•MM
WATER POLLUTION CONTROL RESEARCH SERIES
17O2ODNQO9/69
STUDY OF POWDERED CARBONS
FOR WASTE WATER TREATMENT
& METHODS FOR THEIR APPLICATION
U.S. DEPARTMENT OF THE INTERIOR •FEDERAL WATER POLLUTION CONTROL ADMINISTRATION
-------
WATER POLLUTION CONTROL ADMINISTRATION
The Water Pollution Control Research Reports describe
the results and progress in the control and abatement
of pollution of our Nation’s waters. They provide a
central source of information on the research, develop-
ment, and demonstration activities of the Federal Water
Pollution Control Administration, Department of the
Interior, through in-house research and grants and
contracts with Federal, State, and local agencies,
research institutions, and industrial organizations.
Water Pollution Control Research Reports will be
distributed to requesters as supplies permit. Requests
should be sent to the Planning and Resources Office,
Office of Research and Development, Federal Water
Pollution Control Administration, Department of the
Interior, Washington, D.C. 202142.
-------
STUDY OF POWDERED CARBONS FOR WASTE
WATER TREATMENT AND METHODS
FOR THEIR APPLICATION
by
West Virginia Pulp and Paper Company
Covington, West Virginia 24H26
for the
FEDERAL WATER POLLUTION CONTROL ADMINISTRATION
DEPARTMENT OF THE INTERIOR
Contract Number 14-12-75
September 1969
-------
FWPCA Review Notice
This report has been reviewed by the Federal
Water Pollution Control Administration and
approved for publication. Approval does not
signify that the contents necessarily reflect
the views and policies of the Federal Water
Pollution Control Administration, nor does
mention of trade names or commercial products
constitute endorsement or recommendation for
use.
-------
CONTENTS
Page
ABSTRACT
INTRODUCTION . . . . 1
MATERIALS 3
EXPERIMENTAL PROCEDURES 5
RESULTS 9
CORRELATION OF CARBON PROPERTIES WITH TOC ADSORPTION CAPACITIES . . 18
DISCUSSION 29
CONCLUSIONS 33
REFERENCES 34
-------
ABSTRACT
Eleven comercial and experimental powdered activated carbons were sub-
jected to an intensive laboratory evaluation of their physical and
adsorptive properties to select those best suited to treating municipal
waste water and to gain insight into properties important for this appli-
cation. Measurements were made of TOC (Total Organic Carbon) adsorption
isotherms from Roanoke, Virginia municipal secondary effluent, pore
structure, BET surface area, molasses value, decolorizing index, iodine
value, real density, suspendability, and apparent density. Correlations
were made between capacities for adsorbing organic impurities from munic-
ipal secondary effluent, as measured by TOC adsorptive capacities, and
carbon properties such as iodine value, molasses decolorizing index and
pore structure.
Aqua Nuchar and Hydrodarco were clearly superior to other commercial
grades tested in bC adsorptive capacity per unit cost. The coal-base
experimental carbons, pulverized Nuchar WV-L and Nuchar WV-W, were mdi —
cated by the suspendability test to be more easily removed than other
carbons.
Among many pore structure parameters compared, surface area in pores
greater than approximately 14 Angstroms in radius was found to give
the best correlation with TOC capacity. However, the correlation was
not perfect, the correlation coefficient being 0.91. TOC adsorptive
capacity was indicated reasonably well by the decolorizing index test.
Here, correlation coefficients in the 0.80’s were obtained. It is sug-
gested that the best carbon for adsorbing organics from municipal wastes
have a broad spectrum of pore sizes. Particle size of the powdered car-
bons was found to strongly affect the rate of adsorption.
Aqua Nuchar, Hydrodarco, and pulverized Nuchar WV-L are recommended for
further study.
-------
I NTRODIJCT ION
The studies presented in this report were sponsored by the Federal Water
Pollution Control Administration, under Contract No. 14—12-75.
As commonly practiced, treatment of municipal waste is accomplished by a
combination of settling and biological processes that remove some 85-90%
of the organics present in the raw waste. These are so-called primary
and secondary processes. In many highly populated and industrialized
areas, primary and secondary treatment will not be capable of sufficiently
reducing the organic pollution load to the receiving waters, and more com-
plete treatment will be necessary. Furthermore, as demand for water for
drinking and industry increases, increasing attention is being given to
renovating waste water to remove essentially all impurities present follow-
ing conventional treatment.
The only proven practical means for removing residual dissolved organic
compounds is adsorption on activated carbon.
The ability of powdered activated carbon to remove organics from water is
well recognized. Powdered carbon has been used for many years to remove
the organic compounds which cause tastes and odors in drinking water (1).
Only recently, however, have they been applied to treating waste water
[ 2,3,4). Davies and Kaplan (2) have concluded that the optimum powdered
carbon waste water process is one which uses flocculation-clarification
in a two—stage countercurrent process.
Obviously, the properties of the carbon used are of primary importance.
The carbon must have a high adsorptive capacity for the organic pollutants,
attain a close approach to equilibrium in a reasonable period of time,
and be readily removed by flocculation and clarification.
Some previous work has been performed to determine which basic carbon
properties are important for superior performance in treating waste
water. Studies by O’Conner, et al. , (5) have shown no strong correla-
tion between COD adsorption ability and BET surface area or ability to
adsorb four specific model substances.
One of the critical steps in the powdered carbon process is the removal
of the carbon from the water. Previous studies (2,3,4) indicated the
need for flocculating aids to achieve carbon removal; however, development
of a powdered carbon which settles without flocculating aids could substan-
tially reduce costs.
The objective of the present study was the characterization of commercial
and experimental carbons to select those best suited for municipal waste
water treatment and to determine properties responsible for high adsorptive
capacity for organic pollutants, high rate of adsorption, and ease of remov-
al by settling.
