-RESEARCH

& TECHNOLOGY

II

Using Powdered Activated Carbon:

A Critical Review
Issam N. Najm, Vernon L. Snoeyink, Benjamin
and Jeffrey Q. Adams

W. Lykins Jr.,

Because the performance of powdered activated carbon (PAC) for uses other than taste and
odor control is poorly documented, the purpose of this article is to critically review uses that
have heen reported and to analyze means of employing PAC more efficiently. The extent of
adsorption of synthetic organic chemicals on PAC is strongly dependent on the type of
compound being removed. The reported removals of trihalomethanes and trihalomethane
precursors by PAC range from poor to very good. In selecting the point of addition of PAC,
consideration must be given to the degree of mixing, the contact time between the PAC and
the water, the PAC residence time, and the minimization of interference of adsorption by
treatment chemicals. One of the main advantages of PAC is its low capital cost.
Synthetic and natural organic contaminants are often found in drinking
water sources. Categories of these compounds include taste- and odor- (T&O)
causing compounds, synthetic organic
chemicals (SOCs), pesticides, herbicides,
color, and trihalomethane
(THM) precursors. The low removals of many
undesirable compounds by coagulation,
sedimentation, filtration, and chemical

oxidation require that an adsorption

process be used. Granular activated carbon (GAC) and powdered activated carbon (PAC) can both be used to remove
organic compounds from water. Powdered activated carbon is widely used in
the United States for T&O removal,l~z
and GAC filter-adsorbers arecommonly
used for the same purpose.3 Granular
activated carbon has been the adsorbent
TABLE

Manufacturersspecifications
Parameter

of some
PAC

Iodine number--mg/g
Molasses decolorizing
index
Moisture
as packed--percent
Apparent
density-g/cmg
Ash content--percenf
Passing 100 mesh-@rent
Passing 200 mesh--percent
Passing 325 mesh-percent

l*

of Addition

PAC45

600

z
0.64

3
0.54
6

5
0.74

900
14
10
0.38
3-5
95-100
85-95
65-85

0?8
3-5
95-100
85-95
65-85

of addition

of PAC

disadvantages

time, good mixing

Rapid mix

Good mixing
reasonable

during
contact

JANUARY

99
97
90

98

the

PAC5**

PAC6tt

1,000
18

550
4
0.50
99
ii

Fla.

of different

Advantages

Long contact

preceding

PACs

PAC3$

Intake

Filter inlet
Slurry contactor
rapid mix

available

1,199

TABLE
Point

1
commercially

PAC2t

:;
90

and

A full report of this project, published

under the title Control of Organic Compounds With Powdered Activated Carbon
(catalognumber90581).
isavailablefrom
the American Water W&ks Association
Research Foundation, 66I66 W. Quincy
Ave., Denver, CO 80235. ?_s,

800

*Aqua-Nuchar,
Westvaco,
Covington.
W.Va.
tWPH, Calgon Corp., Pittsburgh,
Pa.
*Aqua, Westvaco,
Covington.
W.Va.
Nuchar
S-A, Westvaco,
Covington,
W.Va.
**Nuchar
SA-20, Westvaco,
Covington,
W.Va.
ttHydrodarco
B, American
Norit, Jacksonville,

Advantages

of choice for removing SOCs, including

volatile organic chemicals (VOCs). There
are inherent advantages to applying
activated carbon in fixed beds, including
lower carbon usage rates for many
applications and ease of spent carbon
regeneration. However, PAC has the
advantages of being a cheaper material
and of requiring minimal capital expenditure for feeding and contacting
equipment; also, it can be applied only
when needed.

rapid mix and flocculation,

time

Efficient
use of PAC
Excellent
mixing for the design contact time, no
interference
by coagulants,
additional
contact time
possible during flocculation
and sedimentation

points

Disadvantages

Some substances
may be adsorbed that would otherwise
probably be removed by coagulation,
thus increasing
carbonusage rate (this still needs to be demonstrated).
Possible reduction
in rate of adsorption
because of interference
by coagulants,
contact time may be too short for equilibrium
to be reached for some contaminants,
some competition
may
occur from molecules that would otherwise
be removed by
coagulation.
Possible loss of PAC to the clearwell
and distribution
system
A new basin and mixer may have to be installed;
some
competition
may occur from molecules
that may otherwise
be removed by coagulation.

1991

ISSAM

Copyright (C) 1991 American Water Works Association

N. NAIM

ET AL

65

TABLE

Isotherms for selected oraanic contaminants

T
t
Compound

Water+

Alachlor

DDW
GW

Aldicarb
Atrazine

DsDw
SW
DDW

Benzene

MGDw

Bromodichloromethane
Bromoform
Carbofuran

MDW
MDW
DDW
SW
MDW
MDW
DDW
SW
MDW

Carbon tetrachloride
Chlorobenzene

Chloroform
Cyclohexane
Dibromochloromethane
Dibromochloropropane

cis-1.2.dichloroethylene

truns-1,2-dichloroethylene
2.4.Dichlorophenol
1,2-Dichloropropane

iii?.
SW
GW
SW
DDW
DD

Dieldrin
Endrin
2,4,5-T ester
Ethylbenzene

Ethylene
chloride
Ethylenedibromide
Geosmin
Heptachlor
Heptachlor
epoxide
Hexachlorobutadiene

::2
0.15
0.12
0.62
0.63
23.3
13.4
0.3
0.4
3.0
5.0

DDW
MDW
DDW
SW
DDW
DDW
DDW
DDW
MDW

16.2
4.5
0.0011
0.9
0.77
13.5
0.18
1.78
11.92
4.64

of PAC

AND

0.10
6.61
3.84
19.1

ii3

The primary characteristic of PAC

that differentiates
it from GAC is its
particle size. Typically, 65-95 percent of
commercially
available PAC passes
through a 325-mesh (44-pm) sieve. For
comparison, Kruithof
et a14 gave the
particle size distribution of two powdered
RESEARCH

0%

L!i?W
MDW

In general, the performance of PAC in

drinking water treatment for uses other
than T&O control is poorly documented.
The purpose of this article is to critically
review those uses that have been reported and to analyze means of using
PAC more efficiently. With this information, it should be possible to determine more easily those situations in
which PAC can be used more effectively
than GAC.

66

80.2
62
10.5
8.4
4.52
38.2
24.4
1.4E-9
3.57
0.10
0.65
16.2
11.9
0.12
0.11
9.0

23.9
MDW
DDW
SW
DDW
SW
DDW
SW

o-Dichlorobenzene

Characteristics

Kt

PAC
c, = 100

l/n

10 rg/L

0.26
0.33
0.38
0.4
0.41
0.29
0.36

0.6
0.7
3.6
4.3
7.7
1.2
1.6
>lOO.OOO
9.6
156.4
20.5
2.2
3.3
108.4
85.7
4.5
4.4
433.6
2.4
105.7
4.2
8.1
2.0
2.8
115.7
137.4
176
51.5
58.2
2.0

0s
0:76
El
0:36
0.84
0.98
0.35
0.31
0.84
0.2
0.93
0.51
0.46
0.38
0.58
0.59
::ii
0.45
0.39
0.29
0.32
0.59
0.48
E
0.79
0.79
0.29
0.53
1.5
0.46
0.54
0.39
0.83
0.92
0.75
0.63

7;
74.5
2::
6Y5
2.8
25j:7
34.7
33.7
2.7
74.0
6.1
1.34
4.5

Dose-mg/L
co = 10 cg/I

i&/L
C,=
1 M/L
1.2
1.6
9.4
11.8
21.9
2.6
4.0
>lOO,OOO
27.7
990
152.3
6.1
8.3
825
900
11.0
10.0
3.300
4.14
990
15.0
25.8
5.2
11.8
495
660
825
160
157
4.2
7.4
330
247.5
33
19.8
28.5
430.4
6.11
902ko

ii0

128.6
7.33
550
55.6
9.6
21.3

activated carbons available in the Netherlands that show 23-40 percent by

mass smaller than 10 pm diameter and
lo-18 percent larger than 74 pm. The
manufacturers specifications, including
particle size distribution, of several PACs
commercially available in the United
States are summarized in Table 1. The
particle size distribution
is important
because the smaller PAC particles adsorb
organic compounds more rapidly than
larger particles.5
Powdered activated carbon is made
from a variety of materials including
wood, lignite, and coal. Its apparent
density, ranging from 0.36 to 0.74 g/cm3
(23 to 46 lb/cu ft), depends on the type of
material and the manufacturing process.
Iodine number and molasses number are
often used to characterize PAC (Table 1).

C,=
1 MI/L
0.1
0.14
0.86
1::
0.24
0.37
>lOO,OOO
2.5
90
13.8
0.6
0.76
8?8
2
309
0.4
90
i::
0.5
1.1
ti
75
14.5
14.3
0.4
0.67
22::
3
1.8
2.5
39.1
0.56
8.:82
10
11.7
0.67
50

C,=
Timet
14
18
11
13
12

days
days
days
days
days

lh
15 min
lh
lh
7 days
7 days
lh
lh
11 days
10 days
lh
15 min
lh
8 days
8 days
11 days
10 days
3 days
31 days
15 days
10 days
10 days
7 days
30 min
3 days
10 days

lh
15 min
3 days
lh
4 days
10 days
5 days
5 days

C%
1.90

lh

Ranges
512-0.6
477-0.2
46-2.6
274-1.3
171-19
785-4.2
317-0.6
8,000-3.000
iO,OOO-3,OOt
57-1.3
90-2
111-2.9
365-2.2
120-2.5
17,500-1,4Of
732-15.4
401-14.1
75-5
io,OOO-3,001
40-1.5
69.2-0.2
85.8-0.2
1.056-14.4
293-59
615-5.5
230-19
214-9.7
415-13.5
406-59
300-l
600-50
196-4.1
126-4.6

!2,500-2,100
io,oOO-3,OuC
565-5.6
65-17
123-0.2
139-1.1
0.05-2.0
0008-0.05
40-0.16
30-0.04
117-1.3

For example, the AWWA standard for

PAC specifies a minimum iodine number
of 500.

