ROBERT A. TAFT WATER RESEARCH CENTER
REPORT NO. TWRC- 7
REGENERATION OF SPENT
GRANULAR ACTIVATED CARBON
ADVANCED WASTE TREATMENT RESEARCH LABORATORY -VII
U.S. DEPARTMENT OF THE INTERIOR
FEDERAL WATER POLLUTION CONTROL ADMINISTRATION
OHIO BASIN REG/ON
Cincinnati/ Ohio
-------
LABORATORY INVESTIGATION OF THE REGENERATION
OF
SPENT GRANULAR ACTIVATED CARBON
by
A. J. Juhola and F. Tepper
for
The Advanced Waste Treatment Research Laboratory
Robert A. Taft Water Research Center
This report is submitted in
fulfillment of Contract No.
14-12-107 between the Federal
Water Pollution Control Ad-
ministration and the MSA Re-
search Corporation.
U. S. Department of the Interior
Federal Water Pollution Control Administration
Cincinnati, Ohio
February, 1969
-------
FOREWORD
In its assigned function as the Nation's principal
natural resource agency, the United States Department of the
Interior bears a special obligation to ensure that our ex-
pendable resources are conserved, that renewable resources
are managed to produce optimum yields, and that all resources
contribute their full measure to the progress, prosperity and
security of America—now and in the future.
This series of reports has been established to pre-
sent the results of intramural and contract research carried
out under the guidance of the technical staff of the FWPCA's
Robert A. Taft Water Research Center for the purpose of
developing new or improved wastewater treatment methods. In-
cluded is work conducted under cooperative and contractual
agreements with Federal, state, and local agencies, research
institutions, and industrial organizations. The reports are
published essentially as submitted by the investigators. The
ideas and conclusions presented are, therefore, those of the
investigators and not necessarily those of the FWPCA.
Reports in this series will be distributed as sup-
plies permit. Requests should be sent to the Office of
Information, Ohio Basin Region, Federal Water Pollution Con-
trol Administration, 4676 Columbia Parkway, Cincinnati, Ohio
45226.
-------
TABLE OF CONTENTS
Page
FOREWORD ii
ABSTRACT iv
INTRODUCTION 1
SUMMARY 3
LABORATORY REGENERATING EQUIPMENT 7
OPERATING PROCEDURE FOR ROTARY TUBE REGENERATION 15
MULTIPLE HEARTH FURNACE 16
MULTIPLE HEARTH FURNACE REGENERATION 18
TEST EQUIPMENT AND PROCEDURES 20
EXPERIMENTAL RESULTS 33
HEARTH FURNACE THERMODYNAMICS 70
DISCUSSION AND CONCLUSIONS 79
RECOMMENDATIONS 83
APPENDIX A, IODINE NUMBER 84
APPENDIX B, MOLASSES NUMBER . 89
LIST OF REFERENCES 91
ill
-------
ABSTRACT
The regeneration of spent granular activated carbon,
used in tertiary treatment of municipal waste water, is a
three-step process, (1) drying, (2) baking and (3) activating.
Wet, drained carbon, as received from the drain bin, is de-
watered in the drying step at carbon temperatures in the range
60°F to 212°F. During the baking step, the organic adsorbate
is carbonized, accompanied by evolution of gases and a for-
mation of free carbon residue in the micropores of the acti-
vated carbon. This occurs in the temperature range 212°F to
1500°F. In the activating step, the goal is to oxidize the
carbon residue with minimum damage to the basic pore structure
and thereby restore the original properties of the virgin car-
bon. In multiple hearth furnace regenerations, the activation
occurs at an activating gas temperature of 1700°F and carbon
temperatures in the range 1500°F to 1650°F. The activating
gas is flue gas with varying amounts of additional steam.
Laboratory studies with an indirect heated rotary
tube furnace demonstrated that the rates at which the drying
and baking steps are performed do not have a significant effect
on the properties of the baked carbon. The controlling step
for the regeneration product is the activation, and the im-
portant parameters are the carbon temperature, length of
activating time, and steam or CC>2 concentration in the acti-
vating gas mixture.
In multiple hearth furnace regenerations, where the
reported carbon losses are 5 to 7% per regeneration, the
characteristics of the activating step are: (1) high activating
gas input rate, (2) relative short activating time, and (3)
relatively low carbon temperature. Operation of the hearth
furnace requires a high activating gas input rate because it
is direct-fired and because of the high heat requirement for
the drying and baking steps. In the laboratory investigations,
a set of operating conditions were found with the rotary tube
regenerator, whereby the regeneration losses were reduced to
2%. The characteristics of the activating step for the rotary
are: (1) low activating gas input rate, (2) longer activating
time, and (3) higher activating temperature. In the hearth
furnace, activating rate is controlled by the carbon temperature
and in the rotary tube regenerator, by the activating gas input
rate. Mean values for these parameters for the two methods of
activation are:
IV
-------
Rotary tube Hearth furnace
N2-std ft3/lt> carbon
C02-std ft3/lb carbon
Steam-std ft3/lb carbon
O2~std ft3/lb carbon
Activating time, min
Mean carbon temperature, °F
Gas utilization (CO + H2O), %
7.4
1.0
1.9
0.0
16
1630
83
27.3
3.3
23.5
0.7
11
1570
18
Thermodynamic calculations made on the hearth fur-
nace regeneration indicate that, by performing only the acti-
vation of a previously baked carbon, the hearth furnace gas
input rate can be reduced drastically and to the point where
it approaches the optimum rate of the rotary tube. Further
laboratory study to determine the feasibility and economics
of this procedure is recommended.
v
-------
INTRODUCTION
The purpose of the work reported herein was to gain
general knowledge of the regeneration process and to determine
the optimum regenerating conditions. Since the direct-fired
multiple hearth furnace is, at present, the most widely used,
the laboratory work was oriented to permit translation of
laboratory results to hearth furnace operations. Construction
of a miniature multiple hearth furnace was not feasible. In
any event, the versatility of such a miniature furnace for
parametric studies would be limited by the direct firing
feature. A continuous-feed, externally-heated, rotary tube
regenerator, which met the requirements of the study, was con-
structed and used. This device permitted independent control
of most of the parameters and, at the same time, permitted
operation at conditions close to those used in the multiple
hearth furnace.
The work described herein was conducted in co-
operation with the Pomona Water Reclamation Plant, Pomona,
California. (The Pomona pilot plant study is a joint project
of the Federal Water Pollution Control Administration and the
Los Angeles County Sanitation District.) For the laboratory
studies, various quantities of virgin carbon and carbons that
had been spent and regenerated were received from Pomona.
The virgin carbon used at Pomona is the Pittsburgh Carbon
Company Filtrasorb 400, a 12x40 mesh coal-based carbon. The
spent carbons were: (1) I7A, spent eight times and regenerated
seven, (2) VIA, spent twice and regenerated once, and (3) once-
spent, virgin Filtrasorb 400. Operating data received from
Pomona on the multiple hearth furnace permitted thermodynamic
calculations and a closer study of the furnace operating con-
ditions.
In the beginning of the experimental studies, it was
recognized that the regeneration process would involve two
steps: (1) that of drying, since the carbon as received from
the contactors is wet and (2) the activation by steam, C02
and possibly O2. Because of the high heat of vaporization of
water, the carbon temperature was expected to remain near 212°F
until most of the water was driven off. At this temperature
no other activity was expected to occur.
Early in the study, it became evident that another
step occurred immediately following the drying step and before
the activating step. In this step, called the baking step, the
adsorbate is carbonized in such a manner that part of it is
evolved as gases and the rest broken down to free carbon. This
process occurs in the 212°F to 1500°F temperature range. The
-------
regeneration was then broken down to three steps; (1) drying,
(2) baking and (3) activating. These steps were then studied
individually as unit operations to determine their effect on
the overall regeneration.
The degree of success in the three regenerating
steps was determined by five measurements, namely:
1. Bulk volume
2. Particle volume
3. Real density or volume of
nonporous carbon
4. Iodine number
5. Molasses number
-------
SUMMARY
Adsorption by granular activated carbon is among the
most promising advanced waste treatment processes for tertiary
treatment of sanitary waste water. Its acceptance depends on
an economical regeneration of the carbon to permit its reuse
over many adsorption-regeneration cycles. At present, a 2.5
MGD facility is in operation at South Lake Tahoe, California
and a 0.3 MGD pilot unit is under study at Pomona, California.
A large-scale facility is in operation at Nitro, West Virginia
to purify river water for domestic use. At each of these
facilities, the regeneration is accomplished in a direct-fired
multiple hearth furnace. The combustion of natural gas with
air is used as the heat source while carbon dioxide, oxygen
and steam, as part of the combustion products, are the acti-
vating agents. Generally, extra steam is added at a rate of
one pound per pound of regenerated product.
Carbon losses of 5% during each adsorption-regen-
eration cycle have been reported at Tahoe and Nitro and 7% at
Pomona. It is highly desirable to reduce these losses since
the cost of the make-up carbon is a large portion of the re-
generation cost. According to estimates made at the Pomona
Water Reclamation Plant, the overall cost of the treatment with
carbon is 8.3C/1000 gal. of water, of which 1.1C is carbon
make-up cost when carbon losses are 10% in an adsorption-
regeneration cycle.
The work reported herein consisted of laboratory
studies to gain general knowledge of the regeneration process
and to determine the optimum regeneration conditions for car-
bon spent in tertiary treatment of waste water. The regen-
erations were accomplished in an externally-heated rotating
tube regenerator with activating gas and carbon contacted in
a continuous counter-current flow pattern. The regenerator
was designed to allow maximum solid recovery, permitting
accurate mass balances to be performed. The mechanical attri-
tion (physical loss under nonregenerating conditions) in the
regenerator is less than 0.2%. The design permits alteration
of both carbon feed rate and degree of mixing of the granules
with the regenerating or activating gas. The activating gas
consisted of nitrogen, carbon dioxide and steam in the pro-
portions found in flue gas. Excess steam was also used. The
degree of carbon loss was computed from weight, bulk and par-
ticle volume changes and the degree of adsorptive capacity
recovery was determined by the decolorizing tests, iodine
number and molasses number.
-------
The studies demonstrated that the carbon regeneration
is a three-step process: (1) drying, (2) baking (pyrolysis of
adsorbates) and (3) activating (oxidation of the carbon residue
from the decomposed adsorbate). The drying occurs predomi-
nantly in the room temperature to 212°F range, the baking from
212°F to 1500°F and activating at temperatures above 1500°F.
These are naturally-occurring steps that also characterize the
multiple hearth regeneration. With Pomona six-hearth furnace
operating data and knowledge obtained from the laboratory
studies, thermodynamic calculations were made on the Pomona
furnace operations. A computed carbon temperature profile for
the furnace clearly defined the three steps, with the drying
occurring on the first three hearths, baking on the fourth and
activating on the fifth and sixth.
The studies carried out on the three steps of the re-
generation process showed that the manner in which the drying
and baking are carried out has no significant effect on the
properties of the baked carbon. Length of baking time was
varied from 4 minutes to 6 hours with slow rise in temperatures
and with a fast rise in temperature, in all cases to 1500°F and
higher. The bulk density of the baked product was always the
same within narrow limits. In every case, the nonvolatile
portion of the adsorbate was reduced to essentially free carbon
as determined by the real density measurements. Carbons baked
under a variety of conditions, when subsequently activated
under identical conditions produced final products of nearly
identical properties. Results of these studies showed that
the carbon dyring and baking conditions were also not decisive
factors in the quality of the product from the multiple hearth
furnace.
Laboratory regeneration studies produced a set of
activating conditions which essentially recovered the physical
properties of the virgin carbon with a carbon loss of about
2.0%. The activating conditions were characterized by a low
gas input per pound of regenerated carbon, an activating time
of 16 minutes, and a relatively high activating temperature of
1630°F. Under these conditions, the activating gas utilization
was 83%. In contrast, the activating conditions in the multiple
hearth furnace are characterized by a high gas input rate, an
activating time of 11 minutes, and a lower activating tempera-
ture of about 1570°F. The gas utilization in the multiple
hearth furnace was about 18%. In the rotary tube, the oxidative
controlling parameter is the activating gas input rate, while
in the hearth furnace, the controlling parameter is the car-
bon temperature. The experimental evidence obtained indicates
that the high activating gas input rate is the cause of the
higher carbon losses in the hearth furnace regeneration.
-------
Mean values of these parameters for the two methods
of activation are given below. The gas input is per Ib of re-
generated carbon.
Rotary tube Hearth furnace
N2-std ft3/lb carbon
C02-std ft3/lb carbon
Steam-std ft^/lb carbon
02-std ft3/lb carbon
Activating time, min
Mean gas temperature, °F
Mean carbon temperature, °F
Gas utilization (CO 4- H2O) , %
7.4
1.0
1.9
0.0
16
1670
1630
83
27.3
3.3
23.5
0.7
11
1700
1570
18
The carbon activated in the rotary tube is once-spent Filtra-
sorb 400 and was regenerated to the initial virgin carbon
properties. The physical properties of this carbon, activated
in the rotary tube, are given below. The data shows the pro-
gression through the three steps of the regeneration. For
comparison, the properties of the initial virgin Filtrasorb
400 are also given.
Dried
Spent
Bulk density, g/cc 0.584
Particle density,g/cc 0.971
Real density, g/cc 1.88
Pore volume,cc/cc p.v.*0.483
Iodine number 673
Molasses number 174
*p.v. = particle volume
Baked Activated
0.532
0.871
2.07
0.578
955
193
0.502
0.825
2.13
0.612
1017
192
Virgin
Filtrasorb 400
0.490
0.810
2.15
0.618
1010
200
When the drying and baking functions are eliminated
from the hearth furnace regeneration (i.e., when the carbon is
not placed in the regenerator until dried or baked) the com-
puted operating conditions begin to approach those of the
laboratory rotary tube activation, as shown below. The colum-
nar captions refer to the conditions of the carbon (i.e.,
pretreatment) before it is charged into the activator.
-------
Hearth Furnace Activating Conditions
Dried Spent CarbonBaked Spent Carbon
Activating time, min 15 15
Mean gas temp., °F 1700 1680
Carbon temp., °F 1500 to 1650 1500 to 1650
N2-std ft3/lb carbon 16.5 14.3
C02-std ft3/lb carbon 2.0 1.7
Steam-std ft3/lb carbon 14.2 12.3
02-std ft3/lb carbon 0.45 0.36
Act. gas utilization, % 29.5 43.0
In the furnace activation of the baked spent carbon, about
41% of the gas-air input is required to heat the carbon to
activating temperature and make up heat losses through the
furnace shell; hence, the activating gas input is still
larger than needed for the activation. Relative to the wet
regeneration, the gas-air input rate is about 50%.
Vent gas analysis was investigated as a method for
monitoring the activation. Although it offered some promise,
it also posed unsolved problems.
In the hearth furnace regeneration, hot carbon is
discharged into water. At the Pomona plant, this quenching
action has been found to reduce the iodine number. Laboratory
investigations during the present study produced a decrease
in iodine number units ranging from 25 to 69 compared to 50 to
100 at Pomona.
-------
LABORATORY REGENERATING EQUIPMENT
Rotary Tube Regenerator Unit
A continuous-feed, indirectly heated rotary tube
regenerator was selected for carbon regeneration studies be-
cause it offers the widest range of versatility with respect
to operating parameters. With this unit, activating gas com-
position and input rate can be varied to any practical limit.
Heat input rate can be varied independently of the activating
gas input rate and carbon residence time can be varied in-
dependently of other parameters.
Figure 1 shows the side view of the unit with some
of the essential components indicated by callouts. These
callouts, further identified by letters, are explained in
more detail below.
a. Furnace - electrically heated, three
heating sections.
b. Rotating tube.
c. Volumetric Feeder - acrison, auger type,
(feed range 0.010 to about 0.10 ft3
carbon/hr, feeds carbon through 1/2 inch
tubing to rotary tube).
d. Steam Generator - electrically heated with
calibrated sight glass.
e. Heated vent line.
f. Motor - Bodine variable speed drive for
tube, (tube rotation range 0 to 2.4 rpm).