Eleven comercial and experimental powdered activated carbons from a
variety of sources were subjected to an intensive laboratory evaluation
1
-------
to determine their physical and adsorptive properties. Emphasis was
placed on the elucidation of the relationship between carbon adsorptive
capacities for pollutants in municipal waste water and other properties
of the carbon.
2
-------
MATERIALS
The eleven carbons studied are described below:
1. Aqua Nuchar: A lignin—base commercial carbon produced by Westvaco
and used for taste and odor control in municipal water treatment.
1967 cost for truckload or carload bag shipment was $150/ton, F.0.B.
plant.
2. Nuchar C-115: A lignin-base commercial decolorizing carbon produced
by Westvaco and used in sugar purification, chemical processing, and
the manufacture of pharmaceuticals. 1967 cost for truckload or car-
load bag shipment was $239/ton, F.0.B. plant.
3. Nuchar C-115 Granular: The same carbon as C—ll5 except that it is
of larger particle size. Cost is not available for this non—standard
carbon.
4. Nuchar C-190: A lignin-base commercial decolorizing carbon produced
by Westvaco and similar to C-115 except that it has higher adsorption
capacity. It has similar uses to C—115. 1967 cost for truckload or
carload bag shipment was $269/ton, F.0.B. plant.
5. Nuchar C-l000: A lignin-base commercial decolorizing carbon produced
by Westvaco and similar to C—115 and C-190, but with exceptionally
high adsorptive capacity. This carbon is used when a minimum carbon
dosage is desirable to obtain maximum adsorption. 1967 cost for
truckload or carload bag shipment was $1,470/ton, F.O.B. plant.
6. Pulverized Nuchar WV-W: This is a pulverized sample of Westvaco’s
Nuchar WV-W 8 x 30, a coal—base granular carbon for municipal and
industrial water treatment. 1967 cost for carload or truckload bag
shipment was $420/ton, F.O.B. plant.
7. Pulverized Nuchar WV-L: This is a pulveri7ed sample of Nuchar WV-L
8 x 30, a coal-base carbon similar to WV-W except that it has a
larger average pore size and a higher adsorption capacity. It is
used for decolorizations, chemical processing, and waste water
treatment. It is produced by Westvaco. 1967 cost for carload or
truckload bag shipment was $520/ton, F.O.B. plant.
8. Pulverized Filtchar: A wood-derived carbon similar to Aqua Nuchar,
but with a somewhat higher density. 1967 cost for truckload or
carload bag shipment was $150/ton, F.0.B. plant.
9. Non-Pulverized Flitchar: The same as pulverized Filtchar except
that it is of larger particle size. No costs are available for
this non—conuiercial carbon.
10. Hydrodarco: A lignite—base comercial carbon for municipal and
industrial water treatment, produced by Atlas Chemical Industries.
1967 cost for truckload or carload bag shipment was $150/ton, F.O.B.
plant.
3
-------
11. Darco G-60:
produced by
and various
carload bag
A lignite—base cormiercial decolorizing type carbon
Atlas Chemical Industries and used in sugar refining
chemical purifications. 1967 cost for truckload or
shipment was $790/ton, F.0.B. plant.
The source of municipal waste water for this study was the Roanoke,
Virginia Sewage Plant. The Roanoke sewage plant serves the entire
Roanoke Valley and has a history of being a well-operated plant. Sec-
ondary treatment is provided by the activated sludge process. The
Roanoke sewage plant is the closest source of domestic activated sludge
effluent to Covington, Virginia, the location of Westvaco’s Carbon Tech-
nical Center. A sunnrary of the Roanoke sewage plant operations for 1967
is given below:
Flow, 18.6 MGD
Population Served, 149,000
Influent BOO, 240—270 mg/i (35% is of Industrial Origin)(a)
BOO Removal, 90%
Suspended Solids Removal, 90%
(a)The amount of industrial waste was estimated assuming a per capita
loading of 0.17 lbs. BOD per day.
4
-------
EXPERIMENTAL PROCEDURES
The experimental procedures to which the carbons were subjected are
described below:
Molasses Value
In the Molasses Value (MV) procedure, the carbon sample is boiled with
a standard molasses solution, filtered, and a (Klett—Sumnierson) colon—
metric reading taken on the filtrate. From a calibration curve, the MV
is read. In the preparation of the calibration curve, eight given weights
of a standard 150 MV carbon are contacted with the standard molasses
solution. The eight given weights of standard carbon are assigned a pre-
determined MV which does not change. The Klett readings obtained when
contacting the eight standard carbon weights with the molasses solution
are used for drawing an isotherm from which is determined the MV of the
unknown sample. The molasses value is reported as being an approximate
indication of pores greater than about 14 Angstroms in radius (6).
Decolonizing Index
In this test, the experimental procedure is the same as for the Molasses
Value test, but the results are given in a fashion that is, in effect, a
relative efficiency. Approximately, it is the ratio of molasses color
capacity of a carbon to that of a standard carbon, times 10. The equation
relating the Decolorizing Index (DI) to the Molasses Value is as follows:
DI = 19.956 log MV - 29.912
Iodine Value
In the Iodine Value test, the weighed carbon sample is contacted with a
standard 2.70 g/l iodine solution and filtered. The amount of iodine
remaining in the filtrate is determined either colorimetrically or by
titration. The iodine value is calculated as the percent of iodine
adsorbed from solution. The iodine value is an indication of pores
larger than about 5 Angstroms in radius (6) and is a rough measure of
total surface area.
Suspendability
The suspendability is intended as a measure of ease of settling for pow-
dered carbons and was first applied some years ago to carbons for municipal
water treatment. For this test, a 200 mg/l concentration of carbon is
mixed in distilled water and placed in a one-liter graduated cylinder.
