, Cincinnati, Ohio (Dec. 1976), 454 pp.
61. Herzing, D.R., Snoeyink, V.L, and Wood, N.F. , "Activated Carbon
Adsorption of the Odorous Compounds 2-Methylisoborneol and Geosmin,
JAWWA, 69_, No. 4, 223-228 (April 1977).
62. Snoeyink, V.L., McCreary, J. and Murin, C.J., "Activated Carbon
Adsorption of Trace Organic Compounds," Final Report on U.S. EPA
Research Grant No. R803473, EPA-600/2-77-223, Cincinnati, Ohio,
March, 1978.
63. Stevens, A. A. and Symons, J.M., "Measurement of Trihalomethane and
Precusor Concentration Changes," JAWWA, 69, No. 10, 546-554 (Oct. 1977).
64. Love, O.T., Jr., Carswell, J.K., Miltner, R.J., and Symons, J.M.,
"Treatment for the Prevention or Removal of Trihalomethanes in
Drinking Water, JAWWA, In Press.
65. Symons, J.M., "Interim Treatment Guide for the Control of Chloroform
and Other Trihalomethanes," Water Supply Research Division, U.S.
Environmental Protection Agency, Cincinnati, Ohio (June 1976), mimeo.
48 pp. plus 4 Appendices.
-------
-A118-
66. Symons, J.M., et al., "National Organics Reconnaissance Survey for
Halogenated Organics in Drinking Water," JAWWA, 67, No. 11, Part 1,
634-647 (Nov. 1975), Update, 67_, No. 12, 708-709 (Dec. 1975).
67. Sylvia, A.E., Bancroft, D.A. and Miller, J.D., "Detection and
Measurement of Microorganics in Drinking Water," In: Proceedings
of the AWWA Water Quality Technology Conference, pp. XXVII-1 to
XXVII-8 (Dec. 1974), American Water Works Association, Denver, CO (1975).
68. Stanier, R.W., Adelberg, E.A. and Ingraham, J., The Microbial World, 4th
Edition, Prentice-Hall, Chapter 29, "Microbial Pathogenicity," p. 791, (1976)
69. Jorgensen, J.H., Lee, J.C. and Pahren, H.R., "Rapid Detection of Bacterial
Endotoxins in Drinking Water and Renovated Wastewater," Applied and
Environmental Microbiology, 32, No. 3, (Sept. 1976).
70. McCreary, J.J. and Snoeyink, V.L., "Granular Activated Carbon in
Water Treatment," J. Am. Water Works Assoc., 69, No. 8, 437-444,
(1977).
71. Guirguis, W.A., Melnyk, P.B. and Harris, J.P., "The Negative
Impact of Industrial Waste on Physical-Chemical Treatment,"
Presented at 31st Purdue Industrial Waste Conf., Lafayette, Indiana,
(May 1976).
72. Directo, L.S., Chen, C.-L., and Kugelman, I.J., "Pilot Plant Study
of Physical-Chemical Treatment," J. Water Poll. Control Fed., 49,
No. 10, 2081-2098 (1977).
73. Monscvitz, J.T. and Ainsworth, L.D., "Detection and Control of
Hydrogen Polysulfide in Water," Public Works, 101. 113 (1970).
74. Eberhardt, M. , "Experience with the Use of Biologically Effective
Activated Carbon," in Translation of Reports on Special Problems
of Water Technology, Vol. 9—Adsorption., H. Sontheimer, Ed.
EPA Report EPA 600/9-76-030, (Dec. 1976).
75. Sontheimer, H., "Biological Treatment of Surface Waters in Activated
Carbon Filters," presented at Seminar on Current Status of Wastewater
Treatment and Disinfection with Ozone, Cincinnati, Ohio, September 15,
1977, Intl. Ozone Inst., Cleveland, Ohio.
76. Kuhn, W., Sontheimer, H. and Kurz, R., "Use of Ozone and Chlorine
in Water Works in the Federal Republic of Germany," Presented at
Workshop on Ozone Chlorine Dioxide Oxidation Products of Organic
Materials, Cincinnati, Ohio, Nov. 1976. Intl. Ozone Inst.,
Cleveland, Ohio.
77. Guirguis, W.A., Jain, J.S., Hanna, Y.A., and Srivastava, P.K.,
"Ozone Application for Disinfection in the Westerly Advanced
Wastewater Treatment Facility," in Forum on Ozone Disinfection,
E.G. Fochtman, R.G. Rice and M. E. Browning, editors. Intl. Ozone
Inst., Cleveland, Ohio, pp. 363-381 (1976).
-------
-A119-
78. Guirguis, W.A., Cooper, T., Harris, J. and Ungar, A., "Improved
Performance of Activated Carbon by Preozonation," Presented at
49th Annual Conf. Water Poll. Control Fed., Minneapolis, Minn.,
October 1976.
79. Scheidtmann, W., "Investigations of the Optimization of Pretreatment
When Using Ozone," in Translation of Reports on Special Problems of
Water Technology, Vol. 9 - Adsorption, op. cit., pp. 98-111.
80. Schalekamp, M., "Use of Activated Carbon in the Treatment of Lake
Water," in Translation of Reports of Special Problems of Water
Technology, Vol. 9 - Adsorption, op. cit.,, pp. 128-159.