Factors affecting

performance

of PAC

Several design and operational parameters affect the performance of PAC

for drinkingwater
treatment. Important
criteria for selecting the point with addition of PAC include (1) the provision of
good mixing or good contact between the
PAC and all the water being treated, (2)
sufficient timeof contact with adsorption
of the contaminant, (3) minimal interference by treatment chemicals with adsorption on PAC, and (4) no degradation
of finished water quality.
The PAC must be added in a way that
ensures its contact with all of the flow.
Addition at locations other than those

TECHNOLOGY

Copyright (C) 1991 American Water Works Association

JOURNAL

AWWA

TABLE 3, continued
6iothems for selected organic contaminants

PAC

Compound
Lindane

Water

DDW
DDW
:z
DDW

Methoxychlor

Methylisoborneol
Parathion
PCB (Aroclor

1254)

Pentachlorophenol

Phenol
Simazine
Styrene
Toluene

TZ
DDW
DDW
DDW
DDW
HA
MDW
DDW
SW
MDW
DDW
GW
DDW
DDW
GW

Toxaphene
Trichloroethylene

2,4,6-Trichlorophenol
m-Xylene
o-Xylene

l/n

4.5
15.4
0.31
14.5
10.0
28.0**
5.83it
5.03
169.3
3.6
11.46
2.7
8.24
42.4
14.4
0.62
30.8
22.3
12.1
15.9
4.5

8%

DDw
HA
MDW
L
DDW
GW
DDW

5.725
2.97
0.66
1.35
0.9
1.0

GF
DDW
DDW

:dp3
21.2
1.3
20.13

DDW
GW
MDW
DDW
GW

p-Xylene

K-f

2
28.15
11.1
7.6

0.74
0.43
0.34
0.39
0.36
0.17
0.46
0.5
1.65
0.26
1.03
0.99
0.42
0.34
0.36
0.51
0.23
0.31
0.48
0.28
0.45
0.37
0.45
0.74
0.81
0.50
0.51
0.51
0.9
0.9
0.32
0.3
0.75
0.27
0.47
0.47
0.16
0.42
0.46

*MDW-mineralized
distilled water, DDW-distilled
deionized
(5 mg/L as TOC)
tFreundlich
K values are listed for C in pg/L and Q in mg/g.
SCIontact time in the isotherm
bottle
Range of remaining
aqueous concentration
in pg/L
**co = 5,000 pg/L
ttc,=
l,~O /a/L

listed in Table 2 may not achieve this

objective.
Sufficient
time of contact is also
necessary, and the time required is an
important function of the characteristics
and concentration of the molecule to be
adsorbed.7 In the absence of competition
and coagulation floe interference, 15 min
is sufficient time for molecules such as
dimethylphenol (molecular weight = 122,
C0=90mg/L,carbondose=250mg/L)to
equilibrate with a 325-mesh (44-pm-diameter) particle if adequate mixing is
used.8 As the molecular size increases,
the rate of diffusion into the pores of the
PAC particle decreases. For example,
rhodamine B dye (molecular weight =
422) requires approximately 5 h to come
to equilibrium,
a lO,OOO-molecularweight fulvic acid requires 17 h, and a
JANUARY

10 fig/L
3.6
2.2
132.7
2.5
3.8
2.16
5.4

Dose-mg/

P&?/L
c,=
1 /ML

c, = 10 fig/1 L

c,=

c,=
1 la/L

22
6.4
319.4
6.8
9.9
3.53
17.0

2.0
0.6
29
0.6
0.9
0.32
1.5

Time$

Ranges

6 days

170-2.1

7 days

149-0.6
460-0.016
l,OOO-100
500-40
0.04-5
0.02-0.04

5 days
5 days
13.7
0.73
3.4
4.2
1.0
2.7
44.8
1.7
2.0
::i
7.1
6.1
3.7
2.9
4.7
43.1
20.6
30.9
11.3
11.3
4.2
2.13
12.3
2.4
4.3
3.1
2.2
3.1
4.1
water,

GW-groundwater.

27.5
8.6
36.7
12.0
2.3
6.9
160
3.2
4.4
8.2
6.22
22
15.7
11.4
17.3
33.33
150
73.3
110.0
99.0
99.0
9.6
4.7
76
5.0
13.9
10
::9
13

2.5
0.8
3.33
1.1
0.2
0.6
14.5
0.3
0.4
0.7
0.57
2
1.4
1.0
1.57
3.0
13.6
6.67
10.0
9.0
9.0
0.87
0.4
6.9
0.45

28 days
28 days
lh
7 days
7 days
lh

12 days
15 min
4 days
15 days
30 min
28 days
28 days
lh
7 days
7 days

7 days
7 days
3 days
15 min
8 days
10 days
lh
9 days
15 days

A:9
0.3
0.8
1.2

SW-surface

50,000-molecular-weight
humic acid requires about two days. The adsorption
kinetics and equilibrium capacity depend
on the type of carbon used, and, therefore,
these values should only be taken as
rough estimates. If insufficient time is
allowed for equilibration, more PAC must
be used to achieve the desired removal.
The effect of PAC particle size on
required contact time in the absence of
floe interference was shown by Najm et
al5for a continuous stirred tank reactor.
In I5 min, 500 pg/L trichlorophenol was
reduced to 25 pg/L by 16pm-diameter
PAC but only to 275 pg/L by loo-pmdiameter PAC. The composite PAC
sample, with an average diameter of
40 pm, reduced the concentration to 180
pg/L. In no case was equilibrium
achieved in 15 min, and the concentration

water,

L
: CW -tap

water,

1Reference

0.17-0.011
0.8-0.058
8,180-220
3.408-0.4
322-5.3
,3,360-2,150
16.5-0.02
177-0.2
148-11.8
5~0,000-3,000
104-2.3
672-2.5
180-14
0.5-0.064
2.2-0.32
63-0.15
31-2.88
32-3.9
l,OOO-6
1 ,OOO-6
350-10
400-4
166-5.6
5 o,oOO-3,000
1,795-6.9
522-10.5
1 2,000-1,500
32.7-1.6
232-2.9

HI4-humic

2:
iii
28
74
74
27
27

2:
:7
ii

ii
38
28
72
28
28
73
28
28
47
12
12
23
23
12
12
28
;:
28
47
28
28

alcid

in the test with the composite sample

was still decreasing after 120 min.

Means of applying PAC

Powdered activated carbon can be fed
as a powder using dry feed machines or
as a slurry using metering pumps.g In a
conventional treatment plant, the common points of PAC addition are the plant
intake, rapid mix, and filter influent.
Another point of addition that should be
considered, although it is not commonly
used, is a continuous-flow
slurry contactor that precedes the rapid mix. The
PAC can be mixed intensely with the
water in the absence of floe to obtain
rapid adsorption kinetics and then incorporated into the floe in the rapid mix
for subsequent removal by sedimentation
andfiltration.5Table2summarizes
some

1991

ISSAM

Copyright (C) 1991 American Water Works Association

N. NNM

ET AL

67

of the important advantages and disadvantages of PAC addition at each of

these points.
When PAC is added at the rapid mix,
incorporation of PAC into the coagulant
floe particles may reduce the rate of
adsorption.@IzGauntlett
and PackhamO
conducted jar tests showing that the
removal rate of chlorophenol by PAC in
the absence of alum addition was most
rapid and that the addition of PAC after
alum addition gave a better rate of removal than when applied just before
alum. Apparently, PAC added after alum
floe had formed adhered to the outer
surface of the floe rather than being
incorporated into the floe. Gauntlett and
Packham argued that the reason for the
reduction in removal rate is that the
adsorbate must diffuse through the part
of the floe surrounding the PAC particle
and then into the particle itself in order
to be adsorbed. Najm et al,** however,
found little reduction in the rate of
adsorption of trichlorophenol on PAC in
spite of the incorporation of the carbon
particles into coagulant floes.
Addition of PAC at the intake has the
advantage of providing extra contact
time, but there is the possibility of the
PAC adsorbing many compounds that
would otherwise be removed by coagulation, flocculation,
and sedimentation.
On-site tests are recommended to determine whether one factor outweighs
the other.
Addition of PAC just before the filter
is advantageous because the PAC can be
retained in the filter and kept in contact
with the water longer, thereby better
using its capacity. The average PAC
residence time is equal to one half of the
time between two successive backwashings, assuming that PAC is continuously
added to the influent water. However,
the PAC must be added carefully to
avoid its penetration into thedistribution
system. The maximum dosage of PAC is
limited by the ability of the filter to
retain the PAC and by the rate of head
loss buildup in the filter, which is expected to increase as PAC dosage increases. Sontheimer et aI3 reported a
lo-20 percent reduction in filter run
time as a result of the addition of PAC.
Dougharty and MorrisI reported that
filter velocities >8-9 m/h resulted in a
breakthrough of PAC when it was added
at a dose of 60 mg/L. In a direct filtration
pilot-plant study, Gifford et al*5 reported
a 40 percent reduction in filter run time
as a result of the addition of 30 mg
PAUL. With a hydraulic loading of 4.9
m/h, no breakthrough
of PAC was
observed in the filter effluent.
Sontheimerll
suggested adding a
separate PAC reactor between the sedimentation basin and the filter to increase
the time of contact and take advantage
of coagulation and sedimentation
for
eliminating competing organics to the
68

RESEARCH

TABLE
PAC

dose required

to reduce

the pesticide

level in distilled

and

PAC
co=
Pesticide

Method

Parathion
2,4,5-T

ester

Endrin

JT

Lindane

JT

Dieldrin

JT
P
*PAC dosage in jar test in which
contact time of 1 h
+PAC dosage in the plant in which
and activated
carbon