The carbon mixture is allowed to settle under quiescent conditions for a
given time. The top 500 ml of liquid is siphoned off and the carbon
remaining in this supernatant is measured. Suspendability is expressed
as the percent of carbon in the supernatant liquid.
5
-------
Apparent Density
In this test, 10 g of powdered carbon are placed in a 100-mi graduated
cylinder and the sample tapped a given number of times on a Tap—Pak
Volumeter. The packed volume is determined and the apparent density
is calculated.
BET Surface Area and Pore Size Distribution by the Nitrogen Method (71
These two properties were calculated from nitrogen adsorption isotherms
(78°K) and dete rmined with a McBain-Bakr adsorption apparatus. Calcula-
tions were performed by Control Data Corporation on a CDC 3600 computer.
Basically, the calculation for surface area determines the quantity of
gas necessary to form a mono-molecular layer on the carbon surface. By
using known values for the area occupied by each molecule, the surface
area of the carbon is computed. The method used was the BET (Brunauer-
Emmett-Teller) method.
Pore size distributions were calculated using Roberts’ (8) technique
applied to the nitrogen isotherms. During adsorption, as the nitrogen
pressure is increased, nitrogen gas condenses in ever larger capillaries,
or pores, of the adsorbent. Measurements are made of the quantity of gas
adsorbed for each increase in pressure. By use of the Roberts’ procedure,
which uses the correlation between relative pressure and size of pores
in which capillary condensation is taking place (Kelvin equation), and
makes correction for nitrogen adsorption on the surface of the pores,
the pore volume in pores between various size intervals is calculated.
By this method, pore radii from 10 to 1 ,000 Angstroms can be measured.
In the McBain adsorption apparatus, the carbons were first outgassed by
simultaneously heating to 300°C and evacuating down to 0.05 microns Hg
pressure. Nitrogen isotherms were measured at liquid nitrogen tempera-
tures (78°K).
Pore Size Distribution by the Mercury Penetration Method
Pore size distributions by the mercury penetration method were determined
with an Aniinco-Winsiow, 15,000 psi model porosimeter. For this test, the
carbon is immersed in mercury and subjected to increasing pressures.
Mercury, a non-wetting material, will enter smaller capillaries as pres-
sure is increased. The amount of mercury forced into the pores of the
carbon is measured at specified pressure intervals. Knowing the contact
angle and surface tension of the mercury, pore sizes nto which the mercury
penetrates can be correlated with pressure. The mercury porosimeter mea-
sures pores from 75 to greater than 100,000 Angstroms in radius. In this
study, it was used to measure pores frcm 1 ,000 to 100,000 Angstroms.
Real Density
Real density is the density of the carbonaceous “backbone” of the carbon
after eliminating the effect of the pores. Real density is frequently
6
-------
determined by helium displacement. In these studies, it was measured by
a combination of nitrogen adsorption and mercury porosimeter data. Pro-
cedures for obtaining nitrogen adsorption and mercury porosimeter data
were discussed above.
In the mercury porosimeter, at the highest pressures attained, pores down
to 75 Angstroms in radius are filled. The volume of mercury used to fill
the container and fill the pores is measured. The volume of the empty
container minus the volume of mercury added gives the volume occupied by
the solid carbon plus the volume in pores smiller than 75 Angstroms in
radius. The McBain nitrogen adsorption data are used to determine the
volume in pores smaller than 75 Angstroms. Knowing the weight of sample
used and volume of solid carbon allows for determination of the real den-
si ty.
TOC (Total Organic Carbon) Adsorption Isotherms
TOC (Total Organic Carbon) isotherms were determined by a batch contacting
procedure. The procedure consisted of contacting 200 ml portions of the
Roanoke municipal secondary effluent with known weights of powdered carbon
followed by filtration and TOC analysis. The carbon and waste water were
contacted for 90 minutes.
Loading on the carbon for each sample is calculated with the expression:
( C 0 - Cf) V
X/M=
where X/M is the TOC loading, C 0 is original (filtered) TOC, Cf is the
TOC after contacting and filtering, V is the volume of Solution (200 ml),
and M is the weight of carbon.
Isotherms were plotted on log-log paper. A straight line on this plot
indicates adherence to the Freundlich equation, as will be discussed
below.
The isotherms were determined within 12 hours following collection of the
waste water. These water samples were stabilized by the addition of ½ ml
of saturated mercuric chloride solution per liter of water. A Beckman
Laboratory Carbonaceous Analyzer was used for the TOC determinations.
Duplicate T0C s were determined for all analyses.
Rate of Adsorption
For the rate of adsorption determination, a six (6.0) g sample of the
carbon was added to a two (2.0) liter sample of Roanoke secondary efflu-
ent. Following this, the carbon-waste water mixture was stirred by a
chain stirrer at 5,000 rpm, and at specified time intervals 25 ml samples
were withdrawn, immediately filtered through fiberglass filter paper, and
TOC’s determined.
7
-------
Particle Size Distribution
Particle size distribu ions were determined by use of an Alpine Jet Sieve.
In this test, a given amount of material is placed on a sieve and a nega-
tive pressure (3.2 inches of water) produced by a vacuum cleaner. The
suction is applied through a rotating slotted nozzle which sucks the
fines-laden air dowrward through the sieve. The amount of material remain-
ing on the sieve is determined and the percent passing calculated. The
operating conditions for each sieve analysis are as follows:
Sieve Opening Initial Weight of Sieving Time
( Microns) Carbon (grams) Minutes
149 10.0000 3/4
74 5.3000 2
44 5.0000 6
2O 0.1000 10
5 a 0.1000 10
a The only way found to clean the 20—micron and 5—micron sieves was with
a sonic bath.