Figure 2 shows additional components of the re-
generative unit.
g. Superheater - gas burner heated for the N2,
CO2 steam mixture.
h. Product Receiver - two quart Mason jar.
i. Water Manometer - indicates pressure in
unit, normal pressure less than 5 mm.
Figure 3 shows the temperature control and vent gas
monitoring instruments. Letters j, k and 1 respectively denote
-------
a. Furnace
b. Rotating tube
c. Volumetric Feeder
d. Steam Generator
e. Heated Vent Line
f. Motor
FIGURE 1 - ROTARY TUBE REGENERATOR, SIDE VIEW
-------
g. Superheater
h. Product Receiver
i. ",/ater Manometer
FIGURE 2 - ROTARY TUBE REGENERATOR, END VIEW
-------
i,k,l Temperature Controllers
rn Tennerature Recorder
n CO Analyzer
o CO2 Analyzer
n H20 Analyzer
q II2 Analyzer
r CO2 Flov; Meter
TJ2 Flow Meter
O2 Flow Meter
:
FIGURE 3 - OPERATING CONTROL PANEL FOR ROTARY TUBE REGENERATOR
-------
West temperature controllers, ammeters and variable trans-
formers to control the temperatures in the three sections of
the furnace shown in Figure 4.
m. Temperature Recorder - twelve point
Speedomax Type G. Five points are
used to monitor the temperature in
the rotary tube, indicated by the
thermocouples TC(1), TC(2), etc.
n. CO Analyzer - Lira Model 300, 0.35% range.
o. CO2 Analyzer - Lira Model 300, 0-15% range.
p. H2O Analyzer - Lira Model 300, 0-35% range.
q. H2 Analyzer - Thermatron, 0-20% range.
r. CO2 Flow Meter - 2.1 std ft3/hr max.
s. N2 Flow Meter - 19.0 std ft3/hr max.
t. O2 Flow Meter - 4.3 std ft3/hr max.
Figure 4 shows the cross sectional side view of the electri-
cally heated furnace, rotary tube, sealed end caps and pro-
duct receiver. The tube assembly is completely sealed to
prevent any loss of gases or carbon. An accurate material
balance is potentially possible although this has not been
accomplished on the gas input and output.
Other pertinent details of the unit are listed below:
Tube size and material - 3.25 in. ID x 65 in. length,
316 stainless steel
Flights - six 1/4 in. deep flights about
the full length of tube. This
depth flight gives carbon a
rolling action rather than a
lift and free fall.
Heating - the heated portion of tube
corresponding to furnace length
is 42 inches, carbon residence
times are reported for the 42
inches rather than the total
65 inches.
11
-------
.Carbon Feed
Tube
-Carbon Seal
Heating Sections of Furnace
\ Rotating
I \ TGas Seal
A in n
I
TCI
II
TC II
III
TC III
Rotating
\ ^- End
TCO}
TCf4) TCC5)
.Gas Outlet
let/ /
Rotary Tub
*-* L Dri
Kf f * ' * ** ''* * * * *
Type
ZJ
Coil Type
Electric Heaters
Gas Inlet
Drive
Sprocket
10
20
30 40
Inches
50
Scale: 1/4" = 2"
Regenerated
Carbon Re-
ceiver - 2 qt
Mason
FIGURE 4 - CROSS SECTIONAL VIEW OF ROTATING TUBE,
END CAPS AND HEATING FURNACE
-------
Tube inclination
Carbon residence time
has been varied from level to
4 in. per 42 in. of tube length.
carbon residence time for 42 in.
of tube length, has been varied
from 10 min to 3.7 hr by varying
inclination and rotation rate.
Operating Characteristics of Rotary Tube Regenerator
The rotary tube regenerator unit was sized to handle
about 2300 cc bulk volume of carbon at feed rate of 0.016 ft3/
hr (450 cc/hr) in a 9-hr work period. In the baking runs, the
feed time for about 2290 cc carbon was about 5 hr and in the
activating runs the feed time for about 2040 cc was about
4.5 hr.
In order to maintain good operating control of the
regenerator and to obtain reliable operating data, the major
part of the carbon throughput must occur under steady-state
conditions. The degree of steady-state operation obtained on
a run depends on the residence time. For the carbon charge
volumes and feed rate mentioned in the previous paragraph,
the degree of steady-state operation achieved for different
residence times are given below:
Residence Time
2.0 hrs
1.5 hrs
1.0 hrs
Steady-state operation as percent
of total run time
Baking
61
71
80
Activating
56
67
78
30 min
15 min
10 min
90
94
98
89
94
98
The more important data in this study were obtained from the
shorter runs of 30 min or less, where the percentages were
90 and over. The greatest deviation from steady-state oper-
ating conditions occurred at start-up when the carbon front
meets the activating gases at full strength over the total
tube length.
13
-------
Another important parameter is the activating gas
residence time. During the study, the gas residence time
varied from 13 to 36 seconds, compared to only 6 seconds for
hearth furnace operations. Other conditions being equal, an
increased residence time increases the utilization of the
activating gases, i.e.,a larger percentage of them are used.
Other operating parameters that should be recognized
are listed below, although their effect on the activation pro-
cess has not been determined.
Parameter Range
carbon loading of rotary tube 1.3 to 15.8% of tube
volume
carbon bed depth at maximum point 0.16 to 0.68 inches
carbon roll over rate 1.1 to 14.5 times/min
14
-------
OPERATING PROCEDURE FOR ROTARY TUBE REGENERATION
Charge and Sample Preparation^
The wet spent carbon received from the Pomona plant
is first dried for 48 hrs at 150°C in a small air convection
oven. About 3 liters of carbon, divided into four 1-liter
beakers, are dried at a time. The dried spent carbon is then
sieved to remove any +14 mesh aggregates and -40 mesh fines.
Next, the carbon is blended to insure uniform particle size
distribution throughout the charge. The blending is done by
separating the carbon into an upper half fraction and a lower
half fraction and then pouring the two fractions simultaneously
back into a common container. This routine is repeated four
times.
The charge for the baking run consists of a weighed
2450 cc of the blended carbon, with volume measured by the
technique described later under test procedures. During the
run, there is usually some segregation of the particles; the
large particles appear to travel through the regenerator
faster than the small ones. After baking,the product is then
blended by the above procedure, reweighed and volume measured.
Samples are taken out for various tests and 2200 cc are charged
to the unit for the activation run.
The feed hopper has a space below the feeding augar
where 160 to 180 cc of carbon charge remains after completion
of the feeding. This carbon is removed by means of a spout
at the bottom of the hopper. After removal, its volume and
weight are subtracted from the volume and weight of the charge
to determine the amount of carbon charge fed to the regene-
rator.
Control of Gas Temperature in Tube
The heating furnace consists of three sections which
can be adjusted individually to maintain the gas temperature
in the tube at thermcouples TC(2) , TC(3) and TC(4), shown in
Figure 4, at the desired temperatures. The reported tempera-
tures by these three thermocouples are recorded in tables
under the heatings TC(2), TC{3) and TC(4).
Vent Gas Analysis
The gas analyzers for CO, C02 and H2 were cali-
brated against standard mixtures of each gas with nitrogen
before the beginning of each run. The water analyzer was not
used during the program because of difficulties with its
15
-------
operation. One and one half cubic feet per hour of gas, dried
by passing through a dry ice trap, was withdrawn from the vent
line and passed through the analyzer. Based on the known
nitrogen flow and percentages of CO, C02 and H2, the flow rates
of these three gases were then computed. A water-cooled con-
denser was installed on the vent line to obtain a measure of
the water vapor remaining in the vent gases.
MULTIPLE HEARTH FURNACE
Figure 5 shows the essential features of the six-
hearth furnace used at the Pomona Water Reclamation Plant.
The furnace consists of a vertical refractory-lined cylindrical
shell containing a series of horizontal hearths positioned one
above the other. The wet spent carbon enters hearth number 1
at the outer edge and is then mechanically moved in a spiral
pattern toward the center of hearth 1 by the rotating rabble
arms. Near the center the carbon drops to hearth number 2
and is mechanically moved to the outer edge of hearth number 2
where it drops to hearth 3. In this manner, the carbon tra-
verses each of the six hearths until delivered from discharge
pipe of hearth number 6. At each of hearths 4, 5 and 6 are
two gas burners and steam inlet tubes, one on each side of the
cylindrical shell (not shown). The hot steam and flue gases
from these burners travel upward through the hearths, counter-
current to the movement of the carbon.
A principal advantage of the multiple hearth furnace,
over the direct-fired rotary type is its greater reaction sur-
face-area-to-furnace-volume ratio, which minimizes capital
costs while maximizing throughput capacity. Compared to in-
direct-heated rotary tube furnaces, the multiple hearth furnace
has the advantage of interior construction of predominantly
refractory materials. The indirect-heated unit's metal tube
is more expensive and not as durable.
A primary disadvantage of any direct-fired furnace
is that the heating requirement cannot be readily separated
from the activating gas requirement. The ranges over which
the operating parameters can be varied independently are thus
greatly restricted. In carbon regeneration, this has a direct
bearing on the carbon losses. In this respect the indirect-
heated tube regenerator has definite advantages.
16
-------
Carb
Gas Out
Hearth
6'
3' Diameter
Rabble Arm
Rabble Teeth
Two Burners and
Steam Inlets at
Hearth 4, 5 and 6,
not shown
Carbon Out
FIGURE 5 - CROSS SECTIONAL VIEW OF MULTIPLE HEARTH FURNACE
USED AT POMONA WATER RECLAMATION PLANT
17
-------
MULTIPLE HEARTH FURNACE REGENERATION
The three steps (drying, baking and activating) of
regeneration are naturally occurring steps that also charac-
terize the multiple hearth regeneration. With operating data
from the Pomona furnace and knowledge obtained from the labo-
ratory studies, thermodynamic calculations were made on the
Pomona furnace operations. The carbon temperature profile, as
given in Figure 6, represents a part of these results. As in-
dicated in the figure, drying occurs on the first three hearths,
baking on the fourth and activating on the fifth and sixth.
Drying requires 15 minutes, baking 4 minutes and activating
about 11 minutes. The gas temperature profile was determined
by direct measurement.
For the above thermodynamic calculations, the re-
generated carbon discharge rate is about 89.4 Ib/hr. The
calculated wet spent carbon feed rate to the top of the re-
generator is 194.0 Ib/hr, of which 77.5 Ib/hr is water, 13.2
Ib/hr is volatiles driven off during baking, 7.6 Ib/hr is
adsorbate carbon residue oxidized during activation and 6.3
Ib/hr is loss of basic carbon structure during activation.
The 6.3 Ib/hr represents the 7% carbon loss during the re-
generation part of the cycle. The initial adsorbate "content"
is 20.8 Ib/hr, of which 13.2 Ib/hr is volatilized during baking,
leaving a 7.6 Ib/hr free carbon residue, which is then oxidized
during the activating.
In terms of heat input, the drying step consumes 40%
of the heat, and the baking 8%. If the drying and baking were
done outside of the hearth furnace, the heat requirement or
gas-air input would be approximately 52% of the amount required
for wet carbon regeneration. The excess gas-air input to
carry the drying and baking functions supplies the activation
step with excess activating gases such that the risk of damaging
the carbon is high. In normal wet carbon regeneration, about
18% of the activating gases are utilized - about 8.5% for use-
ful oxidation of adsorbate carbon residue (7.6 Ibs/hr) and
7.0% for external oxidation of the carbon particles. On suc-
cessive adsorption-regeneration cycles, the iodine number was
observed to decrease from the 1097 for the virgin carbon to
655 for the regenerated carbon on the 8th cycle. Along with
the 7.0% carbon volume loss, a loss of micropore volume is
also occurring.
In view of the high throughput rate of activating
gases, the tendency for external oxidation is greatly moderated
by the short (11-minute) activating time and the relatively low
(1500°F) carbon temperature.
18
-------
3500
3000
2500
2000
3
4J
2 1500
-------
TEST EQUIPMENT AND PROCEDURES
Bulk Volume Measurements
The bulk volume or bulk density is a conveniently
measured property of the carbon that can be used for control
purposes during regeneration. The bulk density of virgin
Filtrasorb 400 has been measured at 0.486, 0.495 and 0.501
g/cc, hence the goal for the regeneration is a density in
this range. With experience the operating parameters can be
adjusted to consistently achieve final density in this range.
A measured change in bulk volume, before and after
regeneration, is a partial measure of the loss of basic car-
bon structure. This method is used at the Pomona Water Re-
clamation Plant. The contactor is initially filled to a
known level; after the carbon has been spent and regenerated,
it is returned to the contactor. The weight of additional
carbon to fill the contactor to the initial volume represents
the loss of carbon on the adsorption-regeneration cycle.
The apparatus for measuring the bulk volume, and
also the bulk density, is shown in Figure 7. It consists of
a funnel-hopper at the top, a one liter graduated cylinder
at the lower end and a glass tube section in the middle con-
taining 4 or 5 screens of 1/4 in. mesh. All parts of the
apparatus must be aligned around the center axis and be per-
fectly plumb to give an even horizontal carbon surface in the
graduated cylinder. The hole in the funnel is about 5/16
inches in diameter, allowing a flow rate of about 400 cc/min
for a 14 to 40 mesh carbon.
The bulk volume apparatus has been found to give
very reproducible results when the carbon is well mixed, i.e.,
is blended. Bulk volume measurements are given in Table 1 for
a complete regeneration run which demonstrates the reproduci-
bility of the measurements and the method for determining the
bulk volume decrease.
The bulk volume apparatus does not give reliable
measurements for volumes below 100 cc. The 30 cc on run 51
was calculated using the weight and 0.489 g/cc bulk density.
To calculate the bulk volume decrease, it is
necessary to measure the carbon remaining in the feeder hopper.
For run 50, it was 180 cc and for run 51, 160 cc.
20
-------
Hopper
4 Screens
1/4" Mesh
•1000 cc Graduate
JL
.Leveling Screws
Scale: 1/4" - 2"
/ / s s ///'/''/'/
FIGURE 7 - APPARATUS FOR MEASURING BULK VOLUME
21
-------
TABLE 1 - BULK VOLUME AND WEIGHT MEASUREMENTS FOR
RUNS 50 AND 51
1. Feed for baking run 50
Charged to feeder hopper,
Measurement
No. Bulk volume, cc
(1)
(2)
(3)
2. Baking run
Measurement
No.
(1)
(2)
(3)
3 . Activating
Charged
Measurement
No.
(1)
(2)
(3)
4 . Activating
Measurement
No.
(1)
(2)
(3)
1000
1000
450
2450
50, product
Bulk volume , cc
1000
1000
248
2248
run 51 feed
to feeder hopper,
Bulk volume, cc
1000
1000
200
dried spent
Weight, g
615.5
615.6
277.6
1408.7
Weight, g
548.5
547.7
134.3
VIA carbon
Density, g/cc
0.615
0.616
0.617
mean 0.616
Density, g/cc
0.549
0.548
0.542
1230.5 mean 0.547
baked product run 50
Weight, g Density, g/cc
548.5
547.7
108.9
2200 1205.1
run 51 product (regenerated)
Bulk volume, cc Weight, g
1000
1000
30
2030
489.4
488.8
15.6
993.8
0.549
0.548
0.545
mean 0.548
Density, g/cc
0.489
0.489
mean 0.489
22
-------
Bulk volume decrease for baking run 50
2450 -(180 + 2248) = 22 cc
or
Bulk volume decrease for activating run 51
2200 -(160 + 2030) = 10 cc
or
10 100 = 0.49%
2040
Bulk volume decrease for regeneration
0.97 + 0.49 = 1.46%
While the above values are not truly additive from a mathe-
matical standpoint, no significant error is being introduced.
An error can occur in measuring the carbon loss in
this manner if the carbon particle size distribution changes
during the regeneration. This is due to possible void space
changes between the particles caused by the change in dis-
tribution of the particles. The void space may remain un-
changed, increase or decrease. The bulk volume decrease may
or may not be, in its entirety, a loss of carbon but also, in
part, a change in void space.