Percentage

removals

i.5
14
1.8
11
2
29
3
18
pesticide is removed

of THMprecursors
cornDared
with

Mode 1
PAC addition, coagulationflocculation-sedimentation,
filtration
Mode 2
PAC addition, coagulationflocculation-sedimentation,
chlorination,
filtration
Mode 3
Coagulation-flocculationsedimentation,
PAC
addition, chlorination,
filtration
*Same

treatment

but with

from

TABLE

of PAC

dosage

distilled

by conventional

of PAC
control*

to New

50 mg

pg/L

treatment

water

Orleans

Removal-

lercent

PAUL

500

mg PAUL

31

84

41

77

27

90

no PA Ca ddition

on the removal

of chloroform

Dose-wag/L
0

10
25
50
100

Loo0

3OC, pH = 10.5-11.2,

c, = 0.05

water

Precursor

precursors

Chloroform
PAC

)&g/L

0.5
0.6
0.9
1.1
1.5
3
3
1.3
i.5
11
23
1.1
2
6
9
1.1
1.7
7
12
by PAC alone, with a

water

river

with the addition

those of an ag2ated

TABLE
Effect

from

5 mg PAUL

Sequence

c, = 0.1

5
10
17
44
14
126
12
70
isi

is removed

THM

Treatment

water

Dose-mg/L

pg/L

17
pesticide

river

c, = 1.0 )&g/L

c, = 0.1

2.5

JF+

Miami

10 pg/L

c, = 1.0 pg/L

JT*

Little

Clzresidual

= 2-3 mg/L,

rcg/L
236
215
170
127
85
48
stored

from

Formation

Ohio

River

water*

Potential
Reduction--percent

I
three

0
9
28
46
GA

days

AND TECHNOLOGY

Copyright (C) 1991 American Water Works Association

JOURNAL

AWWA

Effect 4ofPAC on trihalomethane

T
PAC
Treatment
Settled
Settled
Settled
Cz

*NF-none

(THMFP)

concentration
Removal
of TTHM

CHCl,

TTHM
rmol/L

NF*
27

NF
0.37

0.31
0.36
0.32
0.29
0.22
0.18

16
2
14
22
41
51

+ Cl,
+ PAC +
2
4
8
20
50
100

5
potential

Trihalomethanes-us/L

Dose

mg/L

Type

water
water
water

TABLE
formation

22
16.4
20
16
11
9

15.1
16.4
15.8
16.9
13.0
9.5

8.0
10.2
9.4
12.2
10.0
8.8

percent

found
0.35

1. Preliminary
system size

cost estimates for PAC treatment

Figure

as a function

of dosage and

TABLE
8
Reduction of THMFP and TOC by coagulation and adso@tion on 50 mg/L of two types of PAC in water
from Daytona Beach Aquifer, Fla., and Ilwaco Reservoir, Wash. (after Amy and Chadiks7)
After

Coagulant
Type

Coagulant
Dose

mg/L

After

Coagulation

THMFP

TOC

b&L

mg/L

Carbon

THMFP:TOC

&mg
Daytona

0
;o

Al
Al
Al
Fe
Fe

30
15
30

667
509
418
383
548
401

9.2
6.88
6.54
4.5
7.65
6.76

72.5
74
::8
71:6
59.3

1,087
371
579

8.04
3.33
5.34

135.2
111.4
108.4

894
350
245

6.9
3.0
2.27

129.6
116.7
107.9

Ilwaco

JANUARY

A;

2.5
1.25

Fe
Fe
Fe

2.5
5
7.5

THMFP

TOC

mg/L

P&?/L
Beach

Aquifer,

Adsorption

B
THMFP:TOC

Carbon

THMFP

TOC

THMFP:TOC

l&L

mg/L

Irg/w

Wmg

Fla.
7.96
6.13
4.98
4.22
6.21
5.41

54.4
48.4
4.2
51.7
55.6
47.5

369
309
242
265
321
328

7.31
5.77
4.78
4.18
5.97
5.09

50.5
53.6
50.6
63.4
53.8
64.4

771
240
569

7.11
5.64
2.86

108.4
100.9
83.9

696
258
442

788
347
127

6.83
4.18
2.38

115.4
83.0
53.4

755
321
124

7
5.14
2.91
6.78
3.73
2.06

99.4
86
88.7
111.4
86.1
60.2

433
297
220
218
345
257
Reservoir.

~ I

Wash.

1991

ISSAM

Copyright (C) 1991 American Water Works Association

N. NAJM

ET AL

69

maximum
extent possible. A major
disadvantage, however, is that an additional coagulation step would be required
to remove the PAG.
Improving adsorption efficiency. Various
techniques for applying PAC to improve
its adsorption of large, slowly diffusing
compounds are available. For example,
adding PAC to solids contact clarifiers
has the potential for improved adsorption
efficiency because the carbon can be
kept in contact with the water for a
longer time than when it is added to the
rapid mix of a conventional plant. Kassam et ali6 reported mean carbon residence times ranging from 9 h to 8.5 days
when PAC was added to a solids contact
slurry recirculating clarifier. Hoehn et
all7 studied theaddition of PAC to a pilotscale floe-blanket clarifier. They reported
PAC retention times of one to two days.
In a study of the addition of PAC to a
bench-scale floe-blanket reactor, Najm
et all2 measured steady-state PAC retention times ranging from 9 h to two days.
They determined that the maximum
adsorptive capacity of PAC for 2,4,6trichlorophenol
(TCP) was achieved
during this time. Richardi noted that
the PAC dosage for detergent removal
could be reduced by 25-40 percent if the
carbon was added to the influent of a
floe-blanket clarifier instead of to a
conventional system. More research is
needed to optimize this process and to
determine coagulation procedures that
can be used to increase adsorption
kinetics, such as using a polyelectrolyte
alone instead of alum or ferric salts.Jg
In the Roberts-Haberer
process,zo-zz
buoyant polystyrene spheres l-3 mm in
diameter are coated with PAC. The
spheres are held in the reactor by means
of a screen, and the water to be treated is
passed upflow through the medium.
After saturation, the PAC is removed
from the beads by backwashing
the
medium (downflow) with a high flow
rate. New PAC is then applied to the
beads. The PAC at the entrance to the
process can be equilibrated with the
influent concentration, thereby increasing its adsorption efficiency. Recovery of
the PAC for regeneration may also be
possible. In addition, the Roberts-Haberer filter may be used as a roughing
filter to remove some suspended solids
before filtration.
However,
careful
studies are needed to assess the efficiency
of adsorption for specific organics and
the ability of the polystyrene beads to
retain the PAC in order to determine the
conditions under which the process will
be more cost-effective than GAC.
Sontheimer et alz3 suggested the addition of PAC to a multistage contactor.
They showed that addition of two doses
of PAC, ml and m2, to contactors 1 and 2,
respectively, in series theoretically results in a lower usage rate than addition
of the total carbon dose, mr= ml + m2, to
70

RESEARCH

AND

one contactor. To determine the minimum values of ml and m2 required to

drop the adsorbate concentration from
an initial concentration,
C,, to a final
concentration,
Cz, the intermediate
aqueous concentration, Ci, between the
two reactors was calculated. Assuming
that the PAC reached equilibrium with
the effluent of each contactor and that
the same Freundlich isotherm equation
describes equilibrium in both contactors,
Ci was calculated by

0
Cl

1/n
-flP=l-l

G-

C
(1)
Cl

in which C, = influent concentration to

contactor 1, Ci = effluent concentration
of contactor 1 and influent concentration
to contactor 2, Cz = effluent concentration of contactor 2, and l/n = Freundlich
equation constant.
The PAC dose to each reactor can then
be calculated using Eq 2, described later.
The limitation to this approach is that
equilibrium is not always reached when
PAC is added to conventional contactors,
and the same Freundlich equation may
not apply in each contactor.
When
competing organics are present, the
isotherm
function
depends on PAC
dosage.
Reactions with treatment chemicals and
calciumcarbonate. Careful attention must
be paid to the interaction of PAC with
water treatment chemicals. Activated
carbon will chemically
reduce compounds such as free and combined chlorine, chlorine dioxide, chlorite, ozone,
and permanganate; the demand for oxidants and disinfectants is thereby increased. The reaction of activated carbon
with chlorine will reduce its adsorptive
capacity for compounds such as phenol
and substituted
phenols.24.25 LalezaryCraig et aI26 found a reduction in the
ability of PAC to adsorb both geosmin
and methylisoborneol
in water containing free chlorine and monochloramine. Surprisingly,
the effect of the
monochloramine appeared to be greater
than that of chlorine.
Addition of PAC to a water that is
supersaturated with calcium carbonate
or other precipitates, or treatment that
causes an increase in pH to greater than
the saturation pH just after PAC is
added, such as in lime softening or
aeration of water with a high carbon
dioxide content, may lead to coating of
the PAC particle with precipitates and to
a corresponding decrease in adsorption
efficiency. Also, adsorption at high pH is
often poorer than at low pH because
many organic contaminants
are weak
acids that ionize at high pH, rendering
them more hydrophilic.