8
-------
RESULTS
Table 1 contains a summary of results for Molasses Value, Decolorizing
Index, Iodine Value, Real Density, BET Surface Area, Suspendability for
10 and 30 minutes settling, and Apparent Density for each of the eleven
carbons under study.
Figure 1 presents the TOC adsorption isotherms for Roanoke municipal
secondary effluent determined on December 5, 1967. Figure 2 represents
isotherm results obtained January 30, 1968. The initial filtered TOC’s
for the isotherms of December 5, 1967 and January 30, 1968 were 35.0 and
19.3 mg/i, respectively. An attempt was made to draw the best smooth
curve between the points for each carbon.
Figures 3 and 4 present TOC adsorption rates for Aqua Nuchar, Hydrodarco,
pulverized Nuchar WV-L and pulverized Filtchar, determined March 1968 on
Roanoke municipal secondary effluent. These particular carbons were
selected for rate studies since they represent four entirely different
raw materials and are good candidates for study in later continuous
adsorption studies.
Figure 5 presents particle size distributions of Aqua Nuchar, Hydrodarco,
pulverized Nuchar WV—L and pulverized Filtchar. The Alpine Jet Sieve was
used in these determinations.
Figures 6 and 7 show the pore size distributions in the range of 10 to
100,000 Angstroms radius for the eleven carbons. The pore size distri-
butions in the range of 10 to 1,000 were calculated from nitrogen iso-
therms, and pore size distributions from 1,000 to 100,000 were determined
by the mercury penetration method. Distributions from these two sources
were joined at 1,000 Angstronis radius.
9
-------
NOTES :
(1) Experimental carbon.
(2) Grades not now in production.
(3) Average of duplicate tests.
Table 1
PROPERTIES OF POWDERED CARBONS
Property
Aqua
Nuchar
Nuchar
C-115
Nuch r
C-ll5U)
Granular
Nuchar
C-l90
Nuchar
C-l000
Pulv.
Nuch
Wv-w J
Pulv.
Nuch
WV—L I
Pulv.
Filt—
char
Non
Pulv.
Filt.
Hydro—
darco
Darco
Molasses Value(3)
Decolorizing Index(s)
80
8.1
124
11.9
125
11.9
199
16.0
1,000
30.0
89
9.0
194
15.7
62
5.9
42
2.5
139
12.9
206
16.3
Iodine Value
93.3
95.6
95.4
94.6
95.3
94.2
96.7
81.4
79.9
78.0
70.0
Real Density - g/cc
1.86
1.72
1.80
1.88
1.86
2.54
2.49
1.98
2.18
1.85
1.95
Suspendability -
a. Ten (10) mm.
81.9
80.3
53.6
76.5
78.1
33.1
28.6
73.1
22.9
66.1
41.4
b. Thirty (30) mm.
Apparent Density - # ft 3
BET Surface Area - m /g
60.0
15.2
754
81.2
14.0
811
43.1
9.6
804
65 2
12.7
841
63.7
9.8
998
23.2
43.8
905
22.2
32.4
1138
48.0
21.5
647
18.8
21.2
647
41.9
30.8
523
29.3
23.5
467
-------
0.4
‘ I p11111
AQUA NUCHAR
NIJCHAR C-115
NUCHAR C-115 GRANULAR
NUCHAR C-190
NUCHAR C-bOO
PULVERIZED NUCHAR WV-W
PULVERIZED NUCHAR WV-L
PULVERIZED FILTCHAR
NON-PULVERIZED FILTCHAR
HYDRODARCO
DARCO G-60
I I i Iii ii
4.0 6.0 8.0 10.0
EQUILIBRIUM TOC - mg/i
10
2
ii
8
9
I ilil
20.0 40.0 60.0 80.0 100.0
Figure 1. TOC adsorption isotherrns
Roanoke, Virginia municipal secondary effluent, December 5, 1967
- 0 -i.
—•-- 2.
—A-- 3.
-.-Q-— 4.
—a— 5.
—0-— 6.
—0-— 7.
—0—- 8.
— —.— 9.
— —.-10.
-c-il.
I ‘ I ‘
6
0.2 —
0.1
0.08
0.06
0.04 —
0.02 —
0.01 —
0.008
0.006 —
0.004
0.002
0.001
L)
D)
E
E
u- i
L)
F-
.
.1.
1.0 2.0
11
-------
0
EQUILIBRIUM TOC - mg/l
Figure 2. TOC adsorption isotherms
Roanoke, Virginia municipal secondary effluent, January 30, 1968
L)
E
E
w
C—)
F-
0.
0.
0.
0.
12
-------
A
0
U
2.
3.
PULVERIZED
PULV [ RI ZED
PULVERIZED
NUCHAR WV-L SAMPLE #307
NUCHAR WV-L SAMPLE #312
FILTCHAR
S
16
20
TIME (MIN.)
Figure 3.
Rate of TOC adsorption
1.
CD
‘—4
—4
w
L)
C a .)
cD
F-
C-)
.4
.2
A
2
0
0
w
3
4
8
1
2
24
28
32
36
Roanoke, Virginia municipal s condary effluent
-------
1
AQUA NUCHAR
HYDRODARCO
0
I
12
16 20
TIME (MIN.)
Figure 4.
Rate of TOC adsorption
Roanoke, Virginia municipal secondary effluent
I
•1.
02.
1.
I
‘—4
.
S
0
S
0
2
4
8
24
28
3
-------
100.
go.
80.
70.
60.
50.
40.
30.
20.
10.
9.
8.
7.
6.
5.
4.
3.
2.