81. Van Lier, W.C., Graveland, A., Rook, J.J. and Schultink, L.J.,
"Experiences with Pilot Plant Activated Carbon Filters in Dutch
Waterworks," in Translation of Reports on Special Problems of Water
Technology, Vol. 9 - Adsorption, op. cit., pp. 160-181.
82. Sontheimer, H., "Considerations on the Optimization of Activated
Carbon Use in Waterworks," in Translation of Reports on Special
Problems of Water Technology, Vol. 9 - Adsorption, op. cit.,
pp. 208-214.
83. Van der Kooij, D., "Some Investigations into the Presence and Behaviour
of Bacteria in Activated Carbon Filters," in Translation of Reports
on Special Problems of Water Technology, Vol. 9 - Adsorption, op. cit.,
pp. 348-354.
84. Cornelia, C., and Versanne, D., "Le Role de 1'Ozone dans la Nitrification
Bacterienne de 1'Azote Ammoniacal — Cas de 1'Usine de la Chapelle
Banlieue Sud de Rouene (Seine Maritime), France," Presented at 3rd
Intl. Symp. on Ozone Technol., Paris, France, May 1977. Intl.
Ozone Int., Cleveland, Ohio.
85. Sontheimer, H., Heilker, E., Jekel, M., Nolte, H. and Vollmer, F.-H.,
"The 'Mulheim Process' - Experience with a New Process Scheme for
Treating Polluted Surface Waters," Submitted for publication in
J. Am. Water Works Assoc.
86. Wachs, A., Narkis, N., and Schneider, M., "Organic Matter Removal
from Effluents by Lime Treatment, Ozonation and Biologically Extended
Activated Carbon Treatment," Presented at 3rd Intl. Symp. on Ozone
Technology, Paris, France, May 1977, Intl. Ozone Inst., Cleveland, Ohio.
87. Hopf, W., "Versuche mit Aktivkohlen zur Aufbereitung des Dusseldorfer
Trinkwassers", Wass er/Abwasser, 101, No. 14, 330-336 (1960).
88. Sontheimer, H., Univ. Karlsruhe, Germany. Private Communication (1977).
89. Kuhn, W. and Fuchs, F., "Untersuchungen zur Bedeutung der
Organischen Chlorverbindungen und Ihrer Adsorbierbarkeit," Vom Wasser,
45, 217-232 (1975).
90. Kuhn, W. and Sontheimer, H., "Einige Untersuchungen zur Bestimmung
von Organischen Chlorverbindungen auf Aktivkohle," Vom Wasser, 4_1,
65-79 (1973).
-------
-A120-
91. Kuhn, W. and Sontheimer,H., "Einfluss Chemischer Umsetzungen auf die
Lage der Adsorptionsgleichgewichte an Aktivkohlen," Vom Wasser,
40, 115-123, (1973).
92. Kuhn, W., "Untersuchungen aur Bestimmung von Organischen
Chlorverbindungen auf Aktivkohle," Dissertation, Fak. f. Chemie-Ing.
Wesen, Univ. Karlsruhe, Fed. Rep. Germany (1974).
93. Wolfel, P. and Sontheimer, H., 1974, "Ein Neues Verfahren zur
Bestimmung von Organisch Gebundenem Kohlenstoff in Wasser Durch
Photochemische Oxidation," Vom Wasser, 43, 315-325 (1974).
94. Poggenburg, W., Wasserwerk Dusseldorf. Private Communication (1977).
95. Steiner IV, J. and Singley, J.E., "Methoxychlor Removal from Potable
Water," Engineering Report, Department of Environmental Engineering
Sciences, University of Florida, Gainesville, Florida (1977).
-------
APPENDIX B
ANALYTIC METHODOLOGY
FOR
MONITORING PILOT COLUMN TESTS
Written by
Alan A. Stevens
Reviewed by
Gordon G. Robeck
-------
APPENDIX B
TABLE OF CONTENTS
Introduction Bl
Low Molecular Weight Halogenated Organic Compounds Excluding
Trihalomethanes - Performance Criterion 1 Bl
Organic Carbon - Performance Criteria 2 and 3 B5
Analytic Method B5
Sampling B6
Terminal Summation Trihalomethanes - Maximum Contaminant Level
Requirement B7
General B7
Effect of Time B8
Maintenance of Chlorine Residual B9
Effect of Temperature BIO
Effect of pH BIO
Loss of Volatile Species BIO
Effect of Bromide or Iodide Contamination BIO
Effect of Precursor Contamination Bll
Procedure for Terminal Trihalomethane Determination Bll
Summary B13
Acknowledgments B13
References B14
-------
ANALYTIC METHODOLOGY FOR MONITORING PILOT COLUMN TESTS
INTRODUCTION
The methods needed to monitor the granular activated carbon treatment pilot
studies can be divided into three specific categories:
1. Low molecular weight halogenated organic compounds, excluding
trihalomethanes
2. Organic Carbon
3. Terminal "Summation Trihalomethane"
Some inherent flexibility exists in each of these categories with regard to
specific measurement methods, instruments selected, and so forth. This discussion
will attempt to address some of the options available and relate them as closely
as possible to present, proposed, or future U.S. EPA Standard Methods that are
ultimately needed to meet compliance monitoring requirements.
