WATER POLLUTION CONTROL RESEARCH SERIES • 17020 DAO 07/70
OPTIMIZATION OF THE REGENERATION
PROCEDURE FOR GRANULAR
ACTIVATED CARBON
U.S. ENVIRONMENTAL PROTECTION AGENCY
-------
WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Series describes the
results and progress in the control and abatement of pollution
in our Nationrs waters. They provide a central source of
information on the research, development, and demonstration
activities in the Environmental Protection Agency, through
inhouse research and grants and contracts with Federal, State,
and local agencies, research institutions, and industrial
organizations.
Inquiries pertaining to Water Pollution Control Research Reports
should be directed to the Head, Publications Branch (Water),
Research Information Division, R&M, Environmental Protection
Agency, Washington, D.C. 20460.
-------
OPTIMIZATION OP THE REGENERATION PROCEDURE
FOR GRANULAR ACTIVATED CARBON
Mine Safety Appliances Research Corporation
Evans City, Pennsylvania 16033
for the
ENVIRONMENTAL PROTECTION AGENCY
Project #17020 DAO
Contract #14-12-469
July 1970
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 • Price $1.25
-------
EPA Review Notice
This report has been reviewed by the
Environmental Protection Agency, and
approved for publication. Approval
does not signify that the contents
necessarily reflect the views and
policies of the Environmental Pro-
tection Agency, nor does mention of
trade names or commercial products
constitute endorsement or recom-
mendation for use.
ii
-------
ABSTRACT
Granular activated carbons, spent in tertiary treatment of
waste water, are thermally regenerated in regenerators such
as the multiple-hearth furnace. Wet spent carbon is fed to
the regenerator and undergoes three naturally occurring
steps, namely (1) drying at about 220°F, (2) pyrolysis of
the adsorbed pollutants at 500° to 1550°F and (3) activation
with flue gas and steam at 1600° to 1700°F. For each ad-
sorption-regeneration cycle, the carbon volume loss varies
from 5% to 10% and the activity loss is as high as 13% on
the first cycle but at diminishing amounts on subsequent
cycles.
Laboratory studies of parameters controlling regeneration
show that alkaline and iron nxide ash accumulation in the
carbon. during the use cycles, catalyzes the oxidation of
pores in the 18^ to 28A diameter range, the pores most ef-
fective in pollutant adsorption. By HC1 acid leach, to
yomflyo fho<^o metallic elements from th"e spent CaTPftn"!ETie
subsequenttherrria I regeneration prnrpprU wit^ Tejj^de^""
struct!on of the carbon as measured by the iodine and
molasses numbers. '—•
Steam as activating gas is more effective than C02 in re-
covery of the initial properties of the carbon.
Carbon volume decrease during the baking step averages out
around 2% and during the activation step at 1.8%. Baking
studies indicate that the 2% is not a true carbon loss but
rather an apparent volume decrease due to pyrolysis of
colloidal pollutants on the carbon particle exterior sur-
faces. The true loss of carbon during the laboratory re-
generations is then about 1.8%. The latter appears to
proceed by generation of submicron sized fines from the
carbon particle surfaces rather than due to oxidation.
drying of the wet carbon is more effective than fast
Attempted regeneration by leaching with solutions of NaOH,
NaOCl, ^02 and CC1. were ineffective and/or uneconomical.
This report was submitted in fulfillment of Project Number 17020DAO,
Contract 14-12-469, under the sponsorship of the Environmental
Protection Agency-
Key Words: Carbon regeneration, chemical regeneration,
HC1 acid treatment, thermal regeneration,
tertiary treatment with carbon.
ill
-------
TABLE OF CONTENTS
Page No
I. CONCLUSIONS l
II. RECOMMENDATIONS 5
III. INTRODUCTION 7
IV. EQUIPMENT AND PROCEDURES 9
Rotary-Tube Thermal Regenerating Unit 9
General Procedure for Thermal Regenerations 11
Unit for Leaching Spent Carbons 12
Carbon Test Procedures 14
V. EXPERIMENTAL RESULTS AND DISCUSSIONS 17
Task 1 - Study of Regeneration
Operating Parameters 17
Task 2 - Adsorption-Regeneration
Cycle Studies 46
Task 3 - Determine Feasibility of Low
Grade Carbons as Make Up 48
Task 2A and 2B - Cyclic Adsorption-
Regeneration Studies 51
Task 4 - Chemical Oxidation and Solvent
Extraction 52
Task 5 - Engineering Studies on Furnaces 89
Task 6 - Regeneration Control by Effluent
Gas Analysis 89
Task 7 - Regeneration of 25 Ib Quantities
of Spent Carbon 97
VI. ACKNOWLEDGMENTS 105
APPENDIX A - INPUT AND OUTPUT GAS
COMPOSITIONS OF ACTIVATION RUNS 107
APPENDIX B - GAS ANALYSIS WITH NO CARBON
IN REGENERATOR 113
REFERENCES 115
v
-------
FIGURES
Page
1 Cross Sectional View of Rotating Tube,
End Caps and Heating Furnace 10
2 Unit for Leaching Carbon at Temperatures
Up to 100°C 13
3 Unit for HC1 Leaching of Spent Carbons 15
4 Temperature Profiles for Gas and Carbon
in Multiple Hearth Furnace, Regeneration
of Wet Spent Carbon 21
5 Effect of C0£ Activation on Iodine Number 30
6 Pore Size Distribution, Darco Run 58 and
Filtrasorb 400 45
7 OSF 400 Leached with Low Concentration
Solutions of Caustic Soda and HC1 and
Pure Water; Bulk Density as Function
of Amount Liquid Passed Through Carbon 60
8 OSF 400 Leached with High Concentration
Caustic Soda Solution and HC1 and Pure
Water in Batch Type Process; Bulk
Density as Function of Number of
Batch Treatments 61
9 Ash Content of HC1 Leached OSF 400 68
10 pH of Filtrate From HC1 Leach of OSF 400
When Acid to Carbon Contact Time is 2 Hr 70
11 pH of Filtrate From HC1 Leach When 5 ml of
Acid is Added to 500 cc OSF 400 73
12 pH of Filtrate From HC1 Leach When 20 ml
of Acid is Added to 500 cc OSF 400 74
13 Ash Content of Carbon at Lower End of
Bed When Bed is Treated with Different
Quantities of Acid 77
14 Pore Size Distribution Measured by Water
Adsorption and by Mercury Penetration 81
vi
-------
FIGURES (Continued)
Page
15 Pore Size Distributions of Regenerated
Carbons 82
16 Pore Size Distributions of Regenerated
Carbons 83
17 Cumulative Surface Areas of Regenerated
Carbons 84
18 Cumulative Surface Areas of Regenerated
Carbons 85
19 Particle Volume Decrease During Activation
as Function of Difference in C02
Utilization as Determined by Direct
Measurement and by CO-Ho Analysis 95
20 Iodine Numbers of Regenerated Darco, F 400
and WVP&P Carbons as Function of Bulk
Densities 102
vi5
-------
TABLES
Page
I Regeneration Conditions, Effect of Oxygen 19
II Regeneration Results, Effect of Oxygen 20
III Regeneration Conditions, Wet and Dry OSF 400 23
IV Regeneration Results, Wet and Dry OSF 400 24
V Regeneration Conditions, Effect of
Temperature 25
VI Regeneration Results, Effect of Temperature 26
VII Regeneration Conditions, Steam and C02
Activation 28
VIII Regeneration Results, Steam and C02
Activation 29
IX Baking Conditions Fines Formation Study,
OSF 400 32
X Baking Results Fines Formation Study, OSF 400 33
XI Weight Decrease Determined From Vent Gas
Analyses 35
XII Volume Decrease During HC1 Leach and Fixed
Bed Baking of OSF 400 37
XIII Volume Decrease During HC1 Leach and Rotary
Tube Baking of OSF 400 38
XIV Volume Decrease of Nonleached Carbons
During Baking 39
XV Volume Decrease During Activation 41
XVI Regeneration Conditions, Effect of Particle 42
Size on Regeneration
XVII Regeneration Results - Effect of Particle
Size on Regeneration 43
viii
-------
TABLES (Continued)
Page
XVIII Regenerating Conditions, Tasks 2 and 3 49
XIX Results of Regenerations, Tasks 2 and 3 50
XX Regeneration Conditions, Tasks 2A and 2B,
First Cycle 53
XXI Regeneration Results, Tasks 2A and 2B,
First Cycle 54
XXII Regeneration Conditions, Chemically
Treated OSF 400 55
XXIII Regeneration Results, Chemically Treated
OSF 400 56
XXIV Caustic Solution and Methanol Leach of
Once-Spent Filtrasorb 400, Run 18 58
XXV Regeneration Conditions, HC1 Leach 64
XXVI Regeneration Results, HC1 Leach 65
XXVII Ash Content of Regenerated Carbons Before
and After Dilute HC1 Acid Leach 66
XXVIII Decolorizing Test Results on HC1 Leached
and Nonleached Carbons 66
XXIX Ash Content of HC1 Leached OSF 400 71
XXX Analysis of Filtrate From Experiment 3 71
XXXI Analysis of Filtrate From Experiment 1 69
XXXII Ash Analysis of HC1 Leached Carbons 76
XXXIII Ash Composition of Acid Pretreated and
Nonpretreated Regenerated Carbons 79
XXXIV Iodine and Molasses Numbers as Determined
by Test and Calculated From Surface Area 86
XXXV Surface Area Change at Different Pore
Diameters During Regeneration 87
Ix
-------
TABLES (Continued)
XXXVI C02 Utilization
XXXVII Results of Gas Analysis During Activation
XXXVIII Reactions Associated With Each Gas Analysis
XXXIX K for Tests 1 Through 5
XL Regeneration Conditions, WVP&P and Darco
XLI Regeneration Results, WVP&P and Darco
Page
92
93
96
97
98
99
-------
SECTION I
CONCLUSIONS
1. Thermal regeneration of wet spent carbon from the
tertiary treatment of waste water proceeds by three
naturally occurring steps; (1) drying at about 212°F,
(2) pyrolysis (baking) of the adsorbate at 500° to 1550°F
and (3) activation with steam flue gas mixture at 1650°F to
1700°F. In the Pomona multiple-hearth furnace, Step 1 oc-
curs in 22 min, Step 2 in 9 min and Step 3 in 15 min.
2. Slow drying of the as-received wet spent carbon, as in
an air-convection oven at 150°C for 48 hr, produced regen-
erated carbons of higher iodine number than when drying was
done rapidly in 10 to 20 min in the rotary-tube laboratory
regenerator. Iodine numbers of the regenerated slow-dried
carbons were in the 950 to 970 mg/g range while those of
the fast-dried carbons were in the 900 to 935 mg/g range.
3. Low temperature activation at 1600°F produces lower
activity regenerated carbons than when the activation is
carried out at!650° to 1700°F. The iodine numbers for the
former were in the range 900 to 910 mg/g while for the
latter they were 950 mg/g and slightly higher.
4. Thermal regeneration of graded mesh size fractions in
the 8 to 60 mesh size range showed no significant difference
in rate of regeneration. Spent West Virginia Pulp and Paper
Company carbon of 8 to 20 mesh fraction regenerated at es-
sentially the same rate as the 30 to 60 mesh fraction. The
finer mesh fractions, however, showed a tendency toward
larger carbon losses during the regeneration.
5. The rate of thermal regeneration is more dependent on
the pore structure than on particle size. A carbon such
as 14 to 30 mesh Darco, with 0.60 cc/g of pores in the 30A
to 1000A diameter range, regenerates faster and under much
milder regeneration conditions than Filtrasorb 400 with
0.20 cc/g of pores in the same diameter range. When gran-
ular carbons are sized to smaller particles, pores in the
10.000A to 100,OOOA diameter range are destroyed, but since
these pores do not control the regeneration rate, decrease
in particle size does not increase the regeneration rate.
This generalization applies to processes involving gas dif-
fusion but not to liquid diffusion.
-------
6. Attempts to regenerate spent carbons by leaching with
CC14 or with aqueous solutions of NaOH, NaOCl and \\2®2 were
not effective or economical.
7. HC1 acid leach of spent carbons had a beneficial effect
on the subsequent thermal regeneration. Iodine and molasses
numbers of the regenerated products were closer to the in-
itial values than for the nonleached regenerated products.
8. The effectiveness of the HC1 acid leach was due to the
fact that the acid leached out metallic elements such as Fe,
Na, Ca and K which, if left in the carbon, catalyzed the
oxidation of the pores in which they happen to be located.
Pore structure studies showed that, in nonleached carbons,
enlargement of pores occurred in the ISA* to 28A diameter
range during activation. This fact also implies that pol-
lutant adsorption is predominantly in pores of this size
range since the metallic elements that catalyze the pore
enlargement are initially a part of the pollutants.
9. Studies to optimize the HC1 treatment showed that it
could be performed effectively at ambient temperature, at
acid to carbon contact time of about 2 to 3 hr, at 0.076 Ib
acid/lb carbon (38% HC1 assay) and 1.4% acid concentration
based on weight of HC1. Estimated cost of acid is 0.145<£/lb
of regenerated carbon.
10. Steam activation consistently produced regenerated
carbons of higher iodine number than C02 activation. For
HC1 acid pretreated carbons, steam activation produced car-
bons of 1040 mg/g iodine number while with C02 activation
the iodine number was 960 mg/g. For nonleached carbons, the
results were 940 mg/g with steam activation while with
the iodine number was 880 mg/g.
11. Since the baking is done under conditions that should
not cause oxidation of carbon structure, no carbon volume
loss should be expected. About a 2% volume decrease con-
sistently appeared. The evidence obtained appears to in-
dicate that this volume decrease is due to pyrolysis of
colloidal contaminants on the exterior surface of the carbon
particles and not a true loss of the activated carbon.
12. Carbon volume decrease during the activation step
averages out around 1.8% and appears to be due to loss of
submicron fines broken off the carbon particle exterior
surface rather than oxidation by the steam.
13. Cyclic adsorption-regeneration studies, where the
spent carbon was not acid leached, showed progressive de-
-------
crease of iodine number over two cycles. On the first cycle
of the study, employing acid leach, the initial iodine num-
ber was attained.
14. In direct-fired furnaces the flue gas usually contains
1 to 2% unreacted oxygen. The laboratory studies showed
that oxygen in these quantities has no effect on the regen-
eration process.
15. A vent gas analysis study was undertaken to determine
whether monitoring the vent gases from the activation step
could be used to control the process. The results produced
too many inconsistencies to recommend the method.
16. Vent gas analyses showed that the H20, CO, C02 and H2
react with each other and come to an equilibrium according
to the equation
co~
where the equilibrium constant K varies with temperature.
With excess steam, the utilization of C02 is suppressed and
in some cases C02 has been produced.
-------
SECTION II
RECOMMENDATIONS
The success of the HC1 acid
oratory scale recommends it
scale regeneration systems.
economical relative to acid
to be the difficult problem.
leach pretreatment on the lab-
for application to the large
The pretreatment step is
cost, but corrosion is expected
Further studies are recommended on carbon pore structure
relative to ease of regeneration. Such a study would lead
to a better understanding of the adsorption and regeneration
process and form a basis for the selection of carbons best
suited for tertiary treatment.
Studies of the effect of particle size on ease of regen-
eration demonstrated that regeneration of powdered carbons
could be done effectively in conventional equipment, with
modifications to equipment and procedures of operation.
Further study is recommended for this approach.
-------
SECTION III
INTRODUCTION
On the previous program, Contract No. 14-12-107, regen-
eration equipment was installed and regeneration procedures
worked out for study of the regeneration process. Partial
answers were found for the regeneration process, but also,
new areas of study became apparent which were outside of
the scope of the contract.
One of the more important observations made was that the
overall regeneration, as practiced in the multiple hearth
furnace, consisted of three natural occurring steps; i.e.,
(1) drying, (2) pyrolysis of adsorbate (baking) and (3)
activation. In the activation step, free carbon residue
from the baking step is removed by steam oxidation. These
three steps were studied, but completely satisfactory
answers were not obtained regarding the parameters affect-
ing them.
The present program, Contract No. 14-12-469, was initially
contracted for 14 months, but promising areas of further
investigation were discovered that were outside the scope
of the contract. To investigate these new areas, the con-
tract was extended another 6 months. The continued studies
were divided into seven tasks as identified below.
Task 1 - Study of Regeneration Operating Parameters
Task 2 - Adsorption-Regeneration Cycle Studies
Task 3 - Determine Feasibility of Low Grade Carbon
as Make Up
Task 4 - Chemical Oxidation and Solvent Extraction
Task 5 - Engineering Studies on Furnaces
Task 6 - Regeneration Control by Effluent Gas Analysis
Task 7 - Regeneration of 25 Ib Quantities of Spent
Carbon
After considerable study had been made on the hearth furnace
operating parameters as part of Task 1 studies, a set of
conditions were to be set for the regeneration of one con-
tactor full of carbon at the Pomona, California pilot plant,
-------
a project jointly funded by the County Sanitation Districts
of Los Angeles County and the Federal Water Quality Admin-
istration, U.S. Department of the Interior.
Task 2 had a two-fold purpose. The initial properties of
the spent carbon, as received from the Pomona plant, were
not known exactly, therefore there was some uncertainty in
determining the recovery of the initial properties. This
uncertainty was removed in Task 2 ( and also Task 3) studies
since the properties of the starting material were deter-
mined and the same batch of carbon was carried throughout
the cyclic studies. The other purpose was to study the type
and rate of carbon degeneration over a number of cycles of
regenerati on.
Task 4 involved study of liquid chemical treatment of the
spent carbon as a primary regenerating process, an alterna-
tive to the thermal regeneration. It also involved study
of chemical pretreatment of spent carbon as an aid to the
thermal regeneration.
Task 4 opened up new areas of investigation in that HC1
leach of the spent carbon prior to thermal regeneration
greatly minimized loss of adsorptiye capacity during re-
generation. To investigate the acid leach, and other areas
of promise, the program was extended another six months.