Particle Volume Measurements^
The particle volume measurement excludes the void
space between the particles, but includes the solid carbon
structure and pore volume. From the standpoint of void space
change, the particle volume measurement is a more accurate
method for measuring carbon loss during regeneration.
This measurement is made with mercury displacement
with an apparatus of the type shown in Figure 8. To carry
out the measurement, the (evacuated) volume of the empty
sample flask is first measured by running a measured volume
of mercury from the volumetric bulbs and buret into the flask.
The flask volume starts at the midpoint of the capillary tube
just below the flask. By adjusting the mercury level in the
buret and applying helium pressure or air at atmospheric
23
-------
Null
Manometer,
Dibutyl-
phthalate
Capillary
1 mm ID
Buret-
Vacuum
Pump
Volumetric
Bulbs
Scale: 1/4" - 2'
Mercury
Manometer
Mercury
Leveling
ftulb
FIGURE 8 - APPARATUS FOR MEASURING REAL
AND PARTICLE DENSITIES
24
-------
pressure to the mercury in the buret, the mercury pressure in
the flask is adjusted to 900 mm of mercury. This procedure is
repeated with a weighed amount of dry carbon in the sample
flask; tha mercury volume difference then represents the par-
ticle volume of the carbon sample. About 20 to 30 min evac-
uation time is required to degas the sample sufficiently for
an accurate measurement. A final pressure of 900 mm of mer-
cury is required for the mercury to outline the irregular
carbon particle surfaces. This pressure is not sufficient to
cause mercury penetration into the pores.
The particle density of virgin Filtrasorb 400 is
between 0.807 and 0.820 g/cc; ideally, regeneration should
attain a density in this range. If a change in particle size
distribution occurs in any of the adsorption-regeneration
cycles, it may become impossible to attain, simultaneously, a
bulk density in the 0.486 to 0.501 g/cc range and a particle
density in the 0.807 to 0.820 g/cc range. For the several
spent carbons (dried) which have been regenerated during this
study, the bulk and particle densities are given below with
the calculated bulk volume void space.
For comparison, the void spaces for two virgin
Filtrasorb 400 carbons are 0.398 and 0.389 cc/cc bulk volume.
The void space of the once-spent Filtrasorb 400 (Table 2) is
within the void space range of the virgin carbon, indicating
that the particle size distribution of the two are still the
same. Spent 17A and VIA carbon have smaller void spaces,
indicating that the particle size distributions have shifted.
Sieve analyses indicated that the distributions have shifted
toward the smaller particle size range. This is a reflection
of carbon loss since 17A was regenerated 7 times and spent 8
times and VIA regenerated once and spent twice.
TABLE 2 - BULK AND PARTICLE DENSITIES OF SPENT CARBONS
Bulk density Particle density Void space
Carbon g/cc g/cc cc/cc
17A 0.601 0.970 0.381
VIA 0.616 0.990 0.378
Once spent
Filtrasorb
400 0.584 0.971 0.398
The void space is calculated with the formula
1 000 - bulk.density/ 3/cc _ void space, cc/cc
particle density, g/cc bulk volume
25
-------
To determine the particle volume decrease during re-
generation, four measurements are required for each run, namely
carbon weight before and after the run and particle density be-
fore and after the run. For runs 50 and 51, described earlier
in Table 1, the necessary quantities and calculations are
summarized in Table 3.
TABLE 3 - PARTICLE VOLUME MEASUREMENTS FOR RUNS 50 AND 51
Baking Run 50
Weight fed to regenerator (VIA), g 1401.0
Particle density, g/cc 0.990
Particle volume, cc 1415
Weight of product, g 1230.5
Particle density, g/cc 0.881
Particle volume, cc 1397
Particle volume decrease, cc
Particle volume decrease, %
Activating Run 51
Weight fed to regenerator (raw 50), g 1118.8
Particle density, g/cc 0.881
Particle volume, cc 1270
Weight of product, g 993.8
Particle density, g/cc 0.789
Particle volume, cc 1259
Particle volume decrease, cc
Particle volume decrease, %
Total particle volume decrease
for regeneration 2.14%
A comparison of bulk volume and particle volume de-
creases, as detailed below, show that the particle volume de-
crease is larger for both runs.
particle volume
bulk volume decrease^% decrease, %
Run 50
Run 51
26
-------
Over 20 runs each of baking and activating, it was observed
that the bulk particle volumes fluctuated with one or the
other being larger. The overall averages for the 20 runs each
given below indicate that the two methods are comparable for
a single regeneration.
mean bulk volume mean particle volume
decrease/ % decrease, %
20 baking runs 1.16 1.15
20 activating runs 0.46 0.97
1.62 2.12
Eighteen of these regenerations were of VIA and two were of
once-spent Filtrasorb 400.
Because the whole sample is measured, bulk volumes
can be measured with greater precision than the particle
volumes. Because particle volumes are measured on 25 cc
sized samples, sampling errors can very easily occur. To
avoid sampling errors, the total charge or product is blended
and the sample for testing taken with a spoon rather than by
pouring.
Real Density Measurements
The real density is measured by helium displacement
and includes only the volume of the solid carbon structure.
One of the useful applications of real density is that it
gives an indication of the state of decomposition of the ad-
sorbate on baking or activation. Real density can also be
used, in conjunction with the particle density, for deter-
mining the total pore volume.
Real density measurements are made with the same
apparatus (Figure 8) that was used for particle volume measure-
ments. The sample flask volume is first measured when empty
with helium and again with a weighed carbon sample, the dif-
ference in volume being the volume of the solid carbon struc-
ture. The sample flask volume includes the tubing to the
stopcocks below the flask. The flask, with or without carbon,
is first evacuated from 20 to 30 minutes and then both stop-
cocks below the flask are closed. With mercury levels down,
the volumetric bulbs and buret are filled with pure helium at
atmospheric pressure. Both stopcocks on the null manometer
are opened simultaneously and momentarily to equalize the
pressure in the null manometer with the pressure in the rest
of the system. The lower left stopcock is opened to let the
27
-------
helium enter the flask. The mercury levels in the volumetric
bulbs and buret are raised to increase the helium pressure to
the initial pressure as determined by the null manometer. The
volume of helium then displaced at constant pressure is the
void volume in the sample flask.
The real density of virgin Filtrasorb 400 is 2.15
g/cc, the density a goal of regeneration. This goal, however,
is easily attained since, after the baking step, the real den-
sity is close to 2.15 g/cc. Real density less than 2.15 g/cc
definitely indicates that volatile adsorbates are retained in
the carbon. The real densities and pore volumes, as given be-
low, for runs 50 and 51 are typical.
Real density, Pore volume,
g/cc cc/cc particle vol.
Feed for Run 50 (VIA) 1.88 0.471
Product Run 50 2.14 0.588
Product Run 51 2.16 0.631
Virgin Filtrasorb 400 2.15 0.623
The pore volume is calculated with the formula
1.000 - Particle density, g/cc = pore volume cc/cc
real density, g/cc particle volume
Iodine Number
Measurement of iodine number is a standard decol-
orizing test used in the activated carbon industry for liquid
phase carbons. It is defined as the milligrams of iodine ad-
sorbed by one gram of carbon when the iodine concentration of
the residual filtrate is 0.02 normal. The details of this
test are given in Appendix A.
At the Pittsburgh Carbon Company^'^ the iodine num-
bers of about 34 carbons of several raw materials and several
methods of manufacture were correlated with surface area
measurements. From this work the following equation was de-
rived which relates the cumulative surface area, of pores
larger than 10 A diameter, with the iodine number
Y = 17 + 1.07x
28
-------
where Y is the iodine number and x is the surface area,0in
square meters per gram carbon, of pores larger than 10 A in
diameter. The coefficient of correlation is 0.97 and standard
of error of estimate is 71.5.
The iodine number is used in this program to measure
the extent of recovery of adsorptive capacity or, stated in
other terms, the extent to which the micropores have been
cleared of adsorbate. The micropores are in the 10 to 30 A
diameter range and also contain 95% of the Filtrasorb 400 total
surface area. The iodine number of virgin Filtrasorb 400 has
been found to vary from 985 to 1020 mg/g for different samples.
If J020 mg/g is used, the surface area of pores larger than
10 A diameter as calculated with the above equation is 940 m^/g
while the total surface area as measured by nitrogen adsorption,
using the BET^ method, is 1050 m2/g.
The 985 to 1020 mg/g iodine number range is the
fourth goal to be achieved on regeneration. However, it is not
implied that the iodine number is a direct measure of the ad-
sorptive capacity for biologically nondegradable molecules
present in sewage. Evidence has been presented showing a good
correlation between COD adsorption and iodine number.4 On the
other hand, powdered Darco carbons with low iodine numbers
have been found to have high adsorptive capacities for these
molecules. The iodine test in this program is being used as
a gauge to determine the state of recovery of original pro-
perties of Filtrasorb 400.
Iodine number measurements for runs 50 and 51, show
the progress made during regeneration.
Iodine number
Feed for Run 50, spent VIA 529
Run 50, product 754
Run 51, product 900
The iodine number of this regenerated product falls short of
the virgin carbon iodine number although the bulk and real
densities are at the right level.
Molasses Number
The bulk volume decrease and, more reliably, (when
the loss occurs by way of particle size decrease), the particle
volume decrease measure the carbon loss. Particle size de-
crease can occur by direct oxidation of the particle outer
29
-------
layers by the activating gases. It can also occur from normal
handling attrition if the basic carbon structure has been
weakened by internal overactivation. Some of the attrition
products, or fines, are discarded by back washing in the con-
tactor, or are carried up the furnace stack.
Internal overactivation manifests itself through in-
crease in macropore volume, i.e., increase of pores larger than
30 A diameter. During the process of oxidizing the adsorbate
residue in the micropores, the activating gases come in contact
with the basic carbon structure resulting in enlargement of
micropores to macropore size and further enlargement of macro-
pores. This process thins out the pore walls, making the car-
bon particles more friable or susceptible to attrition.
The molasses number measures the pore enlargement
since it has been found that the molassesenumber relates to
the surface area of pores larger than 28 A diameter (references
1 and 2). The equation showing this relationship is
Y = 129 + x
where Y is the molasses number and x the surface area in square
meters per gram of carbon. Molasses consists of a concentrated
sucrose solution with a considerable concentration of color
bodies. Although the color bodies are more strongly adsorbed
than the sucrose, the large excess of sucrose fills up the
micropores thereby greatly slowing up the color body adsorp-
tion. An exchange of color bodies for sucrose is slow in the
micropores but much faster in the macropores, hence the corre-
lation of macropore area with molasses number.
The molasses number for various virgin Filtrasorb 400
samples has been measured in the range 190 to 216, which is the
fifth goal to be achieved during regeneration. Again, using
the runs 50 and 51, the progression of molasses number during
regeneration is demonstrated.
Molasses number
Feed for Run 50, spent VIA 172
Run 50, product 198
Run 51, product 262
Although the external burn off of the carbon particles was
quite small, the high molasses number of 262 and low iodine
number of 900 indicate that internal damage had been done to
this carbon on this regeneration or the prior one.
30
-------
The details of the molasses test procedure are given in
Appendix B.
Other Measurements
Two other measurements or computations are necessary
to complete the characterization of the regeneration run.
These are the carbon residence time in the regenerator and the
percentage of external burn off.
The residence time is calculated with the following
formula:
v
Vi - Ve
— _ hours residence time in heated
65
zone of regenerator
where:
V- = the bulk volume of total charge fed to
regenerator
V = the bulk volume of product at time
feeding was completed
V = the total bulk volume of product at
completion of run
V- = the feed rate in consistent unit of
volume per hour
42 = the length of the heated zone
65 = total length of the tube
The percent external burn off is calculated with the
following formula:
v x d
w
x 100 = percent external burn off
where:
v = the decrease in particle volume during a run
d = the particle density at the beginning of the run
w =» the total weight decrease during run
31
-------
As examples of this calculation, using data from
Table 3, the percent of external burn off for runs 50 and 51
are:
Run 50
18 x 0.990
1401.0 - 1230.5
x 100 = 10.4%
Run 51
11 x 0.881
1118.8 - 993.8
x 100 = 7.7%
For run 50, the 10.4% is in reality not a "burn off" since no
oxidizing agents were used, but rather indicates that, of the
total volatiles evolved, 10.4% came from the exterior of the
particles. By difference, 89.6% of the volatiles came from
within the pores of the carbon. For run 51, the 7.7% is a
measure of loss of basic carbon from the exterior and by dif-
ference, 92.3% of the weight decrease comes from oxidation of
free carbon adsorbate residue and oxidation of internal carbon
structure.
32
-------
EXPERIMENTAL RESULTS
Parameters Studied
The goals of the regeneration are to remove the ad-
sorbate from the pores of the spent carbon without oxidizing
or burning away any of the outer portions of the carbon parti-
cles or without altering the basic pore structure. The loss
of carbon from the carbon exterior is measured by the change
in bulk or particle volume while change in pore structure is
indicated by the iodine and molasses numbers. In terms of
physical properties of the spent carbon and regenerated carbon,
the goals of the program can also be expressed by the desired
changes as given in Table 4.
TABLE 4 - DESIRED PHYSICAL PROPERTY CHANGE
ON REGENERATION
Initial Regenerated
spent carbon (VIA) carbon
Bulk density,
g/cc 0.616 0.490
Particle density,
g/cc 0.990 0.810
Real density,
g/cc 1.88 2.15
Pore Volume, cc/cc
particle volume 0.471 0.618
Iodine number 529 1010
Molasses number 172 200
The properties of the regenerated carbon are those of the vir-
gin Filtrasorb 400.
To accomplish the desired physical property changes,
the regeneration studies centered primarily around the areas
of activating, baking and drying, with some time devoted to
vent gas analysis and effect of quenching on carbon properties,
The results are presented below under the following headings:
33
-------
Low-Temperature Activation with High Gas Input
High-Temperature Activation with High Gas Input
High-Temperature Activation with Low Gas Input
Regeneration of Once Spent Filtrasorb 400
Comparison of Hearth Furnace and Rotary Tube Activation
Baking Temperature Profile
Particle Volume Decrease During Baking
Drying Studies
Effluent Gas Analysis
Quenching Experiments
Low-Temperature Activation with High Gas Input
Since the adsorbate is predominantly in the micro-
pores, the activating gases come into contact with the basic
carbon structure before they reach the adsorbate or adsorbate
residue. The activating gases first come in contact with the
exterior surface of the carbon particles and then the macro-
pore surface before reaching the micropores, where the bene-
ficial oxidative action can take place. Depending on the
temperature and other operating conditions, varying amounts
of carbon are oxidized from the exterior surface and macro-
pore surface, along with the oxidation of adsorbate or ad-
sorbate carbon residue and micropore surface.
On the basis of thermodynamic calculations, hydro-
carbons are oxidized by steam and carbon dioxide at consider-
ably lower temperatures than free carbon such as formed by
decomposition of adsorbate or the basic structure of the acti-
vated carbon. The extent of this difference is shown in
Figure 9. The letter f designates the fraction of influent
steam consumed before equilibrium is established while oxi-
dizing n-tetradecyl benzene (£20^34) • toluene (C^Hg) or carbon.
These molecules represent the adsorbate in regard to oxidative
susceptibility. At 800°F and equilibrium conditions, very
little of the free carbon is oxidized, while 70% to 95% of the
adsorbates are oxidized. The possibility of carrying out the
regeneration at this temperature was investigated since it
would assure minimum attack on the basic carbon structure. Re-
sults are discussed below.
34
-------
1.0
600
TSoo
FIGURE 9 - FRACTION f OF STEAM CONSUMED AT VARIOUS
TEMPERATURES WHEN REACTION PROCEEDS TO EQUILIBRIUM
35
-------
In the multiple-hearth furnace, the combined baking
and activating time is about 15 minutes and the activating gas
temperature about 1700°F. Since the oxidation rates at the
800°F temperature level were expected to be slower, a consid-
erably longer (2 hours) regeneration time was used. The in-
vestigation, however, demonstrated that the oxidation rates
were quite slow and that the decomposition of the adsorbate
to volatiles and free carbon progressed to a larger degree
than had been expected.