LOW MOLECULAR WEIGHT HALOGENATED ORGANIC COMPOUNDS EXCLUDING TRIHALOMETHANES -
PERFORMANCE CRITERION 1
This group of organic compounds has an operational definition. The intent is
to include such low boiling (low molecular weight) halogenated organics as carbon
tetrachloride, tri- and tetrachloroethylene, vinyl chloride, and others often
found in contaminanted drinking waters. This group is selected because of: 1)
relative ease of analyses, 2) they are common contaminants, 3) evidence suggests
that they break through granular activated carbon adsorption systems earlier than
higher molecular weight toxic materials and therefore as a monitor, place more
stringent operating requirements on the process giving a higher level of protection
to the consumer.
^"Summation Trihalomethanes" is the arithmetic sum of the concentration of the
individual species of trihalomethane found in a given sample. This parameter is
called "Total Trihalomethane" in the proposed Regulation.
-------
-B2-
EPA researchers have determined that over 30 purgeable halogenated organic
compounds can be detected and most resolved by use of the two gas chromatographic
columns under the conditions listed for use in the approved purge and trap trihalomethane
method and when the "multi-purpose" trap containing silica gel is used. Table I
lists these compounds in order of increasing retention time when a 0.2 percent
Carbowax 1500 on Carbopak C column is used.
TABLE I
COMPOUNDS DETERMINABLE ON 0.2% CARBOWAX 1500 ON CARBOPAK C
In Order of Increasing Retention Time
chloromethane
bromomethane
vinyl chloride
chloroethane
methylene chloride
1,1-dichloroethylene
bromochloromethane
1,1-dichloroethane
trans-1,2-dichloroethylene
chloroform
1,2-dichloroethane
1,1,1- trichloroethane
carbon tetrachloride
bromodichloromethane
1,2- dichloropropane — __
2,3-dichloro-l-propene
trans-1,3-dichloropropene
trichloroethylene
1,1,2- trichloroethane
dibromochloromethane —
unresolved
:unresolved
cis-1,3-dichloro-l-propene•
1,2-dibromoethane
2-bromo-l—chloropropane
bromoform— __
1,1,1,2-tetrachloroethane**'
tetrachloroethylene
1,1,2,2-tetrachloroethane
1,4-dichlorobutane
chlorobenzene
unresolved
-------
-B3-
A second column (N-octane on Porasil-c) changes the retention order (see
Table II), gives different separations, and allows detection of some organic
compounds with higher boiling points.
TABLE II
COMPOUNDS DETERMINABLE ON N-OCTANE ON PORASIL-C
In Order of Increasing Retention Time
j-unresolved
vinyl chloride ..
chloromethane —
bromomethane
1,1-dichloroethylene
chloroethane
trans-1,2 dichloroethylene
methylene chloride
carbon tetrachloride
chloroform— .
cis-1,2-dichloroethylene
1,1-dichloroethane
bromochloromethane
1,1,1-trichloroethane
trichloroethylene
bromodichloromethane
dibromomethane
tetrachloroethylene
1,2-dichloroethane
dibromochloromethane
trans-l,3-dichloro-l-properie
1,2-dichloropropane
cis-1,3-dichloro-l-propene
1,1,2-trichloroethane
2-bromo-l-chloropropane
chlorobenzene
1,2-dibromoethane
bromoform
1-chloro-l-hexene
chlorohexane
1,1,2,2-tetrachloroethane
pentachloroethane
£-chlorotoluene
m-dichlorobenzene
hexachloroethane
£-dichlorobenzene
..unresolved
•unresolved
•unresolved
unresolved
:unresolved
•unresolved
.unresolved
unresolved
1,4-dichlorobutane.
j>-dichlorobenzene
hexachlorobutadiene
1,2 ,4-trichlorobenzene
:unresolved
-------
-B4-
Therefore, the operational definition of those low molecular weight halogenated
compounds excluding trihalomethanes related to performance Criterion 1 is; those
compounds that can be detected by use of the U.S. EPA approved trihalomethane
method - purge and trap version. Note: Determining these compounds as part of the
granular activated carbon adsorption system process control program should not be a
large analytic burden when the water utility required to use granular activated
carbon treatment has selected the proper approach for the required measurement of
trihalomethane concentrations.
The purge and trap method is selected for this analysis mainly because gas
chromatographic conditions have not been well defined for resolution of as many
2
compounds by the liquid-liquid extraction method, and because of the difficulty of
solvent interference in the liquid-liquid extraction method with the detection of
many compounds eluting sooner than chloroform. This is not to say, however, that
these conditions could not be developed. Additionally, care must be taken in selecting
a particular purge-and-trap apparatus for introduction of the sample into the gas
chromatograph, as a well designed apparatus is needed to eventually accomplish resolutio
of the same lower boiling compounds (vinyl chloride, bromomethane, and so forth).
Details of this are to begin in a supplemental writeup extending the coverage of the
EPA approved trihalomethane method to this larger group of compounds.
In spite of the resolving power of the two columns used for the purge and trap
trihalomethane method, all possible purgeable organohalides and isomers cannot be
identified unambiguously. Identifications by dual column GC-specific halogen detection
are still presumptive. In order to avoid escalating analytical costs (e.g. - mass
spectrometric (MS) detection), these presumptive identifications are considered
sufficiently accurate and granular activated carbon adsorption treatment should be
adjusted to reduce any of these contaminants to less than 0.5 yg/& based on a
comparison with an analytical standard of the compound of presumed identity, "Performan
Criterion 1." Except in the case of some isomer differentiations, MS confirmation
of compound identity might be useful enough to warrant the effort and could ultimately
lead to a better designed treatment strategy.