-------
SECTION IV
EQUIPMENT AND PROCEDURES
Rotary-Tube Thermal Regenerating Unit
The thermal regenerations, involving the baking and activat-
ing steps, were conducted in a continuous feed, indirectly
heated rotary-tube regenerator. If offers the widest range
of versatility with respect to operating parameters. Acti-
vating gas composition and input rate can be varied to any
practical limit. Heat input can be varied independently of
the activating gas input rate and the carbon residence time
can also be varied independently of the other parameters.
Figure 1 shows the cross sectional view of the rotating tube
and sections of the furnace. Thermocouples TC I, TC II, and
TC III monitor the three sections of the furnace I, II, and
III, which can be temperature controlled independently of
each other. Thermocouples TC (I), TC (2), TC (3), TC (4)
and TC (5) monitor the gas temperature over the carbon bed.
Other pertinent details of the unit are listed below:
Tube size and material -
Flights
Heati ng
3.25 in. dia by 65 in.
length, stainless steel
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 fal1.
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.
Tube inclination
Carbon residence time
has been varied
to 4 inches per
tube length.
from level
42 inches of
The carbon residence time
can be varied from 10 min
to 3.7 hr by varying in-
-------
Carbon feed
Heating sections of furnace
t
End x , L
"" li-TJ
Gas outlet
ube
Carbon Rotating
seal gas seal
/ '' n
^ m 1
1
I | II
1
1
TC I | TC
1 j l'l
i '"
I
II | TC III
i i
IM I !
Rotati ng
gas seal
/End cap
/
^t TC(1) TC(2) TC{3) TC(4) TC(5J 1 ,
^> c
--/,
Rotary /
tube /
Drive
sprocket
1 1
— I ~ — pp — — r^ ~i —
Coll type
electric heaters
I I
I I
A
Gas inlet \
1 i
10
20
30
40
Inches
50
60 65
Regeneratec
carbon re-
ceiver - 2 qt mason
FIGURE 1 - CROSS SECTIONAL VIEW OF ROTATING TUBE, END CAPS AND HEATING FURNACE
-------
Gas input
clination and rotation rate.
The usual residence time for
baking is 30 min and for ac-
tivating 15 min.
The input gas composition and
input rate are varied by met-
ering N2, COp, 0? from com-
pressed gas cylinders and gen-
erating steam by means of a
calibrated boiler. To simu-
late multiple hearth furnace
operations, gas mixtures ap-
proaching the flue gas mix-
ture are usually used, i.e.,
72% N2, 10% C02 and 1
H20.
be varied from
Carbon feed rate
Vent gas analyses
Total""input can __ . _
4 to 40 ft3/hr (stp).
- An auger type volumetric feed-
er is used having a feed rate
range between 0.010 and 0.100
ft-vhr for granular carbons.
The usual feed rate is 0.016
ft3/hr, 453 cm3/hr.
- About 1.0 ft3/hr of gas mix-
ture is withdrawn from the
main vent gas stream during
activation, for C02, CO and
H2 analyses. The side-stream
is passed in succession through
a (1 ) glass wool fi1ter, (2)
water freeze out trap at -80°C
and (3) MSA In-Line-Ultra par-
ticulate filter, before reach-
ing the gas analyzers. The CO
and COo concentrations are ana-
lyzed with MSA Lira Model 300
analyzers and the H2 with an
MSA Thermatron.
General Procedure for Thermal Regenerations
Spent carbons, as received from the tertiary treatment pilot
plant at Pomoma are first dried for 48 hours in an air-con-
vection oven at 150°C, then sieved to 14 by 40 mesh, and
blended as preliminary preparation for the regeneration.
The amount of oversize and undersize particles was always
small and was discarded. The drying is considered the first
11
-------
step in the overall regeneration procedure. Generally 2500
cm-* of the dried spent carbon are measured out for each re-
generation run. The second step is baking,wherein the dried
spent carbon is passed through the regenerator at a rate to
give a 30 min residence time. The gas input during baking
has been varied considerably. On some runs a flue gas mix-
ture was used and on others, nitrogen. The latter, of
course, is not attainable in a direct-fired hearth furnace,
but the runs were made to obtain data on extreme conditions.
The gas temperature on most runs was controlled to vary lin-
early from 800°F at the carbon influent end of the bed to
1550°F at the carbon effluent end. The baking was also tried
with wet carbon feed, i.e., 40% by weight of water, to sim-
ulate conditions in the Pomona hearth furnace even closer.
With wet carbon feed, the baked carbon bulk density was
slightly less than with dried carbon feed. However, with
wet carbon feed, vent line plugging occurred frequently,
hence, dried carbon feed was then standard practice for
most of the baking runs.
During baking, tne adsorbate is carbonized yielding a vola-
tile portion and a free carbon residue which is then removed
by steam oxidation during the activating step. Fifty to
seventy percent of the adsorbate is removed as volatiles.
For the activating step,the gas temperature profile down the
tube is generally 1550°F for section I, 1650°F for section
II and 1700°F for section III. The required activating gas
input rate is calculated from previous experimental results.
For this purpose the bulk density is used as the control
test, i.e., gas input is varied to bring the bulk density
to that of the virgin carbon.
Further details on the regenerator unit and operating pro-
cedures are presented in the final report for Contract 14-
12-1071.
Unit for Leaching Spent Carbons
Task 4 involved leaching experiments with chemical oxidants
and solvents to develop a primary regenerative process or,
if not successful in this effort, to find an oxidant or
solvent that would aid the thermal regeneration process.
Chemicals tried in aqueous solution were H^O?, NaOH, NaOCl
and HC1. In some preliminary experiments, Buchner funnels
were used, but their use did not permit sufficient control
over temperature and flow rate of solution through the car-
bon bed. To correct this problem, the unit shown in Figure
2 was constructed.
12
-------
Glass tube
54 mm 00
47 in. tot,
length
Solution
constant
head
level at
pressure\
Disk with center
hole to mln1- —'
mlze convection
current
Heater
&
Insulation
Filtrate
flow
Carbon bed
Screen
-A
u
X
X*
^
/
X
X
\
*
< '.'X'
••V.';
/•'* •*
1 ."•••
•^ " ^ -
» '." *»
*'".**
^
F
V
v P
K
J
limp
12
1n. x 18 1n.
27 1 cap.
of leach
solution
FIGURE 2 - UNIT FOR LEACHING CARBON
TEMPERATURES UP TO 100°C
AT
13
-------
As shown in the figure, the carbon bed is contained in the
lower half of the glass tube with a constant head of solu-
tion above the bed. The column was heated by resistance
wire wrapped around the tube to maintain temperature of
solution and carbon bed at temperatures near 95°C. Flow
rate was maintained by adjusting the valve attached to
the filtrate tube. Flows of 0.5 to 3.0 1/hr were experi-
mented with, but the flow generally was set at 1.0 1/hr.
The progress of the regeneration was monitored by removing
the carbon from the column at intervals for drying and
weighing. Prior to removing carbon from the column, the
carbon bed was leached with pure water to remove the chem-
icals being used. For HC1 acid leaching of larger quan-
tities of spent carbon at ambient temperature, the unit
shown in Figure 3 was used.
Carbon Test Procedures
To determine the effectiveness of the regenerating process,
various measurements and tests are made on the carbon be-
fore and after the three steps of regeneration. These are
namely bulk density, particle density by mercury displace-
ment, real density by helium displacement, iodine number
and molasses number. The bulk density is based on weight
per unit volume of the container occupied by the carbon, the
particle density is per unit volume of particles and real
density is per unit volume of solids in the carbon particles,
Knowledge of the particle and real densities permits calcu-
lation of the pore volume of the carbon. The bulk densities
are measured during the run for control purpose since they
can be determined quickly.
The iodine number is related to the total surface area of
the carbon and molasses number to the area of pores larger
than 28A" in diameter.
Further details on making these measurements and their sig-
cance are given in the final report of Contract 14-12-
nifjc
107T-
Tables reporting the test results present the bulk density,
percent weight decrease, percent bulk and particle volume
decreases, pore volume, and iodine and molasses numbers.
At the top of each table, property data are given on the
spent carbon, representing starting material, and on the
virgin carbon, representing the goals to be attained by the
regeneration. A comparison of the bulk densities, pore
volumes and iodine and molasses numbers of regenerated car-
bons with those of the virgin carbon, measures the degree of
14
-------
HC1 acid
or
water
Liquid
level
Spent
Carbon _
2500 cc
Glass container,
4 1 cap.
Fine weave
nylon liner
Polyethylene
Contai ner,
3.8 cap.
SS fine mesh
screen
Flow controlled
by stopcock
Graduated
cyli nder
FIGURE 3 - UNIT FOR HC1 LEACHING OF SPENT CARBONS
15
-------
success of the regeneration. The percent decrease in bulk
and particle volumes measures the carbon loss.
16
-------
SECTION V
EXPERIMENTAL RESULTS AND DISCUSSIONS
Task 1 - Study of Regeneration Operating Parameters
Effect of Oxygen - In the direct-fired multiple hearth fur-
nace, the carbon comes in contact with flue gas from com-
bustion of natural gas or flue gas mixed with product gases
from activation. The flue gas can contain 1% to 2% of
oxygen by volume with compositions approximately as given
be!ow:
Composition, %
Component With 6% excess air With 12% excess air
N2 71.7 71.5
C02 9.2 8.7
H20 17.9 17.7
02 1.0 2.1
99.8 100.0
The composition of the gas passing over the hearth changes
after coming in contact with the carbon. For the Pomona
furnace, the calculated composition of gas entering the
fourth hearth, where the baking step occurs, is approxi-
mately as follows:
Component ft3/(hr x Ib carbon) Composi tion,%"'
N2 25.7 46.2
C02 3-8 6-8
H20 18.2 32.7
CO 3.9 7.0
H2 4.0 7.2
55.6 99.9
(!) This is a calculated composition. Due to a gas
phase reaction between the CO and H20, the volumes
17
-------
of CO and H20 are reduced while C02 and H~ volumes
are increased. This reaction is discusses later
in the report.
In the Pomona furnace regeneration, steam is added to the
flue gas, but if steam were not added the calculated com-
position of gas entering the fourth hearth would be closer
to the percentagesgiven below:
Component ft3/(hr x Ib carbon) Composition.%
"2
co2
H20
CO
24.1
2.7
3.8
2.7
2.2
0.2
35.7
67.5
7.6
10.6
7.6
6.2
0.6
100.1
Two activations were run with flue gas mixture of 1% oxygen.
These are Runs 4 and 13, with conditions and results given
in Tables I and II.
In regard to the baking step, no attempt was made to exactly
duplicate the gas compositions of the hearth furnace. It
was rationalized that at the 1550°F maximum temperature, the
presence of CO and H2 would have very little effect. In the
baking steps, flue gas mixtures were used with 1% oxygen
content. These are Runs 1 and 12 in Tables I and II. A
comparison of the results, with and without oxygen, tend to
show no significant difference. Run 16 does not fit the
pattern of results generally obtained on nonoxygen acti-
vations, hence some doubt exists on the validity of the low
(920) iodine number.
The possibility exists that the oxygen reacts with the H.
and CO in gas phase and never reaches the carbon, where
would react exothermally. On subsequent regeneration runs,
the addition of oxygen was discontinued.
Baking Wet and Dry OSF 400 - Figure 4 shows the measured gas
temperature over each hearth of the Pomona furnace. The car-
bon temperatures shown here were calculated from the thermo-
dynamic properties of the system. As indicated in the figure,
18
-------
TABLE I - REGENERATION CONDITIONS, EFFECT OF OXYGEN
Temperature, °F Gas input, ft3/hr (stp)
Run no. Regeneration step
64 (1) baking
65 (1) Run 64 activated
1 baking
4 Run 1 activated
7 baking
13 Run 7 activated
12 baking
16 Run 12 activated
22 baking
23 Run 22 activated
(1) Data on runs from contract 14-12-107.
1
1700
1700
550
1550
700
1550
750
1550
650
1545
2
1700
1700
1250
1650
1200
1650
1250
1650
1255
1650
3
1700
1600
1550
1720
1550
1700
1500
1700
1545
1700
h
4.8
3.6
6.8
1.09
7.9
2.7
6.8
2.7
7.0
3.7
CQ_2
0.00
0.50
0.90
0.14
1.0
0.34
0.90
0.34
0.90
0.47
02
0.00
0.00
0.10
0.015
0.00
0.02
0.05
0.00
0.00
0.00
H20
0.00
1.90
1.71
0.37
1.27
0.47
1.57
0.56
2.11
0.86
-------
TABLE II - REGENERATION RESULTS. EFFECT OF OXYGEN
Run no.
virgin
spent ( ! i
64 (2)
65 (2)
1
4
virgin
spent {•*)
7
13
12
16
22
23
Bulk
density ,
q/cc
0.490
0.570
0.522
0.486
0.500
0.480
0.468
0.583
0.506
0.476
0.503
0.479
0.510
0.480
Bulk
Weight volume
decrease, decrease,
%°/
la
10.41 17 2
7.5J u'*
12.91 0.5|_ , o
5. If 17'4 1.3J 1>8
— _ _ _ _ _
13.51 ,g 5 O.Tl ! 6
6.9J 19'b 0.9J '
14.21 g 7 0.6] . ,
6.4_j 19'7 0.9f K5
13.31 l.ol
7 . 2 J ] y ' ' 1 . 2J 2 . 2
Particle
vol ume
decrease ,
o/
10
1.41 2 2
0.8_j L't-
]*81 i ft
2. Of 3'8
— - _
l-4l 1 9
0.5J '*y
?:i}i-8
°-5l 2.2
l.Tf
Pore
volume,
cc/cc
0.510
0.598
0.625
0.619
0.628
0.650
0.500
0.611
0.633
0.615
0.635
0.600
0.630
Iodine
number ,
mq/g
990
650
960
950
950
1090
630
940
950
940
920
950
950
Mol asses
number
220
210
225
210
220
250
190
211
258
224
253
_ — «
(1) Runs 1, 4, 64 and 65 with this once-spent Filtrasorb 400 (OSF 400)
(2) Data from runs on contract 14-12-107
(3) Runs below with this OSF 400
-------
3500
01
3000
2500
2000
fO
S. 1500
E
O)
1000
500
Calculated flue gas temperature, 3350°F
Calculated flue gas-steam
mixture to hearth 4,
2650°F
Calculated flue gas-steam
mixture temperature to
hearths 5 and 6, 2230°F
Mean gas temperature
Carbon temperature
12345
Hearth number
FIGURE 4 - TEMPERATURE PROFILES FOR GAS AND CARBON IN MULTIPLE
HEARTH FURNACE, REGENERATION OF WET SPENT CARBON
-------
wet carbon fed into the furnace dries while traversing the
first three hearths and then goes through the baking step
on the fourth hearth. Activation occurs over hearths 5
and 6.
To simulate the conditions in the hearth furnace, a number
of drying-baking runs were made in the laboratory rotary
furnace with rewetted OSF 400. It is necessary to dry the
wet spent carbon, as received from Pomona, to permit mea-
surements of its properties. Rewetting consisted of adding
water to the carbon just to the point where it was still
free flowing. On a dry spent carbon basis this occurred at
about 38% water content.
Two regenerations of rewetted carbon were run. The con-
ditions and results are given in Tables III and IV, Runs
2-3 and 8-14. Data on nonwetted carbon runs, Runs 1-4,
and 22-23 are also given for comparison. The bulk densi-
ties of wet-baked carbons were slightly lower than those of
the dry-baked carbons. This would normally be regarded
favorably, except on activation the wet-baked carbons had
slightly lower iodine numbers, i.e., 900 and 935 mg/g com-
pared to 950 mg/g for the dry-baked carbons. In Task 4,
regeneration Runs 9-15, Table XXII, the OSF 400 was wetted
with HpOo and in this case also the iodine number was down,
i.e., 920 mg/g.
Baking of rewetted carbons caused considerable difficulty
by way of vent line plugging. To avoid this problem, sub-
sequent baking runs were made with oven dried OSF 400.
Effect of Temperature - As shown in Figure 4, the gas tem-
perature varies from 600 to 1700°F over hearths 1 to 4 and
holds steady at 1700°F over hearths 5 and 6. Various com-
binations of temperatures on baking and activating were
tried. Some were almost duplicates of the hearth furnace
conditions, some milder and some more severe. Conditions
of regeneration and results which show the effect of tem-
perature are presented in Tables V and VI.
In the regeneration Run 22-23, the baking was carried to the
maximum temperature of 1545°F with relatively high input of
C02 and H^O. Activation was then conducted at 1545°F to
1700°F. The regenerated carbon iodine number was 950 mg/g.
When the baking temperature was increased to 1600°F, as in
regeneration Run 48-51, the regenerated carbon iodine number
dropped to 910 mg/g. Run 50-52 is in part a repeat of Run
48-51 except that the activating temperature was increased
to the 1700°F level. For this run the iodine number dropped
to 900 mg/g.
22
-------
TABLE III - REGENERATION CONDITIONS, WET AND DRY OSF 400
Run
no.
1
4
2
3
8
14
22
23
Regenerating step
and special treatment
dried and baked
Run 1 activated
dried, rewetted, baked
Run 2 activated
rewetted 550 ml H20/
2500 cc and baked
Run 8 activated
dried and baked
Run 22 activated
Temperature,
1
550
1550
680
1550
980
1550
650
1545
2
1250
1650
1250
1650
1270
1650
1250
1650
°F
3
1550
1720
1550
1720
1550
1700
1545
1700
Gas i
N
6
1
6
1
7
2
7
3
2
.8
.09
.8
.99
.9
.7
.0
.7
nput ,
C02
0
0
0
0
1
0
0
0
.90
.14
.90
.26
.00
.34
.90
.47
ffVhr
°2
0.10
0.015
0.10
0.027
0.00
0.00
0.00
0.00
(stp)
?
1.
0.
1.
0.