Table 5 shows the physical property data for two
800°F runs. The feed carbon was 17A that had been dried and
prepared according to the previously described charge pre-
paration procedure. For these two runs, and also for the
other runs given in Table 5 the gas composition and input
rates are given in Table 6. For comparison, the gas compo-
sition and input rates are also given for the Pomona multiple
hearth furnace during a normal wet spent carbon regeneration.
On the basis of per pound of regenerated carbon, the laboratory
activating gas input rate was 12.8 ft3, while for the hearth
furnace it was 27.5 ft3.
TABLE 6 - GAS COMPOSITION AND INPUT RATE
Laboratory rotary Pomona hearth furnace
std ft3/hr std ftyib std ft-Vhf std ftVlb
N2 9.3 18.6 2440 27.3
C02 1.2 2.4 295 3.3
H20 5.2 10,4 2100 23.5
02 67 0.7
Activating gas
totals 12.8 27.5
(N2 not included)
Runs 3 and 5 (Table 5), at 800°F, show a decrease in
bulk density from 0.608 to about 0.570 g/cc and a particle
density decrease from 0.970 to 0.915 g/cc. Considering the
long residence time and large activating gas input rate, this
is a small loss in the carbon density. Later baking studies
definitely indicated that this density decrease was not due
to oxidation of the adsorbate but rather to decomposition of
the adsorbate. For example, 15-gram samples of dried spent
36
-------
TABLE 5 - REGENERATION AT LOW TEMPERATURES, 800 to 1500°F
ui
Feed carbon
or run no.
Dried spent 17A, feed for
runs 3 and 5
Run 3 product
Run 5 product
Virgin Filtrasorb 400
feed for run 4
Run 4
Dried spent 17A, feed
for runs 6, 7 and 13
Run 6
Run 7
Run 8 (run 7 feed)
Run 11 (run 8 feed)
Run 12 (run 11 feed)
Run 13
Temperature
°F
800
800
___
800
___
1000
1200
1200
1400
150
1500
Bulk
density
g/cc
0.608
0.567
0.571
0.501
0.501
0.605
0.566
0.553
0.550
0.521
0.523
0.499
Particle
density
g/cc
0.970
0.915
0.820
0.820
0.970
0.910
0.803
Real
density
g/cc
1.90
2.06
2.15
2.15
1.90
2.07
2.10
Pore volume
cc/cc
particle vol.
0.490
0.555
0.618
0.618
0.490
0.560
0.618
Bulk vol.
decrease
%
0.6
___
0.0
_ — —
0.7
0.5
0.6
0.4
0.05
1.0
Weight
decrease
%
10.0
7.0
— — —
0.4
___
7.7
8.5
1.1
7.4
0.0
18.6
-------
VIA were heated in an oven for 1.5 hr at 850°F in a nitrogen
atmosphere. The samples lost 8.2% in weight, about the same
loss as for runs 3 and 5. The particle density decreased from
0.988 to 0.906 g/cc, which are again comparable to runs 3 and
5 decreases.
Run 4, made with the virgin Filtrasorb 400, showed
no change except for a small (0.4%) weight decrease, indicating,
as could be predicted from Figure 7, that very little oxidation
of basic carbon would occur at 800°F.
On runs 6 to 13, the temperature was increased in
100°F and 200°F intervals to determine whether, at which tem-
perature between 800°F and 1500°F, the spent carbon could be
regenerated at minimum risk to basic carbon structure. The
progression shows that at 1200°F, the regeneration is still
essentially a baking action because the bulk density at 1200°F
is 0.550 g/cc, comparable to 0.545 g/cc, the latter figure
being bulk density of spent carbons baked at a later date at
1600°F. A comparison of runs 7 and 8 show that very little
adsorbate can be driven off at 1200°F by increasing residence
time an additional two hours. When the temperature is in-
creased to 1400°F, (Run 7) an additional amount of adsorbate
is removed in a two-hour period. Part of the removal is due
to the baking action and now partly to steam oxidation since
the bulk density is down to 0.521 g/cc, i.e., below the 0.545
g/cc baked carbon density at 1600°F.
Run 12 at 150°F was made to determine to what degree
the bulk density decreases are due to baking and activating
and to what degree to possible mechanical attrition caused by
the equipment. The zero weight decrease and the small (0.05%)
bulk volume decrease attest to the fact that mechanical attri-
tion loss is nil.
When the temperature was increased to 1500°F (Run
13) the steam oxidation increased considerably, as is indicated
by the final bulk and particle densities which are now slightly
less than those of the virgin carbon. The iodine and molasses
numbers on this run were 520 and 250 respectively. When com-
pared to the 1020 iodine number and 210 molasses number of the
virgin carbon, they indicate that a considerable portion of
the micropores had become enlarged to macropores. Whether this
damage to pore structure was done on this regeneration or on
one of the seven previous regenerations, could not be deter-
mined. At the Pomona plant, on successive adsorption-regen-
eration cycles, the iodine number has been observed to decrease
from 1097 for the virgin carbon to 655 for the regenerated car-
bon on the 8th cycle^, hence the enlargement of pores could
have happened earlier.
38
-------
The low temperature regenerations definitely indicate
that at 1400°F and lower, very little or no steam oxidation of
the adsorbate occurs even at the extended 2-hour regeneration
time and high activating gas input rate. The weight decrease
can be accounted for by decomposition and volatilization of the
adsorbates. In the temperature interval from 1400°F to 1500°F,
the oxidation rate increases rapidly and in a 2 hour regen-
eration time, considerable amount of adsorbate or free carbon
from the adsorbate is oxidized. In the shorter, 11 min, acti-
vation time of the multiple hearth furnace, the amount of oxi-
dation occurring at temperatures below 1500°F would, however,
still be small compared to the oxidation occurring at tempera-
tures between 1500 to 1650°F.
High Temperature Activation with High Gas Input
Because of rapid carbonizing action occurring at
temperatures below 1500°F, the material remaining for the high
temperature activation is essentially free carbon, i.e.,carbon
residue of the adsorbate and the basic carbon structure. The
loss of basic carbon structure cannot then be completely
avoided, hence the best that can be expected is to keep the
carbon loss at as low a level as possible while removing the
adsorbate carbon residue.
Because of the low iodine number (520) on run 13,
further studies were carried out with the main purpose being
to increase the iodine number. The assumption was made that,
at Pomona, on their regeneration of the particular batch re-
ceived by MSAR, the bulk density had been brought down con-
siderably below the 0.500 g/cc of the virgin carbon. Hence,
by regenerating to lower densities the iodine number was ex-
pected to increase.
On runs 14 to 17, the temperature was progressively
increased while maintaining carbon residence time at 2 hrs
and activating gas input as before (see Table 6). The tempera-
tures are recorded in Table 7 and the regenerated carbon pro-
perties in Table 8. As the results show, the iodine number
did not come up to 1020 although the bulk density was brought
down to 0.410 g/cc. The pore volumes were much larger than
the 0.619 cc/cc p.v. for the virgin carbon. The high molasses
number indicated that the pores of the larger size were de-
veloping to the detriment of adsorptive capacity and carbon
hardness.
On runs 18 to 24, the carbon residence time was
progressively decreased from 2.6 to 0.5 hr while still main-
taining activating gas at the prior high input rate and the
39
-------
TABLE 7 - OPERATING CONDITIONS FOR HIGH TEMPERATURE REGENERATIONS
Feed carbon or
run number
Carbon
residence,
hr
Gas temperature, "F
II
III
Weight
decrease
Dried spent VIA feed
for runs 14 to 18 and
20 to 24
Run 14
Run 15
Run 16
Run 17
Run 18
Run 20
Run 22
Run 23
Run 24
Virgin Filtrasorb 400
feed for run 19
2.6
1.4
0.9
0.5
0.5
1500
1500
1500
1550
1700
1700
1700
1700
1700
1500
1500
1550
1650
1650
1650
1700
1700
1700
1500
1550
1600
1600
1550
1600
1600
1600
1600
22.4
25.5
27.0
34.8
37.7
18.3
34.1
30.7
31.6
Run 19
2.0
1700
1650
1550
18.3
-------
TABLE 8 - PROPERTIES OF SPENT CARBONS REGENERATED AT HIGH TEMPERATURES
Feed carbon or
run number
Bulk Particle Real
density, density, density,
g/cc g/cc g/cc
Pore Particle External
volume, volume burnoff,
cc/cc p.v.* decrease,%. %
Iodine Molasses
No. NO.
Dried spent VIA
feed for runs 14
to 18 and 20 to
24
Run 14
Run 15
Run 16
Run 17
Run 18
Run 20
Run 22
Run 23
Run 24
Virgin Filtrasorb
400 feed for run 19
Run 19
0.617
0.481
0.463
0.456
0.410
0.390
0.415
0.418
0.436
0.446
0.495
0.430
0.982
0.755
0.742
0.634
0.698
0.680
0.710
0.731
0.820
0.709
1.88
2.11
2.12
2.18
2.14
2.14
2.15
2.13
2.15
2.15
0.477
0.642
0.649
0.710
0.674
0.683
0.669
0.656
0.619
0.671
6.7
3.8
4.1
4.4
— —
4.3
— — —
17.1
9.8
11.6
13.3
— — —
28.6
529
710
698
690
737
935
945
1001
930
883
995
1160
i/^
306
283
___
388
326
344
379
309
210
320
* p.v. « particle volume
-------
temperatures at near the 1700°F level. In these runs the
iodine number went through a maximum of 1001 at 0.9 hr resi-
dence time and dropped to 883 at 0.5 hr residence time. In
this series of runs, the molasses number stayed at a high
level, considerably above 300 for most of the runs, again in-
dicating a general shift to pores of larger size and also loss
in hardness.
The loss of hardness in these regenerated carbons
was quite apparent because they were very dusty in handling.
Also on the grinding of samples for iodine and molasses, the
samples ground down in less than three hours while the virgin
carbon requires 4.5 hr. (Grinding is done in motor driven
mortar and pestle.)
To further check the effects of high temperature and
high activating gas input on the pore structure of the carbon,
a comparison was made between activity of the carbon particle
exterior layers and the carbon particle interior. This was
done by placing the carbon in a ball mill without balls, but
containing shallow flights. The gentle rolling action over a
24-hour period abraded off some of the outer layers of the
particles. About 1% of fines was abraded off each sample on
each ball mill run. Iodine and molasses numbers were deter-
mined for the fines and interior for each carbon. The results
of these experiments are given in Table 9. The results show
that overactivation proceeds from the outer layers toward the
carbon particle center. In the extreme outer layer, the volume
of small pores, as indicated by the lower iodine number, is con-
siderably less than in the interior. On the other hand, the
activity distribution is more uniform for the virgin Filtrasorb
400. The enlargement of the pores in the outer layers of these
regenerated carbons appears to be the cause of the decreased
abrasion resistance.
High Temperature Activation with Low Gas Input
During the remainder of the study, starting with run
25, the baking and activating, and also the drying, were carried
out as separate runs or operations. In the activating runs,
the gas input was drastically reduced, while the gas tempera-
ture was maintained at the 1700°F level. All carbon residence
times, in the 42 inch heat portion of the tube, were one half
hour and less.
Tables 10 and 11 show operating and carbon property
data on selected runs on spent VIA that are representative of
this study. The baking run is included with the activating
run to give the overall results of each regeneration. As is
42
-------
TABLE 9 - ACTIVITY DISTRIBUTION IN CARBON PARTICLES
U)
Iodine number
Molasses number
Run 20
Run 22
Run 23
Carbon Interior
•asorb 400 1020
:0 945
12 1001
13 930
Fines
970
728
706
727
Difference
- 50
-217
-295
-203
Interior
216
326
344
379
Fines
200
339
350
346
Difference
-16
+ 13
+ 6
-32
-------
TABLE 10 - OPERATING CONDITIONS FOR HIGH
TEMPERATURE BAKING AND ACTIVATING
Run
No.
29
30
31
32
50
51
58
59
Type
function
baking
activating
baking
activating
baking
activating
baking
activating
Carbon
residence ,
min
30
30
90
23
30
33
13
15
Gas input, std
N2
3.7
3.7
3.7
4.8
3.7
2.8
3.7
4.8
co2
0.5
0.7
0.4
0.7
ft3/hr
H20
1.9
2.5
2.0
2.5
Weight
Temperature ,
TC(1)
1700
1700
850
1700
1700
1750
1800
1700
TC(2)
1700
1700
1320
1700
1700
1750
1750
1700
•F
TC(3)
1600
1600
1600
1600
1600
1650
1650
1600
decrease,
%
11.51
9.4J
11. 0\
11. 2 f
12. 2\
11. 2 J
14. 8\
9.3)
19.8
21.0
22.1
22.7
-------
TABLE 11 - PROPERTIES OF CARBONS BAKED AND ACTIVATED AT HIGH TEMPERATURES
Run
No.
29
30
31
32
50
51
58*
59
Bulk
density,
g/cc
0.550
0.500
0.548
0.490
0.547
0.489
0.536
0.485
Particle
density,
g/cc
0.881
0.807
0.877
0.793
0.881
0.789
0.878
0.796
Real
density,
g/cc
2.14
2.13
2.11
2.14
2.14
2.16
2.15
2.14
Pore Particle
volume volume
cc/cc p.v.1 decrease, %
0.589 0.8~l , g
0.622 l.Oj
0.584 0.6"? 2 -
0.629 1.9J
0.589 l.l\ , q
0.634 0.8J •"•'*
0.582 2.3l - _
0.628 0.2J Z-D
External
burnof f ,
6.6
10.4
1.8
19.7
10.4
7.7
14.1
1.6
Iodine
No.
832
864
790
890
754
900
794
880
Molasses
No.
202
239
176
216
198
262
220
259
*Run 58 was revetted with water and fed wet during baking, hence the
lower bulk density than the other baked carbons.
p.v. - particle volume
-------
evident by comparing the gas input rates of Table 10 with those
of Table 6, the gas input rates on these runs were reduced by
more than one half.
As the residence time or temperature was changed,
the gas input rate was changed also so that the final bulk
density always came in the range 0.485 to 0.500 g/cc. In
the density and pore volume measurements, these regenerated
carbons came close to the virgin carbon measurements but the
iodine numbers were too low by about 150 units and the molas-
ses numbers too high by about 35 units. The carbon loss, as
measured by the particle volume decrease, varied between 1.9
and 2.5%, which are substantial decreases from the 5% and 7%
reported for the hearth furnace operations. On run 51, the
gas temperature/in each section of the rotary tube was in-
creased 50°F. The effect of this increase was to raise the
iodine number by about 10 or 20 units to 900 while causing a
very large increase in molasses number. On the overall basis
the runs that came closest to regenerating the carbon to
original virgin state are 31 and 32 as is evident from com-
parison given below:
Virgin
Runs 31-32 Filtrasorb 400
Bulk density g/cc 0.490 0.486-0.501
Particle density 0.793 0.807-0.830
Real density 2.14 2.15
Pore volume
cc/cc particle vol. 0.629 0.623-0.614
Iodine number 890 995-1020
Molasses number 216 190-216
The large divergence here is the iodine number. On all the
runs made on this spent carbon VIA, the iodine and molasses
numbers could never be matched simultaneously with those of
the virgin carbon. It was finally concluded that during the
regeneration at Pomona the pore structure on the batch of car-
bon sent to MSAR had become altered in such a way that the
iodine number was low.
Regeneration of Once Spent Filtrasorb 400
Since the iodine and molasses numbers of regenerated
VIA could never be brought up simultaneously to those of virgin
Filtrasorb 400, we reasoned that some condition in regeneration
was out of line or that a complete regeneration was not possible.
To resolve the problem, quantities of once-spent Filtrasorb 400,
obtained from Pomona, were regenerated under conditions believed
46
-------
to be optimum, as determined from the VIA regeneration. The
first quantity of once-spent Filtrasorb 400 was obtained at a
time midway through the study and was sufficient in quantity
for two regeneration runs. Another larger amount was received
toward the end of the program, but time permitted only one run.