.01 .05 .1 .2 .5 1 2
95 98 99 99.8 99.9 99.99
Figure 5. Particle size distributions of powdered carbons
III I I I I I I I I I
• AQUA NUCHAR
o PULVERIZED FILTCHAR
HYDRODARCO
o PULVERIZED NUCHAR WV-L (SAMPLE NO. 307
! PULVERIZED NUCHAR WV-L (SAMPLE NO. 31
PULVERIZED NUCHAR WV-L (SAMPLE NO. 31
• AQUA NUCHAR (BY ALLEN-BRADLEY COMPANY
I I II
C
L)
5 10 20 30 40 50 60 70 80 90
WEIGHT PERCENT SMALLER THAN STATED SIZE
February, 1968
-------
/
/
/
/
/
/
/
,
/
/
/
V
——
I
,
/
/1
//‘
/
/
V
/,
/
/
Non Pulverized Filtchar
— Hydrodarco
— — — Aqua Nuchar
-- — — Nuchar C-190
Nuchar C-bOO
10
0 .5 1.0 1.5
TOTAL PORE VOLUME-cc/g
Figure 6. Pore volume distribution
.0
100,000
80,000
60,000
40,000
20,000
F
F
/
I
I
I
I
10,000
8,000
6,000
4,000
I
I
I
I
I
I
I
I
LID
I-
LI)
/
1 ,000
800
600
400
/
I
I
I
200
100
80
60
I
/
I
40
I
/
20
16
-------
1 ,000
800
600
400
200
Figure 7. Pore volume distribution
100,000
80,000
60,000
40,000
20,000
10,000
8,000
6,000
4,000
2,000
I .-
=
LiJ
0
1 th
TOTAL PORE VOLUME-cc/g
0
17
-------
CORRELATION OF CARBON PROPERTIES WITH TOC
ADSORPTION CAPACITIES
One of the major aims of the present study has been to add to the under-
standing of the relationship between adsorptive capacities of powdered
carbons for organic pollutants (TOC) and basic properties of the carbons.
Since pore structure is known from previous studies to have a major influ-
ence on adsorptive properties, an extensive study in correlating TOC
capacity and pore structure was undertaken. Correlations were made between
both surface areas and pore volumes within various pore size ranges and TOC
capacities. Also, correlations were made between Iodine Value and Decolor-
izing Index and TOC adsorption capacities.
As a measure of TOC capacity, TOC loadings at 25 and 75% of original fil-
tered TOC concentrations were chosen for correlation. These loadings are
obtained from Figures 1 and 2, which present TOC adsorption isotherms on
municipal secondary effluent. Iodine Value and Decolorizing Index data
are found in Table 1.
Pore volume and surface areas in various pore size ranges were obtained
by the application of Roberts’ method (8) to the nitrogen isotherms.
As a measure of the degree of correlation between the various parameters,
the correlation coefficient was employed (9).
The formula for calculation of the correlation coefficient (r) is as fol-
lows:
— N XY - ( zx) ( zY )
r - (i X)Z] [ N zY - ( () ]
where X and Y = parameters being correlated
N = number of observations (data points)
The larger the correlation coefficient, the greater the correlation between
the parameters with a value of ±1.00 representing a perfect correlation.
The correlation coefficients between TOC adsorption capacity and surface
areas and pore volumes within various pore size ranges are found in Tables
2 and 3.
18
-------
Table 2
CORRELATION COEFFICIENTS BETWEEN TOC ADSORPTION CAPACITY AND SURFACE
AREA IN PORES OF VARIOUS PORE RADIUS INTERVALS
Pore Radius December Isotherms January Isotherms
Intervals - 25% Initial 75% Initial 25% Initial 75% Initial
Angstroms TOC TOC TOC TOC
O lOOO 0.73 0.75 0.43 0.33
10-1000 0.77 0.84 0.54 0.51
14-1000 0.91 0.79 0.63 0.71
16-1000 0.91 0.79 0.68 0.77
18—1000 0.89 0.79 0.70 0.79
20-1000 0.84 0.80 0.72 0.81
25-1000 0.80 0.69 0.58 0.76
30-1000 0.75 0.69 0.58 0.77
14-20 0.88 0.78 0.47 0.49
14—30 0.90 0.80 0.80 0.61
14—40 0.90 0.84 0.82 0.62
16—20 0.90 0.80 0.39 0.45
18—20 0.91 0.75 0.77 0.50
18-25 0.89 0.77 0.69 0.71
18—30 0.91 0.80 0.69 0.73
18—40 0.92 0.79 0.69 0.71
Table 3
CORRELATION COEFFICIENTS BETWEEN TOC ADSORPTION CAPACITY AND PORE
VOLUME IN PORES OF VARIOUS PORE RADIUS INTERVALS
Pore Radius December Isotherms January Isotherms
Intervals — 25% Initial 75% Initial 25% Initial 75% Initial
Angstroms TOC TOC TOC TOC
0-10 0.39 0.53 0.09 0.10
0-30 0.72 0.78 0.45 0.40
0-100 0.80 0.83 0.53 0.52
0—1000 0.81 0.82 0.53 0.52
10-30 0.83 0.83 0.54 0.52
30—50 -0.24 -0.36 0.20 0.07
50-100 —0.48 -0.42 -0.12 0.04
100-1000 0.10 0.16 0.27 0.31
20-100 0.87 0.78 0.67 0.79
20—1000 0.82 0.70 0.64 0.75
19
-------
Correlation coefficients between TOC adsorption capacity, and Iodine
Values and Decolorizing Index values are found in Tables 4 and 5.