1.
0.
2.
0.
,(J
71
37
71
37
75
51
11
86
-------
TABLE IV - REGENERATION RESULTS, WET AND DRY OSF 400
N5
-P-
Bulk Particle
Bulk Weight volume volume Pore
density, decrease, decrease, decrease, volume,
Run no. g/cc % % % cc/cc
virgin 0.490
spent (1) 0.570
1 0.500 12.9
4 0.480 5.1
2 0.500 12.4
3 0.475 6.4
virgin, 0.468
spent (2) 0.583
8 0.507 13. l"
14 0.482 6.5_
22 0.510 13.3
23 0.480 7.2_
0.510
L 17 A °-5l i Q 1-s1?R 0.619
f 1.3J ' -° 2.0J" 0.628
-180 1 • ^ ;_ -i n 2 . 2 '_ /i i 0.613
IB' 1.3! 3'° 1.9; 4'] 0.631
0.650
0.500
18 R °-5L 1 fi °-9' 9 R 0.607
j 1.11 1.9i 0.629
- 19 7 T-0"! 9 9 °-5 oo °-600
1 .2f ^'2 1.7" 2.2 0>630
Iodine
number,
mg/g
990
650
950
950
923
900
1090
630
940
935
950
950
Mol asses
number
220
210
210
220
224
220
250
190
212
229
(1) Runs 1, 4, 2 and 3 with this OSF 400
(2) Runs 8, 14, 22 and 23 with this OSF 400
-------
Ul
TABLE V - REGENERATION CONDITIONS. EFFECT OF TEMPERATURE
o
Temperature, °F Gas input, ft /hr (stp)
Run no. Regenerate ng step
22 baking
23 Run 22 activated
48 baking
51 Run 48 activated
50 baking
52 Run 50 activated
10 baking
41 Run 10 activated
61 baking
66 Run 61 activated
1
650
1545
820
1550
800
1700
800
1560
1700
1700
1250
1650
1350
1650
1350
1700
1250
1650
1700
1700
1545
1700
1600
1700
1600
1600
1550
1700
1600
1600
7.0
3.7
7.0
3.33
7.0
3.24
3.7
8.3
3.7
4.0
0.90
0.47
0.93
0.41
0.93
0.44
0.00
1 .00
0.00
0.55
2.11
0.86
1.68
0.74
1.71
0.56
0.00
2.62
0.00
1.49
-------
TABLE VI - REGENERATION RESULTS, EFFECT OF TEMPERATURE
ro
Run
no.
virgin
spent
22
23
48
51
50
52
10
41
61
66
Bulk
density ,
g/cc
0.468
0.583
0.510
0.480
0.503
0.471
0.500
0.469
0.53
0.476
0.513
0.477
Weight
decrease ,
13.31 -,Q 7
7.2J
15. 2J. 21.4
15.8 01 o
7.2" 21'8
lO-f-20.5
"1:1 ^
Bulk
vol ume
decrease ,
r.2~ 2'2
1.8" or
0.7]
1.6 op
1.2 2'8
1.0 -, 5
0.5" *D
- 3 0
0.7; J'U
Particle
vol ume
decrease ,
%
0.5^ 0 0
1.7
1.9" , R
1.9 lj'b
2.0". 4 7
2.7 4>/
0.1 o o
2.2.~2'3
o!e ^2<5
Pore
vol ume ,
cc/cc
0.650
0.500
0.600
0.630
0.617
0.649
0.623
0.644
0.600
0.636
0.605
0.638
Iodine
number ,
mg/g
1090
630
950
950
910
910
910
900
980
950
880
940
Mol asses
number
250
190
260
245
260
250
204
263
220
240
-------
The reason for the difference in the results given above is
due to an appreciable amount of low temperature activation
during the baking step for Runs 48-51 and 50-52. At tem-
peratures above 1500°F, the rate of activation increases
rapidly with temperature. But increasing the temperature
at the end of the baking step by 50°F for Runs 48-51 and
50-52 caused a considerable amount of low temperature ac-
tivation to occur at 1600°F. Low temperature activation
appears to be detrimental to the recovery of iodine number,
i.e., regeneration of micropores to original state.
To support the above interpretation, regeneration Run 60-61
was run at the 1700°F level but nitrogen sweep gas was used
during baking to avoid any activation. In this case the
iodine number was 940 mg/g; up considerably from the 900
mg/g as obtained for Run 50, also activated at the 1700°F
level.
Regeneration Run 10-41 was performed under a similar tem-
perature schedule as Run 22-23 except that N~ sweep was used
during baking. For this run, the iodine numBer of the re-
generated carbon was 950 mg/g, the same as for Run 22-23.
The conclusion is that at temperatures of 1550°F and below,
C02 and HoO vapor are relatively inert toward the basic car-
bon structure.
In regard to the temperature during the activating step, it
is desirable to maintain the temperature at or close to the
1700°F level. In those runs where the temperature profile
was from 1500° to 1700°F, most of the activation occurs in
the third section of the regenerator where the temperature
is 1700°F. When the gas reached sections II and III at the
lower temperatures, the COo and H20 contents are partially
depleted, hence reaction rates are considerably slower.
Steam and C02 Activations - Task 6 results indicated that
when C02 was utilized to any appreciable extent in the ac-
tivation step, the iodine number decreased. To verify this
trend, regenerations were then conducted in which the acti-
vation step was solely with C02-N2 mixture or steam-N2 mix-
ture. The C02-N2 activation was also carried out on a HC1
pretreated carbon.
The conditions of regeneration and results are presented in
Tables VII and VIII. Included for comparison are data of
Run 61-66, which was activated with a steam-C02-N2 mixture
but the C02 did not contribute to the activation. Figure 5
presents graphically the results of Table VIII and results
obtained in the Task 6 studies. The iodine number of the
C02-activated, nonleached carbon is 840 mg/g compared to
27
-------
TABLE VII - REGENERATION CONDITIONS, STEAM AND CO? ACTIVATION
N3
CO
Run no. Regenerating step
62 baking
67 Run 62 activated
64 baking
68 Run 64 activated
61 baking
66 Run 61 activated
HC1 leach
95 baking
96 Run 95 activated
Temperature, °F Gas input, ft3/hr (stp)
1
1700
1700
1700
1700
1700
1700
850
1550
2_
1700
1700
1700
1700
1700
1700
1350
1650
3.
1600
1600
1600
1600
1600
1600
1550
1700
N.2
3.7
4.0
3.7
2.0
3.7
4.0
7.0
14.0
C0.2
0.00
0.00
0.00
2.91
0.00
0.55
0.93
4.00
H.20
0.00
2.81
0.00
0.00
0.00
1.49
1.71
-------
TABLE VIII - REGENERATION RESULTS, STEAM AND C0 ACTIVATION
10
Run no,
virgin
spent
62
67
64
68
61
66
1 each
95
96
Bulk Wei
Bulk Particle
ght volume volume Pore
density, decrease, decrease, decrease, volume,
g/cc %
0.468
0.583
0.518 13.3J
0.467 11. Ij
0.516 13.l!
0.478 9.1;
0.513 13. ll
0.477 8.0
0.555 6.3"
0.498 11.7
0.486 4.6
01 of r r- / r r
h io L L / L L
0.650
0.500
?o a 2.4 _ o 7 2.6 ,0 0.603
"'y 1.3 6ml 1.2, 0.647
21 0 2-2- 3.2 2-3- 3.7 °-604
1.0 1.4: 0.645
?n 1 2-3l 3 0 ]-9- ? 5 °'605
- 20.1 Q>7^ 3.0 Q>6j 2.5 Q>638
1.21 2.6] 0.528
-21.2 1 .7r 5.0 0.3- 4.8 0.624
2.1 1.9j 0.640
Iodine
number,
mg/g
1090
630
880
940
900
840
880
940
_ _ _
900
960
Molasses
number
250
190
220
260
220
250
220
240
250
240
-------
1100
en
en
E
S-
01
JD
EE
T3
O
1000
900
800
0
© MCI leach, 1550° to 1700°F act.
Runs Qnonleach, 1550° to 1700°F act.
36 &
79 Q nonleach, 1700° to 1600°F act.
0.0 0.5 1.0
Mol C02/mol carbon oxidized
FIGURE 5 - EFFECT OF C02 ACTIVATION ON IODINE NUMBER
30
-------
940 mg/g for the steam activated. For the HC1 pretreated
carbon, the iodine number of C02-activated carbon is 960
mg/g compared to 1040 mg/g for the steam activated. A 100
unit drop in iodine number can be expected when activation
is solely by C02.
Study of Particle Volume Decrease During Baking - The carbon
loss during regeneration is measured by the particle volume
decrease. It occurs during both the baking and activating
steps. The volume decrease during the activating step is
understandable and expected since the activating gases would
also be expected to act on the basic carbon structures and,
thereby, reduce its volume. The particle volume decrease,
however, was unexpected during the baking step when a nitro-
gen sweep gas is used. A considerable effort was made to
determine whether the volume decrease was due to actual loss
of basic carbon structure or whether it was due to two other
possible effects, i.e., (1) expansion of carbon during ad-
sorption and contraction during removal of adsorbate or (2)
decomposition and volatilization of colloidal adsorbate
clinging to the exterior surface of the carbon particles.
A satisfactory answer to this question was not found, al-
though some of the experimental evidence appeared to favor
the colloidal adsorbate hypothesis. Because this question
was not resolved, there was uncertainty as to the true car-
bon loss suffered during the regeneration. Was it about 1.5%,
assuming true loss occurring only during the activating step,
or was it more like 3.0%, assuming true loss occurring dur-
ing both steps?
Vent gas analyses have indicated that, at least in part, the
particle volume decrease during activation is due to loss of
fines of submicron diameter. If fines formation is the
mechanism by which particle volume decreases during baking,
it would be profitable to determine whether there are con-
ditions by which it could be minimized.
One method that suggested itself was to carry out the baking
under extremely mild conditions. This can be done in a fixed-
bed oven. The carbon would not be agitated as in a rotary
tube to cause possible attrition of weakened particle surface
layers. The heatup time for the baking step can be very
gradual and over an extended time of at least 50 hr. Evolu-
tion of gases would not be rapid enough to carry off any
fines. To determine this point, four fixed-bed oven baking
runs were made. The baking conditions and test results are
given in Tables IX and X. Each spent carbon was HC1 acid
leached prior to baking according to a standardized procedure
given in Task 4.
31
-------
TABLE IX - BAKING CONDITIONS FINES FORMATION STUDY. OSF 400
Temperature, °F Gas input, ft /hr (stp)
Run no.
0)
_ — —
(2)
_ — —
(3)
80
— — —
(4)
92
93
94
97
Treatment
HC1 leach
Fixed bed
HC1 leach
Fixed bed
HC1 leach
Fixed bed
Rotary tube
HC1 leach
Fixed bed
Rotary tube
Rotary tube
Rotary tube
Rotary tube
1
60-^
« ~ *
60^
_ mm _
60 ->
850
_ — _
60-^
850
850
850
850
2
1550
_ — _
1550
_ _ _
1550
1350
_ — —
1550
1350
1350
1350
1350
3
in 48 h r
— — —
in 48 hr
_ — _
in 48 hr
1550
_ — _
in 48 hr
1550
1550
1550
1550
No COo HoO
1.25
_ _ .
1.25
„ .. _
1.25
3.7
— • —
1.25
3.7
3.7
3.7
3.7
-------
TABLE X - BAKING RESULTS FINES FORMATION STUDY. OSF 400
Lo
Ru
n no.
OSF 400
HC1 leach
(1)
HC1 leach
HC
HC
(2)
1 leach
(3)
80
1 leach
(4)
92
93
94
97
Bulk
Bulk Weight volume
density, decrease, decrease,
q/CC % %
0
0
0
0
0
0
0
0
0
0
0
0
0
.583
.555 5.8]
.506 10. 8j
-i
.560 4.8
.507 11.1
.562 4.4~
.511 10.0
.510 0.8_
.552 6.1"
.505 9.9
.504 0.4_
.506 0.6
0.6
.502 0.1
i r r\ 'J •
- 1 K 1 1
1 O * \) n
t- •
ISA °-
IS. 4 ^
0.
-14.7 1.
0.
0.
-15.8 1.
0.
0.
8l, i
-f I
3JJ- '
"\
8 f\ r
— / n
8
6]
1 j-2.1
4
8
4 r 2.3
1 j
8
0.5
Particle
vol ume-
decrease ,
1 .
1.
1.
0.
1 .
0.
0.
1 .
0.
0.
0.
0.
1.
2j
6 in
— 1 LI
3
o"
2-1.6
4..
2~
3-1.9
4J
4
7
1
Po
vol
cc/
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
re
ume ,
cc
501
527
617
520
614
521
598
608
519
622
618
621
616
617
Iodine
number,
mq/q
630
600
880
880
_ -. ..
930
930
850
920
900
Molasses
number
190
- — —
_ .. _
_ _ _
220
240
— — —
220
-------
In the fixed-bed Run (1), a considerable amount of ash was
found on the surface of the baked carbon particles, indi-
cating that oxygen from some source had reached the carbon.
This was unexpected since nitrogen was passed through the
oven during heat-up and cool-down. On Run (2), the amount
of ash formed was considerably less, but still considered
too much for the purpose of the run. On Runs (3) and (4),
panels of granular carbon, between stainless steel screens,
were placed around the OSF 400 carbon beds to shield out
external oxygen. Also new stainless steel screens and trays
were used to eliminate oxidized surfaces which could re-
lease oxygen on contact with hot carbon. On these two runs
no ash appeared on the baked carbon surface.
The results of Runs (3) and (4), which are the most valid
ones, present conflicting evidence. Particle volume de-
crease was very small, of the order of 0.3%, while the bulk
volume decrease was much larger, i.e., 1.1% and 1.4%. It
appears as though the 0.3% decrease in volume of the par-
ticles then permitted the particles to fit closer together,
hence the larger decrease in bulk volume.
When the fixed-bed runs were then rebaked in the rotary-
tube regenerator, again in an inert atmosphere, the bulk
volume and particle volume decrease tended to be equal and
of the order of 0.4% over repeated baking runs, Runs 80 and
92 to 97. A 0.4% to 0.8% weight loss also occurred on each
successive baking run. Vent gas analyses, Table XI, in-
dicate that the weight loss is essentially due to evolved
CO and COo. However, the amount of oxygen in the ^ was not
sufficient to account for the amount of CO and CO? evolved,
hence other sources of oxygen pick-up were investigated.
A possible chemisorption could occur when the baked carbon
is exposed to air during handling and bulk volume measure-
ments. On each baking run, the chemisorbed oxygen would
then be evolved as CO and C02. To test this hypothesis, the
product of Run 93 was kept under nitrogen and fed to Run 94.
In this run, also, there was the usual weight and volume
loss, and evolution of CO and C02- The conclusion arrived
at was that the oxidized surface of the rotary tube yielded
oxygen to the carbon. This could mean that on every baking
run, whether in inert atmosphere or not, a 0.4% particle
volume decrease occurs. This loss is characteristic of the
manner in which the unit is operated, i.e., on a long con-
tinuous run this loss would not occur after the metal sur-
faces have become reduced.
The conclusion at this point of the study was that, on MCI
leached spent carbons, there is only a negligible loss of
carbon volume during baking.
34
-------
TABLE XI - WEIGHT DECREASE DETERMINED FROM
VENT GAS ANALYSES
Run no.
92
93
94
97
Weight
decrease ,
g/hr
1 .05
1.38
1.44
0.82
Gas release rate,
mol /hr
CO
0.035
0.035
0.010
0.019
C02
0.010
0.051 (])
0.025
0.009
Weight of
gas released ,
g/hr
1.42
3.32 (])
1.38
0.93
(1) This value appears to be in error
35
-------
The study was continued by considering the carbon losses in-
curred during the acid leach. Table XII separates out the
pertinent data from Table X. For Runs (2), (3) and (4), the
particle volume decrease is about 1.3% while the bulk volume
decrease is about 0.7%. This is the reverse of the volume
decreases occurring during baking. Apparently, the larger
particle volume decrease causes a very small change in pack-
ing efficiency of the particles.
If a further examination is made of HC1 leached-rotary tube
baked carbons (see Table XIII), the same pattern is observed
on bulk and particle volume decreases as in Table XII. Here,
also, the particle volume decrease during baking of previous
HC1 leached spent carbons is of the order of 0.2%. Acid
leach reduced the particle volume by about 2.0%. These bak-
ing runs were made in a flue gas atmosphere, and since the
particle volume decrease is essentially the same as for the
fixed-oven runs, the flue-gas then must also act as an in-
ert gas.
A second conclusion, arrived at from this study, is that col-
loidal material from the waste water adheres to the exterior
surfaces of the carbon particles, thereby enlarging them.
The HC1 acid leach removed this colloidal material, hence
the particle volume decreases during leaching but with little
or no particle volume decrease during baking.
When nonleached spent carbon is baked, the colloidal mater-
ial is subject to heat treatment which will cause it to de-
compose and volatilize. In reviewing the past runs, evidence
has been found which shows that the degree of decomposition
and volatilization varies with the conditions of baking.
These data are presented in Table XIV.
In the early part of the program, when runs up to and in-
cluding number 31 were made, Section I of the rotary tube
was heated to 750°F. On these runs, the particle volume de-
crease was on the order of 0.5%. At this temperature, liquid
water collected in the carbon influent end of the tube caus-
ing operational difficulties. To correct this, the temper-
ature of Section I has since been operated at 800°F and
higher. Except for Run 44, the particle volume decrease also
rose, now in the 1.4% to 3.3% range. The explanation that
appears most reasonable is that when Section I is 750°F, the
carbon temperature rise is gradual as it proceeds down the
rotary tube into higher temperature sections. Under a
gradual temperature rise, the colloidal matter stays fixed
to the carbon surface and then decomposes, leaving a deposit
of free carbon. The free carbon deposit lessens the volume
loss from what would occur if all the colloid had been re-
moved.