-------
-B5-
ORGANIC CARBON - PERFORMANCE CRITERIA 2 and 3
Analytic Method
A method for the determination of organic carbon must include a step to
remove or adjust for the presence of inorganic carbonate. Instruments on the
market today, using manufacturers recommended procedures, do this adequately and
efficiently. Many procedures, however, effect the removal of carbonate by acidification
with mineral acid followed by exhaustive purging of carbon dioxide to waste. In
this process, a purgeable fraction of organic carbon (POC) is lost. Ideally, this
purgeable fraction should be included in any total organic carbon (TOC) measurement.
Research in WSRD laboratories, however, has shown that, in general, the POC concentration
is only a small fraction of the TOC concentration and little organic carbon (< 10
yg/&) would be lost by measuring only the non-purgeable organic fraction (NPOC).
Only occasionally have POC concentrations been found in the range of 0.1 - 0.3
mg/£. When the insignificance of POC can be demonstrated, only NPOC need be measured.
Note: For these cases NPOC determinations can be used to satisfy the TOC requirements
of "Performance Criteria 2 and 3." This can simplify the sampling requirements and
analysis and expand the options for methods and instruments that can be used.
Because of the requirements of the Performance Criteria 2 (measure a change of
0.5 mg/£) and the commonly found low concentrations of organic carbon in a granular
activated carbon column effluent (< 0.1 mg/£) the method used to measure organic
carbon must be more precise and accurate when compared to methods and instruments
used in the waste water treatment field, where concentrations of organic carbon are
often in the hundreds of milligrams per liter. Required precision and accuracy for
measurement of organic carbon in granular activated carbon column effluents therefore
must be in the range of 0.1 mg/£ + .05 mg/& or better. Although no EPA standard
procedures now exist for this purpose, instruments are available that can be operated
-------
-Be-
according to manufacturers instructions to achieve these goals. At least one of
these is capable of measuring both POC and NPOC independently.
The instrument (and method) must also be capable of measuring TOG concentrations
as high as those expected to occur in the granular activated carbon column influent
(2-10 mg/£). This is well within the range of instruments capable of the required
low level analysis.
Sampling
Because of the variability with time of concentrations of organic carbon in
actual treatment situations, determining what constitute a representative "baseline"
(TOG ) sample for all systems is difficult. Additionally, activated carbon fines
emanating from new granular activated carbon adsorbers have been observed to influence
organic carbon determinations. For these reasons, the first organic carbon measurements
are recommended to be performed after one week of adsorber operation and be performed
at least in triplicate twice during a 2-day period (six samples each on adsorber
influent and effluent). The average of effluent values not influenced by activated
carbon fines (activated carbon are not visible, or samples do not have a NPOC value
considerably higher than others of the set) is then the "baseline" (TOG ) effluent
concentration against which future effluent samples are compared, Performance
Criterion 2, and determines the initial organic carbon removal percentage, Performance
Criterion 3, when compared with a "baseline" influent concentration.
Sample containers should be the same as those used for sampling for trihalomethane
analysis when the POC fraction is to be measured. Headspace requirements are not
as important when the POC fraction has previously been shown to be insignificant.
-------
-B7-
TERMINAL SUMMATION* TRIHALOMETHANES - MAXIMUM CONTAMINANT LEVEL REQUIREMENT
General
A detailed method for determining Terminal Trihalomethane (Term THM) concentrations
3
has been described by Stevens and Symons. The principle of Term THM measurement
"Terminal Summation Trihalomethane" is the special case where the individual species
Term THM concentrations in the units yg/£ are arithmetically summed.
Because their formation is not instantaneous, trihalomethane (THM) concentrations
increase in the water as it flows through a water treatment plant (unless removed
during treatment) to reach some value higher than that which would be observed if
an analysis for THM species was performed immediately after sampling at the first
point of chlorination. Further, the consumer is likely to receive water with THM
concentrations higher than those leaving the plant because the reaction proceeds in
the distribution system. Additionally, not only are the concentrations of THM time
dependent, but the rate of the reaction is dependent on pH, precursor concentration,
nature of precursors, temperature and to some degree free chlorine concentration
early in the chlorination process. Finally, the ratio of chloroform to other trihalomethanes
is highly dependent on the bromide content of the source water.
The Term THM concentration is defined as the concentration of THM that occurs
at the termination of the measurement of this parameter. To measure Term THM
concentration, chlorine-precursor reaction conditions are selected according to the
treatment practiced at the particular plant being evaluated. In general, to
determine Term THM a sample of water is chlorinated under these plant conditions,
and chloroform and other THM species concentrations are determined after some
preselected holding period or periods.
*As noted previously "Total" is used in the proposed Regulation, rather than
"Summation" to mean the arithmetic sum of individual trihalomethane concentrations.
-------
-38-
Th e Term THM concentration is an important parameter for evaluating drinking
water quality because it is an estimate of the concentration of THM reaching the
consumer at various points (residence times) in the distribution system. Therefore,
determining the change with time in the Term THM concentration in the effluent of a
granular activated carbon adsorption system is a good estimate of whether or not
the effluent would meet the summation trihalomethane maximum contaminant level
required by the Interim Primary Drinking Water Regulations.