PAC performance
Many studies have been conducted to
assess the adsorption of organic com-

pounds on PAC under both laboratory

and field conditions. Various types of
organics, activated carbons, and waters
were included in these studies. The most
common means of evaluating adsorption
on activated carbon, especially PAC, has
been through the use of adsorption isotherms. The experimental bottle-point
technique is widely used to achieve an
isotherm relationship between theaqueous adsorbate concentration
and the
carbon surface loading (see Randtke and
Snoeyinkg for a description of this test).
Because these isotherms are conducted
under laboratory conditions,
the accuracy of performance predictions based
on them needs to be examined.
Table 3 lists some of the compounds
for which isotherms have been determined along with their Freundlich isotherm constants. Also listed are the
predicted carbon dosages required to
reduce the concentration of each of these
compounds from 100 pg/L to 10 pg/L
and 1 pg/L, and from 10 pg/L to 1 pg/L.
Each of these calculations used the isotherm of that compound and assumed
that equilibrium is achieved at the final
concentration. The isotherm will change
as the type of carbon and the type of
water treated change. The following
equation was used to calculate the carbon
dosages required:

*=-z- cc, - C,)

Qe

cc, - c,,

(2)

K Ccl

in which m is the carbon dosage (mg/L),

C,, and C, are the initial and equilibrium
concentrations (pg/L), respectively, and
K and n are the Freundlich equation
constants.
All the isotherm data listed in Table 3,
except those of Lalezary et a127and Najm
et al, were conducted
using GAC
pulverized to a smaller size to enhance
kinetics and reduce the time required for
equilibration.
Some of the isotherms of
Miltner et a128were conducted with 50 X
200-mesh pulverized GAC (74-297-pm
diameter), whereas the rest of the isotherms were conducted using pulverized
GAC of ~200 mesh.
A major concern with some of the
isotherm data listed in Table 3 relates to
the contact time allowed between the
carbon and the water before sampling.
Although higher adsorption rates are
achieved when GAC is pulverized, some
large-molecular-weight
organic compounds may not reach equilibrium in
less than oneor twodays. The isotherms
for large molecules determined with one
hour or less of contact time can be used
as rough estimates of adsorbability but
probably indicate a capacity lower than
achievable at equilibrium.
Caution should be exercised when
using the bottle-point isotherm of a trace
compound conducted with natural water.
The competitive effect of the background

TECHNOLOGY

JOURNAL

Copyright (C) 1991 American Water Works Association

AWWA

natural organic matter on the adsorption

isotherm of the trace compound on activated carbon has to be carefully considered when the isotherm is being used to
predict the performance of an activated
carbon process. When conducted in a
multisolute system, the bottle-point isotherm of a trace compound depends on
the initial concentration
used. Frickzg
determined that the isotherm capacity
of activated carbon for p-nitrophenol
(PNP) decreased with decreasing initial
PNP concentration in the presence of a
constant background concentration of phydroxybenzoic acid (HBA) and phenol4sulfonic acid (PSA). With an initial
concentration of 5 mg/L of each of HBA
and PSA, a decrease in initial PNP concentration from 10 to 2 mg/L resulted in
a 50 percent reduction in the carbons
capacity for PNP. Similarly, the authors
research showed that the carbon capacity
for TCP present in natural water and
determined using the bottle-point isotherm technique depended on the initial
concentration of TCP. At a certain TCP
equilibrium concentration,
the carbon
capacity decreased with decreasing initial concentration
as a result of the
competitive effect of the background
natural organics.
Therefore, the choice of the proper
isotherm significantly affects the evaluation of the performance of PAC in a
continuous process. Although the effect
of competitive adsorption behavior on
the kinetic models is not clearly understood, the authors believe that the isotherm to be used should be that conducted with an initial concentration close
to the concentration of the influent to
the continuous process.
Published reports of PAC performance
are discussed subsequently and, when
possible, are compared with performance
predictions based on equilibrium
isotherms. The data on PAC performance
are categorized according to type of
compound as well as according to impact
on drinking water quality.
T&O-causingcompounds. Powdered activated carbon is widely used in the United
States for T&O control. It has been
estimated that about 90 percent of water
utilities worldwide that use activated
carbon use it in the form of PAC.ii.30In
1977 PAC was used in 25 percent of the
683 US water utilities surveyed, which
included the 500 largest utilities.1 The
number of utilities using PAC is constantly on the rise.2 Compounds causing
T&O problems vary widely from the
strongly adsorbed, low-molecular-weight
compounds, such as 2,4-dichlorophenol
(2,4-DCP), to the weakly adsorbed, highmolecular-weight
compounds, such as
some types of humic substances. This
has resulted in a wide variation in PAC
adsorption efficiency.
The concentration
of organic compounds that cause a T&O problem is
JANUARY

TABLE
Quantity
System
Size
tnsd
0.1
0.5
1.0
2.5
5.0
::
25
50
75
100
150

ojPAC

Quantity
5 mg PAWL

required as a junction
of PAC

10 mg PAUL

2.9
14.6
29.2
73.0
146.0
292.0
438.1
730.1
L460.3
2.190.4
2,920.5
4.380.8

1.0
2.5
5
10
15
25
50
75
100
150

876.1
584.1
1,460.3
2.920.5
4,380.8
5.841.0
8,761.5

43.8
219.0
438.1
1,095.2
2.190.4

2,920.5
4.380.8
7,301.3
14.602.6
21,903.g
29,205.2
43.8073

6,571.2
4,380.8
10,951.g
21.903.9
32855.9
43.807.8
65.711.8

10

Cost

ojPAC

of PAC-.$/year

5 mg PAUL

10 mg PAUL

25 mg PAUL

50 mg PAUL

75 mg PAUL

2.132
426

4,263
853

10,658
2,132

4P63
10,658
21,316
42,632
63,963
106,595
213,204
319.798
426,393
639,597

8,526
21,316
42,632
86,279
127,911
213,204
426,393
639,597
852,786
1,279,179

21,316
53,290
106.580
213,204
319,798
532,988
1,065,989
1.598,977
2,131,979
3,197,969

4,263
21,316
42,632
106,580
213,204
426,393
639,597
1,065,989
2,131,979
3.197969
4,263,959
6,395,938

6,395
31,974
63,963
159,899
319,798
639,597
959,395
1,598,977
3,179,969
4,796,961
6,395,938
9,593,923

TABLE

0.1
0.5
:i

75 mg PAUL

29.2
146.0
292.0
730.0
1,460.3

2,190.4
1.460.3
3,650.6
7.301.3
10,951.9
14.602.6
21,903.g

cost estimates jorpurchase

Annual
System
Size
msd

50 mg PAUL

14.6
73.0
146.0
365.0
730.0

TABLE

i::

Required-lb/day

2.5 mg PAUL

5.84
29.2
58.4
146.0
292.0

Annual
System
Size
rnsd

of dosage and system size

Capital
5 mg PAUL
2,125
2,420
5,095
3,270

5
10
E

7,555
11.215
20,295
14,080

50
75
100
150

32,105
43,230
48,980
68,495

11

cost estimates of PAC feed equipment

and O&M

10 mg PAUL
2,135
2,620
3,655
6,005
8,765
13,620
18,290
26,425
39,940
54,290
61,795

85,885

highly dependent on the type of compound. The threshold odor number

(TON) is widely used as a measure of
taste and odor in drinking water. The
TON produced by a given concentration
of contaminant is usually different for

Costs

of Feed

25 mg PAUL
2.165
3,155
4,715
7,820
12.680
21,260
27,740
37,580
58,995
80,000
95,155
133,725

Rquipment+/year
50 mg PAUL

75 mg PAUL

2,365
4,065
5,925
11,130
18,810
29,095
38,800
53,315
86,115
119,520
138,685
186,525

2,595
4,665
6,980
14.180
23,480
36,990
48.290
67,865
113,065
152,320
174,905
249,135

each compound. For example, geosmin is

known to cause odor problems in the low
nanograms-per-litre
range of concentrationsz7 whereas 2 pg/L of PNP causes
objectionable odor, and 8 pg/L causes a
taste problem.31 Sigworth32 presented a

1991

ISSAM

Copyright (C) 1991 American Water Works Association

N. NAJM

ET AL

71

list of T&O-causing
herbicides and
pesticides alongwith theconcentrations
that caused detectable odor. The concentrations
varied from 17 pg/L for
aldrin to 4.7 mg/L for methoxychlor.
The concentration of PAC required to
remove T&O varies widely, consistent
with the diversity of T&O-causing compounds. Researchers33 found that 1 mg
PAC/L reduced the TON by 1-24 units
in the water supplies of four US cities,
indicating a wide variability in adsorbability of T&O-causing
compounds.
Bench-scale studies of three types of
carbon showed that 75 mg/L of the best
carbon was required to achieve humic
odor reduction from a TON of 15 to 3,
and 30 mg/L of the best PAC was required to reduce the concentration of 2,4DCP from 1.75 too.24 mg/L, corresponding to a TON reduction from 22 to 3.34
The isotherms of Najm et all2 and Aly
and Faust35 (see Table 3), however, predict PAC doses of 13 and 20 mg/L,
respectively, to achieve the same removal
of 2,4-DCP.
A full-scale study was conducted on
the performance of PAC for the removal
of T&O at the East Drive Well in North
Miami Beach, Fla.36 Doses of PAC up to
27 mg/L failed to remove a hydrogensulfide-like odor. In another full-scale
study,s7 an average PAC dose of 23 mg/L
reduced the TONof influent water to the
Nitro, W.Va., water plant from 30-325 to
14.6, which was still much higher than
the acceptable value of 3 or less. The
specific compounds causing the T&O
were not determined, so comparison with
predicted dosages based on the isotherm
is not possible.
Laboratory tests were conducted by
Lalezary et alz7 to study the efficiency of
two types of PAC (PAC l* and PAC 2t)
for the removal of the following odorcausing compounds-geosmin,
2,3,6-trichloroanisole (TCA), 2-isopropyl-3-methoxypyrazine, 2-isobutyl-3-methoxypyrazine, and 2-methylisoborneol (MIB)-at
the nanograms-per-litre
concentration
level. The effects of chlorination, coagulation, and filtration on the adsorption of
these compounds on the two types of
PAC were investigated. In a pilot study
that followed, Lalezary-Craig
et alz6
investigated the effects of initial concentration, chlorine and chloramine residual, presence of background organics,
contact time, and filtration rate on the
removal of geosmin and MIB by PAC 2.
The authors reported that carbon doses
as low as 10 mg/L reduced 66 rig/L each
of geosmin and MIB to 2 and 7 rig/L,
respectively, for surface coverages of
approximately
6 pg/g. However, data
are not available in this range to determine whether equilibrium
had been
achieved. The presence of chlorine and
chloramine, as well as background organics, was found to decrease the removal of the two compounds by PAC.
72