36
-------
TABLE XII - VOLUME DECREASE DURING HC1 LEACH
AND FIXED BED BAKING OF OSF 400
Run no.
Volume decrease,^
Bulk
Particle
Spent leached
Fixed bed baked, No.(1)
Spent leached
Fixed bed' baked, No.(2)
Spent leached
Fixed bed baked, No.(3)
Spent leached
Fixed bed baked, Mo.(4)
means, excluding No.(l)
2.3
0.8
1.8
o.el
3
J'
2.6
i.i
0.8
1 .4
r 1-7
2.2
1.6
0.3
i.o"
0.2
- 1.9
-1.2
' - 1 5
0.3 ''^
2.2
1.5
37
-------
TABLE XIII - VOLUME DECREASE DURING HC1 LEACH
AND ROTARY TUBE BAKING OF OSF 400
Volume decrease,
Run no. Buik Parti cTe
Spent leached O.ol 0 7 Loo
45 2-7J J
Spent leached 0.4] 9 r 1.9
49 2.1_j ^>;J 0.2r '•1
Spent leached 0.2
63 2.4j"
Spent leached O.ol , Q 1 . 7|_ -, r
78 1.9J l>y -0.27
Spent leached 1.2 0 Q 2.6_ 9 Q
95 1.7j'y 0.3^-y
means 2.5 2.4
38
-------
TABLE XIV - VOLUME DECREASE OF NONLEACHED
CARBONS DURING BAKING
Run no
Section 1
Temperature,
. _y_glume decrease. %
BuikParticle
12
22
29
31
37
38
44
48
50
62
64
71
72
750
650
750
750
850
850
810
820
800
1700
1700
800
800
0.6
1 .0
1.2
0.8
2.0
2.0
0.4
1.8
1.6
2.4
2.2
1.8
1.6
0.3
0.5
0.8
0.3
1.6
1.4
0.0
1.9
2.0
2.6
2.3
3.3
2.9
39
-------
When the temperature of Section I is raised to 800°F and high-
er, the increased temperature causes an increase in expulsion
rate of volatiles from within the pores and also in the rate
of volatilization of the colloidal matter. The net result
is that a greater portion of the colloidal matter is removed
with the vent gases, and hence the larger decrease in particle
volume.
When these carbons are activated, the prior treatment does
not appear to greatly affect the volume decrease. As shown
in Table XV, the mean particle volume decrease is about 1.8%.
If the particle volume decreases, during HC1 leach and bak-
ing is accepted as being due primarily to removal of col-
loidal matter from the exterior surface of the particles,
then the true carbon loss is closer to 1.8% rather than the
3.5% to 5.0% as previously reported.
Effect of Particle Size on Ease of Regeneration - A parameter
that needed to be investigated was the effect carbon mesh
size may have on the ease of regeneration. It is known that
powdered carbons regenerate much more rapidly than coarse
granular carbons. If the regeneration rate is appreciable
within the range of mesh sizes normal in commercial carbons,
it is evident that the small particles could be over-acti-
vated and thereby suffer excessive particle volume loss
while the large particles are still underactivated. A large
difference in ease of activation can lead to a new investi-
gation in optimizing mesh size, contacting procedures and
regeneration. The ease of regeneration of the open pore
structured Darco carbons has demonstrated that wide vari-
ations can exist amongst carbons. On the Darco, complete
regeneration was always effected by the baking step alone.
This study was conducted primarily with West Virginia Pulp
and Paper Company spent carbon since it consisted of coarser
mesh sizes than the OSF 400. The original 8 to 30 mesh was
sieved and sized to obtain four mesh size fractions, 8 to 12,
20 to 30, 30 to 60 and -60. A powdered -100 mesh OSF 400
and a powdered -100 mesh Nuchar Aqua A were later added to
the study. Prior to thermal regeneration, each carbon was
HC1 acid leached and dried according to procedure given in
the Task 4 section of this report.
The conditions of regeneration and results are given in
Tables XVI and XVII. The results of the study indicate that
the size of the particle does not control the rate of re-
generation. In the WVP&P series, no trend is indicated in
the iodine number. The iodine number of the OSF 400, Run 85,
is in line with those of the baked acid leached granular
OSF 400 carbons. However, decrease in particle size appears
40
-------
TABLE XV - VOLUME DECREASE DURING ACTIVATION
Volume decrease, %
Run no. Pretreatment Bulk Particle
29-30 HC1 leach 0.4 0.4
31-32 HC1 leach 1.4 2.6
37-39 HC1 leach 1.4 2.0
38-40 HC1 leach 1.3 1.5
48-51 HC1 leach 0.7 1.9
50-52 HC1 leach 1.2 2.7
means 1.1 1.8
12-16 none 0.9 1.5
22-23 none 1.2 1.7
29-30 none 0.4 0.4
31-32 none 1.4 2.6
37-39 none 1.4 2.0
38-40 none 1.3 1.5
44-46 none 1.4 2.5
45-47 none 1.1 1.8
49-53 none 0.7 1.7
63-69 none 0.7 1.9
78-79 none 0.9 2.0
means 1.0 1.8
41
-------
N3
TABLE XVI - REGENERATION CONDITIONS, EFFECT OF PARTICLE
SIZE ON REGENERATION (1)
Run
no.
81
88
82
89
83
90
84
85
Carbon an
mes
WVP&P,
WVP&P ,
WVP&P,
WVP&P,
WVP&P,
WVP&P,
WVP&P,
OSF 400
h size
8 to 1
8 to 1
20 to
20 to
30 to
30 to
-60
d
Regenerating
step
2
2
30
30
60
60
, powdered
b a k i
Run 81
baki
Run 82
baki
Run 83
baki
baki
ng
act.
ng
act.
ng
act.
ng
ng
Temperature ,
1
800
1555
800
1555
800
1550
800
1700
2
1300
1650
1300
1650
1300
1650
1300
1700
°F G
3
1500
1700
1500
1700
1500
1700
1500
1600
as i
No
2.
3.
2.
3.
2.
3.
0.
3.
nput,ft3/hr (stp)
C02
0
5 0.43
0
5 0.43
0
5 0.43
5
0
H?U
1.13
_ _ _
1.14
_ _ _
1.25
_ _ _
-100
87 Nuchar, powdered
-100
baking
830
1320
1500
2.0 0.18 0.13
(1) All spent carbons HC1 leached prior to regeneration
-------
TABLE XVII - REGENERATION RESULTS - EFFECT OF PARTICLE SIZE ON REGENERATION
Run
no.
Carbon
Spent WVP&P (8 to 30)
OSF 400
Spent Nuchar (-100)
Virgi
Virgi
Virgi
81
88
82
89
83
90
84
85
87
n WVP&P (8 to 30)
n F 400
n Nuchar (-100)
WVP&P
8 to 12
WVP&P
20 to 30
WVP&P
30 to 60
WVP&P -60
OSF 400 -60
Nuchar, aqua A
Bulk Particle
Bulk volume volume Weight Iodine
density, decrease, decrease, decrease, number,
g/cc % % % mg/g
0.597
0.583
0.304
0.504
0.469
0.320
0. 538 1 .6J. 2 3 0.4
0.519 0.7J ' 0.2.
630
390
1070
1090
760
-0.6 7.3l_n.7 1000
4.7J 1050, 1070
Molasses
number
180
190
250
290
0.532 2. it , ? 0.6~1 o i 8.2l 1? c 950
0.519 1.1J 1.5J 4.6J 1010, 1030
0.478 1 .8J_ 4 n 0.7
0.457 2.2J ' 1.3.
0.630 9.6
0.694 5.7
0.294 9.0
2 n 9-3"l 14 9 1Q00
6.1J ' 1060
13.1 1000
12.3 870, 860
15.1 770
230
270
-100
-------
to accelerate the carbon loss as measured by bulk volume and
particle volume decreases.
A more important factor in rate of regeneration may be the
pore size distribution. Granular carbons such as the Darco,
of the type used in tertiary treatment, regenerates more
readily than WVP&P and Filtrasorb 400, granular or powdered.
Pore size distribution curves in Figure 6 of F 400 and the
Darco show differences that can explain the greater ease with
which the Darco can be regenerated. The Darco has an abun-
dance of pores in the 30A to 1000& diameter range which offer
passage ways for the gases to diffuse out of or into the
small pores. Also a greater portion of the Darco surface
area is in larger pores than that of F 400. In F 400, the
number of passage ways or pores of 30A to 1000A diameter are
considerably less, hence gas diffusion out of or into the
20^ pores is then slower. The major portion of the surface
area of F 400 is in pores less than 20A diameter, which can
hold the adsorbate molecules more strongly.
The experimental evidence indicates that pores of 1000& to
100.000A diameter do not appreciably slow the rate of gas
diffusion while pores below 1000A, particularly in the 30A
region, may be rate controlling. When granular F 400 or
Darco are pulverized, some of 1000A to 100,OOOA diameter pore
volume becomes interparticulate volume. Hence pulverizing
does not appreciably change the gas diffusion rates in the
particles. Nuchar Aqua A is expected to have a pore size
distribution similar to Darco. Nuchar Aqua A was easily
regenerated to the virgin carbon adsorptive capacity at 15%
weight decrease and 9.0% vulk volume decrease.
The evidence for the large pores becoming interparticulate
volume is the bulk density change that occurred when OSF 400
was sized. In the 14 to 40 mesh granular form, the bulk den-
sity is about 0.560 g/cc, but after pulverizing, the bulk
density increased to 0.750 g/cc. The bulk density of pow-
dered carbon was based on the minimum volume obtained by
tapping a graduated cylinder containing a weighed amount of
the carbon until the volume stopped decreasing.
General Cone! usions - Aside from the conclusions drawn fro;n
thevarious parameter studies, one important result was not
explicitly stated although quite apparent, i.e.,, that the
iodine numbers of the regenerated carbons were all consider-
ably lower than the iodine number of the virgin carbon given
at the top of each table. The properties of the virgin car-
bon were assumed to be initially those of the OSF 400, hence,
were the goals to be attained when complete regeneration was
accomplished. This assumption was based on the results at-
44
-------
1.20
Ul
10 30 100 1000 0 10,000 100,000
Pore diameter, A
FIGURE 6 - PORE SIZE DISTRIBUTION, DARCO RUN 58 AND FILTRASORB 400
-------
tained on the OSF 400 when regenerated with caustic and then
baked in Task 4. The parameter study of Task 1 definitely
indicates that manipulations of the temperature, gas compo-
sition and gas input rates during thermal regeneration do
not create the necessary conditions for a satisfactory re-
generation.
Task 2 - Adsorption-Regeneration Cycle Studies
After repeated regenerations, a gradual change in pore struc-
ture can be expected to occur which will ultimately cause
disintegration of the carbon particles and cause the carbon
to lose its adsorptive capacity. Inorganic compounds are
also expected to accumulate in the pores and thereby further
diminish adsorptive capacity.
To study these possible effects, an adsorption-regeneration
cyclic study was carried out with the cooperation of the
Pilot Plant Pomona. Starting with 2500 cc bulk volume of
Filtrasorb 400, 14 by 40 mesh carbon, the adsorption phase
was carried out at the Pomona Plant, with secondary treated
waste water, and the regenerating and testing at the MSA Re-
search Corporation laboratory at Evans City, Pennsylvania.
Carbon loss occurring on each cycle was made up with virgin
Filtrasorb 400.
The fact that the properties of the initial carbon were
known at the outset of the cyclic studies avoided one of the
problems that had caused considerable uncertainty on all
previous regeneration runs.
It had been the original intent to perform ten cycles, but
after the third one it was apparent that the carbon was de-
generating much faster than had been expected. This trend
is shown by the decrease in iodine number and increase in
molasses number, as below:
Cycle Iodine number, mq/g Molasses number
initial 1090 250
1st 1040 310
2nd 935 290
3rd 940 355
It was also noted that the activating step on the second and
third cycles were greatly accelerated. To prevent over acti-
vation, it was necessary to greatly reduce the activating
gas input rate- Below are input rates which show the large
46
-------
decrease necessary on the second and third cycles.
Activating gas Input rate. ft3/hr (stp)
C.y c 1 e jjgCO^ F£o
1st 3.6 0.36 0.99
2nd 1.27 0.17 0.18
3rd 1.27 0.17 0.12
For each cycle the other activating conditions were essen-
tially the same and, in each case, the final regenerated
carbon density was 0.468 g/cc, the bulk density of the
virgin carbon.
The overall carbon loss for the first cycle was 5.4%, 2.8%
for the second, and 1.2% for the third, as measured by de-
crease in particle volume. The 5.4% loss is larger than
normally experienced while the 2.8% for the second cycle
i s normal.
In the light of improved results on thermally regenerating
HC1 leached spent carbon, it became apparent that leaching
could have greatly reduced the degeneration rate. The
accelerated degeneration rate appears to be associated with
the increased ash contents on each successive cycle. Data
on ash increase are as follows:
Cycle Ash,%
virgin 5.7
1st 7.6
2nd 8.6
3rd 9.5
Estimates were made to determine whether the decrease in
iodine number was accompanied by a decrease in adsorption
of waste water contaminants. The calculations were made
using the formula
(Density of spent carbon)-(denslty of reg carbon) Y lnn
(density of reg carbon) A IUU =
% adsorbate.
For the two completed cycles and the adsorption phase
third, the calculated results are as follows:
47
-------
Cycle %, Adsorbate
1st 17.1
2nd 14.1
3rd 15.0
A decrease in adsorptive capacity may be indicated, although
the decrease may also be due to fluctuations in adsorbate
concentration of the secondary treated waste water.
The regenerations were carried out under conditions that can
be attained in a multiple hearth furnace. Further details
on the cyclic studies are presented in Tables XVIII and XIX.
Task 3 - Determine Feasibility of Low Grade Carbons as
Make Up
The procedure for this task was identical to Task 2 except
that a lower grade Filtrasorb 100 was used as make-up carbon
This study was based on the premise that a lower activity
carbon would be upgraded during regeneration and, because
of the price difference between Filtrasorb 400 and 100,
would lead to an economic advantage. The 1969 prices were
28 l/2<£/lb for Fi 1 trasor b 400 and 20 l/2<£/lb for Filtrasorb
100. On the two cycles, the pattern was the same as in
Task 3. Any benefit that might have accrued from use of
lower grade make-up carbon was obscured by the rapid degen-
eration. Test results showing the degeneration of the car-
bon are given below:
Cycle Iodine number, mg/g Molasses number
Initial 1090 250
1st 1040 310
2nd 920 310
As in Task 2, the activating gas input had to be decreased
considerably on the second cycle to prevent over activating,
as i ndi cated below:
Activating gas input rate, ft3/hr (stp)
Cycle N_2 C£? -2—
1st 2.5 0.32 0.72
2nd 1.27 0.17 0.21
48
-------
TABLE XVIII - REGENERATING CONDITIONS, TASKS 2 AND 3
Run
no.
24
25
54
56
26
27
55
57
91
98
Task
2
2
2
2
3
3
3
3
2
2
Cycle
1
1
2
2
1
1
2
2
3
3
Regenerati ng
step
bake
activation
bake
activation
bake
activation
bake
activation
bake
activation
Temperature
1
770
1550
825
1550
760
1550
840
1550
850
1550
2
1350
1650
1350
1650
1340
1660
1350
1650
1350
1650
, °F
3
1560
1700
1550
1700
1560
1700
1550
1700
1550
1700
Gas input, ft3/h
N2
7.0
3.6
7.0
1.27
7.0
2.5
7.0
1.27
7.0
1.27
COo
0.90
0.36
0.93
0.17
0.90
0.32
0.93
0.17
0.90
0.17
r (stp
H20
1.47
0.99
1 .68
0.18
1.57
0.72
1.57
0.21
1 .43
0.12
-------
TABLE XIX - RESULTS OF REGENERATIONS. TASKS 2 AND 3
Run
Task
virgin
spent
24
25
reg
spent
54
56
reg
spent
91
98
virgin
spent
26
27
reg
spent
55
57
reg
spent
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
Regenerating
step
to Pomona
adsorption
bake
acti vati on
to Pomona
adsorpti on
bake
acti vation
to Pomona
adsorpti on
bake
activation
to Pomona
adsorption
bake
activation
to Pomona
adsorption
bake
acti vati on
to Pomona
adsorpti on
Bulk
density,
g/cc
0.468
0.548
0.493
0.468
0.474
0.541
0.480
0.469
0.472
0.549
0.482
0.473
0.468
0.546
0.484
0.464
0.476
0.538
0.481
0.469
0.478
0.544
+ 13.0"
-12.8
- 4.2
-4.0
-5.6
+ 14.6
-15.0 - -5.8
- 3.3
+ 11.8
-11 .7
- 4.1
h -5.1
Bulk Particle
volume volume Iodine
change, change, number,
% % mg/g
3.36J -1.12]
2.52T 7'° ---1
1.38.1 ---J
0.81"
1 .73
1.12.
o.r-
1.0
1.1 .
+ 0.84
- -3.6 -3.33
-0.33.
-oT
- -2.6 -0.0
-1.2 J
0.60J -0.72"
1.81k -3.9 -0.73
1.54J -0.86.
0.97
1.31
1 .51
-1.98"
- -3.8 -1 .24
-0.33
1090
- -5-4 III
1040
- -2.8 915
935
- -1 .9 834
940
1090
L -2.3 HI
1 1040
...
- -3.5
920
Molasses
number
250
310
- - -
270
290
280
355
250
310
- - -
310
+ 13.2
-0.40
-1 .48
-------
For_this task, the conclusions are the same as for Task 2.
It is felt that removal of part of the inorganic compounds
by HC1 leach would have greatly reduced the rate of de-
gradation. The ash content increased in the same manner as
in the Task 2 carbons.
Cycle
Initial
1st
2nd
Ash,%
5.7
7.9
8.7
Overall carbon loss on the first cycle was 2.3% and on the
second 3.5%.