The operating conditions and regenerated carbon properties for
these three regeneration runs are given in Tables 12 and 13.
To establish a standard, virgin Filtrasorb 400 was put through
the same baking and activating steps.
Physical property comparisons were made on the re-
generated products of these three runs and those of the virgin
Filtrasorb 400, as given below:
Means of runs Virgin
34 and 36 Run 65 Filtrasorb 400
Bulk density, g/cc 0.503 0.486 0.486-0.501
Particle density, g/cc 0.830 0.805 0.807-0.830
Real density, g/cc 2.14 2.15 2.15
Pore volume
cc/cc particle vol. 0.612 0.625 0.623-0.614
Iodine number 1014 960 995-1020
Molasses number 196 225 190-216
Except for the iodine and molasses numbers for run 65, all the
other properties fall within the ranges that had been measured
for the virgin Filtrasorb 400. The results definitely indicate
that the spent carbon can be regenerated to its original state
in one reg eneration as far as it is possible to measure this
within limits of experimental accuracy. The regenerations of
runs 34 and 36 were accomplished with 2.8% and 2.9% particle
volume decrease, respectively/and run 65 with 2.2%.
Another potential carbon loss not measured previously
was that due to formation of fines which, in the adsorption
part of the cycle, would be flushed out during the back-washing,
This loss was expected to be nil on the low gas input acti-
vations. A measure of fines formation was made on the regen-
erated product of run 65 by sieving out the -40 mesh fraction.
By weight this was 0.15%. Hence, by including the fines loss,
the maximum particle volume decrease amounted to about 3.0%.
The good recovery of the original properties on runs
34 and 36 adds support to the earlier conclusions that basic
carbon structure of VIA is not that of the virgin carbon.
Comparison of Hearth Furnace and Rotary Tube Activation
Various comparisons have been worked out on measured,
calculated and estimated operating parameters to show the de-
47
-------
TABLE 12 - OPERATING CONDITIONS FOR BAKING AND ACTIVATING
ONCE SPENT FILTRASORB 400
Feed carbon
on run no .
Spent carbon
feed for run 33
run 33
run 34
Spent carbon
feed for run 35
*» run 35
00 run 36
Virgin Filtra-
sorb 400 for
run 37
run 37
run 38
Spent carbon feed
for run 64
run 64
run 65
Carbon
Type residence,
function min
baking
activating
baking
activating
baking
activating
baking
activating
88
23
78
30
76
30
12
13
Gas input, std ft3/h*
N CO H°
2.0
3.7 0.5 0.9
2.0
3.7 0.5 0.9
3.7
3.7 0.5 0.9
4 . 8
3.6 0.5 1.9
Temperature ,
TC(1)
600
1700
600
1700
600
1700
1700
1700
TC(2)
1100
1700
1100
1700
1100
1700
1700
1700
•F
TC(3)
1600
1600
1600
1600
1600
1600
1600
1600
Weight
decrease,
%
11.7")
6.6J
10.2")
5-5J
l.-f)
*-7J
10. 4~\
7.5J
17.6
15.7
6.1
17.2
-------
TABLE 13 - PROPERTIES OF BAKED AND ACTIVATED ONCE SPENT FILTRASORB 400
VO
Feed carbon
on run no.
Spent carbon
feed run 33
run 33
run 34
Spent carbon
feed run 35
run 35
run 36
Virgin Filtra-
sorb 400, feed 37
run 37
run 38
Spent carbon
feed run 64
run 64
run 65
Bulk
density.
g/cc
0.584
0.532
0.502
0.574
0.527
0.504
0.486
0.487
0.463
0.574
0.522
0.486
Particle
density.
g/cc
0.971
0.871
0.825
0.955
0.866
0.835
0.807
0.799
0.767
0.949
0.863
0.805
Real
density.
g/cc
1.88
2.07
2.13
1.89
2.09
2.15
2.15
2.15
2.16
1.94
2.15
2.15
Pore
volume
cc/cc p.v*
0.483
0.579
0.613
0.494
0.585
0.612
0.623
0.629
0.644
0.510
0.598
0.625
Part icle
volume
decrease , %
1-A •> «
1.5J 2'8
0.9*)
2.0J 2'9
0.3*1
0 . 8 J 1 • 1
—v
o!aj 2-2
External
burnoff
%
12.4
19.5
8.3
36.5
21.0
16.5
13.6
10.7
Iodine
No.
673
955
1017
960
1010
1020
1030
1097
960
Molasses
No.
174
193
192
190
200
210
207
244
225
* p.v. - particle volume
-------
gree of difference between the hearth furnace activating con-
ditions and the optimum rotary tube activating conditions.
Carbon Temperature Profiles
Figures 10 and 11 show the measured temperatures of
the three furnace sections and gas temperatures in the rotary
tube for the activating runs 32 and 34 (23 minutes residence
time) and run 59 (15 minutes residence time). The estimated
carbon temperature is given by the broken line curve in each
figure.
In the first three runs of the program, both carbon
and gas temperature were measured, but the measurement of car-
bon temperatures was discontinued because of frequent breakage
of the thermocouples. At the 800°F temperature regeneration,
run 3, it was found that at the thermocouple positions TC (1) ,
(2), (3) and (4) (see Fig. 4), the temperature was about 40°F
above the carbon temperature and at position TC (5) the gas
temperature was 80°F below the carbon temperature. These
temperature differences were used as a guide to locate the
carbon temperature profiles in Figures 10 and 11.
In Figure 12, the carbon temperature profiles of
Figures 10 and 11 are now plotted as function of carbon resi-
dence time. For comparison, the calculated carbon temperature
profile for the hearth furnace activation obtained from Figure
6, is included. The profiles indicate that on runs 32 and 34
(23 minutes residence time) the carbon is at temperatures
1500°F and above for 16 minutes, while on run 59 (15 minutes
residence time) the carbon is at temperatures 1500°F and above
for 10.5 minutes. The carbon in the hearth furnace is at
temperatures 1500°F and above for 11 minutes. The important
difference is that the carbon in the rotary tube is near the
1650°F level for a much longer time than the carbon in the
hearth furnace. At the 1650°F, the oxidation rate is con-
siderably faster, hence, in regard to temperature, the rotary
tube activation is carried out under more severe conditions.
Another difference is that, in the hearth furnace,
the heat is delivered to the carbon solely by the flue gas-
steam mixture while in the rotary tube the heat is delivered
to the carbon also by direct transfer through the tube wall.
The result is that the rotary tube operates with a smaller
temperature difference between gas and carbon. The tempera-
ture difference is estimated to be around 40 to 50°F in the
activating zone, while in the hearth furnace the temperature
difference varies from 70 to 200°F.
50
-------
2000
0 12 24 36 48 60
Distance From Carbon Influent End of Tube, Inches
FIGURE 10 - GAS AND CARBON TEMPERATURE PROFILES IN ROTARY
TUBE FOR 15 MIN CARBON RESIDENCE TIME, RUN 59
2000
1500
M
Siooo
0)
a
6
0)
E-i
500
T
1745°F
TC 1
-O
1820°F
MTC 3
189
TC 3
12 24 36 48 60
Distance From Carbon Influent End of Tube, Inches
FIGURE 11 - GAS AND CARBON TEMPERATURE PROFILES IN ROTARY
TUBE FOR 23 MIN CARBON RESIDENCE TIME, RUNS 32 AND 34
51
-------
1800
oi
K>
Hearth Furnace
1200 I
12
14
2 4 6 8 10
Activating Time, Min
FIGURE 12 - ACTIVATING TIME MULTIPLE HEARTH FURNACE AND ROTARY TUBE
RUN 32, 34 AND 59
16
-------
Gas Input and Utilization Rates
Table 14 gives the nitrogen and activating gas input
rates for five of the rotary tube activating runs and, for
comparison, those of the hearth furnace. The activating gas
input rate for the hearth furnace is 4.3 to 9.5 times greater
than it is for the rotary tube activations. On the other hand,
the rate of activating gas utilization for the rotary tube is
2.7 to 4.0 times greater than for the hearth furnace.
TABLE 14 - GAS INPUT AND ACTIVATING GAS UTILIZATION
RATES DURING ACTIVATION
N7 std ft3/lb
C02 std ftVlb
H20 std ft3/lb
02 std ft3/lb
Activating gas
totals, ftVlb
Gas utilization
% of input
Carbon weight loss
% (based on initial
weight)
Run
32
9.6
1.4
5.0
0.0
6.4
59
11.2
Run
59
9.6
1.4
5.0
0.0
6.4
49
9.3
Run
34
7.4
1.0
1.9
0.0
2.9
78
6.6
Run
36
7.4
1.0
1.9
0.0
2.9
66
5.5
Run
65
7.2
1.0
3.8
0.0
4.8
51
7.5
Hearth
furnace
27.3
3.3
23.5
0.7
27.5
18
13.4
As brought out in the previous section, the rotary tube acti-
vating occurs at a higher temperature but the amount of oxi-
dation is moderated by controlling of activating gas input
rate. At the low input rates, the gas utilization rate is
high, 47% and greater, which slows down the reaction rate
since both the C02 and H2O reactions with carbon are revers-
ible, i.e.,
C0
H2°
2CO
+ C
In the hearth furnace the amount of oxidation in the presence
of the large excess of activating gases is controlled by main-
taining a lower carbon temperature at which the oxidation rate
is slower. Only 18% of the activating gas is utilized. An
undesirable characteristic of this method of activation rate
control is that the exterior surfaces of the carbon particles
are exposed to a high activating gas concentration for the
53
-------
total activating time. By mass action, external burn-off of
the carbon particles is accelerated while internal burn-off is
relatively slower because of finite diffusion time required for
the gases to enter the pores. In the rotary tube, because of
the higher temperature, the diffusion rate into the pores is
faster thereby accelerating internal burn-off. Because of the
low activating gas input rate, the gas composition changes
rapidly, thereby decreasing the activating gas concentration
in contact with the external surface. External burn-off is
thereby minimized.
The carbon weight loss on the once-spent carbons,
runs 34, 36 and 65, are in the 5.5 to 7.5% range, compared to
11.2% for run 32 and 13.4% for the hearth furnace activation
weight loss. This is due to the smaller initial amount of ad-
sorbate on the once-spent carbons. For the once-spent carbon,
the initial density was about 0.587 g/cc and the overall weight
decrease on regeneration 17.6%, while for the spent VIA acti-
vated in run 32 and also the type activated in the hearth fur-
nace, the initial density was 0.615 g/cc and the total weight
loss on regeneration, 21.0%. This points out the necessity of
knowing the adsorbate content of the spent carbon feed to per-
mit proper control of the activating gas input rate.
Activating Gas Composition
Table 15 gives compositions of activating gases used
in the Pomona hearth furnace and many of the rotary tube runs.
TABLE 15 - ACTIVATING GAS COMPOSITIONS FOR HEARTH
ROTARY TUBE REGENERATIONS
Composition, %
Flue gas Runs Runs Runs
Flue gas plus steam 3 to 24 32 & 65 34 & 36
N2 71.5 49.8 59.2 60.0 71.8
C02 8.7 6.0 7.6 8.8 9.7
H20 17.7 42.9 33.1 31.2 18.5
02 2.1 1.3 0.0 0.0 0.0
Column two gives the flue gas composition of California natural
gas burned with 12% excess air of 80% relative humidity. With
the 12% excess air, the flue gas contains 2.1% oxygen. Column
three gives the mean composition of the flue gas-steam mixture
as used in the Pomona hearth furnace. The amount of steam
added is based on 0.8 Ibs per pound of regenerated carbon.
54
-------
The activating gases used in the rotary tube are essentially
synthetic flue gas mixtures (minus the free oxygen) plus steam
added, but not in the amount added at Pomona. For runs 34 and
36, the activating gas is essentially a synthetic flue gas when
the natural gas is burned with stoichiometric amounts of air.
Two activations were run in which the activating gas
compositions were drastically changed. In run 47, no steam was
used and in run 49, no C02. The gas compositions used were:
Run 47 Run 49
N- 60.8 65.8
C02 39.2 0.0
H20 0.0 34.2
These runs were made with spent VIA carbon to gain general
knowledge on the activation with these two gases and also as
an attempt to increase the iodine number of this carbon. The
operating conditions of these runs are given in Table 16 and
physical properties in Table 17. The corresponding baking
runs are included to give the overall results.
Although both activations brought the bulk densities
within the virgin carbon range, neither one brought up the
iodine number but both brought up the molasses numbers to an
excessively high value. On these two runs, the CC>2 activation
fared the worst. These results are comparable to the previous
runs (Table 11) on VIA, when a synthetic flue gas-steam mixture
was used as activating gas. When the water content in the flue
gas-steam mixture is at the 30% level, the activating gas, to
all practical purposes, acts as a N2~steam mixture. The steam
is a faster acting oxidizing agent and, being in excess,carries
the major part of the oxidation load. Vent gas analysis of
the N2-steam activation showed that C02 was generated along
with the CO and H2 hence, CO2 oxidation can be expected to be
suppressed when it is present as one of the activating in-
fluent gases.
Baking Temperature Profile
During the regeneration series consisting of runs 14
to 24, it became apparent that the adsorbate was carbonized
while the carbon temperature was rising from room temperature
to about 1500°F, the temperature at which the activating rates
become rapid. The carbonization produced a volatile component
which was evolved and a free carbon which remained in the pores,
In the following activating step the primary objective was to
oxidize the carbon residue of the adsorbate with minimum attack
on the basic carbon structure.
55
-------
TABLE 16 - OPERATING CONDITIONS FOR BAKING AND ACTIVATING VIA
01
Carbon
Run
No.
46
47
48
49
Run
No.
46
47
48
49
Type
residence,
function
baking
activating
baking
activating
Bulk
density,
g/cc
0.547
0.498
0.547
0.487
TABLE 17
Particle
density,
g/cc
0.873
0.796
0.877
0.776
mm
30
35
33
44
Gas input, std
N2 C02
4.8
4.8 3.1
4.8
4 . 8
ft3/hr
H2O
_»•»
2.5
Temperature , *F
TC(1) TC(2)
1700 1700
1700 1700
1700 1700
TC(3)
1600
1600
1600
Weight
decrease ,
*
13. 4\
9.3J 21'4
il'.lj 23'6
- PROPERTIES OF BAKED AND ACTIVATED VIA
Real
density
g/cc
2.12
2.13
2.14
2.17
Pore
, volume , ^
cc/cc p.v.
0.588
0.626
0.590
0.642
Particle
volume ,
decrease ,
2.1
0.5
2.4
0.3
Act . gas
utilization,
r% %
_—
49
_——
72
Iodine
No.
810
834
802
877
Molasses
No.
219
262
215
288
* p.v. - particle volume
-------
At this stage of the study, it was thought that the
manner in which the baking step, i.e., carbonization, was
carried out would cause the carbon residue to deposit in dif-
ferent forms. It is apparent that if the carbon residue were
in a finely divided form, it would be more easily oxidised
while if it formed aggregates or assumed craohiLic plates, it
would be more difficult to oxidize. The ease or difficulty of
the residue oxidation would inversely determine the severity
of the required activating conditions and thereby, have an
effect on carbon losses.
The study involved heating dried spent VIA at various
rates to temperatures above 1500°F and also holding carbon at
the maximum temperature for varied lengths of time. The baked
materials were then tested and activated under relatively
standard conditions to determine ease of activation.
Figures 13 and 14 show temperature profiles for the
baking experiments. The profile designated as feed for runs
40 and 41 were baked in an oven in an inert atmosphere. The
dried and sieved (14x40 mesh) spent VIA was placed into the
oven in shallow trays and heated in each case to 1650°F,
according to the time schedule shown in the profile. A nitro-
gen flow of 5 std ftVhr was maintained through the oven during
the entire heating and cooling period. The other temperature
profiles corresponding to the run numbers were baked in the
rotary tube with nitrogen flow, usually 3.7 std ft3/hr, passed
through the tube. Operating and other data are given on some
of these runs in Tables 10, 11, 16 and 17.