RESEARCH

AND

The removals of geosmin and MIB

achieved by PAC in the pilot plant compared well with those predicted by the
bench-scale studies for similar initial
concentration, chlorine dosage, and coagulation conditions.26,27
Pesticides
and herbicides.
A growing
problem facing many water treatment
facilities is theincreasingpublicconcern,
and the related regulations, regarding
the concentrations
of pesticides and
herbicides (collectively termed biocides)
in drinking water sources. Biocides are
transported from agricultural fields into
lakes and rivers by water runoff. Miltner
et a13sreported the presence of several
biocides in the Sandusky River in northwestern Ohio during a survey from May
to July 1984. These biocides included
alachlor, metolachlor, atrazine, cyanazine, metribuzin, carbofuran, and simazine at mean concentrations
ranging
from 0.44 pg/L for simazine to 8.17 pg/L
for metolachlor. The maximum concentrations, however, ranged from 1.15 to
22.3 pg/L.
Six biocides-endrin,
lindane, methoxychlor, toxaphene, 2,4-D, and 2,4,5TP-are included in the current drinking-water standards.3g New standards,
proposed in May of 1989,40 do not include
endrin, but include eleven additional
biocides. Table 3 lists isotherm data for
some of these compounds, showing them
to be strongly adsorbed. Table 3 shows
that a PAC dose of ~5 mg/L should be
capable of reducing the concentration of
any of the b&ides listed from 10 to ~1
pg/L and that the type of water affects
the isotherm capacity, an effect probably
caused by natural background organics.
A comparison of the carbon dosages
required to reduce the pesticide concentration to a predetermined value in a jar
test using distilled and natural water is
shown in Table 4.41 The data show that
the adsorptive capacity of PAC in raw
river water is 50-93 percent lower than
that in distilled water.
Miltner et a138investigated the removal
of several pesticides that were included
in a national pesticide survey. Alachlor,
atrazine, carbofuran, linuron, metolachlor, metribuzin, and simazine were
found in the influent water to three
water treatment
plants in Ohio. The
performance of PAC for their removal
was investigated in full-scale as well as
bench-scale studies. The isotherms for
these compounds on pulverized GAC are
listed in Table 3. Although the adsorption
capacities of PAC in jar tests and fullscale treatment were in good agreement,
they were approximately two orders of
magnitude lower than those predicted
by the pulverized GAC isotherms listed
in Table 3. One possible reason for this
difference is that the initial concentrations in the isotherm studies were much
higher than those used in the jar tests
and full-scale studies. This would lead to

a greater reduction in biocide adsorption

capacity in the jar tests and full-scale
plants, as discussed earlier in this article.
Other possible reasons for thedifference
include (1) the short contact time allowed
(1.5 h$ compared with several days for
the pulverized GAC isotherms) and (2)
some interference by coagulants.
The study of Miltner et a138concluded
that all pesticides detected in the Sandusky and Maumee rivers in Ohio will
not be removed by conventional water
treatment processes unless an activated
carbon adsorption process is employed.
Similar conclusions were reached by
Richard et a142for the removal of atrazine
from the Des Moines, Iowa, water supply.
Poor removals of the pesticide atrazine
by PAC, however, were observed at New
Orleans, La.43144
SO&. Synthetic organic compounds
are the products of a wide range of
manufacturing
processes, and some are
classified as hazardous to animal and
human health. Occasionally, these compounds find their way into natural water
sources by means of accidental spills,
dumping, or leaching to groundwater
aquifers. The US Environmental
Protection Agency (USEPA) has set maximum contaminant
levels (MCLs) for
several VOCS,~~ and many other compounds are currently being added to the
list (see Federal Register46 for the proposed listing). Many of these compounds
are included in Table 3.
The adsorbability of SOCs is highly
dependent on the type and properties of
the compound being removed. A benchscale study on PAC adsorption of four
SOCs present in Mississippi River water
was conducted at New Orleans, La.43
The removals of 1,2-dichloroethane,
benzene, toluene, and atrazine by conventional treatment processes (coagulation, flocculation, sedimentation, filtration, and chlorination) were investigated,
along with their adsorption on 5,50, and
500 mg PAC/L. The raw water concentrations of the compounds, however,
never exceeded 1 pg/L and continuously
varied throughout the period of the study.
These facts made it difficult to draw
solid conclusions concerning the removal
of these compounds. The data, however,
suggest that removal did not occur unless
PAC was present.
The removal of TCP and trichloroethylene (TCE) from groundwater
by
PAC added to a bench-scale floe-blanket
reactor was investigated by Najm et a1.i2
With the long carbon residence times
supplied (one to two days), complete
utilization of the adsorptive capacity of
the carbon, as indicated by the isotherm,
was achieved.
*Aqua Nuchar, Westvaco Corp.. Covington, W.Va.
tWPH. Calgon Corp., Pittsburgh,
Pa.
$The totalcontact
timeincludedaperiodofsedimentation
when mixing was not employed. The amount of adsorption
that occurred during this time was not established.

TECHNOLOGY

JOURNAL

Copyright (C) 1991 American Water Works Association

AWWA

TABLE

12

Total cost estimates of PAC treatment

PAC
System
-

Size

5 mg PAUL

Cost

Estimates-$/l,OOOgal
50 mg PAUL

25 mg PAUL

75 mg PAUL

A*

0.1
0.5

0.083
0.019

0.017
0.017

0.100
0.036

0.083
0.021

0.033
0.033

0.116
0.054

0.085
0.025

0.083
0.083

0.168
0.108

0.092
0.032

0.167
0.167

0.259
0.199

0.101
0.036

0.250
0.250

0.351
0.286

i.5
5
10
15
25
50
75
100
150

0.008
0.013
0.006
0.004
0.004
0.003
0.003
0.002
0.002
0.002

0.017
0.017
0.017
0.017
0.017
0.017
0.017
0.017
0.017

0.030
0.025
0.023
0.021
0.021
0.020
0.020
0.019
0.019
0.019

0.014
0.009
0.007
0.005
0.005
0.004
0.003
0.003
0.002
0.002

0.033
0.033
0.033
0.033
0.033
0.033
0.033
0.033
0.033

0.042
0.047
0.040
0.038
0.038
0.037
0.036
0.036
0.035
0.035

0.012
0.018
0.010
0.008
0.007
0.006
0.005
0.004
0.004
0.003

0.083
0.083
0.083
0.083
0.083
0.083
0.083
0.083
0.083

0.095
0.101
0.093
0.091
0.090
0.089
0.088
0.087
0.087
0.086

0.023
0.017
0.015
0.011
0.010
0.008
0.007
0.006
0.005
0.005

0.167
0.167
0.167
0.167
0.167
0.167
0.167
0.167
0.167

0.184
0.190
0.182
0.178
0.177
0.175
0.174
0.173
0.172
0.172

0.022
0.027
0.018
0.014
0.013
0.011
0.009
0.008
0.007
0.006

0.250
0.250
0.250
0.250
0.250
0.250
0.250
0.250
0.250
0.250

0.272
0.277
0.268
0.264
0.263
0.261
0.259
0.258
0.257
0.256

mgd

*A-capital

cost

plus O&M

feed equipment

cost,

A major spill of carbon tetrachloride

(CCIJ in a tributary of the Ohio River
during early 1977 resulted in several
studies of its removal by PAC. Benchscale studies at the USEPA laboratory in
Cincinnati utilized untreated river water
containing approximately 70 pg CCL&.
The water was dosed with varying
amounts of PAC, mixed rapidly for 1
min, and then contacted for 1 h in a jar.
The highest PAC dose of 30 mg/L resulted in CC& reduction of only 20
percent. The amount adsorbed on the
carbon surface, 1.86 mg/g, was higher
than the equilibrium capacity predicted
by the l-h isotherm of Dobbs et aP7given
in Table 3. During this period, 9.6 mg
PAC/L was used at the Cincinnati, Ohio,
water treatment plant. Using conventional treatment and this PAC dosage,
the CC& concentration in the finished
water was virtually equal to that in the
influent water to the plant.48
Following the CC& incident, another
spill of toxic chemicals occurred in the
same river at a different location. The
spill consisted of hexachlorocyclopentadiene and octachlorocyclopentadiene.
Although the concentrations of the compounds were quite low (cl pg/L), up to
30 mg PAC/L was added as a precaution
at the Evansville, Ind., plant.49 Both
compounds were virtually absent (x0.1
pg/L) in the finished water.
After consumers complaints
of a
pesticidelike taste and odor in the drinking water of the Sunny Isles water
treatment plant, North Miami Beach,
Fla., in September 1977, laboratory
analysis of the raw and finished water
revealed the presence of 42 SOCs ranging in concentration
from 0.01 to 73
~g/L.36.50~5i A study of the removal of
natural total organic carbon (TOC) and
SOCs by 7.5,15, and 30 mg PAC/L was
initiated. The PAC was added at the well
intake to achieve a maximum of 2 h of
contact time. A finished water concentration of 40 pg/L of volatile halogenated
JANUARY

Treatment

10 mg PAC/L

B-PAC

material

cost only,

C-total

cost

aliphatics, includingdichloroethenes
and
chloroethanes (l,l-dichloroethane
and
l,l,l-trichloroethane),
was virtually
unaffected by the addition of up to 26 mg
PAC/L. The chlorobenzene concentration in the finished water was reduced
from 0.5 to ~0.1 pg/L with the addition
of 10 mg PAC/L. Nonvolatile synthetic
organic chemicals were reduced from a
combined finished water concentration
of 45 to ~5 pg/L with the addition of 7.5
mg PAC/L. Comparison of the removal
data, however, with those predicted by
the isotherms listed in Table 3 was not
possible because the concentrations of
the compounds were far lower than
those used in the isotherm tests.
THMs and THMFP. The discovery that
THMs are formed during drinking water
treatment52*53 has caused major concern
because of the potential health hazard
associated with thesecompounds. Rook52
attributed the formation of these compounds to the reaction between natural
organic matter and chlorine. Organic
compounds that undergo this reaction
are referred to as THM precursors. The
amount of THMs that will be produced if
the reaction goes to completion is referred
to as the THM formation
potential
(THMFP). In 1979 the USEPA established an MCL for total THMs at 100
pg/L as an annual average.54 This MCL
is currently being reviewed by USEPA.
Adsorption
of THMFP. The reduction of THMFP can be determined by
measuring THMFP directly or by measuring TOC. Although there is usually a
good correlation between the TOC and
THMFP for a given type of water, there
is nocorrelation that applies toall waters
because of the differences in the characteristics of the organics. For example,
Lange and Kawczynski55 observed sig
nificant removal of TOC with virtually
no reduction in THMFP. Hentz et a156
observed similar results for another
water. Amy and Chadik,57 however, reported a substantially greater reduction