The estimated adsorbate on the carbon on each cycle was as
follows :
Cycle
1st
% Adsorbate
16.7
2nd
3rd
13.0
13.8
Further details on Task 3 studies are also presented in
Tables XVIII and XIX.
Task 2A and 2B - Cyclic Adsorption-Regeneration Studies
In view of the improved thermal regeneration of HC1 leached
carbons, the cyclic studies were resumed during the six
months extension period. For these studies, two new 2500 cc
batches of virgin Filtrasorb 400 were prepared. Batch 2A
was as-received except for sieving out the -40 mesh fraction
and batch 2B was leached in succession by HC1 acid, pure
water, hot caustic solution and pure water. This was done
to remove all of the F203, CaO, MgO, Na20 and K2Q and part
of the Si02 and A1203, see Task 4 on HC1 leach. The final
ash content was 3.5% compared to 5.7% for the virgin carbon.
These carbons were spent at Pomona as was done in Task 2 and
then regenerated at MSA Research laboratories. Prior to the
thermal regeneration, each batch of spent carbon was HC1
leached.
51
-------
The regeneration conditions and results of these two regen-
erations are presented in Tables XX and XXI.
For both regenerated carbons, the iodine numbers are higher
than those of the original by a considerable margin although
the final bulk density in each case matched the original.
This is the first time this had occurred in the regeneration
program.
The molasses numbers did not change significantly which is
as desired.
The carbon losses are larger than had been expected. The
actual regeneration losses are 4.3% for 2A and 4.4% for 2B
as measured by particle volume change. For 2B, the carbon
loss during the adsorption phase was 1.3%, by particle
volume change, while there was an actual increase of 0.52%
in bulk volume. For 2A, the carbon loss during the adsorp-
tion phase was very large, being 7.2% by bulk volume change
and 8.5% by particle volume change.
The bulk and particle volume changes occurring during each
step of the regenerations, for these two carbons, do not fit
the general pattern as established in Tables XIII and XV.
The bulk volume changes of 2B come closest to fitting the
general pattern. In view of the studies made on particle
volume decrease during baking in a previous section, most of
the apparent baking losses should not be included in the
overall regeneration loss. If 2.4% is allowed for the ap-
parent baking loss, the overall adsorption-regeneration cycle
loss for 2B is 3.3%. This percentage figure agrees well with
the 3.7% overall weight loss.
Ash analyses on 2A and 2B after regeneration were 4.7% and
3.5%.
This was only the first adsorption-regeneration cycle but
the results look promising in that original activity is re-
covered without excess carbon loss.
Task 4 - Chemical Oxidation and Solvent Extraction
H20? Pretreatment - Two pretreatment experiments were per-
formed with 3% H202. In Run 9, Tables XXII and XXIII, 2500 cc
of dried OSF 400 were wetted with 550 ml of 3% H202 and then
baked in the regenerator. This treatment produced no ob-
servable effects on the baked carbon. In Run 11, 2500 cc of
dried OSF 400 were leached with 4800 ml of 3% H20? in Buchner
funnels and the leached carbon dried and baked. The leach-
ing reduced the bulk density from 0.583 g/cc to 0.569 g/cc.
52
-------
TABLE XX - REGENERATION CONDITIONS, TASKS 2A AND 2B, FIRST CYCLE
Temperature, °F Gas input. ft3/hr (stp)
Run Regeneration step 1 2 3 N9 CO? H_?£
t 2A
leached
03
04
it 2B
1 eached
02
05
dried
baked
activated
dried
baked
activated
— — C. C- —t-
860 1360 1550 7.00 0.90 1.58
1550 1650 1700 2.19 0.37 1.12
860 1360 1550 7.00 0.90 1.39
1550 1650 1700 2.19 0.37 0.81
-------
TABLE XXI - REGENERATION RESULTS, TASKS 2A AND 2B, FIRST CYCLE
Ln
Run
Initial 2 A
Spent 2A
HC1 leached
103
104
Initial 2B
Spent 2B
HC1 leached
102
105
Bulk
d e n s i ty ,
g/cc
0.477
0.540
0.527
0.492
0.474
0.473
0.546
0.533
0.494
0.475
Bulk Particle
Weight volume volume Pore
change, change, change, volumes
% % % cc/cc
0.659
+ 5.0] -7.2j -8.5l 0.568
-3.1_ n R -0.6L n 1 -0.5L_i? Q 0.592
-8.8f "'8 -2.3 ' -1.4^ ^ 0.631
-4.8J -1.3] -2.8 0.649
0.669
+ 15.9 +0.51 -1.3J 0.551
-3.7 -37 -1 .4i 41 -0.8_ c 7 0.574
-8.6 " -1.3 ' -3.0 ' 0.631
-5.6J -1.8J -0.7 0.649
Iodine
number>
mg/g
1105
1150
1065
1150
Molasses
number
290
290
235
300
-------
TABLE XXII - REGENERATION CONDITIONS. CHEMICALLY TREATED OSF 400
Run
no.
Ui
Ln
8
14
9
15
11
17
18
20
21
Regenerating step,
special treatment
baker feed rewetted
with 550 ml H20 per
2500 cc carbon
Run 8 activated
baker feed wetted
with 550 ml 3% H202
per 2500 cc carbon
Run 9 activated
Temperature, °F
baker feed leached
with 4800 ml 3%
H202 per 2500 cc
carbon
Run 11 activated
500 cc carbon
with caustic,
Run 18 baked
Run 20, second
leached
HC1 , etc.
bake
1
980
1550
1000
1550
860
1700
485
550
2
1270
1650
1250
1650
1250
1700
1230
1240
I
1550
1700
1550
1700
1550
1600
1550
1550
Gas
*2
7.9
2.7
7.9
2.7
7.9
2.1
2.0
2.0
input, •
cp2
1.0
0.34
1.0
0.34
1.0
0.27
0.00
0.00
Ft3/hr
02
0.0
0.00
0.0
0.02
0.0
0.00
0.00
0.00
(stp)
HzP-
1.60
0.47
1.32
0.72
1.50
0.39
0.00
0.00
-------
TABLE XXIII - REGENERATION RESULTS, CHEMICALLY TREATED OSF 400
Ul
Run no.
virgin
spent
8
14
9
15
11
17
18
20
21
Bulk
Bulk Weight volume
density, decrease, decrease,
g/cc % %
0.469
0.583
0.507 13. V
0.482 6.5_
0.507 13.5
0.477 7.1J
0.504 14.5]
0.480 5.9J
0.481 18.6"
0.467 5.5
0.460 0.9
- 18.8 0 • 5~L 1 . 6
l.lj
-i
-197 0* °\- 1 9
iy./ K1j i.y
1Q fi °'9l 1 7
- 19.6 Q>8J. 1.7
2.4"
- 21.9 1.8 - 4.9
0.7.
Parti cl e
volume
decrease ,
0.9] 2.8
1.9J
0.81 , 9
l.lj '
0.9"1 , ,
2 . 2f '
6.4'
8.9
2.7
Pore
vol ume ,
cc/cc
0.650
0.500
0.607
0.629
0.611
0.629
0.610
0.629
0.616
0.634
lodi ne
number,
mg/g
1090
630
940
935
940
920
960
950
910
1030
Mo 1 asses
number
250
190
212
229
211
237
213
251
260
-------
On baking, the bulk density reduced to 0.504 g/cc, which is
slightly lower than the 0.507 g/cc for untreated baked car-
bons. The iodine numbers of both carbons after activation
were at the same level as for the unpretreated carbons, Run
14 as example, hence no benefit was derived. Additional
Ho02 treatment would not create more favorable results since
trie \\2$2 cost already for the above leaching pretreatment is
7.2<£/Tb carbon.
Caustic Leach - Three caustic leach runs were performed to
determine the economic feasibility of caustic leaching as
an alternate method for spent carbon regeneration.
The first leach run, designated as Run 18, was exploratory
to determine the critical parameters. Concentration, flow
rate, and temperature were varied. Methanol, caustic-methanol
solution,, and periodic pure water and HC1 acid leach were
also tried. Attempts were made to follow the effectiveness
of the treatment by grading the color density of the filtrate,
but this method proved to be unreliable since the color kept
changing in tone. The only trustworthy method of monitoring
was by periodic bulk density determinations.
Table XXIV gives a chronological record of the carbon treat-
ment. The initial volume and weight were 500 cc and 292 g,
respectively. About 350 1 of the various solutions were
passed through the column before the filtrate became essen-
tially colorless. At this point the bulk density was 0.481
g/cc, still above the 0.468 g/cc bulk density of virgin
Filtrasorb 400.
The leached carbon was then baked twice in a nitrogen at-
mosphere to volatilize as much of the adsorbate as possible.
The baking treatment brought the density down to 0.460 g/cc
and iodine and molasses numbers to high values of 1030 mg/g
and 260, respectively. These test results are fairly close
to the properties of the virgin carbon received from Pomona
on December 9, 1968. On this basis it was assumed that the
spent carbon used in most of the regeneration runs of this
program had initially the properties of this virgin carbon.
During the leaching, it was noted that if the carbon bed was
allowed to boil, even gently, by raising the temperature to
100°C, the filtrate started to carry carbon fines. The
movement of the carbon bed may have caused abrasion of the
carbon particles. The particle volume decrease during the
leach was 6.4%. During baking, it was 2.7%, which is high
compared to results on previous OSF 400 baking runs. The
higher particle volume decrease here may be another indi-
cation of generation of fines during the leaching process.
57
-------
TABLE XXIV - CAUSTIC SOLUTION AND METHANOL LEACH OF
ONCE-SPENT FILTRASORB 400. RUN 18
CO
Vol ume
of
leach
liq, 1
0
46
9
3
6
6
9
6
3
12
3
Fl ow
rate* Type
1/hr liquid
2.5-3.
2.5
2.5
2.5
2.5
2.5
2.5
2.5
4.0
4.0
0 5.0% NaOH
1 .0% NaOH
Water
0.25% NaOH
0.50% NaOH
1.0% NaOH
5.0% NaOH
0.25% NaOH
0.10% NaOH
Water
Temp, oH of
°C filtrate
...
50-80
1
Above leaching done in Buchner funnel
Below leaching continued in. heated column, uni
1
4
11
50
3
4
12
4
1
10
4
1
12
1
32
2
2
1
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
2.0
2.0
2.0
2.0
1:8
0.10% NaOH
0.10% NaOH
0.05% NaOH
0.10% NaOH
1.0% NaOH
0.10% NaOH
Water
Methanol
1 .0% NaOH in
methanol
Methanol
Water
1.3% HC1
Water
0.10% NaOH
0.10% NaOH
0.10% NaOH
1.3% HC1
0.10% NaOH
0.10% NaOH
95-100
11
i
10.7
60-65 8.8
10.6
95-100
80
95-100 5.3
95-100
80
95-100
Bulk Bulk
den of volume
Color of carbon, decrease,
filtrate g/cc %
0.584
Dark yel low
Light yellow 0.529 0.0
t shown in Figure
100 color unit
60
17
3
4
2.5
0.513 1.8
Light pink
Light yellow
Light pink 0.509 1 .8
Light yellow
Green
Colorless 0.492 2.2
70
11
..
Col orless
30
-------
In the second caustic leach run, Run 19, the caustic con-
centration was kept very low, at 0.025% and 0.013%. The
carbon bed was kept at near 100°C, but actual boiling or
movement of the bed was avoided as much as possible. At
intervals, the carbon bed was leached with dilute HC1.
About 550 1 of caustic and HC1 solutions were passed through
the bed, but filtrate still had low level of color. After
450 1 of solution had been passed through the bed, the car-
bon was taken out, dried and the bulk density measured. No
change had occurred in bulk volume (still at 500 cc) but
the density was still high, 0.576 g/cc.
Figure 7 shows the decrease in bulk density as function of
amount of solution passed through the carbon for each Run
18 and 19. In general, the concentrations of caustic were
5- to 10-fold greater in Run 18 than Run 19. The higher
concentration required less solution but reduction in water
required was considerably less than 5- to 10-fold.
A third run was made with 25% caustic soda and HC1 solution
treatments in a batchwise process. Each batch treatment
consisted of a routine of caustic soda leach, water leach,
HC1 leach and water leach; the routine was done twice. In
the caustic leach, 235 g (regenerated weight) of spent car-
bon was placed in a flask with a caustic solution and heated
at boiling point temperature for one hour. The caustic
solution was filtered from the carbon in a Buchner funnel
and the carbon, on the funnel, then leached with six liters
of hot pure water, followed by one liter of hot 6% HC1
solution and finally with four liters of hot pure water.
Each batch treatment required overall 600 g caustic soda,
30 cc concentrated HC1 and 24 liters of water.
The results of this run are given in Figure 8, which show
the change in bulk density after each batch treatment. At
the fifth treatment the effectiveness of the treatment was
very small. An extrapolation of the curve indicates that
inordinately large quantities of caustic would be required
to bring the bulk density down to 0.468 g/cc. At ten-batch
treatment, the amount of caustic required would cost 25<£/lb
of carbon. If recovery of caustic, for reuse, were possible,
the recovery process would have to be better than 98% ef-
ficient to avoid excessive material cost.
On the basis of the three caustic soda and HC1 leach runs of
widely different agent concentrations, the conclusion is that
caustic leaching is not competitive with thermal regeneration,
Inordinately large amounts of water would have to be used to
regenerate the carbon to initial activity. Over 8% of the
water purified per pound of carbon would be reused for leach-
59
-------
0.60 i—
0.55
o
o
"•-.
en
#t
>>
j«:
3
03
0.50
0.46
50
TOO
150 200 250
Gal/lb carbon
300
350
400
FIGURE 7 - OSF 400 LEACHED WITH LOW CONCENTRATION SOLUTIONS OF CAUSTIC SODA AND HC1,
AMD PURE WATER; BULK DENSITY AS FUNCTION OF AMOUNT LIQUID PASSED THROUGH CARBON
-------
0.60
o
o
en
0.55
O)
•a
_i^
3
0.50
0.46
1
0
1
1 2 3
1 ! 1
4
Number
of
5
times
1
6
each
ed
7 8
9 10
FIGURE 8 -
OSF 400 LEACHED WITH HIGH CONCENTRATION CAUSTIC SODA SOLUTION
AND PURE WATER IN BATCH TYPE PROCESS; BULK DENSITY AS
FUNCTION OF NUMBER OF BATCH TREATMENTS
AND HC1
-------
ing purposes. The caustic and HC1 acid costs would also be
excessive. For Run 18, more than 0.84 Ib of 76% caustic
and 0.15 Ib of 37% HC1 acid would be required. At 5.35<£/lb
for the caustic soda and 1.95<£/lb for the acid, the agent
costs would be more than 4.8<£/lb of regenerated carbon. For
Run 19, the agent costs would be considerably over 3.3<£/lb
of carbon before regeneration could be effected to initial
properties.
Of academic interest, it was observed that the leaching ef-
ficiency of the caustic solution increases with increased
temperature. Periodic HC1 acid treatments appear to accel-
erate the caustic leach.
NaOCl Leach - Two short leaching runs were made with NaOCl
solution in the leaching column with 500 cc quantities of
OSF 400. The first run was with 35 1 of 1.0% NaOCl solution
at about 90°C and 1.0 1/hr flow rate. The carbon density
decreased from 0.580 to 0.556 g/cc, but bulk volume decreased
from 500 to 452 cc. The filtrate was black from finely sus-
pended carbon.
In the second run, 40 1 of 0.25% NaOCl solution were run
through the column. In this case also, there was consider-
able disintegration of the carbon. Density decreases from
0.580 to 0.550 g/cc and volume from 500 to 420 cc.
Because of the high carbon losses, NaOCl was judged an un-
satisfactory regenerating agent.
CC14 Extraction - Fifty cc of OSF 400 were extracted in a
300 ml Soxhlet extractor for 40 turnovers of the CCl^. The
amount of CCl^ used in the extractor was 210 ml. After the
extraction, the treated carbon was air-oven dried at 150°C
for several hours and again heated under vacuum at 110°C in
preparing samples for CC1* adsorptive capacity test. The
CC1* adsorptive capacity, under static vacuum test, was 36%
compared to 38% for the untreated OSF 400. Estimated iodine
numbers based on the CC1. capacity tests were 550 and 640 mg/g
for the treated and untreated carbons, respectively.
No benefit was derived from the CCl^ extraction.
HC1 Leach Pretreatment, Effect on Iodine Number - HC1 leach
was first conceived as a pretreatment to assist in the re-
moval of adsorbate during the baking step. The rationali-
zation was that organometal1ic compounds of Fe, Ca and Na
adsorbed from waste water would resist volatilization and
thereby leave a larger carbon deposit. HC1 leached OSF 400
carbon, however, showed no greater adsorbate loss than un-
62
-------
leached carbons but a large increase in iodine number oc-
curred after the activating step. It is now believed that
in_the unleached carbon, Fe203, CaO and Na20 catalyzed the
oxidation of the base carbon structure and thereby lowered
the iodine numbers and increased the molasses numbers of
the regenerated carbons from what they should have been.
HC1 leach removed these metals and the thermal regeneration
then restored the carbons closer to their initial properties.
Tables XXV and XXVI present regenerating conditions and re-
sults of a number of runs made under a variety of conditions
but in all the runs HC1 acid pretreatment was employed. Re-
generation Runs 33-34, 35-36, 45-47, 49-53 and 78-79 were
performed under conditions that can be attained in a directly
heated multiple hearth furnace. The iodine numbers of these
regenerated carbons varied between 1020 and 1050 mg/g com-
pared to 900 to 950 mg/g for thermal regenerated unleached
carbon. Since the iodine number of the initial carbon is
believed to be 1090 mg/g, there is still room for improve-
ment.
Regeneration Runs 63-69 and 65-70 were performed under con-
ditions that cannot be duplicated in directly heated multiple
hearth furnaces because of the inert gas sweep used during
baking. The high temperature over the baking step would also
be difficult to attain. The inert gas sweep during baking
and the higher temperatures are now of academic interest
since they produced regenerated carbons of lower iodine
numbers.