The results of this study are summarized in Table 18,
which gives the heating times to raise the carbon temperature
from about 200°F to 1500°F and the lengths of time the carbons
were held at 1500°F or above. Although the heating time was
varied from 210 to 4 minutes, no trends were indicated in the
properties of the baked carbon nor in the iodine and molasses
numbers of the corresponding activated products. The maximum
deviation from the mean for the densities and pore volumes was
of the order of one percent. For the iodine and molasses
numbers, the maximum deviation is considerably larger. The
means and deviations are as follows:
Baked .
Bulk density, g/cc0.546 - 0.004
Particle density, g/cc 0.876+* 0.007
Real density, g/cc 2.13 - 0.02
Pore volume, cc/cc particle volume 0.590 x 0.005
Iodine number 798 + 44
Molasses number I98 ~ 22
57
-------
2000 r
Feed for
Runs 40
& 41
0 2 4 6 8 10
FIGURE 13 - TEMPERATURE PROFILES DURING BAKING OF SPENT VIA
2000
1500
fu
e
(0
a
0)
1000
500
uns 27,29
44,46,48
& 50
I
I
I
I
10
40
50
20 30
Minutes
FIGURE 14 - TEMPERATURE PROFILES DURING BAKING OF SPENT VIA
58
-------
TABLE 18 - PHYSICAL PROPERTIES OF BAKED SPENT VIA
m
vo
Run
No.
Feed
40
Feed
41
31
42
27
29
44
46
48
50
53
54
62
Heat
time
min
210
210
50
50
11
11
11
11
11
11
6
6
4
Time at
, 1500'F,
min
90
90
18
18
20
20
20
20
20
20
10
10
7
means
Bulk
density,
g/cc
0.544
0.548
0.548
0.548
0.546
0.550
0.546
0.547
0.547
0.547
0.546
0.545
0.542
0.546
Particle
density,
g/cc
0.861
0.877
0.877
0.876
0.873
0.881
0.878
0.873
0.877
0.881
0.883
0.879
0.872
0.876
Real
density ,
g/cc
2.15
2.13
2.11
2.16
2.11
2.14
2.14
2.12
2.14
2.14
2.14
2.12
2.13
2.13
Pore
volume A
cc/cc p.v.
0.599
0.588
0.585
0.595
0.586
0.589
0.589
0.592
0.589
0.589
0.588
0.586
0.592
0.590
After Activation
No.
780
820
790
777
842
832
776
810
802
754
774
820
794
798
Molasses
NO.
184
202
176
176
205
202
171
219
215
198
221
213
209
198
.
897
890
890
892
878
864
891
834<1>
877
900
879
886
Mol.
No.
235
259
216
224
235
239
242
288
262
290
249
(1) CO, activated, not included in mean values
* p.v. - particle volume
-------
Activated
Iodine number 886 - 14
Molasses number 249 ± 33
The mean real density of 2.13 g/cc indicates that the
adsorbate residue in the pores is essentially free of carbon
since the real density of the virgin carbon is 2.15 g/cc. The
mean pore volume of 0.590 cc/cc particle volume is 0.028 cc/cc
particle volume smaller than the mean 0.618 cc/cc particle
volume for the virgin carbon (see Table 4). The 0.028 cc/cc
particle volume is an approximate measure of the carbon resi-
due volume. The mean molasses number (198) for the baked car-
bon is already very close to the mean virgin carbon molasses
number and is a good indication of enlarged pores in spent VIA.
The 6 and 4 minute heating-up times for runs 53 and
62, respectively, are comparable to the 4 minute baking period
calculated for the hearth furnace (see section on multiple
hearth furnace regeneration). One conclusion that can be
drawn from these experiments is that the manner in which the
carbon is baked, whether in the rotary tube or hearth furnace,
has no significant effect on the baked or activated product.
Particle Volume Decrease During Baking
During baking, a particle volume decrease occurs
although the neutral atmosphere precludes possible oxidation of
carbon particle outer layers. Other causes for the particle
volume decrease have been investigated, i.e., (1) possible
shrinkage of the carbon particle due to adsorbate removal and
(2) removal of sewage residue adhering to the exterior surface
of the particles. In either case, the particle volume de-
crease as measured is not a true measure of carbon loss.
The list of representative runs given in Table 19
shows a mean particle volume decrease of 1.3% for the spent
VIA while the particle volume decrease for virgin Filtrasorb
400 is 0.3%. By difference, the particle volume decrease due
to the presence of adsorbate is 1.0%. This amount, subtracted
from the previously reported overall particle volume decreases
of 1.8% to 3.0%,would give the true carbon losses as 0.8% to
& • V <& •
Efforts to determine whether the decrease is shrink-
age or residue removal were conclusive but did demonstrate the
fact that the volume decrease is not due to mechanical attri-
tion in the rotary tube or formation of fines by rapid ex-
pulsion of adsorbate from the carbon. The fines thus formed
could be carried out of the regenerator by the sweep gas.
60
-------
TABLE 19 - PARTICLE VOLUME DECREASE DURING
BAKING IN ROTARY TUBE
Run Particle volume Run Particle volume
No. decrease, % No. decrease, %
25 1.4 52 1.1
27 1.9 53 1.2
29 0.8 54 0.8
31 0.6 mean 1.3
42 0.7
44 1.6 virgin Filtra-
46 2.1 sorb 400 0.3
48 2.4
50 1.1
Two experiments related to particle volume decrease
have already been reported in part in the sections "Regen-
eration at Low-Temperature and High-Gas Input" and "Baking
Study". The first experiment involved the two 15g samples of
spent VIA carbons which were heated in a stationary oven to
850°F for 1 1/2 hrs in an atmosphere of nitrogen. No particle
volume decrease was observed. The weight decrease was 8.2%
and the real density was 2.01 g/cc. The low real density in-
dicates that the adsorbate was not completely carbonized,
hence, the lack of particle volume decrease may be due to
sufficient adsorbate remaining in the pores to prevent volume
decrease. In this case, any sewage residue adhering to the
outside surface of the carbon particles would be more free to
volatilize than the constricted adsorbate in the pores. Since
no particle volume decrease was observed, the implication is
that the particle volume decrease is not due to loss of ex-
terior residue.
The other experiment involved the feed for runs 40
and 41. In this case, larger quantities of spent VIA were
baked in the oven to a maximum temperature of 1650°F in an at-
mosphere of nitrogen. The results of these baking experiments
were:
Real Particle Weight
density, volume decrease,
g/cc decrease, % %
feed to run 40 2.15 0.5 11.5
feed to run 41 2.13 1.1 H-6
61
-------
The 11.5 and 11.6% weight decreases in these two runs
indicate that considerably more adsorbate was carbonized than
during the prior 850°F baking experiment. The real densities
of 2.15 and 2.13 g/cc also indicate that the residue from the
adsorbate is essentially free carbon. The 0.5 and 1.1% par-
ticle volume decreases are comparable to those of Table 19,
all of which were baked in the rotary tube. The experiment
indicates that the particle volume decrease is associated with
temperature above 850°F, but does not differentiate between
shrinkage and residue removal.
In these experiments there was no possibility of
loss by mechanical attrition since the carbon was held sta-
tionary in trays. The slow heat-up rate (210 minutes) is not
conducive to rapid expulsion of adsorbate volatiles, thereby
creating fines and aiding in their expulsion from the trays.
Loss of fines by carry-over with the nitrogen sweep gas would
be nil because of the low (5 ft3/hr) flow rate relative to
the large oven volume (4.3 ft3). The decrease of particle
volume due to these causes had been considered a distinct
possibility with the rotary tube. Since the particle volume
decreases by the two methods are comparable, the carbon losses
due to these causes for the rotary tube are now considered
quite small.
Drying Studies
In the multiple hearth furnace, the wet carbon is
dried on the first three hearths in 15 minutes by contact with
gases at 600 to 1300°P. The final temperature of the carbon
at the end of the drying period is about 212°F.
Most of the drying for the laboratory rotary tube
runs was done in a small laboratory hot air convection oven
set at 150°C. About 3 liters of carbon, as received from
Pomona, were equally divided between four 1-liter beakers and
kept in the oven for about 48 hours. After 24 hours, the car-
bon in each beaker was turned over so that the carbon that was
at the bottom during the first 24 hours was at the top for the
second 24 hours. The spent carbon as received from Pomona was
sometimes wet and sometimes partially dried. No check was
made as to the efficiency of this drying procedure, but it
appeared satisfactory because the bulk densities of the dried
carbons were very consistent.
Two runs were made wherein the drying and baking
were combined to greatly accelerate the drying. In each case,
the spent carbon was first dried in the preceding manner and
the necessary bulk volume and weight measurements made. A
quantity of water was added to each charge which would fill
62
-------
the pores yet leave the carbon in a free flowing state. The
water content was 31% on the wet-weight basis. For comparison,
the wet carbon fed to the Pomona furnace is about 40% water on
the wet-weight basis.
The operating conditions and the carbon properties
of these two runs and also the corresponding activating runs
are given in Tables 20 and 21. The combined drying and baking
time for run 55 was 17 minutes and for run 58, 13 minutes.
This is shorter than the combined time for the hearth furnace
which is 19 minutes, hence, the drying times in the rotary
tube can be assumed to be considerably shorter than 15 minutes.
In run 55, the gas temperatures in the three sections I, II
and III were 1700, 1700 and 1600°F respectively. At these
baking temperatures, the bulk density of the baked product was
0.538 g/cc and the particle density was 0.863 g/cc. For the
oven dried spent carbons baked at the same temperature, these
densities were normally 0.546 and 0.876 g/cc (from Table 18).
Rapid drying then caused a greater evolution of absorbate
volatiles. On run 58, an attempt was made to further increase
the volatilization of adsorbate by raising the gas temperatures
in the three sections to 1800, 1750 and 1650°F and shortening
the residence time to 13 minutes. Section I of the rotary
tube, where the gas temperature is 1800°F, is the carbon in-
fluent end, hence the wet carbon in 5 minutes comes in contact
with gas at 1800°F and a tube wall that may be over 1800°F.
The increase in severity of baking, however, produced no in-
creased adsorbate volatilization. In regard to iodine and
molasses numbers for both the baked and activated products,
there was no improvement. These are essentially at the levels
shown in Table 18 for the oven-dried spent VIA.
During wetting and drying of the spent carbon feed,
the carbon lost weight on each wetting and drying cycle. This
treatment was repeated in the hopes that it would lead to a
non-destructive way of regenerating spent VIA. It is important
to know the nature of the basic carbon structure of VIA since
a great deal of work had been done with it.
In this study, an initial 2,450 cc bulk volume of
oven-dried carbon was alternately wetted with 670 cc water and
dried for 12 cycles. On the 12th cycle, the density appeared
to level off at 0.570 g/cc, far from the 0.490 final density
so the study was discontinued at this point. The results of
the series of experiments are plotted in Figure 15. On each
cycle, the wet period or equilibration time was two days and
the drying time one to two days in the laboratory oven set
at 150°C.
63
-------
TABLE 20 - OPERATING CONDITIONS FOR RAPID DRYING OF WET SPENT VIA
Run
No.
55
57
58
59
Run
No.
55
57
58
59
Type
function
dry fc bake
activating
dry & bake
activating
TABLE 21
Bulk
density,
g/cc
0.538
0.488
0.536
0.485
Carbon
residence
min
17
30
13
15
, Gas
N2
3.
4.
8.
4.
- PROPERTIES OF
Particle
density,
g/cc
0.863
0.776
0.878
0.796
Real
density
g/cc
2.13
2.15
2.15
2.14
input, std ftj/hr
C02 H2O
7
2 0.6 1.4
0
8 0.7 2.5
Weight
Temperature decrease ,
I II
1700 1700
1700 1700
1800 1750
1700 1700
BAKED AND ACTIVATED VIA AFTER
Pore
, volume ,
cc/cc p.v.*
0.594
0.638
0.591
0.628
Particle
volume
decrease , %
0.6
___
2-3l 2 5
0.2J 2'5
III %
1600 11.27\ ,. .
1600 9.3 J
-------
-------
On the second cycle, the carbon became covered with
a white powder and the bulk volume increased to 2.740 cc, i.e.,
by 11.8%. The powder, leached from the interior of the par-
ticles, adhered to their exterior and caused an apparent par-
ticle volume increase. The powder persisted to varying degrees
up to the seventh cycle. On the eighth and subsequent cycles
the carbon was leached on a Buchner funnel with 3.5 liters of
water before drying. The leaching on the eighth cycle brought
the bulk volume down to 2,433 cc, which was below the initial
volume by 0.7%. On each of the subsequent cycles, the bulk
volume decreased by a diminishing amount until, on the 12th
cycle, the volume decrease was 2 cc or 0.04%. The final bulk
volume decrease from the original was 28 cc or 1.1%.
Although this experimental study did not lead to a
non-destructive means for regeneration of VIA, it produced the
first directly measured evidence of change in carbon volume
caused by sewage residue adhering to the exterior surface of
the carbon particles.
Effluent Gas Analyses
Analysis of vent gas offers a means for monitoring
the activating step of the overall regeneration process. In
the hearth furnace, the gas sample for activation control
should be withdrawn from the fourth hearth at a point where
none of the adsorbate volatiles are present but where the
activation is essentially completed.
Vent gases have been analyzed on most of the rotary
tube activation runs. Some results have been encouraging al-
though problems still exist. As examples of the type of re-
sults obtained, calculations are presented on three typical
activation runs - runs 45, 47 and 49.
On run 45, the effluent gases are not in correct
proportions to represent any possible equilibrium condition
with the carbon. This becomes apparent from the following
calculations.
TABLE 22 - VENT GAS ANALYSIS OF ACTIVATION RUN 45
Input, Output (dry),
Gas gram moles/hr gram moles/hr
N2 6.07 6.07
C02 0.88 0.61
H2 2.01
H2O 3.16
CO 2.00
66
-------
On this run, the carbon weight decrease is 30 g/hr or 2.5 gram
moles/hr. This weight decrease had to go by way of the follow-
ing two oxidation reactions:
H2O + C = CO + H2 (Eq. 1)
CO2 + C = 2CO (Eq. 2)
By inspection of the data it is apparent that there is insuf-
ficient CO in the vent gas. If only reaction (1) occurred,
the amount of CO in the vent gas should be 2.5 gram moles/hr.
If only reaction (2) occurred, the amount of CO should be 5.0
gram moles/hr. Since both reactions are known to occur, the
amount of CO should be between 2.5 and 5.0 gram moles/hr in-
stead of the 2.00 gram moles/hr as measured.
On run 47, where only CO2 and N2 were fed to the re-
generator, the input gas composition and dry output gas analysis
are as follows:
TABLE 23 - VENT GAS ANALYSES OF ACTIVATION RUN 47
Input, Output (dry),
Gas gram moles/hr gram moles/hr
No 5.95
C02 3.92
CO
5.95
1.45
3.89
On this run, the carbon weight decrease is 23 grams/hr or 1.92
gram moles/hr. In this case, the gram moles of carbon oxidized
per hour agrees with the amount of CO formed, i.e., 2 x 1.92 -
3.84 grams moles/hr of CO. However, the amount of CO2 consumed
indicates that 4.94 gram moles/hr of CO should have been formed
instead of the 3.89 gram moles/hr as measured.
On run 49, where only H2O and N, were fed to the re-
generator, the input gas composition and dry output gas analysis
are as shown in Table 24. The weight decrease on this run is
25 6 grams/hr or 2.13 gram moles/hr. In this case, the water
reacts with carbon to form CO and C02 according to equation (1)
and to equation (3), below:
2HoO + C = CO-, + 2H-, (Eq. 3)
According to equation (1), the gram moles of carbon oxidized
and CO generated are the same, i.e., 1.68 gram moles. By
67
-------
equation (3) another 0.42 gram moles is oxidized, giving a total
of 2.10 gram moles/hr which is in good agreement with the ob-
served 2.13 gram moles/hr carbon burnoff. In this case, the
amount of H2 generated is short of the amount that should have
been generated according to the amounts of CO and CO, generated.