The selected conditions for the Term THM measurement must be the same as
those experienced at the water treatment plant or distribution system and must be
reproducible from sample to sample. Critical conditions to consider are time of
reaction (time elapsed before halting the halogenation reaction with a reducing
agent) maintenance of a chlorine residual, temperature, pH, prevention of loss of
the volatile products during the time of contamination and avoidance of contamination
of reagents.
Effect of Time
Although a single measurement of THM concentrations after a storage
period of several days in a bottle under appropriate conditions can give a
useful determination of the Term THM concentration for that specified time, much
more information can be gained from the reaction-rate curves obtained by plotting
THM concentrations vs. time. The rate curves obtained by periodic measurement of
THM concentrations in properly stored finished water can be used to estimate the
future THM concentrations at any given time after water leaves the treatment plant,
as required by the Interim Primary Drinking Water Regulations. This is particularly
important when the goal is to estimate ultimate consumer exposure to THM at different
points along the distribution system. The THM concentration-vs-time curve
-------
-B9-
is especially useful where the utility has a large variation in the time that water
is in various parts of the distribution system.
When analyzing a granular activated carbon adsorption system effluent, simply
measuring a single point on the THM growth curve such as one representing maximum
residence time in the distribution system would place a requirement on adsorber
performance more stringent than required under the Interim Primary Drinking Water
Regulations. To be consistent with that part of the Regulation, generation and use
of a complete THM growth rate curve is recommended to properly evaluate adsorber
performance. Specifically, the effluent water from the granular activated carbon
adsorption unit should be chlorinated in a manner consistent with treatment plant
practice where free chlorination is used (see below). Aliquots of that water are
then stored for various times from "0" (corresponding to clear well) to "T" (corresponding
to maximum time in distribution system) . The intermediate time samples _(T_, _T, and
~2 4"
so forth) determining the shape of the THM growth curve and, therefore, expected THM
concentrations at intermediate distribution system residence times.
A minimum of five points on the curve (0, T, and 3 in between) are selected
consistent with the sampling plan for THM MCL compliance monitoring. The average
Term Summation THM concentration obtained from those five values (of Term Summation
THM at respective times) is then used to judge adsorber performance.
Maintenance of Chlorine Residual
In conventional U.S. water treatment practice, maintenance of a free
chlorine residual throughout the distribution system often is recommended
or required. The continued reaction of precursor with chlorine to yield
trihalomethanes depends on the maintenance of a free chlorine residual.
Thus for evaluation of systems, where free chlorination is practiced, a
chlorine-residual measurement always must be performed at the time of
THM analysis to ensure that a free residual is present.
-------
-BlO-
Effect of Temperature
Upon chlorination of a natural water, approximately twice as much chloroform
can be formed in a given period of time at 25°C as is formed at 3°C. This
range of temperature is not uncommon, summer to winter, in U.S. surface
waters. A need for close temperature control during the determination of Term
THM concentration, therefore, is indicated. Because temperature largely is
controlled seasonally, this temperature effect must be taken into account, if
extrapolations to summer operating conditions from winter pilot studies and
vice versa are made.
Effect of pH
The trihalomethane formation rate has been shown to increase with an
increase in pH. Because pH is a factor determining rate of THM formation and
therefore Term THM values, the pH should be controlled near that found in the
distribution system.
Loss of Volatile Species.
To prevent misleading losses of trihalomethanes produced during the
reaction period, the reactions must be carried out in sealed, head-space-free
containers. Container materials should be all glass or glass with PTFE-lined
caps. Standard glass-stoppered reagent bottles filled to overflowing so as to
wet the stopper surface or the PTFE-septum-sealed serum vials, used for sampling
for Inst. THM determinations have been found suitable.
Effect of Bromide or Iodide Contamination
As mentioned earlier, bromide or iodide present in the water can cause formation
of THM species other than chloroform, as a result of first reacting with chlorine.
In the case of bromide, the relative amounts of THM species formed has been shown
to be highly dependent on the bromide content of the water and the chlorine dose,
presumably because these determine the ratio of bromine to chlorine available for
completing reactions. This change of product ratio, because bromine-containing
-------
-Bll-
species are heavier than chloroform, could cause a dramatic change in calculated
Terminal Summation Trihalomethane values.
Preliminary work indicates that equal amounts of bromine and chlorine substitution
as trihalomethanes would be expected if the original bromide concentration is as
little as 2 percent of the chlorine dose. Clearly, any bromide or iodide contamination
of reagents used will cause a different ratio of THM species to be formed than
would occur normally on chlorination of that water under plant conditions and must
be avoided.
Effect of Precursor Contamination
In the WSRD laboratory, distilled, deionized, activated carbon-filtered water
has been used for "blank" water for reagent preparation. At pH 7, the contribution
of precursor in reagents has been small. At higher pH, however, blank values tend
to be higher. Care should be taken to minimize volumes of reagents used in Term
THM measurements in order to avoid this contribution to the THM concentrations
obtained.
Procedure for Terminal Trihalomethane Determination
A test for Term THM concentrations can be standardized in approach, but the
conditions for sample treatment and storage will vary widely from system to system
depending upon distribution-system residence time, total clorine demand of the
sample, ambient temperature of the system, and pH of the finished water in the
particular system under investigation. These variables must be chosen to match
those in the system.