The HC1 leach for these regenerations was done on Buchner
funnels with acid solution and leach water heated to near
boiling point before being poured over the carbon. About
30 ml of 37% HC1 diluted to 500 ml water was used to treat
500 cc of carbon. About 5 1 of water was then used to leach
out the chloride ion, i.e., until a silver nitrate test was
negative. At this point, the pH was still 3.5.
HC1 Leach Posttreatment, Effect on Iodine Number - Prior to
the HC1 pretreatment studies, tests had been carried out to
determine the effect HC1 leach of regenerated carbons would
have on the iodine and molasses numbers. The results showed
that the ash content had no effect on molasses number while
removal of ash lowered the iodine number in some cases. The
results of these tests are summarized in Tables XXVII and
XXVIII, This information confirms the fact that the in-
crease in iodine number of regenerated carbons, pretreated
with HC1, is due to the effect the pretreatment had on the
thermal regeneration rather than being due simply to removal
of the ash.
63
-------
TABLE XXV - REGENERATION CONDITIONS. HC1 LEACH
Run Temperature, °F Gas input,ft3/hr (stp)
no. Regenerating step, special treatment _]_ 2_ 3_ N_2 £0.2 —2—
leach, 150 ml HC1 per 2500 cc OSF 400, hot --- ---
33 above carbon baked 800 1300 1550 7.0 0.90 1.67
34 Run 33 activated 1550 1650 1700 7.2 0.90 1.84
leach, 200 ml HC1 per 2500 cc OSF 400, hot --- ---
35 above carbon baked 800 1300 1550 7.0 0,90 1.29
36 Run 35 activated 1575 1675 1700 8.3 1.00 2.25
leach 100 ml HCL + 50 g NaOH/2500 cc
OSF 400 --- ---
45 above carbon baked 800 1350 1600 7.0 0.90 1.78
47 Run 45 activated 1550 1650 1700 5.8 0.77 1.70
leach, 150 ml HC1 per 2500 cc OSF 400, hot ---
49 above carbon baked 810 1350 1600 7.0 0.93 1.71
53 Run 49 activated 1550 1650 1700 3.12 0.41 0.93
leach same as for Run 45
63 above carbon baked 1700 1700 1600 3.7 0.00 0.00
69 Run 63 activated 1700 1700 1600 9.2 1.10 2.10
65 OSF 400 baked 1700 1700 1700 3.7 0.00 0.00
Run 65 HC1 leached
70 above activated 1700 1700 1600 9.2 1.10 2.69
leach same as for Run 49
78 above carbon baked 800 1350 1550 7.0 0.93 1.96
79 Run 78 activated 1550 1650 1700 6.0 0.80 1.94
-------
TABLE XXVI - REGENERATION RESULTS, HC1 LEACH
Ul
Run no.
virgin
spent
1 eached
33
34
1eached
35
36
leached
45
47
1 eached
49
53
1eached
63
69
65
1eached
70
leached
78
79
Bulk Uei
Bulk Particle
ght volume volume Pore
density» decrease, decrease, decrease, -volume,
g/cc
0.468
0.583
0.557 4.2'
0.508 10.2
0.478 7.7.
0.560 3.8'
0.510 10.7
0.484 6.9.
0.563 3.7"
0.500 13.7
0.468 7.7.
0.555 5.2'
0.495 12.5
0.472 5.6.
0.544 6.01
0.506 9.5
0.472 7.8.
0.517 13.1 "
0.507 2.1
0.475 7.7.
0.551 4.6"
% *, % cc/cc
0.650
0.500
0.1 1 0.0'
- 20.6 1 .7 !- 3.7 0.5
1.8J 1.3.
O.Ol O.l"
- 20.2 2.0 r 3.1 0.5
1.1 J 3.3.
O.O"1 0.0 "
• 23.4 2.7
1.1.
0.4 '
- 21.8 2.1
0.7.
0.2 1
- 21 .6 2.4
0.7 j
2.5'
-21.5 0.0
1.3 .
0.0 '
0.504 10.9 i- 20.0 1 .9
0.480 5.8
0.7 .
-3.8 2.8
1.8 .
0.3 '
-3.3 1.9
1.7 .
•
-3.3
1 .9 .
2.0 '
-3.8 1.0
2.1 j
0.527
-1.8 0.604
0.627
0.524
-2.7 0.611
0.639
- - -
-4.6 0.612
0.638
0.527
-3.9 0.629
0.645
_ _ _
-4.3 0.614
0.642
0.614
-5.0 0.615
0.642
1.7 1 0.519
-2.6 -0.4 h3.3
2.0J 0.632
lodi ne
number ,
mg/g
1090
630
660
910
1020
_ _ —
930
1040
...
940
1050
- - -
970
1020
- - -
900
990
820
970
— — —
1000
1040
Mol asses
number
250
190
219
213
234
_ - -
210
215
_ _ _
240
230
_ - _
210
250
_ _ _
230
210
210
220
.. _ _
230
230
-------
TABLE XXVII - ASH CONTENT OF REGENERATED CARBONS
BEFORE AND AFTER. DILUTE HC1 ACID LEACH
Ash after
Ash after regeneration
Run no. Buik density, g/cc regeneration, & dil HC1 leach,
(regenerated) spent regenerated % %
Virgin F 400 --- (0.469) (5.7) (5.1)
56 M) 0.615 0.488 8.4 4.5
63 (1) 0.616 0.445 9.2 4.8
65 0.574 0.486 7.5 5.8
27 0.546 0.464 7.9 5.3
32 0.582 0.467 7.3 4.4
(1) Runs from contract 14-12-107
TABLE XXVIII- DECOLORIZING TEST RESULTS ON HC1
LEACHED AND NONLEACHED CARBONS
Iodine number, mg/g ;1 glasses number
Run no. no leach HC1 leach no leach HC1 leach
Virgin F 400 1090 1060 240 250
56 880 790 290 280
63 850 840 320 320
65 960 920 220 230
27 1040 1040 310 280
32 950 940 --- 270
66
-------
HC1 Leach Pretreatment. Effect of Temperature - The success
of the HC1 acid pretreatment in improving the thermal regen-
eration prompted further study to optimize the pretreatment
conditions. A rapid preliminary study was made, according
to the procedure used previously to determine whether an ef-
fective leaching could be performed at ambient temperature.
The leaching was done by passing 500 ml HC1 solution through
500 ml of OSF 400 in a Buchner funnel. The acid filtrate was
recycled twice and discarded. Pure water was then passed
through the bed until the filtrate was essentially neutral
and free of chloride ion. About 2 liters of water was used
in each case to leach out the acid.
The results of this study are presented graphically in
Figure 9. Leaching was done at near boiling point and at
ambient temperature. A good measure of contact time is not
obtainable with this procedure, hence, the data obtained is
not suitable for scale-up purposes. Ash analysis, by ig-
nition was made on the original OSF 400 and each leached
batch. In order to put the ash analyses on a comparable
basis, the measured ash content was multiplied by the ratio,
0.468/density of spent carbon. This gives the would-be ash
analysis after regeneration to 0.468 g/cc bulk density.
The curves in Figure 9 show that there is not a significant
difference between cold and hot leach and that in the acid
to carbon contact time employed, the ash content cannot
readily be reduced below 4.5%. Since it is considerably
better to operate at ambient temperature, because of reduced
corrosion and avoidance of heating costs in large scale op-
erations, the hot leach studies were discontinued.
HC1 Leach Pretreatment, Effect of Concentration - With know-
ledge obtained from the preliminary studies, another test
procedure was devised so that reliable information could be
obtained on contact time versus amount of acid needed. Five
glass tubes were set up to permit five simultaneous leaching
experiments. Tube dimensions were: I.D. = 4.8 cm, length =
63 cm. Five hundred cc of OSF 400 was placed in each tube,
giving carbon bed depth of 34 to 36 cm. The carbon in each
case was then submerged in water so that the water level
came to the top of the carbon bed. (This would simulate the
condition of a contactor at Pomona, after a backwash, in
case the acid leaching could be dOne in the contactors.)
Five hundred ml acid were then added to each of four tubes
and pure water into the fifth. Acid concentrations used
were: 5, 10, 20 and 30 ml concentrated HC1 diluted to 500 ml.
Flow through each bed was controlled by means of a stopcock
at lower end of tube and, for this set of experiments, flow
was set at 100 ml/hr. This gives a contact time of about
67
-------
oo
4.0
2.0
0
0
E Ambient temp,
0 90-95°C
j_
40 80 120 160 200
ml of cone. HC1/500 cc carbon
FIGURE 9 - ASH CONTENT OF HC1 LEACHED OSF 400
-------
2 hr. After the acid level reached the top of each carbon
bed, water was added to the tube and flow continued at prev-
ious rate. The pH, measured on each 50 ml of filtrate, are
plotted in Figure 10 as function of the filtrate volume.
Elemental analyses were also made on the filtrate of the
20 ml acid experiments. Qualitative chloride tests were
made on each filtrate with silver nitrate.
After the acid leach was completed, carbon samples were
taken from the top, midsection and bottom for ash analyses.
The results of the ash analyses are given in Table XXIX.
Also, in calculating these results, the measured ash con-
tent was multiplied by the ratio, 0.468/bulk density of
spent carbon, to put the results on a comparable basis.
On analyzing the data obtained thus far, several trends were
noted. A part of metallic ions comes off very easily. Pure
water passed through the bed, reduced the ash content from
8.2 to 6.8%. The pH rose from about 5.0 to 6.0 at the same
ti me.
When 5 ml of acid were added, the ash content reduced to
5.1% at the top part of the bed but was still 6.2% at the
bottom. In this experiment, also the pH rose from 5.0 to
6.0 indicating that no free acid penetrated the bed. The
filtrate contained Ca and Na ions, but no other metal ions,
in concentrations given in Table XXX. These facts signify
that the 5 ml of acid was completely consumed within the
bed. For the 500 cc carbon (235 g regenerated weight), 5 ml
of acid is definitely too small to give uniform removal of
the metallic elements.
When 10 ml or more of acid was added, the ash content
dropped to the 4.8% level with variation between 4.5 to
5.4%. This variation appears to be experimental error since
no logical trend is indicated. The pH of the filtrates
passed through a minimum, below 1.0, indicating penetration
of acid.
Elemental analyses were made on the filtrate of Experiment 1
(20 ml acid) at several filtrate volumes. The results of
the analyses are given in Table XXXI. The %2®3 1S mainly
TABLE XXXI - ANALYSIS OF FILTRATE FROM EXPERIMENT 1
Filtrate Element content of filtrate, mg/ml
volume, ml Na £a_ E.2^-3
700 0.055 1.98 3.78
800 0.045 1.77 2.18
1000 0.020 trace 0.46
2000 °-11 m9/ml
total solids
69
-------
3.0
2.0
1.0 —
0.0
1000
2000
Filtrate volume, ml
3000
4000
FIGURE 10 - pH OF FILTRATE FROM HC1 LEACH OF OSF 400 WHEN ACID TO
CARBON CONTACT TIME IS 2 HR
-------
TABLE XXIX - ASH CONTENT OF HC1 LEACHED QSF 400
Experiment
no.
5
3
2
1
4
Acid used,
Ash content, %
ml
0
5
10
20
30
top
6.8
5.1
4.6
4.7
4.8
middle
4.9
4.7
5.1
4.5
bottom
r
6.8
6.2
5.4
4.9
5.0
Initial ash content of unleached OSF 400 = 8.2%
Ash content of virgin F 400 = 5.7%
TABLE XXX - ANALYSIS OF FILTRATE FROM EXPERIMENT 3
Filtrate
volume, ml
Element content of filtrate, mg/ml
Residue on
Na. C_§_ C_L evap. to dryness
500
1000
0.070 0.225
0.010 0.170 0.36
2.14
1.32
71
-------
Fe with smaller amounts of Cr, Mn and Al. If the mean_
atomic weight of R is assumed to be 56 g/mol, that of iron,
then the metallic elemental composition of leached ash is
approximately 1% Na, 49% Ca and 50% R. It comes out that
one percent of ash reduction is chemically equivalent to
2.6 g HC1. Pure water reduces the ash content from 8.2% to
6.8%, while 10 ml of acid reduces it down to about 4.8%
hence, the water aids the acid considerably. If it is as-
sumed the acid contributes to the ash reduction from 6.8 to
4.8%, i.e., 2.0%, then the amount of HC1 required is 5.2 g.
Ten ml of concentrated HC1 contains 4.5 g HC1, hence over
10 ml are required for the reaction.
HC1 Leach Pretreatment. Effect of Contact Time - A series of
leaching experiments were conducted at nominal contact times
of 1, 2 and 4 hr. Since the contact time varies with void
volume in the carbon bed and the void volume in turn depends
on the manner the carbon was loaded into the contactor, the
actual contact time can vary for a given carbon loading.
Void volume can vary from about 40% to 55%, hence a contact
time of 2 hr calculated for 40% void could in reality be
2.9 hr for an expanded bed of 55% void. To avoid this error,
the contact times are reported in terms of liquid flow rate
per unit weight of carbon, thus:
ml liquid/ gal. liquid/ Contact time
(g carbon x hr) (Ib carbon x hr) range, hr
0.850 0.102 1 to 1.4
0.426 0.051 2 to 2.7
0.213 0.025 4 to 5.5
Progress of the leaching experiments was measured by deter-
mining pH of the filtrate and chloride ion concentration with
silver nitrate. When chloride ion concentration became very
low, the water leach phase of the treatment was then stopped
and carbon then dried and analyzed.
Figures 11 and 12 show the effluent pH curves at the various
contact times for 5 and 20 ml acid dosages. Figure 11 shows
that at the 5 ml acid dosage, the chloride ion does not pen-
etrate to any extent as a free acid even at the shortest
contact time. Most of the acid is converted to salts. At
the 10, 20 and 30 ml acid dosages, free acid penetrated the
carbon bed, indicating that more than the stoichiometric
amount of acid was added. Curves of Figure 12 also show no
definite trends regarding contact time, which indicates that
the chemical reaction and acid diffusion are relatively fast.
72
-------
7.0 —
6.0
5.0
= 4.0
OL
3.0
2.0
1.0
0.0
0
0.426 ml/(g x hr)
1000
2000 3000
Filtrate volume, ml
4000
4500
FIGURE 11 - pH OF FILTRATE FROM HC1 LEACH WHEN 5 ml OF ACID
IS ADDED TO 500 cc OSF 400
-------
0.852 ml/(g x hr)
1000
2000 3000
Filtrate volume, ml
4000 4500
FIGURE 12 -
pH OF FILTRATE FROM HC1 LEACH WHEN 20 ml OF ACID
IS ADDED TO 500 cc OSF 400
-------
On the 0.426 experiment of Figure 12, the carbon was ana-
lyzed for chlorine after the acid treatment. The amount
of chlorine found was 0.2% as compared to 0.04% of the
virgin Filtrasorb 400. Although the filtrate was free of
chloride ion, according to the silver nitrate test, some
chloride still remained in the carbon. At present, it is
not known whether the chloride will be liberated from the
carbon during regeneration. Ash analyses of the HC1 treated
carbons have shown only traces of chloride.
Table XXXII presents ash analysis data on the leached car-
bons ^sampled from three positions in the carbon bed. Figure
13 gives the ash analyses for the lower end of the carbon bed.
The results indicate that the 10 ml dosage is not sufficient
to bring the ash content down to the 4.8% level, for the
235 g carbon being leached. The optimum amount of acid may
be 14 to 15 ml for a 0.426 ml/(g x hr) flow rate, as indi-
cated by the broken line curve.
HC1 Leach Pretreatment, Effect on Ash Content - There is a
logical reason why the acid leach reduces the ash content to
a lower limit of about 4.8% regardless of amount of acid
used and contact time. Pittsburgh Carbon Company has found
that HC1 leach reduces the ash content of Filtrasorb 400
type carbon by about 1.0%, i.e., for initial ash content of
5.7% the ash content after leach is about 4.7%.
The elemental analysis of ash from Filtrasorb 400 type car-
bon, as determined by Pittsburgh Carbon Company, is approxi-
mately as follows:
Component %_
SiOo 41.4 85 4
A1263 44.0 Bb'4
Fe203 7.4
CaO 2.9
MgO 2.4 14.6
KoO 1.0
Na20 0.7
99.8
In the analysis the two relatively insoluble components, Si02
and A1203, constitute 85.4% of the ash. The others are quite
soluble in HC1. By taking the ratio of 85.4/99.8 times 5.7%
we get 4.9% as the nonsoluble portion of the ash on the car-
bon. This means that the acid leach has removed all the in-
organic compounds retained by the carbon plus 15% of the ash
initially on the carbon and leaving essentially an Si02-Al203
75
-------
TABLE XXXII - ASH ANALYSES OF HC1 LEACHED CARBONS
Experiment
no.
as received
5
• _ _
3
6
8
2
10
11
1
13
7
4
12
9(Run 79)
14(F-400)
15(F-400)
Acid ,
ml
0
5
5
5
10
10
10
20
20
20
30
30
30
0
15
30
Flow rate >
ml/hr
100
50
100
200
50
100
200
50
100
200
50
100
200
100
200
100
ml/(g x hr)
0.425
0.213
0.425
0.850
0.213
0.425
0.850
0.213
0.425
0.850
0.213
0.425
0.850
0.425
0.850
0.425
Ash
top
8.2
6.8
. _ _
5.0
4.9
4.7
4.7
4.7
4.8
4.7
4.7
5.0
4.9
4.8
•. « _
5.0
content
center
8.2
—
...
4.9
...
4.7
4.9
5.1
5.0
...
...