of THMFP than of TOC with the addition

of PAC.
Duringconventional
water treatment,
part of the natural organic matter comprising the THMFP is removed by chemical coagulation.
Although
these removals are usually low,55 some have
been reported to be as high as 86 percent.58 Amy and Chadik57 reported a
reduction in the THMFP of Ilwaco
Reservoir (Wash.) water from 1,087 to
371 pg/L with an alum dose of 2.5 mg/L
(a 66 percent reduction) and to 245 pg/L
with an optimum ferric dose of 7.5 mg/L
(a 77 percent reduction).
The removal of THM precursors by
adsorption
on activated carbon has
varied from poor to very good. Many
studies have reported the need for very
high PAC dosages to achieve reductions
in THMFP. In a pilot-scale study of
conventional treatment processes, Love
et a15g investigated
the reduction of
THMFP by PAC (Table 5). A PAC dose of
100 mg/L was required to reduce the
THMFP from 53 to 25 pg/L. In a benchscale study at New Orleans, La.,43 researchers determined that 90 percent
reduction of THMFP could be achieved,
compared with that obtained by conventional treatment, with a PAC dose as
high as 500 mg/L (Table 6). Blank runs
of conventional treatment with no PAC
added resulted in THMFP reduction
efficiencies varying from 27 percent in
the spring to 54 percent in the fall. Hentz
et aP6 determined that PAC doses ranging
from 75 to 125 mg/L were required to
achieve a slight reduction of 7.5-29 percent of a THMFP of approximately 425
pg/L. Their results agree with those of
Wood and Demarco,60 Symons,Gi and
Lange and Kawczynski,55 who also found
PAC to be ineffective for the removal of
THM precursors.
The use of PAC was evaluated for the
removal of chloroform precursors in a
study involving 15 of Kentuckys larger
water utilities.62 After the Ohio River

1991

ISSAM

Copyright (C) 1991 American Water Works Association

N. NAJM

ET AL

73

water had been chlorinated in a jar with

a dose of 2-3 mg chlorine/L and then
stored for three days, the terminal
chloroform concentration was 236 erg/T.
(Table 7). As shown in Table 7, a PAC
dose of 50 mg/L applied prior to chlorination resulted in a 50 percent decrease
in the terminal chloroform concentration. With a PAC dose of 1,000 mg/L,
however, 20 percent of the chloroform
formation potential (47.5 pg/L) was still
not reduced. Addition of PAC to a pilotscale floe-blanket reactor for the control
of THMFP was investigated by Hoehn et
al. The addition of 25 mg/L of PAC* to
a water containing 362 pg THMFP/L
improved the reduction efficiency by a
mere 4 percent over the 56 percent
reduction achieved by coagulation alone.
However, the addition of 21 mg/L of
another PAC.t a higher-grade carbon.
resulted in an added 1%percent reduction
in THMFP.
The PAC adsorption of THM precursors present in six US waters was
investigated by Amy and Chadik.57 The
raw water THMFP varied from 313 E.rg/L
in the Mississippi River (La.) to 1,087
pg/L in the Ilwaco Reservoir (Wash.).
With optimum alum doses of 15 and 2.5
mg/L as Al, THMFP in the Mississippi
River and Ilwaco Reservoir waters was
reduced to 130 and 370 &L, respectively.
The THMFP was then reduced to 81 and
196pg/L, respectively, with the addition
of 50 mg PAC/L. Table 8 lists the results
of Amy and Chadik5 for reduction of
THMFP in water from the Daytona
Beach Aquifer (Fla.) and the Ilwaco
Reservoir by using two types of PAC
after coagulation. In the aquifer water,
the value of THMFP:TOC for the raw
water, 72.5 rg/mg, was virtually unchanged after coagulation. The ratio,
however, dropped to an average of 50
pg/mg after treatment
with 50 mg
PAC/L. This suggests that the portion of
the natural organic matter that produces
THMs is selectively removed by PAC. As
for the Ilwaco Reservoir water, Table 8
shows that the THMFP: TOC value
dropped from 135 pg/mg for the raw
water to 108 pg/mg with the addition of
7.5 mg ferric chloride/l, followed by a
drop to 53 rg/mg after treatment with
50 mg PAC/L. This also suggests the
selective removal of THM-forming
natural organics by PAC.
A review of these findings clearly sug
gests that the extent of THMFP reduction by PAC largely depends on the types
of PAC and water being used. In the
absence of a reliable surrogate parameter, the THMFP of a water sample
needs to be measured directly. A standard
method has been proposed for the determination of THMFP.63
Adsowtion
of THMs. Prechlorination is commonly practiced for the control
of bacterial and algal growth in the
treatment plant. In such a case, the

THM formation reaction starts prior to

the addition of PAC. By the time PAC is
added, the water will contain some remaining THM precursors, THMs, and
some residual chlorine. However, the
oxidation-reduction
reaction between
chlorine and the surface of the activated
carbon will result in the reduction of the
adsorption capacity of activated carbon
for phenolic compounds, as well as the
destruction of the residual chlorine.@%
The same decrease in capacity may
occur for other compounds. Therefore,
the effect of PAC addition on the finished water THM concentration will not
necessarily be due to the adsorption of
the THM precursors or the THMs; the
decrease in the residual chlorine concentration will result in lower THM formation,56 which can be misinterpreted as
higher THM removals by PAC. On the
other hand, the destruction of the carbon
surface will result in lower than expected
removals of THMs per unit mass of PAC
because of the reduction in carbon
capacity.
Anderson et ala investigated the control of THMs by PAC addition to the
influent to the flocculation basin of a
40-mgd water treatment plant. In the
first part of the study, chlorine was
added to the raw water prior to the point
of addition of PAC. In the second part,
the chlorination point was moved to the
flocculation
basin effluent, allowing
some contact between the water and the
carbon in the absence of chlorine. With
prechlorination, the THM concentration
in the filter effluent was reduced from
40.2 to27.5 pg/L with theadditionof
7.3
mg PAC/L, and from 37.4 to 23.8 pg/L
with the addition of 21.6 mg PAC/L.
When the chlorine was added after the
PAC, the THM concentration in the
filter effluent dropped from 51.4 rg/L
when no PAC was added to 28.2 ccg/L
with 19.4 mgPAC/L added. The removals
were higher than those predicted using
the isotherms of Dobbs et al47 listed in
Table 3. However, a good comparison
cannot be made because chlorine was
present during the adsorption of THMs.
In another full-scale study, Singley et
al51 investigated the performanceof PAC
added to a softening plant for the control
of THMs. Prechlorination
of the raw
water was also practiced during part of
this study. The finished water THM
concentration of 40 I*g/L was virtually
unchanged after the addition of 14.3 and
26.6 mg PACK.. The three-day THMFP,
however, decreased from 135 to 40 rg/L
with the addition of the two carbon
dosages, and the average TOC of the
water was reduced from 14 to 10.4 mg/L
(26.6 percent reduction). The THM results may have been affected by the
removal of chlorine by the PAC. When
the chlorination point was moved to the
recarbonation units, the finished water
THM concentration was still unchanged

with the addition of 7.1 mg PAC/L.

However, the three-day THMFP was
reduced from 135 to 110 pg/L, and the
TOC was reduced from 12.6 to 4.4 mgiL
(65.1 percent reduction). The much
higher removals of TOC achieved when
the chlorination point was moved (65.1
percent), compared with those achieved
during prechlorination
(26.6 percent),
are consistent with a reduction in PAC
capacity for organic compounds in the
presence of free chlorine.
Cost analysis of PAC use
The most important component of the
cost of using PAC is the cost of the PAC
itself. Over the last 30 years, the price of
PAC has risen from about $O.O7/lb to as
high as $.45/lb, an increase of >500
percent. During this time, the wholesale
and producers price indexes increased
by about 200 percent. The largest increase in PAC costs occurred in the mid1970s following improvements in PAC
products. In addition to the cost of the
carbon, treatment costs for PAC usage
include capital and operation and maintenance requirements for process feed
equipment.67-70$ Preliminary cost estimates for PAC treatment have been
developed for a range of system sizes and
PAC dosage rates. Table 9 presents the
quantity of PAC required (lb/day) as a
function of system size and dosage,
assuming 70 percent utilization of plant
capacity and continuous application of
PAC. Design requirements
for feed
equipment were based on the maximum
flow rate and PAC dosage rate. Tables 10
and 11 present annual cost estimates
($/year) for the purchase of PAC and
process feed equipment, respectively.$
Table 12 summarizes the unit cost estimates for PAC treatment ($/l,OOO gal).
Figure 1 presents unit cost estimates
for PAC treatment ($/l,OOO gal) as a
function of dosage rate for four system
sizes. A significant decrease in cost
occurs with increasing size for small
systems up to about 1 mgd in capacity.
Systems with capacities in the range of
l-100 mgd show only small gradual
decreases in unit costs with increasing
size, given a fixed PAC dosage. For
example, assuming a PAC dose of 5
mg/L, the unit treatment cost drops
from about $0.10 to $0.03/1,000 gal with
increasing system size from 0.1 to 1 mgd.
By increasing system capacity from 1 to
100 mgd, the unit costs gradually decrease from $0.03 to $0.02/1,000 gal. In
this analysis, unit costs increase nearly
linearly with respect to increasing dosage
rate, given a fixed system size.
A typical maintenance dose of PAC for
taste and odor control is approximately 5
*Aqua Nuchar. Westvaco Corp.. Covington. W.Va.
tNuchar SA. Westvaco Corp.. Covington. W.Va.
$All cost figures were updated to 1988.
5An interest rate of 10 percent and an equipment life span
of 20 years were assumed for this analysis.