-------
-B12-
In work at the WSRD laboratory, a large (1-3 liter) sample of water is collected,
and the pH adjusted to that selected with an appropriate inorganic (e.g. phosphate
or borate) buffer. The final buffer strength is about 0.01 M, although the strength
is not critical as long as the desired pH is maintained. The sample then is chlorinated,
if needed, by the addition of a previously standardized chlorine or hypochlorite
solution. Sufficient chlorine is added at this time to maintain a free residual
for the duration of the test period.
Several sample bottles are filled and capped head-space-free, two bottles for
each point to be determined on the THM growth rate curve. For example, ten bottles
are needed for the five points on the curve to be determined. One of the zero-time
sample bottles contains sodium thiosulfate to immediately reduce the chlorine so
that the "0" time THM concentration is measured. The other zero-time sample has no
reducing agent and is used for measurement of the chlorine residual. This entire
sequence from sample collection to the capping of the bottles should be done as
quickly as possible to avoid loss of THM during the manipulations.
The samples, except zero-times, are stored at the selected temperature.
After the preselected times one sample bottle is opened and an aliquot is transferred
by pouring into a smaller bottle containing sodium thiosulfate to prevent further
reaction of precursor with chlorine. This smaller bottle is quickly sealed head-
space-free to await THM analysis. This measurement determines a THM concentration
for the respective time on the rate curve. A second bottle is opened at the same
preselected time and the chlorine residual is measured.
The actual measurement of THM concentrations that are arithmetically summed to
produce the "summation" or "total" value can be performed by use of either of the
two U.S. EPA approved procedures for that determination (purge-and-trap or liquid-
1 2
liquid extraction). ' The calculated sum of the THM concentrations by the EPA
approved procedure is the Terminal Summation Trihalomethane concentration for the
respective time period.
-------
-Bis-
Brief ly, for the purge and trap analysis, the sealed sample is brought to
25°C prior to opening in order to obtain reproducible purging efficiencies. A 5-m£
aliquot then is removed and transferred to a glass purging apparatus, wherein the
trihalomethanes are stripped from the aqueous phase by passage of a flow of helium
upward through the sample. The trihalomethanes stripped in this manner are collected
on a sorbant, porous polymer material contained in a stainless steel trap that is
placed in series with the purging device. The trihalomethanes are desorbed thermally
from the trapping material onto a gas chromatographic column. Finally, temperature-
programmed gas chromatography is carried out, and the concentrations of trihalomethanes
are measured by use of a halogen-specific detector. The liquid-liquid extraction
2
method involves the extraction of a small volume of water with an even smaller
volume of organic solvent, followed by gas chromatographic analysis of the extract
using an electron-capture detector.
Summary
Weekly analysis of the effluent from a granular activated carbon adsorption
system by this procedure will provide a good estimate of whether or not the delivered
drinking water will be in compliance with the trihalomethane maximum contaminant
limits as specified in the Interim Primary Drinking Water Regulations.
ACKNOWLEDGMENTS
The author and compiler wish to express their sincere thanks to Ms. Maura M.
Lilly, who typed this Appendix so quickly and accurately, such that the entire
Interim Guide could be finished on time.
-------
-B14-
REFERENCES
1. "The Analysis of Trihalomethanes in Finished Waters by the Purge and
Trap Method," U.S. Environmental Protection Agency, Environmental
Monitoring and Support Laboratory, Cincinnati, Ohio 45268, Sept. 9, 1977.
2. "The Analysis of Trihalomethanes in Drinking Water by Liquid/Liquid
Extraction," U.S. Environmental Protection Agency, Environmental Monitoring
and Support Laboratory, Cincinnati, Ohio 45268.
3. Stevens, A.A. and Symons, J.M., "Measurement of Trihalomethane and
Precursor Concentration Changes," JAWWA, 69, No. 10, 546-554
(Oct. 1977).
-------
APPENDIX C
DESIGN OF PILOT GRANULAR
ACTIVATED CARBON COLUMNS
Written by
0. Thomas Love, Jr.
and
Kenneth L. Kropp
Reviewed by
Gordon G. Robeck
-------
APPENDIX C
TABLE OF CONTENTS
Introduction Cl
Selection of Materials Cl
Adsorption Columns C2
Establishing Test Conditions C2
Operation C7
Parts List C9
Acknowledgments Cll
-------
DESIGN OF PILOT GRANULAR ACTIVATED CARBON COLUMNS
Introduction
The "Interim Treatment Guide for Controlling Organic Contaminants in
Drinking Water Using Granular Activated Carbon" describes two adsorption
schemes. The first, is to retrofit an existing water treatment plant by
replacing the filtration media with granular activated carbon, and the
second, called post-filter adsorption, requires separate contactors following
te filtration step. The ability of either adsorption scheme to reduce
the concentration of trace organics should be investigated on-site using
a small manageable system before expending large sums of money on a process
design that may give marginal performance. Trial experimentation (i.e.
pilot studies) permits selecting the appropriate concept through improved
engineering judgement. This Appendix describes the design and operation
of the EPA Water Supply Research Division's experimental adsorption systems
used in organic removal studies. This material is presented for guidance
as an example of one approach to gathering adsorption performance data.