— _ _
4.9
, %
bottom
8.2
6.8
— — -•
6.2
6.4
5.2
5.4
5.5
4.6
4.8
5.0
4.6
4.8
4.7
4.6
4.9
4.5
76
-------
7.0
c
Ol
4->
c
o
o
to
6.0
O 0.852 ml/(g x \\r)
0 0.426 ml/(g x hr)
A 0.213 ml/(g x hr)
10 15 20
ml concentrated HC1
25
30
FIGURE 13 - ASH CONTENT OF CARBON AT LOWER END OF BED
WHEN BED IS TREATED WITH DIFFERENT QUANTITIES OF ACID
77
-------
residue. To remove these compounds requires a caustic leach
in addition to the HC1 leach. This investigation was started
in Task 2B but since this study progressed only through the
first cycle, no definite conclusion could be drawn.
A more complete study was made by determining the ash content
for Run 67, a nonpretreated regeneration run, and Run 79, an
acid pretreated run. The analyses are given in Table XXXIII
with those of a virgin Filtrasorb 400. Run 67 shows a con-
siderable increase in the oxides, Fe203, CaO, MgO, H20, Na20
and Cr203. As is evident from Run 79, these oxides are con-
siderably reduced by the HCl acid leach, hence one or more
of these oxides contribute to the lowered recovery of iodine
number and over recovery of the molasses number.
HCl Leach Pretreatment. Effpr.t on Pore Structure - In pore
structure studies performed at Pittsburgh Carbon Company1^>ij
it was found that the iodine number was proportional to the
surface area of pores larger than 10A in diameter and the
molasses number was porportional to the surface area of pores
larger than 28A in diameter. Equations 1 and 2 below show
these relationships.
I2 No. = 17 + 1.07 x (s.a. of pores >loA) (1)
Molasses No. = 129 + (s.a. of pores />28A) (2)
With reference to the iodine and molasses numbers given in
previous experimental results, these equations indicate that
acid pretreatment minimizes the decrease in surface area of
the smaller pores and also prevents increase in surface area
of pores larger than 28A diameter. Ash build up, as occurs
on successive regenerations without acid leach, accelerates
these changes.
A further study was conducted to determine the manner in
which the surface area changes occurred. Pore size distri-
bution curves were determined for selected carbons, using the
water adsorption method^ and mercury porosimetry. In pre-
paration for the water isotherm determinations (and also por-
osimetry), the carbons were HCl acid and pure water leached
to remove hydrophilic compounds from the carbon surface.
The validity of the water adsorption method depends on the
water being adsorbed by capillary condensation, with negli-
gible monolayer adsorption. The adsorption method is ap-
plicable to maximum pore diameter of 500A to 1000A\
Pores in the larger diameter range are measured by mercury
porosimetry5. These measurements cover the diameter range
30A to lOO.OOOA, hence the two methods overlap in the range
78
-------
TABLE XXXIII - ASH COMPOSITION OF ACID PRETREATED AND
NONPRETREATED REGENERATED CARBONS
Component
Si02
A1203
Fe203
CaO
MgO
K20
Na20
Ti02
Cr203
Total ash
in carbon
virgin
F 400
2.36
2.51
0.42
0.17
0.14
0.06
0.04
0.03
0.01
5.7
Ash composition, '
nonpretreated
Run 67
2.27
1.94
0.80
1.23
0.34
0.40
0.15
0.12
7.2
K
pretreated
Run 79
2.24
1.62
0.59
0.05
0.09
0.20
0.07
0.12
5.0
79
-------
30& to 500JL For some of the pore size distributions, the
agreement in the overlap range is good while for others,
some discrepancy exists. The trend is for the mercury por-
osimeter to measure a larger pore diameter at a given volume.
Figure 14 presents two distribution curves, Run 68 showing
the best agreement and F 400 the poorest. A possible explan-
ation for the discrepancy is that the carbons are slightly
compressed by the mercury at the higher pressures. Maximum
pressure at 30A is 60,000 lb/in2.
Figures 15 and 16 present the complete pore size distribution
curves of the selected samples. The overlap portion as
measured by mercury porosimetry was left out in each case.
Those in Figure 15 show the effects of acid leach, steam ac-
tivation and C02 activation. The virgin F 400 is uppermost
in pore volume at the 28A* diameter, and then Runs 79, 67 and
68 in descending order. The iodine numbers were respectively,
1090, 1040, 940 and 840 mg/g, while the molasses numbers were
quite close to each other.
Those in Figure 16 also show effect of acid leach; Run 36 be-
ing uppermost was pretreated, while Runs 46 and 67 were not.
Runs 46 and 67 are presented together because they have iodine
numbers close to each other, i.e., 920 and 940, respectively,
but have greatly different molasses numbers, 320 and 260, re-
spectively.
The relationship between the decolorizing tests and pore
structure is not readily apparent by visual inspection of
the pore size distribution curve, but comes more discernable
when the cumulative surface areas at different pore sizes
are compared. The cumulative surface area can be calculated
with the equation
AA =
4 AV
where AA is an increment of
increment of pore volume AV
summing up AA over the pore
pores larger or smaller than
calculated.
in Figure 17
This has been done
and for carbons of
surface area associated with
with mean diameter D. By
volume, the surface area of
any specified D can then be
for the carbons of Figure
Figure 16 in Figure 18.
15
From Figure 17, the surface areas of pores larger than 10A
in diameter for the carbons F 400, Runs 79, 67 and 68 are
965, 925, 860 and 805 m2/g, respectively. The surface areas
of pores larger than 28A* are respectively 120, 100, 110 and
100 m2/g. Likewise from Figure 18, the surface areas of
pores larger than 10A" in diameter for the carbons Runs 36, 67
80
-------
1 .00
00
5000 10,000
100,000
Pore diameter, A
FIGURE 14 - PORE SIZE DISTRIBUTION MEASURED BY WATER ADSORPTION
AND BY MERCURY PENETRATION
-------
1 .00
00
0.80
cr.
o
° 0.60
*>
-------
1.00
oo
0.80 —
CT>
0.60 —
O)
QJ
S-
o
O-
0.40 —
0.20 —
10 20 30 50 100
FIGUPsE 16 - PORE SIZE DISTRIBUTIONS OF REGENERATED CARBONS
500 1000 0 5000 10,000
Pore diameter, A
100,000
-------
I 1 1
1000
00
CD
Ol
J_
O!
-------
1000
00
Ln
800
en
CVJ
to
o>
03
-------
and 46 are 915, 860 and 840 m£/g, respectively. Surface
areas of pores larger than 28A in diameter are respectively
85, 110 and 150 m2/g. When these surface areas are sub-
stituted into equations 1 and 2, the calculated decolorizing
numbers agree with test results within * 5% in the iodine
numbers and within 13% in the molasses numbers. These re-
sults are given in Table XXXIV.
TABLE XXXIV - IODINE AND MOLASSES NUMBERS AS
DETERMINED BY TEST AND CALCULATED FROM SURFACE AREA
Carbon
Filtrasorb 400
Run 79 (HC1)
Run 67
Run 68 (C02)
Run 46
Run 36 (HC1)
(Run 58 [Darco]
Iodine number, mg/g
test
1090
1040
940
840
920
1040
580
calc
1040
1010
940
880
920
1000
620
Molasses number
test
250
230
260
250
320
230
370
calc
250
230
240
230
280
210
420
Further study of the surface area curves also indicates where
the major portion of the adsorption may be occurring. By
inspection of the curves in Figure 17, it is observable that
mo§t of the change in core structure occurs in the pores from
14A to 28A. Beyond 28A, there can be considerable change in
pore volumes, as is apparent from inspection of curves in
Figures 15 and 16. The surface area of these larger pores
is, however, too small to be effective. The same is true
for Runs 67 and 38, of Figure 18, but Run 46 is an exception.
For this carbon considerable change occurs at the 28A diameter
region.
To show the change in pore structure, the difference in sur-
face area between that of F 400 and each of the other car;
bons has been calculated at pore diameters from 10A to 50A.
These calcualted areas are given in Table XXXV. For Runs 79,
67, 68 and 36, the surface area changes, as already noted,
occur mostly in the 14A to 28A diameter range. The result;
indicate a shift in the pore sizes. Pore volume in the I
to 28A range decreases while the volume in the 14A to 20A in-
creases. It appears as though the pores that were originally
ISA to 22A diameter have decreased to 14A to 18A diameter.
The net effect, however, is that the cumulative surface area
of pores larger than 10A diameter decrease on regeneration.
Further study is desirable to clarify the exact changes that
do really occur.
86
-------
TABLE XXXV - SURFACE AREA CHANGE AT DIFFERENT
PORE DIAMETERS DURING REGENERATION
Surface area di
F
h
F
K
F
400
400
400
400
400
- 79
- 67
- 68
- 46
- 36
(HC1)
(C02)
(HC1)
10A
40
105
160
130
30
12)1
35
110
160
115
40
14ft
30
105
155
90
40
16A
80
160
165
155
79
fference,
ISA*
200
270
370
170
195
20fl
150
215
130
115
150
m2/q
28A
20
10
20
-30
30
soft
16
00
00
-10
15
The results definitely do show that HC1 pretreatment mini-
mizes pore structure changes. The results also suggest that
the regeneration action takes place mostly in pores of 18A
to 28A diameter. These pores must then also be the most ac-
tive in the adsorption process.
HC1 Leach Pretreatment. Standardized Procedure - Based on the
parameter studies, a standardized procedure was adopted for
treatment of carbons as for Task 2A, 2B and 7. The procedure
was as follows carried out in the unit shown in Figure 3.
Carbon Bed - 2500 cc of dried spent carbon are placed
in the contactor.
Wetting of Carbon Bed - Pure water is run upflow until
water level is at top of carbon bed.
HC1 Acid - 75 ml of 37.9% HC1 acid is diluted to 2500 ml
and passed downflow through carbon bed at 500 ml/hr
flow rate.
Water Leach - After HC1 solution has passed through
carbon bed, 16 1 of pure water are passed downflow
through carbon bed at 500 ml/hr flow rate.
Drying of Carbon - Treated carbon is removed from
contactor and dNed in air convection oven at 150°C
for about 48 hr. The carbon is ready for thermal
regenerati on.
Based on the preceding laboratory data, the scaled-up con-
ditions for acid leaching of carbon bed of size in the
Pomona contactor are given below.
87
-------
Tamperature - Ambient
Carbon volume - 269 ft3 (6 ft dia by 9.5 ft
depth)
Carbon weight - 6650 Ib
HC1 acid per unit carbon - 0.076 Ib acid/lb carbon
(37.9% assay)
Total HC1 acid - 505 Ib (0.076 x 6650)
Flow rate per unit - 0.426 lb/(lb x hr)
carbon weight - 0.051 gal/(lb x hr)
Flow rate of acid
solution and water - 340 gal/hr (0.051 x 6650)
Leach acid solution
volume - 1700 gal (340 x [500 x 100])
Leach water volume - 10,200 gal
pH of filtrate at end of
leaching operation - 3.0 to 3.5
Chloride ion concentration
in fi1trate at end of
leaching operation - nil
Ash content of leached
spent carbon
"- 4.1% based on spent carbon
density of 0.548 g/cc
- 4.8% based on reg. carbon
density of 0.468 g/cc
Overall leaching time - 35 hr
Acid-carbon contact time - 2 to 3 hr
To perform the acid leach, the contactor is given its final
back-wash. Then the water level in the bed is lowered so
that it is level with the carbon surface. Acid solution is
then pumped through the bed at rate and quantity stated
above, followed by the water leach. From this point on the
carbon is handled in the usual manner.
At the 0.076 Ib acid/lb carbon rate, the acid cost is about
0.145<£/lb of regenerated carbon when 37.9% assay technical
grade acid sells for 1.95<£/lb.
-------
Task 5 - Engineering Studies on Furnaces
The anticipated need for engineering studies was based on the
premise that the regeneration is controlled by the parameters;
temperature, gas composition, gas input rate and carbon
residence time as investigated in Task 1. At the time the
program was outlined, difficulties were expected in attain-
ing the precise conditions in a direct-fired multiple hearth
furnace and that indirect-heated rotary furnaces may have to
be used to carry out the regeneration in whole or in part.
As the program progressed, it was found out that the main
obstacle in the way of satisfactory regenerations was the
metallic elements. When these elements were leached out with
HC1, it became apparent that the multiple hearth furnace
could perform an adequate regeneration and the need for en-
gineering studies of various other furnaces was no longer
necessary.
Task 6 - Regeneration Control by Effluent Gas Analysis
The analysis of effluent gas from the activating step ap-
peared to offer an instant monitoring method for the ac-
tivation. When the baking and activating are run in suc-
cession as in the multiple hearth furnace, it is essential
that the activating effluent gas in withdrawn from the re-
generator before any appreciable mixing has occurred with
the gas released from the baking step. In the pomona fur-
nace the point of withdrawal would be at the fourth hearth.
In order for the effluent gas analysis to be meaningful,
an analysis of the input gas composition and also a good
control of the input rate are essential.
Vent gas analyses were conducted on 25 activation runs.
However, an answer as to the possibility of this method for
monitoring purposes cannot be given, as yet, because of a
number of unexplained deviations that have appeared in the
results. These deviations, however, appear to be giving
information about the activating process.
The concept of vent gas analysis for activation monitoring
was based on the premise that the carbon entering the activat-
ing step is essentially free of combined oxygen, sulfur and
hydrogen, or at least present in quantities too small to in-
validate the results for monitoring purposes. It was also
assumed that the gas entering the activating step is essen-
tially N2' CC>2» H20 and an insignificant amount of 02- The
activating reactions should then be basically:
89
-------
C02
+ C
+ C
2CO
CO +
H2
(1)
(2)
The input flow rates of N2, C02 and H20 are known and of these
the N2 flow rate remains unchanged on passage through the regen-
erator. Since either reaction seldom goes to completion, the
vent gas is usually a mixture of N2, C02, H/>0, CO and Ho. A
partial stream of the vent gas was passed through a cold trap
at -80°C to remove the water, hence the percentage compositions
of C02, CO and H2 in N2 were analyzed. Since N2 makes up the
balance of the gas and its flow rate is known, it is then pos-
sible to calculate the flow rate of each component.
The gas utilization is then a simple subtraction of gas out-
put rate from the gas input rate, thus:
C02 utilization
(C02 ft3/hr input) — (C02 ft3/hr output) =
(C02 ft3/hr utilization) (3)
H20 utilization
Is equal to H2 ft3/hr output (4)
H20 unreacted
(H20 vapor ft3/hr input) — (H2 ft3/hr output) =
(H20 ft3/hr unreacted) (5)
The C02 utilization can also be calculated from the CO and H2
output rates, thus
(CO ft3/hr output) — (H2 ft3/hr utilization) =
2
(C02 ft3/hr utilization) (6)
All gas measurements were converted to standard temperature
and pressure (stp).
Input and Output Gas
Resul ts_ - For the
summarized
'2
record, the input and
in Appendix A. In
ratio was that of the
output gas compositions are
most activations, the input C02 and N
flue gas mixture. The H20 vapor concentration was varied
considerably relative to the N2-C02 mixture. In Run 67, no
C02 was added, only N2 and H20 vapors. In Run 68, no H20
vapor was added, only N2 and C02. However, for this run the
C02 concentrations were above trie analyzer capacity and,
therefore, no analyses are recorded for this run.
90
-------
^ Phase Equilibria - According to thermodynamics, a gas
mixture consisting of H20, CO and C02 and H« can be expected
to react at the activating temperatures according to the
equation,
H20 + CO < > C02 + H2 (7)
and come to an equilibrium, which varies with temperature.
The equilibrium equation and values of the equilibrium con-
stant K are given below for several temperatures of interest
(C02) (H2)
(H20) (CO) = K
The quantity in each bracket is the mole percent of the gas.
Temperature °F K_
1400 1.36
1500 1.04
1600 0.86
1700 0.73
1800 0.63
If, after reaching equilibrium at the elevated temperature,
the gas mixture is cooled rapidly, it can retain the ele-
vated temperature equilibrium composition. That this happens
to some degree during the activation runs is indicated by
the K values given in the last column of Table XXXVII. They
vary from 0.61 to 1.42, suggesting approach to equilibria at
temperatures from 1400°F to 1800°F. Since the temperature
of zone one in the regenerator was either near 1550° or 1700°F.,
it had been expected that K for the 1550°F runs would stabi-
lize near 1.0 and for the 1700°F runs near 0.7. However, the
temperature of zone one had no bearing on the K value. At
present no explanation can be offered for the observed vari-
ation in values of K.
COp Uti1j zati on - The C02 utilizations as calculated from
the COp input and output rates, Equation 3, were for most
actiyation runs less than when calculated from the CO and H2
output rates according to Equation .6. The CO? utilizations
and the differences are reported in Table XXXVI.
In the attempt to determine the significance of the difference
in C02 utilization, it was found that a low order of correlation
exists between this difference and the percentage particle
volume decrease. The percentage figures for the particle
volume decreases are given in the last column of Table XXXVI
91
-------
TABLE XXXVI - CO? UTILIZATION
C02 utilization,ft3/hr (stp)
ro
Run
32
34 (HC1)
36 (HC1)
39
40
41
47 (HC1)
51
52
53 (HC1)
66
67 (HC1)
69 (HC1)
70
76 (WVP&P)
77
79 (HC1)
88 (HC1)
89 (HC1)
90 (HC1)
98 Task 2
104 Task 2A
(HC1)
105 Task 2B
(HC1)
106 (WVP&P)
(HC1)
108 (WVP&P
(HC1)
Direct
0.22
0.00
-0.13
-0.07
0.14
-0.12
-0.07
0.15
0.20
-0.02
0.11
-0.53
-0.09
-0.10
0.08
0.07
-0.08
-0.01
-0.01
-0.11
0.00
-0.14
-0.06
-0.03
-0.15
CO-H2 anal
0.23
0.02
-0.02
0.02
0.16
-0.01
0.01
0.18
0.30
0.06
0.07
-0.52
0.02
-0.05
0.11
0.03
-0.03
0.03
0.00
-0.04
+ 0.10
-0.03
+0.07
+0.01
-0.05
CO-
Di fference
utilization, ft3/hr
-0.01
-0.02
-0.11
-0.09
0.02
-0.11
-0.08
-0.03
-0.10
-0.08
+ 0.04
-0.01
-0.11
-0.05
-0.03
+0.04
-0.05
-0.04
-0.01
-0.07
-0.10
-0.11
-0.13
-0.04
-0.10
Particle volume
decrease, %
2.6
1 .3
3.3
2.0
1.5
2.2
1.8
1.9
2.7
1.7
0.6
1.2
1.9
2.1
1 .6
0.5
2.0
0.2
1.5
1.3
1.2
2.8
0.7
1.7
2.1
-------
TABLE XXXVII - RESULTS OF GAS ANALYSIS DURING ACTIVATION
Gas utilization, %
Run no.