CORRELATION COEFFICIENTS
Table 4
BETWEEN TOG ADSORPTION CAPACITY AND IODINE VALUE
% Equilibrium TOG
25
50
75
100
December Isotherms
0.66
0.70
0.66
0.51
January Isotherms
0.45
0.34
0.27
0.33
Table 5
CORRELATION COEFFICIENTS BETWEEN
TOG ADSORPTION CAPACITY AND MOLASSES DECOLORIZING INDEX
% Equilibrium TOG
25
50
75
100
December Isotherms
0.87
0.83
0.78
0.72
January Isotherms
0.62
0.79
0.80
0.80
The surface area data giving two of the higher correlation coefficients
are presented in Figures 8 and 9. The data giving the best two correla-
tion coefficients when considering pore volume in pores of various sizes
are presented in Figures 10 and 11. Figures 12-15 present plots of Iodine
Value and Decolorizing Index versus TOC adsorption capacities. These
graphs also show the lines of regression.
The points in Figures 8-15 are labeled according to the carbons they
represent. The numbers code for the carbons are:
1. Aqua Nuchar
2. Nuchar C-115
3. Nuchar C-115 Granular
4. Nuchar C—190
5. Nuchar C-l000
6. Pulverized Nuchar WV-W
7. Pulverized Nuchar WV-L
8. Pulverized Filtchar
9. Non-Pulverized Filtchar
10. Hydrodarco
11. Darco G—60
20
-------
0 .01 .02 .03 .04
TOC ADSORPTION
( ‘4
E
L i i
L ii
(-)
L i
=
U)
200
1 60
120
80
40
0
.10 .15
TOC ADSORPTION
240
(‘4
E
Lii
0
.05
.20
Figure 8. TOC adsorption versus surface area in pores of radii 14-40 angstroms
-------
201
161
0 ,
E
N)
Li
L)
U-
U,
81
40
0 .05
.10 .15
TOC ADSORPTION
241
• DECEMBER TOC ADSORPTION ISOTHERMS-
75% EQUILIBRIUM TOC
• JANUARY TOC ADSORPTION ISOTHERMS-
75% EQUILIBRIUM TOC
Si •i
.02
TOC ADSORPTION
.20
Figure 9. TOC adsorption versus surface area in pores greater than 16 angstroms in radius
-------
.6
.5
.4
U
U
3
U i
N) =
(A) J
C D
>.
U i
0
.1
0
0 .01
02
TOC ADSORPTION
5
.03
04
0
.05
.10
TOC ADSORPTION
.15
.20
Figure 10. TOC adsorption versus pore volume in pores of radii 10-30 angstronis
-------
U
U
w
w
C
0
TOC ADSORPTION
.
I
DECEMBER TOC ADSORPTION ISOTHERMS-
25% EQUILIBRIUM TOC
JANUARY TOC ADSORPTION ISOTHERMS-
25% EQUILIBRIUM TOC
.5
.4
S
DECEMBER TOC ADSORPTION ISOTHERMS-
75% EQUILIBRIUM TOC
JANUARY TOC ADSORPTION ISOTHERMS-
75% EQUILIBRIUM TOC
U
U
U i
=,
-j
C
Ui
C
cL
Sb
.1
57
Ni.
0
0
.01
.7
Iii
TOC ADSORPTION
Figure 11. TOC adsorption versus pore volume in pores of radii 20-100 angstroms
-------
.0125 .0250 .0375
TOC ADSORPTION
100
•5 1
w
-J
U,
C)
C)
95
90
85
80
75
70
1
7 1
><
U i
C)
(D
NJ
C)
-J
C)
L)
U i
0
8
.0125
.0250 .0375
TOC ADSORPTION
DECEMBER TOC ADSORPTION
JANUARY TOC ADSORPTION I
.0500
0
.0500
Figure 12.
TOC adsorption versus IV and DI for 25% equilibrium TOC
-------
1
TOC ADSORPTION
><
w
(D
N4
-J
C
L)
w
0 .05 .10
.15 .20
LU
-J
N.) LU
cD
C
0 .05 .10 .15 .20
TOC ADSORPTION
Figure 13. TOC adsorption versus IV and DI for 50% equilibrium TOC
-------
I
10
DECEMBER TOC
JANUARY TOC
I
S
.7
13
ADSORPTION ISOTHERMS
ADSORPTION ISOTHERMS
12
16
.
S
12
•4
Ii
I’
DECEMBER TOC ADSORPTION ISOTHERMS
JANUARY TOC ADSORPTION ISOTHERMS
LU
-j
N)
LU
1
57
><
LU
‘-4
‘-4
‘-4
-j
L)
LU
D
110
110
0
.05 .10 .15 .20
TOC ADSORPTION
0 .05
.10 .15
TOC ADSORPTION
Figure 14. TOC adsorption versus IV and DI for 75% equilibrium TOC
-------
100
w
=
-J
N)
w
0
95
90
85
80
75
70
><
w
(D
N . J
0
-J
0
L)
U i
0
TOC ADSORPTION
TOC ADSORPTION
Figure 15.
TOC adsorption versus IV and DI for 100% equilibrium TOC
-------
Drscuss loN
It is seen from Figures 1 and 2 that many of the isotherms do not fit
the Freundlich equation over the entire concentration range and, in
fact, some become practically vertical. The Freundlich isotherm is
described by the relation:
X/M kCl/n
where X/M is the adsorption capacity in terms, for example, of weight
adsorbed per unit weight of carbon, C is the concentration of adsorbing
substance at equilibrium (bC in this case), and k and n are constants.
If the Freundlich equation is obeyed, a log-log plot of X/M versus C
produces a straight line. Adsorption of simple substances and simple
mixtures onto activated carbon at low concentrations corriiionly follow
the Freundlich isotherm.