Selection of Materials
One precaution taken by EPA in trace organic studies is to construct
pilot scale equipment with stainless steel, Teflon, and glass whenever
possible to minimize contamination from structural materials during
experimentaiton. Whether or not materials such as rubber, acrylic resin,
polyvinylchloride (PVC), polyethylene tubing, or similar products would
compromise the experimental results is not known. To avoid this possibility,
however, these materials should not be used if possible.
-------
-C2-
Adsorption Columns
Because of a limitation of available water, the WSRD pilot scale
adsorbers consist of 3.8 cm (1.5 inches) diameter glass columns 153 cm
(60 inches) in length. A schematic of the experimental system with
details on fabrication are shown in Figures 1-4. The columns are arranged
so that both modes, sand replacement and post-filtration adsorption,
can be studied simutaneously if desired. Granular activated carbon is
placed in column 1 (see Figure 1) to a depth allowable in the existing
filter boxes at the water treatment plant. Settled water should be
applied to this unit at an approach velocity similar to that in the
existing plant. Approximately 15 cm (6 inches) of graded gravel should
be placed in the bottom of the columns as an aid in distributing the
backwash water. Some type of surface scrubbing (air or water scour)
should be incorporated in the system because with granular activated
carbon, like sand, most of the filtration occurs in the upper few centimeters
of the bed and vigorous scouring insures adequate cleansing. The surface scrubber
shown in Figure 1 (see also Figures 2 and 4) is intended to slide on a Teflon
ferrule so that it can be located close to the granular activated carbon surface
for effective agitation. If headloss monitoring is desired, a U-tube
manometer or a sensitive pressure gage can be included in the design. If, however,
headloss is not monitored, a backwashing schedule such as that used within the existing
plant can be employed.
Establishing Test Conditions
Like disinfection, the process of adsorption is very dependent upon
contact time (see Figure 24, Appendix A, page A72). The term "empty bed
contact time" (EBCT) is commonly used to characterize this adsorption
variable and is calculated by dividing the volume of media "V" by the
hydraulic loading "Q" (i.e. — = EBCT). Assuming a filter box contains from
76 to 122 cm (30 to 48 inches) of media and the hydraulic loadings range
2
from 4 to 8 m/hr (2 to 4 gal/min/ft ) then the EBCT in a sand replacement
-------
-C3-
INFLUENT
(PUMPl
DETAIL B
•
a.
UI
O
SURFACE SCOUR
DETAIL B.
BACKWASH EFFLUENT
EFFLUENT
(V) COL. 1
TO WASTE
DETAIL A
r
T
D
1
I
J
(
t
\)
•
!
(;
i)
•
T
1
%
1
y
A
(j
z
0
K
u
o
Ui
u
X
^
z
0
* (
j- ^VGI
INFLUEN1
•-GRADED GRAVEL
COL. 4
COL. 3
COL. 2 ®
BACKWASH INFLUENT
FIGURE 1 PILOT GRANULAR ACTIVATED CARBON COLUMNS
-------
-G4-
ITEM 1
ITEM 2
ITEM 4
ITEM 5
ITEM 6
ITEM 7
Slip joint for
surface scrubber
-COLUMN
INFLUENT
ITEM 3
ITEM 8
ITEM 6
ITEM 10
— ITEM 9
FIGURE 2 DETAIL A
-------
-C5-
ITEM 3
I EDO
ITEM 5
ITEM 6
ITEM 7
ITEM 8
ITEM 6
ITEM 10
— ITEM 9
FIGURE 3 DETAIL B
-------
-C6-
1/8" STAINLESS STEEL TUBING
"316"
SLOTS OR HOLES
Q^ END OF TUBING CRIMPED
CLOSED
FIGURE 4 SURFACE SCOUR
-------
-C7-
mode could typically vary between 1.5 to 15 minutes, with an average of 7 to 8
minutes.
The additional columns shown on Figure 1 can be operated concurrently in a
post-filter adsorption mode to examine the effects of longer empty bed contact
times. For example, columns 2, 3, and 4 (Figure 1) can be charged with granular
activated carbon, exposed to filtered rather than settled water, and sampled in
series to monitor the breakthrough (wave front) of specific organics ("Performance
Criterion 1") and the total organic carbon ("Performance Criterion 2 and 3"). On
the other hand an investigator may desire to use column 2 as a sand filter receiving
settled water and columns 3 and 4 as adsorbers. Another column in series would
further increase the EBCT. Additional columns and the appropriate plumbing modifications
would allow a utility to investigate in parallel, rather than in sequence, performance
of different brands of commercially available granular activated carbon, should
that be desirable. Finally, the choice as to whether or not to apply disinfected
water to the adsorber will depend on how the final treatment scheme will be arranged.
The engineer should have some experimental "breadth" to select the most satisfactory
adsorption design.
Operation
The granular activated carbon must be wetted before it is put into service.
This is accomplished by backwashing the material similar to the initial steps in
using any granular media. Gently tapping or bumping the columns might be necessary
to insure that the media is wetted. Sufficient freeboard should exist to permit 50
percent bed expansion during backwash. The frequency of backwashing during the
experimental study will vary, depending upon the same factors influencing full-
scale filtration (e.g., headloss, turbidity, carryover floe from settling, etc.)