74 RESEARCHANDTECHNOLOGY

JOURNALAWWA
Copyright (C) 1991 American Water Works Association

mg/L. At this rate, quantities of PAC

required for systems of l.O- and 100-mgd.
size are about 1,000 and 1,070,OOO lb/
year, respectively. Many systems must
increase the PAC dosage rate for a period
of time to treat seasonal and occasional
variations in water quality. A utility
may have to increase the PAC dose from
5 to as much as 75 mg/L for several
weeks to reduce the concentrations
of
synthetic organic contamination resulting from an occasional chemical spill or
seasonal agricultural runoff. For example, if a lo-mgd system increases its
typical maintenance PAC dose rate from
5 to 50 mg/L for four weeks during the
year, the additional PAC required is
about 73,600 lb above the usual maintenance dosage. This results in a total
PAC requirement of about 180,200 lb/
year. This short-term but significant
increase in the quantity of PAC required
affects the treatment cost somewhat
proportionally.
The total cost estimate
for the system, assuming only the 5mg/L dose rate, is about $53,85O/year.
By including the four-week increase in
PAC dose to 50 mg/L., the total cost
increases to about $84,66O/year, an increase of about 60 percent. The reduction
in the cost of applying PAC can be
calculated in a similar manner if there
are periods of time when no PAC is
required.

Summary and conclusions

Powdered activated carbon can be
applied at different points in drinking
water plants. In selecting the point of
addition of PAC, consideration must be
given to the degree of mixing and to
minimizing
short-circuiting,
the adequacy of the time of contact between the
PAC and the water, the PAC residence
time, and the minimization
of interference of treatment chemicals with the
adsorption process. For example, good
mixing can be achieved in the rapid mix
and flocculation basin, but PAC in the
sedimentation basin may not beeffective
because of poor mixing. When PAC is
added to some treatment processes, such
as floe-blanket reactors, significantly
longer carbon residence times are
achieved compared with conventional
treatment. These longer residence times
increase the removal of slowly adsorbing
compounds and decrease the carbon
usage rate.
The efficiencies of removal of tasteand odor-causing compounds by PAC
vary, depending on the type and concentration of the compound causing the
problem. High efficiencies of removal of
compounds such as 2,4-dichlorophenol,
geosmin, and 2-methylisoborneol
were
achieved with relatively low PAC doses
(lo-25 mg/L). Doses of PAC ranging
from 75 to 620 mg/L, however, were
required to remove p-nitrophenol and
humic odor.

Powdered activated carbon has been

effective for the removal of several
biocides, depending on their type and
concentration
as well as the type of
contactor. However, PAC doses as high
as 475 mg/L were required to reduce the
concentration of metolachlor from 50 to
1 fig/L.
The extent of adsorption of SOCs on
PAC is strongly dependent on the typeof
compound being removed. In general,
PAC has a relatively low capacity for
VOCs compared with nonvolatile organics. It should be emphasized, however, that the means of application of
PAC, as well as the contact time available, largely determine the percentage
removal of these and other trace organics
from drinking water.
The reported removals of THM precursors and THMs from drinking water
by PAC range from poor to very good.
The fraction of natural organic matter
that forms THMs varies in different
waters. Some studies have shown that
the ratio of THMFP to TOC of a natural
water decreases after treatment with
PAC, indicating the preferential removal
of THM precursors by PAC. However,
depending on the type of water being
treated, few studies have reported reduction in the TOC of the water with an
insignificant
drop in THM precursor
concentration.
Existing data indicate
that no good correlation exists between
the concentration of THM precursors
and TOC in different types of water.
One of the main advantages of PAC is
its low capital cost. Cost estimates of the
addition of PAC at various dosages to
different plant sizes are presented in
this article. The annual cost of PAC
decreases with increasing plant size. For
a lo-mgd plant operating at full capacity
and using a continuous PAC dose of 10
mg& the total cost is approximately
$0.038/1,000
gal, which amounts to
$380/day or $138,7OO/year.

Acknowledgment
This work was supported by Lyonnaise
des Eaux, Paris, France; the AWWA
Research Foundation; and the US Environmental Protection Agency. The opinions expressed in this article are thoseof
the authors and not necessarily those of
the supporting agencies.

References
AWWA CommitteeRePort. Measurement

and Control of Organic Contaminants

by
Utilities. JOUY. AWWA 69:5:267 (May
1977).

AWWA. 1984 Utility Operating Data.

Denver, Colo. (1986).
GRAESE, S.L.; SNO~YINK,
V.L.; & LEE,
R.G. Granular Activated Carbon FilterAdsorber Systems. Jour. A WKA, 79112~3
(Dec. 1987).
KRUITHOF, J.C. ET AL. Selection of Brands
of Activated Carbon for Adsorptive Prop
erties. Activated Carbon in Drinking

6.
7.

8.
9.

Water Technology.
KIWA-AWWARF
Rept. AWWARF, Denver, Cola. (1983).
NAJM, LN. ET AL. Effect of Particle Size
and Background
Organics on the Adsorption Efficiency of PAC. JOUY. A WWA,
82:1:65 (Jan. 1990).
AWWA. AWWA Standard for Powdered
Activated Carbon. B600-78. Denver, Cola.
(1978).
MEIJERS, J.A.P. & VANDERLEER,
R.C. The
Use of Powdered Activated Carbon in
Conventional
and New Techniques.
KIWA-AWWARF
Rept. AWWARF, Denver, Colo. (1983).
RANDTKE. SJ. & SNOEYINK, V.L. Evaluating GAC Adsorption
Capacity. Jour.
A WWA. 75:8:406 (Aug. 1983).
AWWA. Water Q,al&
and Treatment.
McGraw-Hill,
New York (3rd cd., 1971).

10. GAUNTLETT.
R.B. & PACKHAM.
R.F. The
Use of Powdered Activated Carbon in
Water Treatment.
Proc. Conf. on Activated Carbon in Water Treatment, Univ.
of Reading, Water Res. Assn., Medmenham, England, Apr. 1973.
11

SONTHEIMER,

H. The

Use

of Powdered

Activated Carbon. Translation of Reports

on Special Problems of Water Technology,
Volume 9-Adsofltion.
EPA-600/g-76-030
(Dec. 1976).
12. NAIM, I.N. ET AL. Powdered Activated

Carbon in FloeBlanket Reactors. 1989

AWWA Ann. Conf., Los Angeles, Calif.
13. SONTHEIMER,
H.: K~LLE, H.; & SPINDLER,
P. Rohiil und Trinkwasser, Vertifx de..
Bereichs fiir Wasserchemie. Univ. Karlsruhe, FRG (1967).
14. DOUGHARTY,
I.D. & MORRIS, R.L. Studies
on the Removal of Actinomycete Musty
Tastes and Odors in Water Supplies.
Jour. A WWA, 59:10:1320 (Oct. 1967).
15. GIFFORD, J.S.; GEORGE, D.B.; & ADAMS,
V.D. Svnergistic
Effects of Potassium
Permanganate and PAC in Direct Filtration Systems for THM Precursor Removal. Water Res.. 23:lO (Oct. 1989).
16. KASSAM, K. ET A;. Accumulation
and
Adsorption Capacity of Powdered Activated Carbon on a Slurry Recirculation
Clarifier. 1989 AWWA Ann. Conf., Los
Angeles, Calif.
17. HOEHN,
R.C.
ET AL. THM
Precursor
Control WithPowderedActivatedCarbon
in a Pulsed-Bed, Solids Contact Clarifier.
1987 AWWA Ann. Conf., Kansas City,

MO.

18. RICHARD, Y. Personal communication

(1986).
19. LETTINGA. G.; BEVERLOO, W.A.; & VANLIER,
WC. The Use of Flocculated Powdered
Activated Carbon in Water Treatment.
P~ogy. Water Technol., lo:537 (1978).
20. HABERER, K. & NORMANN,
S. Untersuchungen zu einer Neuartigen Pulverkohle-Filtrationstechnik
fib die Wasseraufbereitung.
Vom Wassey, 49:331(1979).
21. HABERER, K.&Z NORMANN, S. Entwicklung
eines Kurztalt-Filtrations
Verfahrens
zum Einsatz vor Puiverkohle
in der
Wasseraufbereitung.
Gas und Wasserfach, Wasser/Abwa.&er, 118:393 (1980).
22. HOEHN. R.C. ET AL. A Pilot-Scale
Evaluation of the Roberts-Haberer
Process for
Removing Trihalomethane
Precursors
From Surface Water With Activated
Carbon. 1984 AWWAAnn. Conf., Dallas,
Texas.
23. SONTHEIMER,
H.; CRITTENDEN,
J.C.;
&
SUMMERS,

JANUARY 1991

R.S.

Activated
ISSAM

Copyright (C) 1991 American Water Works Association

N. NAJM

Carbon
ET AL

for

75

Water Treatment.
DVGW-Forschungsstelle, Engler-Bunte-Institut,
Universitit
Karlsruhe. FRG (2nd ed. lEnelish1 1988).
24.

25.

26.

nn
I.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.
40.

41.

42.

43.
76

SNOEYINK,.V.L.

& SUDAN,

CT.