One possible explanation for deviation from the Freundlich isotherm in
this case is a positive constant error in the TOC analysis. This seems
unlikely. Another, and more probable, explanation is that the waste
water has some components which are much less strongly adsorbed than
other components leading to quite different behavior at high and low
carbon dosages. At high carbon dosages, necessary to produce equilib-.
rium concentrations in the lower range, a residual TOC is left which is
not removed by addition of further carbon. It is noticed that the ten-
dency to leave a residual bC not removed by further carbon is greatest
for pulverized and granular Filtchar. This carbon, as indicated from
its pore size distribution, has most of its pore volume in extremely
small pores and thus apparently lacks large pores providing adsorptive
capacity for higher molecular weight substances. One would expect the
best carbon for treating municipal waste water to have a broad spectrum
of pore sizes.
The pore size distributions in Figures 6 and 7 show some quite distinctive
differences in pore structure between these powdered carbons. Pulverized
Nuchar WV-L, Nuchar C-l90, and Nuchar C—1000 are seen to have a quite
broad distribution of pores in the size range responsible for adsorption
equilibrium properties (these are micropore, roughly 0 to 100 Angstroms
in radius). Filtchar has a very narrow distribution with most of its
micropore volume in pores smaller than 10 Angstroms in radius. The other
carbons have distributions of intermediate breadth.
It is interesting to note the characteristic extremely large volume pos-
sessed by the lignin based carbons, Aqua Nuchar, Nuchar C-115, Nuchar
C-19O, and Nuchar C-1000, in the pore size range 1,000 to 100,000 Ang-
stronis in radius.
A comparison of the data in Tables 2 and 3 reveals that generally much
higher correlation with TOC capacity is found for surface areas in.
various pore size intervals (Table 2) than for pore volumes in various
intervals. Among the intervals tested, the highest correlation is seen
with surface area in pores greater than about 14 Angstroms in radius and
29
-------
with surface area in pores in the approximate range of 14-40 Angstroms.
(Surface area in pores greater than 1,000 Angstroms is negligible.) It
is not possible to tell which, if either, of these ranges is of primary
importance. It may be, for example, that the 14—40 Angstrom range cor-
relates well with ICC capacity only because there is a high correlation
between surface area in the 14—40 Angstrom range and surface area in the
40-1,000 Angstrom range, or vice versa. In other words, they are not
statistically independent. Even though the December isotherms almost
always gave higher correlation coefficients than the January isotherms,
the same trends are seen for both sets of data.
The high correlation of surface area in pores greater than a given size
can be easily explained by a physical model which visualizes only the
surface area in pores larger than the adsorbing materials being accessi-
ble for adsorption. Surface area in pores smaller than the adsorbing
molecules contribute nothing to the adsorptive capacity, since the mole-
cules cannot enter these pores.
In correlations of pore structure with TOC capacity (Figures 8 through
11), it is seen that Aqua Nuchar gives consistently higher bC capacities
than expected from the line of regression, while C-l000 gives a lower
than expected capacity. From the pore size distributions. Aqua Nuchar
is known to have a small average pore size while Nuchar C-1000 has a
large average pore size. The same trend is seen in other carbons, those
with small average pore size giving higher than expected capacity. This
is perhaps not too surprising since it is known that the tenacity of
adsorption is greater in small pores than in larger ones, thus for equal
available surface area the smaller pored carbons might be expected to
have a higher capacity.
Correlations of TOC adsorption capacity and Iodine Value and Decolorizing
Index are found in Tables 4 and 5, respectively. The Iodine Value and
Decolorizing Index tests were chosen for correlation studies since both
are comon standard carbon tests and are relatively simple to perform.
A high correlation was found to exist for the Decolorizing Index test
and a low correlation for the Iodine Value test, as is apparent from
these data. In view of the pore structure results, the Decolorizing
Index test would, in fact, be expected to show a high degree of correla-
tion with bC, since this test is known from previous work to indicate
surface area in pores greater than approximately 14 Angstroms in radius.
These are, of course, essentially the same pores which gave the highest
correlations with TOC capacity.
It is not clear whether the imperfect correlation of TOC capacity with
pore structure is due to errors in measurement of TOC capacity, pore
structure parameters, or a discrepancy in the model. Certainly there
is a wide variety of compounds contributing to TOC in municipal effluent.
It is interesting to note that a better correlation can be obtained between
Decolorizing Index and pore structure than between TOC capacity and pore
structure. A correlation coefficient of 0.97 was obtained for the correla-
tion with pore volume in pores between 20 and 100 Angstroms radius.
30
-------
The suspendability was selected as an indication of carbon settling char-
acteristics and was performed on all carbons. The results are shown in
Table 1. Increased speed of settling could influence strongly the size
of clarifiers and need for flocculating aids. It was encouraging to note
that pulverized Nuchar WV-W, pulverized Nuchar WV-L and non-pulverized
F-iltchar gave much lower suspendability results than the other carbons.
These carbons should be tested in a continuous adsorption system.
The rate of adsorption curves for Aqua Nuchar, Hydrodarco, pulverized
Nuchar WV-L and pulverized Filtchar, Figures 3 and 4, shows that at
least 95% of the TOC ultimately removed was adsorbed within four minutes.
The particle size distributions for Aqua Nuchar, Hydrodarco and pulverized
Filtchar, Figure 5, show these carbons to have medium particle sizes of
about 10 microns. The pulverized Nuchar WV-L samples were found to be
somewhat coarser with median particle sizes of 55, 19 and 13 microns for
samples numbered 307, 310 and 312, respectively. The difference in par-
ticle size of the pulverized Nuchar WV-L, numbers 307 and 312, is reflected
in the rate of adsorption studies. The coarser WV-L, sample number 307,
was considerably slower than WV-L, sample number 312. The observation of
slower rate of adsorption for large particle size is in agreement with the
observation of many previous workers in granular carbons. It is not sur-
prising that the same should be seen in the powdered carbons.