Details on monitoring the system are given in Appendix B, however, a routine operation
schedule would include the following:
-------
-C8-
Parameter Frequency
Flow Adjustment, Q Daily
Temperature, pH, Turbidity Investigator's discretion —
Based on variability of ap'plied water
Low Molecular weight
halogenated organic compounds Weekly
Organic Carbon Weekly
Terminal Summation THM Weekly
At most field installations no restriction in available water volume would
exist as it does in the WSRD pilot plant. Therefore larger, at least 10 cm
(4 inch) diameter columns are recommended for use in the treatability studies
proposed in the Interim Treatment Guide (see pages 23-26.)
The EPA Water Supply Research Division is currently designing
an experimental adsorption system using 10 cm (4 inch) diameter glass columns. The
system will be installed in at least one existing water treatment plant, and the
experimental results along with the problems encountered with pumps, maintenance,
and operation will be reported on when available. Details may be obtained by
writing to the Director, Water Supply Research Division, Municipal Environmental
Research Laboratory, 26 West St. Clair Street, Cincinnati, Ohio 45268.
-------
ITEM NUMBER
1
2
3
4
5
9
10
V
R
PUMP
C-9
PARTS LIST
DESCRIPTION* QUANTITY
STAINLESS STEEL MALE CONNECTOR 3 EACH
1/8" TUBE x 1/4" PIPE THREAD
BORED THRU WITH TEFLON FERRULES
(EXAMPLE - SWAGELOK FITTING
# SS-200-1-4-BT WITH TEFLON
FERRULES)
1/4" STAINLESS STEEL PIPE TEE 3 EACH
STAINLESS STEEL MALE BRANCH TEE 8 EACH
1/4" TUBE x 1/4" PIPE THREAD
(EXAMPLE - SWAGELOK FITTING
# SS-400-3-4TTM)
1/4" STAINLESS STEEL CLOSE NIPPLE 3 EACH
1/2" x 1/4" STAINLESS STEEL REDUCING 8 EACH
BUSHING
1-1/2" CORNING CONICAL FLANGE 16 EACH
STYLE 1, ALUMINUM, # 72-9061
1/2" TEFLON SHEET SHAPED LIKE THE 8 EACH
FLANGE IN ITEM 6, BOLT HOLES
CLEARENCE DRILLED FOR 5/16 X 18
BOLTS, CENTER DRILLED AND TAPPED
FOR 1/2" PIPE THREAD
1-1/2" GASKET, STYLE 1-2, SOLID 8 EACH
TFE, TYPE T, CORNING NUMBER 72-9255
1-1/2" CORNING CONICAL MOLDED 8 EACH
INSERT (HARD), # 72-9057
1-1/2" X 72" CORNING PYREX 4 EACH
CONICAL PROCESS GLASS PIPE
# 72-7501
ALUMINUM 5/16 X 18 x 2" BOLTS W/NUTS 24 EACH
STAINLESS STEEL FORGED BODY SHUT-OFF 17 EACH
VALVE WITH VEE TYPE STEM AND 1/4"
TUBE FITTINGS (EXAMPLE - WHITEY
VALVE # SS-1VS4)
STAINLESS STEEL FORGED BODY REGULATING 2 EACH
VALVE WITH 1/4" TUBE FITTINGS
(EXAMPLE - WHITEY VALVE # SS-1RS4)
STAINLESS STEEL AND TEFLON GEAR PUMP 2 EACH
WITH MAGNETIC COUPLING AND INTERNAL
BY-PASS (EXAMPLE - MICROPUMP # 12-50-316)
* MENTION OF COMMERCIAL PRODUCTS DOES NOT CONSTITUTE
ENDORSEMENT BY USEPA.
-------
C-10
PARTS LIST (CONT.)
ITEM NUMBER DESCRIPTION* QUANTITY
STAINLESS STEEL MALE CONNECTOR 4 EACH
1/4" TUBE x 1/8" PIPE THREAD FOR
USE WITH MICROPUMPS (EXAMPLE -
SWAGELOK FITTING # SS-400-1-2)
STAINLESS STEEL 1/4" UNION TEE 8 EACH
(EXAMPLE - SWAGELOK FITTING
# SS-400-3)
STAINLESS STEEL TUBE REDUCER 3 EACH
1/8" x 1/4" WITH 1/8" TEFLON
FERRULES (EXAMPLE - SWAGELOK
FITTING # SS-200-R-4 WITH
TEFLON FERRULES) TO BE USED
WITH THE SHUT-OFF VALVES ON THE
SURFACE SCOUR LINE.
1/8" STAINLESS STEEL TUBING "316" 20 FEET
FOR SURFACE SCOUR. SHAPED AS
IN FIGURE 4 AND PLACED THRU
ITEM 1 (DETAIL A) AND CONNECTED
TO THE SURFACE SCOUR VALVE
USING THE STAINLESS STEEL TUBE
REDUCER.
1/4" STAINLESS STEEL TUBING "316" 60 FEET
FOR ALL CONNECTIONS OTHER THAN
SURFACE SCOUR.
1/2" TEFLON TAPE USED FOR SEALING 2 SPOOLS
ALL PIPE THREAD CONNECTIONS
* MENTION OF COMMERCIAL PRODUCTS DOES NOT CONSTITUTE
ENDORSEMENT BY USEPA
-------
-Cll-
Acknowledgments
The authors wish to acknowledge the assistance of Ms. Maura M. Lilly
who typed Appendix C.
»US GOVERNMENT PRINTING OFFICE 1978—757-140/6652
-------
-'4
-------