The amount of hydrogen should have been 2.52 gram moles/hr in-
stead of the 1.89 gram moles/hr as measured.
TABLE 24 - VENT GAS ANALYSES OF ACTIVATION RUN 49
Input, Output (dry),
Gas gram moles/hr gram moles/hr
N2 5.95 5.95
H2O 3.02
C02 0.42
CO 1.68
H2 1.89
Probable causes of some of the discrepancies ob-
served above are (1) reaction of the vent gases with the copper
vent lines and (2) inaccuracies in C02 and N2 influent measure-
ments .
Quenching Experiments
At the Pomona plant, it has been observed that the
iodine number of quenched regenerated carbons is lower than
that of regenerated carbons which have been cooled more grad-
ually in an inert gas atmosphere. According to Figure 2 of the
May, 1967 Pomona Monthly Report, this difference varies from
50 to 100 iodine number units.
Quenching experiments were performed with virgin
Filtrasorb 400 to check out the loss in iodine number and re-
late it to possible changes in pore structure. Seven experi-
ments were conducted, the first four with a 16x20 mesh sieve
fraction and the last three with a 14x40 mesh sieve fraction.
In each experiment, the carbon was heated to about 1750°F in
a retort in a nitrogen atmosphere and then let fall into water
at room temperature. In one experiment the hot carbon was
dropped into a container flushed with nitrogen to get a com-
parison of the effects produced by slow cooling.
The results of the experiments are given in Table 25.
They show a general decrease of 25 to 69 iodine units for the
water quenched samples. An exception is experiment II, where
no decrease was observed. Nitrogen quenching (experiment IV)
68
-------
caused a 23 unit increase in iodine number. The molasses
numbers show a maximum deviation of t 15 about the mean 200
value but indicate no trend or correlation with the iodine
number. The first four samples show some pore volume fluctu-
ations which, however, do not correlate with the iodine numbers
or molasses numbers. In the last three samples, the pore
volume shows little or no change due to the quenching. At
present, there is no explanation for the foregoing results.
They are offered here solely as recorded observations.
TABLE 25 - RESULTS OF QUENCHING EXPERIMENTS WITH FILTRASORB 400
Sieve
fraction
16x20
16x20
16x20
16x20
14x40
14x40
14x40
Weight,
g
30.00
29.84
- 0.16
30.00
29.90
- 0.10
30.00
29.50
- 0.50
Pore volume
cc/g
0.742
0.716
-0.026
0.742
0.757
+0.015
0.759
0.772
+0.013
0.760
0.749
-0.011
0.747
0.747
0.000
0.740
0.741
+0.001
0.742
0.749
+0.007
Iodine
No.
Molasses
No.
185
220
+ 35
185
210
+ 25
205
186
- 19
205
210
214
191
- 23
194
205
+ 11
192
- 69
(1) Letter b indicates initial sample before
quenching, (a) after quenching, and (c)
change in properties.
(2) Sample IV cooled in nitrogen.
69
-------
HEARTH FURNACE THERMODYNAMICS
Material and Heat Balance
On the basis of the operating data obtained from
Pomona on the regeneration of wet spent carbon and the infor-
mation obtained from the laboratory rotary tube runs, the
material and heat balance and the carbon temperature profile
have been calculated for the regeneration of:
1. Wet spent carbon
2. Dried spent carbon
3. Baked spent carbon
The thermodynamic calculations were made for 2 and 3 above to
determine the effect decreased heating loads would have on the
required heating gas and steam input rates. In the wet spent
carbon regeneration/ the gas-steam input rate is excessive in
terms of the activation step such that the activation rate
must be kept under control by keeping carbon temperature low.
The following are the measured operating conditions
and assumptions used in the calculations:
1.
2.
3.
4.
5.
Carbon discharge rate
Natural gas input rate
Air input rate:blower cap.
Steam input rate 0.81 Ibs/
Ib carbon
Steam generator operating
pressure,
estimated steam tempera-
ture at hearths
89.4 Ib/hr
270 ft3/hr @ 60°F
10,800 ft3/hr @ 60°F
71.5 Ib/hr
10 psi
220°F
6.
Natural gas composition:
Component mole fraction
CH4
Cfl6
CO2
°2
N2
He
Sulfur
0.8721
0.0806
0.0178
0.0016
0.0048
0.0000
0.0229
0.0002
0.18 grains per 100 SCF
70
-------
7. Measured gas temperatures on hearths
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
Hearth
1
2
3
4
5
6
600
800
1000
1680
1710
1720
Shell temperature 300 to 400°F.
Carbon effluent temperature, (assumed)
Effluent gas temperature
Carbon influent temperature
Oxygen in flue gas, 12% excess air
Burners 4' 5 & 6th
Steam distribution to hearths 20% to 4,
and 40% to
Flue gas distribution to hearths 1/3
Weight water evaporated during drying
Weight decrease during baking
Weight decrease on oxidation of
adsorbate residue
Carbon loss due to activation
Composition of adsorbate volatiles
(from rotary tube vent gas analysis)
77
13
7
6
1650°F
600°F
70°F
2%
hearths
40% to 5
6
to each
.5 Ib/hr
.2 Ib/hr
.6 Ib/hr
.3 Ib/hr
Component
CO
C02
Ho
Composition,%
35.6
16.4
48.0
21.
Estimated heat of adsorbate decomposition
at 1400°F - +900 Btu/lb
Bv means of well known principles of thermodynamics
far^tloA'anf ^7% fro. tE,-tea, inPf About
23 SS^SS^.'S"
vation and heating of carbon.
71
-------
TABLE 26 - MATERIAL AND HEAT BALANCE FOR REGENERATION
OF WET SPENT CARBON
Material
Spent carbon, dry
Water in carbon
Activating steam
AH,
Natural gas
AH0
Air
Regenerated carbon
Water from carbon
AHV
Volatile adsorbate
AH..
nd
CO
co2
Gas from activation
AH,
A
No
*
CO,
H20
AH,
CO
H2
Ib/hr
M*l&a»^»
116.5
77.5
71.5
12.9
237.8
516.2
89.4
77.5
Input
Ib moles
/hr
3.97
0.714
8.26
Output
4.30
Temp. , 5€u
•F /hr
70 300
70 800
220 5,200~)
75,800J
60 000
290,800
60 000
372,900
1650 65,300
600 18,500*)
82, 100 J
Heat,
0.1
0.2
21.7
77.9
9TT9"
17.5
27.0
18,700""l
7.4
5.1
0.7
180.5
42.5
82.4
27.4
2.0
514. *
0.26
0.12
0.35
6.45
0.96
4.58
0.98
0.98
Heat loss through shell and rotating shaft
(by difference)
600 1,000 I
600 700 ( 5.5
600 300J
26,600
600 23,900
600 5,300
600 19,700
87,400
600 3,700
600 3,500j
356,700
16,200
372,900
45.6
4.3
9979
72
-------
If the wet carbon is dried outside of the hearth
furnace and the dried spent carbon is then baked and activated
in the hearth furnace, the required heat input is 60.5% of the
heat input used in the wet carbon regeneration. The calculated
material and heat balance is given in Table 27. If a further
step is taken and the baking is also done outside the hearth
furnace, the required heat input for the hearth furnace acti-
vation is 52.4% of that required for the wet spent carbon. The
calculated material and heat balance for the activation is
given in Table 28.
The material and heat balance for each type of re-
generation (activation) indicates the self consistency of the
data and manner in which the calculations were carried out.
The important information obtained from these calculations is
the amount of steam-gas-air mixture input required for each
type of regeneration. When converted to flue gas-steam input
per pound of regenerated product, comparisons can be made with
rotary tube activations.
Activating Gas Utilization
The calculated gas input per pound of regenerated
product is given in Table 29 for each of the three types of
hearth furnace regenerations.
TABLE 29 - ACTIVATING GAS INPUT PER POUND OF
REGENERATED PRODUCT, HEARTH FURNACE
Gas input, ft3/lb
Component wet. dried baked
N2 27.3 16.5 14.3
CO? 3.3 2.0 1.7
H70 23.7 14.3 8.6
02 0.75 0.45 0.36
The gas input rates for the baked carbon activation are
approaching those of run 32, which are:
No - 9.6 ft3/lb
CO, - 1.4 "
H20 - 5.0
In other operating conditions the hearth furnace activation of
the baked carbon and run 32 are quite close. These are the
73
-------
TABLE 27 - MATERIAL AND HEAT BALANCE, DRIED SPENT CARBON
Input
Material
Fl 01 wC ^ i O A
Spent carbon dry
Activating steam
AHy
Natural gas
AH
c
Air for combustion
Regenerated carbon
Volatile adsorbate
AHd
CO
£°2
Gas from activation
AHa
C02
H2O
/\ U
AHy
CO
U
H2
Ib/hr
116.5
43.3
7.8
144.0
89.4
7.4
5.1
0.7
108.2
25.9
41.2
29.4
2.1
309.4
Ib moles
/hr
2.38
0.432
Output
0.26
0.12
0.35
3.90
0.59
2.29
1.05
1.05
Heat loss through shell and cooling air
for rotating shaft
Temp. , Btu
•F /hr
70 300
220 3,100^
45,900J
60 000
176,000
000
225,300
1650 65,300
18,700"")
600 1,000 (
600 700 |
600 300 J
43,500*^
600 14,400
600 3,200
600 9,800
43,700
600 4,100
600 3,700^
16,200
224,600
Heat,
*
0.1
21.7
78.1
9TT9
29.1
9.2
)> 54.5
7.2
iffoTo
74
-------
TABLE 28 - MATERIAL AND HEAT BALANCE, BAKED SPENT CARBON
Input
Material
Baked
Activating steam
Natural gas
AHc
Air for combustion
Regenerated carbon
Gas from activation
AH,
a
COj
H2
AHv
CO V
H2
Heat loss through shell
Ib/hr
103.3
37.4
6.8
124.6
272.1
89.4
95.0
21.5
32.9
30.1
2.2
271.1
Ib moles Temp., Btu
/hr "F /hr
70 300
2.08 220 2,700")
39,600J
0.374 60 000
152,200
000
194,800
Output
1650 65,300
47,200^
3.39 600 12,500
0.49 600 2,700
1.83 600 7,900
34, .900
1.08 600 4,200
1.08 600 3,800_^
and rotating shaft 16,200
194,700
Heat,
%
0.2
21.7
78.1
100.0
33.5
58.1
8.3
99.9
75
-------
activating time and amount of carbon oxidized during the acti-
vation. The numbers are given below:
Hearth, baked carbon Run 32
Activating time, rain 15 16
Ib carbon oxidized/lb
of product 0.155 0.127
In this case, the activating time is the time the carbon is at
1500°F or above. The higher value of the carbon oxidized in
the hearth furnace reflects the 7% carbon loss while the par-
ticle volume decrease for run 32 is 1.9%. For run 32, the
overall particle volume decrease is 2.5%.
The activating gas utilization for the three types
of hearth furnace activations are:
% utilization
wet spent carbon 17.8
dried spent carbon 29.5
baked spent carbon 43.0
For run 32, the activating gas utilization is 59.2%.
The closeness of approach of the activating con-
ditions of the hearth furnace activation and run 32 indicates
that activation of baked carbon in the hearth furnace may be
feasible under conditions of rotary tube activation. Whether
this will result in a decrease in carbon loss can be deter-
mined to some degree by further activation studies in the
rotary tube regenerator under conditions closer to those of
the hearth furnace. Whether a successful activation of baked
carbon in the hearth furnace is more economical than wet car-
bon regeneration is a subject for economic study.
Carbon Temperature Profiles
The flame temperature, influent gas temperature and
carbon temperature profiles have already been presented in
Figure 6 for the regeneration of wet spent carbon. In con-
structing the carbon temperature curve, it was assumed that a
sharp break occurs between each step. At the end of the drying
step the carbon temperature is 212°F and then rises rapidly in
hearth 4 to 1500°F during the baking step. From the labora-
tory rotary activator work, it was found that activation at
temperatures below 1500°F was slow even at high activating
gas throughput rates, hence; at the much shorter residence time
76
-------
in the multiple furnace, very little activation can be ex-
pected at temperatures below 1500°F. Above 1500°F, the acti-
vating rate becomes more rapid as the temperature rises to
1650°F, the assumed effluent temperature for the regenerated
carbon.
Figure 16 shows the possible temperature profiles
for the carbon and gases during regeneration of dried spent
carbon. Among other assumptions, it is assumed that the vent
gas is at 600°F, and the regenerated carbon leaves the 6th
hearth at 1650°F. The two circles represent the calculated
points. For regeneration of dried spent carbon the heat input
is 60.5% of the heat input for the wet spent carbon. The gas
flow through the hearths is much slower. It can be expected
to give up a higher percent of its heat content to the carbon.
This will bring the gas and carbon temperatures closer.
In Figure 17, where activation is the only step,
the same pattern of heat exchange occurs as in Figure 16. The
gas and carbon temperatures are even closer. In this case, the
heat input is 52.4% of that required for the wet spent carbon.
For the latter two regenerations, the activation
step has increased from 11 minutes, for the wet spent carbon,
to 15 minutes. How this will affect the carbon losses is not
known at present.
77
-------
/ouu
2400
CM
* 2000
o>
2 1600
10
0)
0,1200
§
E-i
800
400
1 1 1 1 1
— —
— 1820*F 1700°F —
- ^ Z3— ~— — =
G -X***^* X°~~"~
X* X^ Carbon
^X Baking . Activating
^ 1 I 1 I 1
3 4
Hearth Number
FIGURE 16 - TEMPERATURE PROFILES FOR GAS AND CARBON IN
MULTIPLE HEARTH FURNACE, REGENERATION OF DRIED SPENT CARBON
2800
2400
^ 2000
^
V
^ 1600
m
K
»^
^••^ ^^^ ^^ ^~^—
^^^ ^^
^^ ^^
x^" / Carbon
-X
^^ Heat Carbon Activating
*x I I I
123456
Hearth Number
FIGURE 17 - TEMPERATURE PROFILES FOR GAS AND CARBON IN
MULTIPLE HEARTH FURNACE, ACTIVATION OF BAKED SPENT CARBON
78
-------
DISCUSSION AND CONCLUSIONS
The laboratory studies demonstrated a number of fac-
tors in the regeneration of granular activated carbon. Some
loss of the basic carbon structure will be unavoidable since
the final activation step occurs at temperatures of 1500°F and
above, where carbon is readily attacked by the steam and C02.
During attempted regenerations at temperatures below 1500°F
and with high activating gas input rates, the adsorbate car-
bonized, evolving a volatile fraction and leaving a free carbon
residue in the pores. About 60% of the adsorbate is evolved as
volatiles. The remaining 40% of carbon residue oxidized very
slowly at temperatures below 1500°F, hence the necessity of
carrying out the final activation at 1500°F and above.
The overall regeneration of the spent carbon was
found to proceed by way of three naturally occurring steps;
(1) a drying step wherein the wet carbon as received from the
contactors is dried, (2) a baking step wherein the adsorbate
is carbonized to the volatile fraction and free carbon residue
in the micropores and (3) an activating step wherein the free
carbon residue is removed from the micropores by steam and C02
oxidation. Laboratory investigations of steps 1 and 2 have
not revealed any optimum procedures that would reduce carbon
losses. The success of the overall regeneration is then de-
termined by the manner in which the activating step is per-
formed .
In the laboratory tube regenerations, the progression
of properties through the baking and activating step follow the
pattern as given below for two of the spent carbons studied.
Spent VIA
Bulk density, g/cc
Particle density, g/cc
Real density, g/cc
Pore volume, cc/cc
particle volume
Iodine number
Molasses number
Dried
0.616
0.990
1.88
0.473
529
172
Baked
0.546
0.876
2.13
0.590
798
198
Activated
0.493
0.789
2.15
0.633
886
249
79
-------
Once-Spent Filtrasorb 400
Dried Baked Activated
Bulk density, g/cc
Particle density, g/cc
Real density, g/cc
Pore volume, cc/cc
particle volume
Iodine number
Molasses number
0.584
0.971
1.88
0.484
673
174
0.532
0.871
2.07
0.578
955
193
0.502
0.825
2.13
0.613
1017
192
The properties of the activated once-spent Filtrasorb 400
essentially fall into the property ranges of the virgin Filtra-
sorb 400, indicating a good recovery of all the essential pro-
perties. Attempts to bring the iodine number of spent VIA to
the 1000 level have all failed. It was finally concluded that,
at the Pomona plant, this particular carbon quantity which was
sent to MSAR had been activated in a manner to change the
physical properties of the basic carbon.