Figure 5 shows that the particle size distribution of Aqua Nuchar is in
agreement with distributions of this material by several other methods.
The median particle size for Aqua sample was 12 microns and is in close
agreement with the medium particle size of 11 microns found by other
methods (3).
Other things equal, the carbon requiring the least expenditure to treat
municipal waste water would be the one which has the highest TOC adsorp-
tive capacity/cost ratio. The capacity should be measured at the TOC of
the water passing from the first to the second stage, in a two-stage
countercurrent system, since this will be the loading on the spent carbon
removed for disposal or regeneration.
An average TOC adsorptive capacity was determined from the December and
January adsorption isotherms at a TOC concentration of 15 mg/i. As men-
tioned above, loading at the intermediate concentration is the critical
loading to use in a two-stage countercurrent process, since this will
give an indication of carbon usage or rate of exhaustion.
Table 6 shows this ratio calculated from the present data.
31
-------
Table 6
TOC ADSORPTIVE CAPACITY/COST RATIO
TOG Adsorptive Capacity/Cost
Carbon ( q TOG Adsorbed/ )
Hydrodarco 5.42
Aqua Nuchar 5.00
Nuchar C-115 3.59
Nuchar C-190 2.48
Nuchar C-l000 1.71
Darco G-60 0.48
Pulverized Nuchar WV-W
Pulverized Nuchar WV-L
Pulverized Filtchar
Non-Pulverized Fi 1 tchar
In terms of adsorptive capacity per unit cost, it is seen that Aqua Nuchar
and Hydrodarco are by far the superior carbons with ratios of 5.00 and
5.42, respectively. These two carbons are followed in effectiveness by
Nuchar C-115 and Nuchar C-l9O. Costs were not calculated for pulverized
Nuchar WV-L, Nuchar WV-W, and pulverized and non-pulverized Filtchar,
since these are non-comercial carbons.
32
-------
CONCLUSIONS
In terms of TOC adsorptive capacity per unit cost for a two-stage counter-
current adsorption system to treat municipal waste water, Aqua Nuchar and
Hydrodarco, of the commercial powdered carbons evaluated, are by far the
superior carbons.
The experimental coal-based carbons, pulverized Nuchar WV-L and Nuchar
WV-W, showed clear superiority in settling characteristics, as measured
by the Suspendability Test. This is probably due to the much higher
density and somewhat greater particle sizes of the coal-based carbons.
Pulverized Nuchar WV-L and Nuchar C-bOO were seen to have quite broad
pore size distributions in the micropore range, 0—1000 Angstrom radius.
Filtchar A had a very narrow distribution. The other carbons had inter-
mediate distributions.
A fairly good correlation was found between TOC adsorptive capacity and
pore structure. The pore structure parameter giving the best correlation
was surface area in pores greater than about 14 Angstroms in radius.
Surface area in pores 14-40 Angstroms in radius gave a good correlation
as well. It was seen that carbons with small average pore sizes gave
better than expected TOC capacities while carbons with large average
pore sizes gave poorer than expected capacities. This may be explained
by the greater tenacity of adsorption known to exist in smaller pores.
The ideal carbon should have a broad spectrum of pore sizes so as to
accommodate the wide variety of molecules present in waste waters.
Perhaps mixtures of carbons of different pore structure characteristics
would be most efficient.
Decolorizing Index gave practically as good correlation with TOC adsorp-
tive capacity as pore structure.
In batch rate of adsorption experiments, Aqua Nuchar, Hydrodarco, pul-
verized Nuchar WV-L and pulverized Filtchar were seen to adsorb 95% of
the ICC ultimately removed within 4 minutes. As expected from previous
studies on granular carbons, rate of adsorption of a powdered carbon was
seen to be strongly influenced by its particle size.
The lignin—based carbons were seen to have unusually large volumes in
pores in the 1,000 to 100,000 Angstrom radius range.
33
-------
REFERENCES
1. Baylis, John R., Elimination of Taste and Odor in Water , McGraw-Hill
Book Company, Inc., New York, 1935.
2. Davies, D. S. and Kaplan, K. A., “Activated Carbon Eliminates Oroanics,”
Chem. Epg. Progr. , 60, (12), 46 (1964).
3. Davies, D. S. and Kaplan, K. A., “Removal of Refractory Organics
from Waste Water with Powdered Activated Carbon,” J. Am. Water Works
Assoc. , 38, 442 (1966).
4. U. S. Government Memoranda from A. N. Masse on the Powdered Carbon
Adsorption Plant, Lebanon, Ohio; July 20, August 3, August 19,
November 22, and December 6, 1965.
5. O’Connor, B., Dobbs, R. A., Griggs, S. H., Villiers, R. V., and
Dean, R. B., “Activated Carbon for Waste Water Renovation: I.
Removal of Dissolved and Colloidal Organic Material by Powdered
Activated Carbon,” 149th National Meeting of the American Chemical
Society, Detroit, Michigan (April 6, 1965).
6. Pittsburgh Activated Carbon Company, “Basic Concepts of Adsorption
on Activated Carbon,” Calgon Center, Pittsburgh, Pennsylvania.
7. Orr, Clyde and Dallavalle, J. M., “Fine Particle Measurement,”
Chapters 7-10, The MacMiller Company, New York, 1959.
8. B. F. Roberts, Paper Presented at 145th National Meeting of the
American Chemical Society, New York City, Sept. 8-13, 1963.
9. Croxton, Frederick E., Cowden, Dudley J., and Klein, Sidney, Applied
General Statistics , Chapter 19, Prentice-Hall, Inc., Englewood Cliffs,
New Jersey, Third Edition.
34
------- |