&chl&

rination by Activated Carbon and Other

Reducing Agents. Disinfection: Waterand
Wastewater U.D. Johnson, editor). Ann
Arbor Sci. P&l., Arm Arbor, Mich.il975).
MCGUIRE. MJ. & SUFFET. I.H. Aoueous
Chlorine/Act&ated
Carbon Interactions.
Jour. Envir. Engrg. Div.-AXE,
119:
3:629 (June 1984).
LALEZARY-CRAIG, S. ETAL. Optimizing the
Removal of Geosmin and 2.Methylisoborneol by Powdered Activated Carbon.
Jour. A WWA, 80:3:73 (Mar. 1988).
LALEZARY. S.; PIRBAZARI, M.; & MCGUIRE,
M.J. Evaluating
Activated Carbons for
Removing Low Concentrations
of Tasteand Odor-Producing
Organics. Jour.
A WWA, 78:11:76 (Nov. 1986).
MILTNER, RJ. ET AL. Final Internal Report
on Carbons Use Rate Data. USEPA,
Cincinnati. Ohio (Tune 1987).
FRICK, B. Adsorptionsgle&hgewichte
zwischen Aktivkohle
und organischen
Wasserinhaltsstoffen
in Mehrstoffgemischen bekannter
und unbekannter
Zusammensetzung.
Doctoral dissertation, Univ. of Karlsruhe, FRG (1980).
HANSEN,
R.E. The Costs of Meeting the
New Water Quality Standards for Total
Organics and Pesticides. 1975 AWWA
Ann. Conf.. Minneanolis. Minn.
BURTSCHEL~, R.H. ET AL. Chlorine Derivatives of Phenol Causing Taste and
Odor. Jour. A WWA, 51:2:205(Feb. 1959).
SIGWORTH, E.A. Identification
and Removal of Herbicides and Pesticides. Jour.
A WWA, 57:8:1016 (Aug. 1965).
Westvaco Corp. Taste and Odor Control
in Water Purification.
Westvaco Corp.,
Covington, W.Va. (1970).
BROWN. G.N.: HYNDSHAW. D.B.: & VAN
DONGE~. D.B: Powdered Activated Carbon Removal of Tastes and Odors Plus.
Proc. of AWWA Ann. Conf.. 97:2 (1977).
ALY, O.M. & FAUST, S.D. Removal of 2,4Dichlorophenoxy-acetic
Acid Derivatives
From Natural Waters. Jour. AWWA,
51:2:221
(Feb. 1965).
SINGLEY, J.E. & ERVIN, A.L. East Drive
Well Field Powdered Activated Carbon
Study. Rept. to the City of North Miami
Beach, Fla. (Sept. 1978).
SYMONS, J.M. ET AL. Interim Treatment
Guide for Controlling
Organic Contaminants in Drinking Water Using Granular
Activated Carbon. Water S<pply Res.
Div.. USEPA. Cincinnati.
Ohio (Ian.
MIL~NER, R J. ET AL. Treatment of Seasonal Pesticides in Surface Waters. Jour.
AWWA, 81:1:43 aan. 1989).
Fed. Reg., 43~28 (Feb. 1978).
USEPA. National Primary and Secondary
Drinking
Water Regulations; Proposed
Rule, 40 CFR Parts 141, 142, 143, Vol.
549722062. Cincinnati, Ohio(May 1989).
ROBECK, G.G. ET AL. Effectiveness
of
Water Treatment Processes in Pesticide
Removal. lour. AWWA. 57:2:181
(Feb.
1965).
*
RICHARD, J J. ET AL. Analysis of Various
Iowa Waters for Selected Pesticides:
Atrazine, DDE, and Dieldrin--1974.
Pesticides Monitoring Jour., 9:3 (Dec. 1975).
USEPA. Powdered Activated
Carbon
Treatment for Removal of Trace Organic
RESEARCH AND TECHNOLOGY

44.

45.
46.
47.

48.

49.
50.

51.

52.

53.

54.
55.

56.

57.

58

59.

60.

61.

62.

Compounds From theNew Orleans Water

Supply. USEPA Grant R-804404 (Mar.
1976).
DIFILIPPO, J.D.; COPELAND, L.G.; 8~ PEIL,
K.M. Evaluation of Powdered Activated
Carbon for the Removal of Trace Organics
at New Orleans, La. EPA-600/2-81-027.
USEPA, Cincinnati, Ohio (1981).
Fed. Reg., 52:130 (July 1987).
Fed. Reg., 54:97 (May 1989).
DOBBS, R.A.; MIDDENDORF, R.J.; & COHEN,
J.M. Carbon Adsorption
Isotherms for
Toxic Organics. MERL, USEPA, Cincinnati, Ohio (May 1978).
SEEGER, D.R.; SLOCUM, C.J.; & STEVENS,
A.A. GC-MS Analysis of Purgeable Contaminants in Source and Finished Drinking Water. Ann. Conf. on Mass Spectrometry and Applied Topics, St. Louis,
MO., May 28-June 2,1978.
REXING, M. Personal communication
(1978).
SINGLEY, J.E. ET AL. Minimizing
Trihalomethane Formation in a Softening Plant.
MERL, USEPA, Cincinnati, Ohio (1977).
SINGLEY, J.E.; BEAUDET, B.A.; & ERVIN,
A.L. Use of Powdered Activated Carbon
for Removal of Specific Organic Compounds. Proc. Seminar on Controlling
Organics in Drinking Water, 1979AWWA
Ann. Conf., San Francisco, Calif.
RooK,JJ. Formation of Haloforms During
Chlorination
of Natural Waters. Proi
Water Trtmt. Exam.. 23:2:2X4 (1974).
BELLAR, T.A.; LICHTENBERG,
J.J.;
&
KRONER, R.C. The Occurrence of Organohalides
in Chlorinated
Drinking
Water. lour. A WWA, 66:12:703
(Dec.
1974). Fed. Reg., 44:231 (Nov. 1979).
LANGE, A.L. & KAWCZYNSKI. E. Trihalomethane Studies, Contra C&a County
Water District Experience. Proc. Water
Trtmt. Forum VII. California-Nevada
AWWA Section, Palo Alto, Calif., Apr.
1978.
HENTZ,L.H.JR.;HOEHN,R.C.;&RANDALL,
C.W. Haloform Formation in Water From
Peat-Derived Humic Substances and the
Reducing Effects of Powdered Activated
Carbon and Alum. Proc. 1980 AWWA
Ann. Conf., Atlanta, Ga.
AMY, G.L. & CHADIK, P.A. PAC and
Polvmer Aided Coaeulation
for THM
Control. Proc. Natly Co& on Envir.
Engrg. ASCE (July 1982).
HOEHN, R.C. ET AL. Chlorination
and
Water Treatment
for Minimizing
Trihalomethanes
in Drinking Water. Proc.
COT& on Water Chlorination:
Environmental Impact and Health Effects, Vol. 2
(R.H. Jollev. H. Gorchev. and D.H. Hamilion, e&to&). Ann Arbo;, Mich. (1977).
LOVE, O.T. ET AL. Treatment
for the
Prevention or Removal of Trihalomethanes in Drinking Water. DWRD, USEPA,
Cincinnati, Ohio (1976).
WOOD, P.R. & DEMARCO, J. Effectiveness
of Various Adsorbents
in Removing
Organic Compounds From Water. 176th
ACS Mtg.: Activated Carbon Adsorption
of Organics From the Aqueous Phase,
Miami Beach, Fla., Sept. 10-15, 1978.
SYMONS, J.M. ET AL. Interim Treatment
Guide for the Control of Chloroform and
Other Trihalomethanes.
DWRD, USEPA,
Cincinnati, Ohio fJune 1976).
ZOGORSKI, J.S.; ALLGEIER, G.D.; & MUL
LINS. R.L. JR. Removal of Chloroform

63.

64.

65.

66.

67.
68.

69.

70.
71.

72.

73.

74.

From Drinking
Water. Res. Rept. 111.
Univ. of Kentucky Water Resources Res.
Inst., Lexington aune 1978).
Standard Methods for the Examination of
Water and Wastewater. APHA, AWWA,
and WPCF. Washington, D.C. (17th ed.,
1989).
SUIDAN, M.T.; SNOEYINK, V.L.; & SCHMITZ,
R.A. Performance
Predictions
for the
Removal of Aqueous Free Chlorine by
Packed Beds of Granular Activated Carbon. Water-1976:
I. Physical Chemical
Wastewater Treatment. AICHE Sym. Ser.,
73:18 (1976).
SUIDAN, M.T.; SNOEYINK, V.L.; & SCHMITZ,
R.A. Reduction of Aqueous HOC1 With
Granular Activated Carbon. Jour. Envir.
Engr. Div.-ASCE,
103:677
(1977).
ANDERSON, M.C. ET AL. Control of Trihalomethanes
Usine Powdered Carbon.
Natl. Conf. on Envi; Engrg. ASCE (July
1980).
LEE, R. Personal commun. (Nov. 1988).
GUMERMAN, R.C.; RUSSELL, L.; & HANSEN,
S.P. Estimating Water Treatment Cost,
Vol. II, Cost Curves Applicable to 1 to200
rngd Treatment Plants. EPA-600/2/79l&b (1979).
GUMERMAN, RX.; BURRIS, B.E.; & HANSEN,
S.P. Estimation of Small Svstem Water
Treatment Costs. USEPA 6bO/S2-84-184
(Mar. 1985).
FISHER, J.L. Personal communication
(Nov. 1988).
DOBBS, R.A. & COHEN, J.M. Carbon
Adsorption Isotherms for Toxic Organics.
EPA-60018-80-023 (Apr. 1980).
EL-DIB, M.A.; MOURSY, A.S.; & BADAWAY,
M.I. Role of Adsorbents in the Removal of
Soluble Aromatic Hydrocarbons
From
DrinkingWater.
WaterRes., 12:1131(1978).
COHEN, J.M. ETAL. Effects of Fish Poison
on Water Supplies, Part 1, Removal of
ToxicMaterials. Jour. AWWA,52:12:1551
(Dec. 1960).
STEINER, J. IV; SINGLEY, J.E.; & EDWARD, J.
Methoxvchlor
Removal From Potable
Water. journal AWWA, 71:5:284
(May
1979).

About the authors:

At the time of this
;a~study, Issam N. Najm
,,-Zwas a graduate
re search assistant at the
University
of Illinois
at Urbana, where he
received MS and Ph D
degrees in environmental engineering.
He is currently
a
senior engineer
with James M. Montgomery, Consulting
Engineers
Inc., 250
N. Madison
Ave., Pasadena, CA 91109.
Najm
is a member
of AWWA
and
IA WPRC,
and his articles have been
published
in several technical journals.
Vernon L. Snoeyink is professor of environmental
engineering,
University
of
Illinois,
205 N. Mathews Ave., Urbana,
IL 61801.
Benjamin
W. Lykins Jr. is
chief of the Systems and Field Evaluation
Branch,
US Environmental
Protection
Agency, 26 W. Martin
Luther King Dr.,
Cincinnati,
OH 45268. Jeffrey Q. Adams
is an environmental
engineer in USEPA S
Systems and Field Evaluation
Branch.

Copyright (C) 1991 American Water Works Association

JOURNAL AWWA