A particle volume decrease has been observed during
the baking step although the baking atmosphere is nonoxidative.
Furthermore, the baking temperatures are sufficiently low to
rule out oxidation by water vapor or CO2 released from the ad-
sorbate. The investigations which have attempted to attribute
the particle volume decrease to carbon shrinkage and/or re-
moval of sewage residue on the outer surfaces of the carbon
particles have not been conclusive. The results, however,
appear to favor the sewage residue concept. The estimated
particle volume decrease, by either shrinkage and/or sewage
residue removal, is about 1.0% of the 1.3% average measured
on about 20 baked carbons.
Carbon loss can show up on the first reactivation
when the activating conditions are sufficiently severe to
cause decrease in particle size, i.e., external burnoff. It
can show up on later reactivations if most of the basic struc-
ture burnoff is internal. By internal burnoff, the pore walls
thin out making the exterior layers of the carbon less re-
sistant to abrasions. Even on normal handling, the carbon
particles break down forming fines of less than 40 mesh. These
fines being back washed out of the contactors measure up as
carbon loss.
Multiple-hearth furnace reactivations, as performed
at the Pomona plant, have suffered an average carbon loss of
7% as measured by bulk volume decrease in the contactors. In
the laboratory work, conditions have been found by which once-
spent Filtrasorb 400 has been regenerated to virgin carbon
80
-------
properties with overall particle volume decreases of 2.8% and
2.9% on two regeneration runs. In these runs the particle
volumes decreased during baking by 1.3% and 0.9%, respectively.
When allowances are made for the possible baking shrinkage and
residue removal, the particle volume decreases that represent
true carbon losses are then 1.8% and 2.0%. Sieve analyses of
activated products show 0.15% of -40 mesh fines. When this
loss is added to the above particle volume decreases the best
measured carbon losses on these two runs are 1.9% and 2.1%.
Spent VIA carbon has repeatedly been regenerated to the virgin
carbon bulk density level but the virgin carbon adsorptive
properties were not recovered. For this carbon, the overall
mean particle volume decrease for 19 regenerations was 2.9%.
When the baking and fines corrections are applied the mean car-
bon loss is 1.3%.
The large difference in carbon loss between the
multiple hearth furnace and the rotary tube regenerator appears
to be due to the different activating conditions. In the
hearth-furnace regeneration of wet spent carbon, the furnace
functions as a dryer, baker and activator. Since the heating
requirement cannot be separated from the activating gas re-
quirement, the large heating load due to the drying and also
baking requires an excess of heating gas relative to the acti-
vating step. The hearth furnace activation is characterized
by a large activating gas input rate, a short activating time,
low activating gas utilization rate and relatively low carbon
temperatures. In the rotary tube regenerator, where the acti-
vating gas input rate can be regulated independently of heat
requirements, the activation is characterized by low gas input
rate, slightly longer activating time, higher activating gas
utilization rate and higher carbon temperatures. In both acti-
vations the gas temperatures were at the 1700°F level. The
operating data on the hearth furnace and two laboratory runs
show this difference:
Run 32 Run 36 Hearth Furnace
N, std ft3/lb 9.6 7.4 27.3
CO? std ftVlb 1.4 1.0 3.3
H20 std ftVlb 5.0 1.9 23.5
Oj std ft3/lb 0.0 0.0 0.7
Carbon temperature,°F 1630 1630 1570
Activating time, min 16 16 11
Gas utilization, % 59 66 18
Further laboratory work is required wherein the hearth furnace
activating conditions are closely duplicated to determine
whether the above stated operating conditions are the real cause
81
-------
of the higher carbon losses. By performing the drying and
baking steps separately, the calculated gas input rate for a
hearth furnace activation comes much closer to the optimum
rates used in the rotary tube. The degree of effectiveness
of this change can also be determined with the laboratory re-
generator .
Vent gas analysis as a means for monitoring the
activation has shown promise but considerable difficulties
have been experienced in getting consistent results.
Carbon quenching experiments have shown some de-
crease in iodine number but not as large as reported at the
Pomona plant. No correlation with pore structure or molasses
number was observed.
It appears now quite feasible to perform the drying
and baking steps in a multiple hearth furnace by adjusting the
operating conditions so that the carbon temperature never ex-
ceeds 1500°F and that the carbon discharge temperature is
1500°F. On the second pass through the hearth furnace the
carbon is brought up to the activating temperature with gas in-
put rate near that given in Table 29.
82
-------
RECOMMENDATIONS
Work in several areas of the spent carbon regeneration
field are recommended on the basis of the funding of this pro-
gram. These recommendations are listed as follows:
1. Study of Activating Parameters
In deciding the best method of regeneration that
should be used, it is essential that the exact cause of the
hearth furnace regeneration losses be determined. Work done
so far indicates that it is due to the high gas through-put
rate. By duplication of the hearth furnace conditions this
can be verified.
Parameter studies to establish operating limits are
required for engineering of the process and equipment and to
subsequently determine the relative economics of regeneration
by (1) indirect heated rotary tube furnace, (2) direct fired
multiple hearth furnace,(3) a combination of the two and (4) a
two pass regeneration using the multiple hearth furnace.
2. Adsorption-Regeneration Cycle Studies
It can be expected with a fair degree of certainty
that, on each regeneration, some damage will occur to the pore
structure even under the optimum regenerating conditions. For
the economic evaluation, it is essential to determine the ex-
tent of this damage. This study requires the performance of
successive adsorption-regeneration cycles, starting with virgin
carbon, and evaluating the product on each regeneration.
3. Determination of Feasibility of Low Grade Carbon as Make Up
Damage to carbon structure manifests itself as re-
duced abrasion resistance and consequent fines losses. A lower
activity carbon has further to go before it comes friable,
hence, cheaper lower grades may present an economic advantage.
4. Chemical Oxidation
Treatment of spent carbon with a chemical oxidant
such as hydrogen peroxide is known to remove some of the adsor-
bate. In conjunction with the baking step, this procedure may
lower the carbon residue in the micropores. The smaller car-
bon residue will require less severe activating conditions and
thus reduce carbon loss.
83
-------
APPENDIX
IODINE NUMBER
The Iodine Number is defined as the milligrams of iodine ad-
sorbed by one gram of carbon when the iodine concentration of
the residual filtrate is 0.02 normal.
Procedure
1. Grind a representative sample of carbon until 90% or
more will pass a 325-mesh sieve (by wet screen analysis)
2. Dry the sample for a minimum of three hours in an
electric drying oven maintained at 150°C.
3. Weigh 1.000 gram of dried pulverized carbon
(see note 2).
4. Transfer the weighed sample into a dry, glass-stoppered,
250-ml Erlenmeyer flask.
5. To the flask add 10 ml of 5%-wt HC1 acid and swirl
until carbon is wetted.
6. Place flask on hotplate, bring contents to boil and
allow to boil for only 30 seconds.
7. After allowing flask and contents to cool to room
temperature, add 100 ml of standardized 0.1 normal
iodine solution to the flask.
8. Immediately stopper flask and shake contents vigorously
for 30 seconds.
9. Filter by gravity immediately after the 30-second
shaking period through an E & D folded filter paper.
10. Discard the first 20 or 30 ml of filtrate and collect
the remainder in a clean beaker. Do not wash the resi-
due on the filter paper.
11. Mix the filtrate in the beaker with a stirring rod and
pipette 50 ml of the filtrate into a 250 ml Erlenmeyer
flask.
12. Titrate the 50 ml sample with standardized 0.1 normal
sodium thiosulfate solution until the yellow color has
almost disappeared.
84
-------
13. Add about 2 ml of starch solution and continue titra-
tion until the blue indicator color just disappears.
14. Record the volume of sodium thiosulfate solution used.
15. Calculate the Iodine Number as follows:
X _ A - (2.2B x ml of thiosulfate solution used)
m ~ weight of sample (grams)
_ NT x ml of thiosulfate solution used
_ T
—
_
50
Iodine Number = — D
m
X/m = mg iodine adsorbed per gram of carbon
N]_ = normality of iodine solution
N7 = normality of sodium thiosulfate solution
A = N! x 12693.0
B = N2 x 126.93
C = residual filtrate normality
D = correction factor (obtained from attached graph)
Notes on Method
1. The capacity of a carbon for any adsorbate is dependent
on the concentration of the adsorbate in the medium con-
tacting the carbon. Thus, the concentration of the
residual filtrate must be specified, or known, so that
appropriate factors may be applied to correct the con-
centration to agree with the definition.
2. The amount of sample to be used in the determination is
governed by the activity of the carbon. If the resid-
ual filtrate normality (C) is not within the range
0.008N to 0.035N, given on the Iodine Correction Curve,
the procedure should be repeated using a different
size sample.
3. It is important to the accuracy of the test that the
potassium iodide to iodine weight ratio is 1.5 to 1 in
the standard iodine solution.
Reagents and Equipment
Hydrochloric acid, 5% wt - To 550 ml of distilled water
add 70 ml of reagent-grade concentrated
hydrochloric acid.
85
-------
Sodium Thiosulfate, 0.1 normal - In a one-liter volumetric
flask dissolve 24.82 grams of reagent-grade
sodium thiosulfate crystals (NaoSoO^-SI^O)
in distilled water. Add about 0.1 gram of
reagent-grade sodium carbonate and dilute to
the one-liter mark. This solution should be
allowed to stand for a few days before stan-
dardizing.
Standardize with reagent-grade metallic
copper. Dissolve about 0.2 grams of copper,
weighed to the nearest 0.1 mg, in 5 ml of
concentrated nitric acid and boil gently
to expel brown fumes. Dilute to about 20 ml
with distilled water and add ammonia water
dropwise until the solution is a deep blue
color. Boil again until the odor of the
ammonia is faint. Neutralize with acetic
acid until the precipitate which forms with
the acid dissolves and add 5 or 6 drops in
excess. Again bring to boiling. Cool to
room temperature. Add solid potassium
iodide in sufficient amount to redissolve
the copper iodide precipitate which forms.
Titrate with sodium thiosulfate until the
iodine fades to a light yellow color. Add
starch indicator and continue the titration
by adding the thiosulfate dropwise until a
drop produces a colorless solution. Calcu-
late the normality of the sodium thiosulfate
as follows:
Normality of _ weight copper
sodium thiosulfate" ml thiosulfate x 0.06354
Iodine Solution - Dissolve 127 grams of reagent-grade
iodine and 191 grams of potassium iodide in
distilled water. (See Note 3). Dilute to
one liter in a volumetric flask. To stan-
dardize the iodine solution, pipette 25.0 ml
into a 250 ml Erlenmeyer flask and titrate
with the standardized O.lN sodium thio-
sulfate. Use the starch indicator when the
iodine fades to a light yellow color. Then
finish the titration by adding the thio-
sulfate dropwise until a drop produces a
colorless solution.
86
-------
Calculate the normality of the iodine
solution as follows:
Normality of _ ml thiosulfate x normality thiosulfate
iodine solution 25
Starch Indicator - Mix one gram of soluble starch with a
few ml of cold water. Pour the mixture into
one liter of boiling water and allow boiling
to continue for a few minutes. This solution
should be made up fresh daily for best re-
sults.
Filter Paper - E & D, folded filter paper, 18.5 cm, No. 192.
Burrell Corporation Catalog No. 34-390.
87
-------
1.20
I ' I ' I ' I
1.15
£1.10
M
O
4J
U
1.05
4J
U
Q)
§1.00
u
0.95
0.90
I • I - I • I
1
0.010 0.014 0.018 0.022 0.026 0.030
Residual Filtrate Normality (C)
FIGURE 18 - IODINE CORRECTION CURVE
88
-------
APPENDIX
MOLASSES NUMBER
Molasses solutions are treated with pulverized activated carbon
of unknown decolorizing capacity and with a standard carbon of
known Molasses Number. The optical densities of the filtrates
are measured and the Molasses Number of the unknown is calcu-
lated from the ratio of the optical densities and the standard
value.
Procedure
1. Grind a representative sample of carbon until 90% or
more will pass a 325-mesh screen (by wet screen analysis)
2. Weigh 0.46 gram portions of pulverized standard carbon
of known decolorizing capacity and unknown carbon and
transfer to separate 400 ml beakers.
3. Add 50.0 ml of molasses solution to each beaker and
stir until the carbon is thoroughly wetted.
4. Place the beakers on hotplate and allow to boil for
30 seconds.
5. Immediately after boiling filter the samples by
vacuum through a Buchner funnel, using Whatman No. 3
filter paper which has been previously coated by
filtering 50 ml of the filter paper suspension (see
Note 2). Discard the first 10 to 15 ml of the sample
filtrate and collect remainder.
6. Using a 2.5 mm effective light path and a 425 mu (blue)
filter, compare the optical densities of the samples
against distilled water in a Fisher Electrophotometer
or other suitable instrument (see Note 3).
7. Calculate the Molasses Value as follows:
Molasses Value = K x **
A
K = Molasses value of standard carbon
A = Optical density of filtrate from carbon
being tested
b = Optical density of filtrate from standard
carbon
89
-------
Notes on Method
1. Any grade of commercial molasses which may be purchased
will vary considerably in depth of color. The dilution
of the initial solution to give the same final filtrate
color with the standard carbon compensates for such
variations. Once such an adjustment has been made on a
given lot of molasses, the proper dilution may be made
routinely.
2. The Buchner funnels and precoated filter paper should
be prepared beforehand so that no delay is encountered
in filtering the samples. Discard the filtrate from
the filter paper suspension before filtering the
samples.
3. Since the Fisher Electrophotometer is not equipped with
2.5 mm-cells, the cell holder was modified to accom-
modate the 10-mm Klett Summerson cell. A 2.5-mm
effective light path is obtained by placing in the
10-mm cell a 7.5-mm glass immersion plate.
Reagents and Equipment
Standard carbon - Small quantities of primary standard
carbon are available from the Activated Carbon
Division, Pittsburgh Coke & Chemical Company.
Molasses solution - Prepare by diluting 146 grams (see
Note 1) of blackstrap molasses with one liter
of distilled water. The weight of molasses
to be used varies with the particular lot of
molasses and is adjusted, if necessary, so
that the standard carbon produces a filtrate
with an optical density of 0.38 to 0.42. The
molasses solution is stored in a refrigerator
and any unused portion discarded after 24
hours. A suitable grade of Plantation black-
strap molasses can be obtained at any local
health food store.
Filter paper suspension - Prepare by mascerating 16
circles of Whatman No. 3, 7-cm filter paper
in one liter of distilled water.
Absorption cell - Klett Summerson, No. 902, 10-mm.
Immersion plate - Klett Summerson, No. 903, 7.5-mm.
90
-------
LIST OF REFERENCES
1. Juhola, A.J., Matz, W.H., and Zabor, J.W., "Adsorptive
Properties of Activated Carbon-Physical Structure and
Adsorption in Aqueous Phase", a paper presented at the
American Chemical Society Meeting, Division of Sugar
Chemistry, April 1, 1951.
2. Grant, R.J., "Basic Concepts of Adsorption on Activated
Carbon", Pittsburgh Activated Carbon Company.
3. Brunauer, S., Emmett, P.H., and Teller, E., J. Am. Chem.
Soc. £0, 309 (1938).
4. Communication from Arthur N. Masse, FWPCA Water Research
Laboratory, Cincinnati to Charles W. Carry, Los Angeles
County Sanitary District on observations made by Carl
Brunner, January 11, 1968.
5. Iodine and molasses decolorizing tests are standard tests
used in the activated carbon industry to determine activity
of liquid phase carbons. Origin of these test procedures
was not available.
91
------- |