32
34 (HC1)
36 (HC1)
39
40
41
47 (HC1)
51
52
53 (HC1)
66
67
69 (HC1)
70 (HC1)
76
77
79 (HC1)
88 (HC1)
89 (HC1)
90 (HC1)
98
104 (HC1)
105 (HC1)
106 (HC1)
108 (HC1)
C02
47.0
2.5
-2.0
2.0
25.0
0.0
1.3
43.9
68.2
14.6
12.7
2.0
-4.1
26.8
3.5
-6.0
7.0
1.0
-9.3
58.8
-5.5
12.5
1 .8
-6.3
H20 C02+H20
74.8
40.0
42.7
63.0
74.0
59.0
43.5
73.0
82.5
65.5
69.8
73.0
46.0
35.2
61 .6
42.9
43.3
57.5
47.4
50.4
66.7
55.4
50.6
36.2
33.5
65.9
28.3
28.8
43.6
54.0
41 .6
31.3
62.6
76.2
50.0
54.4
54.0
31 .4
23.0
49.5
31.9
28.8
43.6
34.7
35.1
62.1
39.6
40.7
25.7
23.1
Rate carbon oxi
By gas anal
16.8
11 .4
15.3
20.8
13.0
23.3
11.4
10.9
11 .7
10.1
16.8
23.2
15.9
13.6
9.1
17.0
12.3
7.7
8.2
9.5
2.3
8.9
7.3
8.4
13.0
dation, q/hr
By weighi ng
21.
17.
16.
21 .
16.
26.
15.
14.
16.
11.
20.
27.
17.
17.
13.
24.
13.
12.
11.
13.
9.
13.
12.
11 .
15.
5
0
5
6
3
3
4
5
3
4
0
7
6
0
0
5
5
8
8
4
2
9
7
0
7
Di f f erence
4.
5.
1.
0.
3.
3.
4.
3.
4.
1.
3.
4.
1 .
3.
3.
7.
1 .
4.
3.
3.
6.
5.
5.
2.
2.
7
6
0
8
3
0
0
6
6
3
2
5
7
4
9
5
2
9
6
9
9
0
4
6
7
Weight 0)
decrease ,
g/h
5.
3.
7.
3.
3.
5.
a.
3.
4.
3.
1.
4.
4.
4.
3.
0.
4.
0.
3.
2.
3.
8.
1 .
4.
5.
r
6
7
4
3
2
1
5
8
7
2
3
6
0
2
4
6
9
4
9
8
6
2
5
2
3
K
0.86
0.67
0.93
1.32
1.36
1.14
0.85
0.78
0.87
1.12
0.86
1.42
0.95
0.85
0.79
0.61
0.86
0.60
0.72
1 .00
1 .21
1.15
0.79
0.70
0.74
3.9
(1) Weight decrease calculated from particle volume decrease
-------
and the correlation is shown graphically in Figure 19. No
logical explanation can be offered for this correlation.
Rate of Carbon Oxidation - The rate of carbon oxidation was
calculated from the rates of COo and steam utilizations,
with COo utilization rate calculated according to Equation 6.
The oxidation rates based on gas analysis were then compared
to oxidation rates based on direct weight of carbon before
and after the activating step. It was found that the oxi-
dation rates based on weight measurements were consistently
larger and that the differences were approximately equal to
the weight decreases associated with particle volume decreases.
The results of these calculations are summarized in Table
XXXVII. The mean weight of carbon oxidized, as obtained by
difference, was 3.7 g/hr while the mean weight based on
particle volume decrease was 3.9 g/hr. This fact suggests
that particle volume decreases may be due to formation of sub-
micron size fines which are then carried out of the regen-
erator by the vent gas.
For Runs 36, 70 and 79, the calculated C02 utilizations were
negative, indicating that COo was actually produced rather
than consumed. This is possible if steam input rate is large,
in view of Equation 7 and equilibrium constants K.
Gas Analysis Without Carbon in Regenerator - Because of some
of the apparent discrepancies in the gas analyses, a number
of tests were conducted in which gas mixtures were passed
through the regenerator under activating conditions but with-
out carbon. These tests were primarily a check on the ac-
curacy of the gas concentration measurements.
The results of the measurements are reported in Appendix B,
and Table XXVIII presents an analysis of the results. Ex-
cept for Test 5, these show a self consistency of 0.01 to
0.02 ft^/hr in the consumed and produced gas volumes, hence
the apparent discrepancy observed in the gas analyses of
the activation runs are not due to errors in measurements.
Tests 3, 4 and 5 show formation of C02 when CO is added to
the gas mixture, indicating that with high steam and CO con-
centration, C02 can be formed as per Equation 7.
Table XXXIX presents the K values for these tests. Except
for Test 1, they fall in the range 0.65 to 1.36. Because of
the very small CO concentration in Test 1, the high value
of K = 1.69 is due to lack of instrumental accuracy.
94
-------
VD
Ul
4.0i—
3.0
Ol
to
ro
-------
TABLE XXXVIII - REACTIONS ASSOCIATED WITH
EACH GAS ANALYSIS
Test 1
COp + metal—-metal oxide + CO
C02 consumed 0.92 - 0.93 = -0.01 ft3/hr
CO produced 0.01 - 0.00 = 0.01 ft3/hr
H20 + metal—> metal oxide + H2
H2 produced 0.02 - 0.00 = 0.03 ft3/hr
\\2 consumed = (-.03 ft3/hr)
(from H2 prod)
Test 2
H2 + C02—* CO + HoO
Hp consumed 0.58 - 0.79 = -0.21 ft3/hr
C02 consumed 0.72 - 0.93 = -0.21 ft3/hr
CO produced 0.20 - 0.00 = 0.20 ft3/hr
H20 produced
(from H2 prod)0.20 - 0.00 = (0.20 ft3/hr)
Test 3
CO + H20 —> C02 + Ho
CO consumed 0.49 - 0.90 = -0.41 ft3/hr
H20 consumed = (-0.41 ft3/hr)
(from CO cons)
COo produced 1.35 - 0.93 = 0.42 ft3/hr
H20 produced 0.38 - 0.00 = 0.38 ft3/hr
Test 4
CO + H20 •--> COp + H?
CO consumed 6.23 - 0.50 = -0.27 ft3/hr
H20 consumed = (-0.27 ft3/hr)
(from CO cons)
C02 produced 1.21 - 0.93 = 0.28 ft3/hr
H2 produced 0.25 - 0.00 = 0.25 ft3/hr
Test 5
CO + HoO — > C02 + H2
CO consumed 6.10 - 0.20 = -0.10 ft3/hr
H20 consumed = (-0.10 ft3/hr)
(from CO cons)
C02 produced 1.00 - 0.93 = 0.07 ft3/hr
H2 produced 0.17 - 0.00 = 0.17 ft3/hr
96
-------
TABLE XXXIX - K FOR TESTS 1 THROUGH 5
Test K
1 1.69
2 0.98
3 0.72
4 0.97
5 1.06
Task 7 - Regeneration of 25 Ib Quantities of Spent Carbon
The object of this task was to carry out larger quantity re-
generations after the optimum operating conditions had been
worked out in Task 1. Regenerations of Darco and WVP&P were
to be carried out. Darco is a 14 x 30 mesh granular carbon
manufactured by Atlas Powder Company and the WVP&P is an
8 x 30 mesh granular carbon manufactured by West Virginia
Pulp and Paper Company.
As has already been learned, the work of Task 1 during the
first 12 months of the program did not yield optimum con-
ditions which would reproduce the initial properties in the
regenerated carbon, hence confirming runs of larger duration
could not be carried out during that period.
Exploratory regeneration runs were, however, carried out
with the WVP&P and Darco carbons. The Darco regeneration
followed a drastically different pattern from that of Fil-
trasorb 400; it regenerated to the virgin bulk density or
lower during the baking step, although the iodine number
did not always come up to the initial value- No activating
step appeared necessary, but variations were tried in baking
condi ti ons.
The regeneration conditions and results for these two car-
bons are given in Tables XL and XLI.
For the Darco carbon, in Run 58, a gas mixture approximating
a flue gas was used as sweep gas and carbon residence time
was 34 min. This brought the bulk density below the virgin
carbon bulk density and the iodine number came up to the
level of the virgin carbon. On Run 59, the residence time
was shortened to 21 min in an attempt to raise the bulk
density. In Run 60, N2 sweep gas was used and the baking
temperature was raised. In this case, the bulk density
matched the virgin carbon density but the iodine number was
now below that of the virgin carbon. Run 75 is similar to
Run 59 but with lower temperature; the results were not sat-
isfactory. Run 73 was HC1 pretreated and produced a regen-
97
-------
TABLE XL - REGENERATION CONDITIONS. WVP&P AND DARCO
VO
00
Run no .
1
1
1
_
1
1
58
59
60
. _
73
75
74
76
99
_ _ _
00
01
06
_ _
07
08
Regenerati ng
step
Darco, baked
Darco, baked
Darco, baked
spent Darco HC1
1 eached
above, baked
Darco, baked
WVP&P, baked
Run 74 act
spent Darco HC1
thermal reg
Large Darco
spent Darco HC1
thermal reg
spent WVP&P, HC1
above baked
Run 101 act
Large WVP&P
spent WVP&P, HC1
above baked
Run 107 act
Tempera
1
850
850
170C
800
850
800
1550
850
run -
880
880
1550
run -
880
1550
1
1
1
-
1
1
1
1
1
10
-
1
1
1
15
-
1
1
ture ,
2
350
350
700
--
350
250
350
650
360
.9 Ib
--
400
350
650
.5 Ib
--
350
650
°F
1
1
1
-
1
1
1
1
1
Gas input, ft3/hr (stp)
3
550
550
600
--
550
450
550
700
600
feed, 21.7
-
1
1
1
--
600
550
700
feed, 48.3
(i
-
1
1
ncl udes
—
550
700
N2
7.0
7.0
3.7
7.0
7.0
7.0
3.3
7.0
hr
7.0
7.0
4.0
hr
baki
2.0
7.6
co2
0.90
0.90
none
—
0.93
0.93
0.93
0.41
...
0.93
thermal
0.93
0.93
0.67
thermal
H?0
1.49
1.71
none
—
1.50
2.05
1 .96
0.78
...
2.14
regenerati
1 .76
1.67
1 .49
regenerati
Residence
time, min
34
21
35
38
34
30
16
...
27
on time
34
26
9
on
ng and activating)
0.93
1 .00
1.74
2.68
-------
TABLE XLI - REGENERATION RESULTS, WVP&P AND DARCO
v£>
Run no.
spent Darco
spent WVP&P
virgin Darco
virgin WVP&P
58
59
60
Darco HC1
leach
73
75
74
76
spent Darco
HC1 leach
99
spent Darco
HC1 leach
100
spent WVP&P,
lot 1
HC1 leached
101
106
spent WVP&P,
lots 2 to 6
HC1 leached
107
108
Bulk
densi ty ,
g/cc
0.450
0.597
0.408
0.504
0.380
0.390
0.408
0.433
0.387
0.402
0.531
0.509
0.502
0.472
0.419
0,496
0.471
0.415
0.582
0.568
0.535
0.517
0.582
0.568
0.531
0.512
Weight
decrease ,
of
...
17.2
16.0
13.1
H^} 18'°
13.7
12.41
5.5-f
6 . 3 1 on 3
14.9J
l^}20"2
_ _ _
2.3]
7.8k 14.1
4.6J
2.81
9.5h 17.5
6.lJ
Bulk
vol ume
decrease ,
%
...
—
3.2
3.2
4.2
to}3'9
2.0
1.3l
o.ej
0 . 4J_ A n
3.6J
_ - _
0.21.4.4
4.2f
0.0"!
2.2H 2.9
0.7J
0.2]
3. IF 4.7
1.4J
Particle
volume
decrease ,
%
...
---
3.7
7.0
6.8
4-1"L 8.0
6.0
2 • 9~L 4
i.ej-
0 . 3|_ 2 1
1 .8J
I:?}2-1
-0.3J
2.2h 3.6
1.7J
-0.5]
3.9k 5.5
2.1J
Pore
volume,
cc/cc
0.585
0.522
0.686
0.669
0.664
0.592
0.677
0.662
0.603
0.619
0.562
0.581
0.667
0.562
0.585
0.665
0.554
0.574
0.617
0.628
0.554
0.576
0.618
0.637
Iodine
number ,
mg/g
310
570
1070
580
540
510
630
530
590
970
330
360
660
330
360
700
800
950
1070
805
1120
Mol asses
number
. . .
180
366
330
300
310
320
210
250
260
270
305
250
270
310
230
250
280
-------
erated carbon with a low bulk density but also a high iodine
number. This is a favorable position since it will now be
possible to adjust baking conditions to increase bulk density
while decreasing the iodine number to match those of the
virgin carbon.
After the effectiveness of the HC1 leach procedure had been
established, regeneration of these two carbons was resumed
during the extension period, but on new batches of carbons
received from Pomona. The conditions of regeneration and
results are included in the lower halves of Tables XL and
XLI, Runs 99, 100, 101, 106, and 107 and 108.
About 13.7 Ib of (dried) spent Darco were available for the
extended regenerations which now consisted of (1) an HC1
pretreatment according to the standardized procedure and
(2) thermal regeneration normally designated as the baking
step. Of the 13.7 Ib, 2.8 Ib was used in Run 99 to estab-
lish the regeneration conditions. The remaining 10.9 Ib
were then regenerated in a continuous run over a 22 hr
period.
Other conditions of operation not given in Table XL are:
Feed rate - 0.016 ft3/hr; 450 cc/hr
Tube rotati on -1.1 rpm
Tube inclination - 3.75 in. per 42 in. length
No problems that could not be easily corrected were encount-
ered during the run. The carbon losses, 4.4% by bulk volume
and 2.1% by particle volume decrease, are lower than had been
expected as based on results of the previous runs. Runs 59,
60 and 73. The iodine number is higher than the value re-
ported for the virgin carbon (of lower density). Some of
these differences may be due to different initial properties
of the carbons, since these results are now on a different
batch from Pomona.
The great ease with which the Darco can be regenerated is due
to its open pore structure, i.e., a large portion of the
total pore volume is in pores greater than 30A diameter.
Carbons with a large portion of large pores also tend to be
softer and therefore subject to greater attrition and this
is reflected in the generally larger decrease in particle
volume during the regeneration. For Darco the particle volume
decrease ranged from 3.7% to 8.0% for the first batch, while
on the extended regeneration run of the second batch, the
particle volume decrease was only 2.1%.
100
-------
The regeneration of WVP&P carbon was expected to follow a
pattern very similar to that of Filtrasorb 400 since they
are both activated by a similar method from bituminous coal.
In the spent WVP&P regenerations, a baking and an activating
step are necessary and also the required conditions are very
similar to that for Filtrasorb 400. Run 74-76 of Tables XL
and XLI shows the regeneration results when no acid pretreat-
ment was employed. The iodine number was 970 relative to
1070 mg/g for the reported virgin carbon. Table XVII, Runs
88, 89 and 90, shows the effect of HC1 acid pretreatment.
For these regenerated carbons the iodine numbers were in the
1010 to 1070 mg/g range. On the new batch acquired for the
extended run, Runs 107 and 108, the spent carbon was HC1 pre-
treated as routine procedure. Iodine numbers of 1070 and
1120 were attained. A large excess of steam was used on Run
108 in the attempt to bring the bulk density down to 0.504
g/cc, the density reported for the virgin carbon. The excess
steam did not accomplish this. However, the steam utilization
was low, about 23.5%, and particle volume decrease was ab-
normally large, 5.5%.
Other conditions not given in Table XL for WVP&P runs are:
Feed rate - 0.016 ft3/hr
Tube rotation - 1.2 rpm for bake; 24 rpm for
activation
Tube inclination - 3.75 in. per 42 in. length
Figure 20 shows the various carbons studied in perspective.
The Darco carbons have much lower densities and iodine num-
bers than Filtrasorb 400 and WVP&P carbons. There is some
indication that absorbate pick up may be slightly less for
the Darco since the weight loss of the Darco during regen-
eration ranged from 13% to 20% while the weight loss of
Filtrasorb 400 varied between 18% and 24%. This may not be
an entirely valid comparison since the absorbate concentration
in the waste water varies and the Darco may have been spent
at a lower concentration.
HC1 leach of the spent Darco, prior to baking, increased the
iodine number from 560 to 630 mg/g for the first batch re-
ceived from Pomona. This increase is proportionately the
same as observed with acid leached Filtrasorb carbons. The
iodine numbers of the regenerated second batch are much high-
er.
In the Filtrasorb 400 regenerations, the attempt has been to
regain the 0.469 g/cc bulk density and 1090 mg/g iodine num-
101
-------
1200
1100
1000
900
800
0>
^
01
OJ
T3
O
700
600
500
400
300
200
100
0
— Darco
WVPRP
F11trasorb
400
X
0 00
0 0
O
0
O
El
C02
act.
0
2nd hatch
©
1st hatch
0 • IIC1 leached
HB Monlpached
X V I rgln carhons
Filled In synhols, bake and act.
at 1700" to 1600°F
Open symbols, bake and act. at
1550" to 1700"F
I 1 I I
I I
0.35
0.40 0.45
Bulk density, g/cc
0.50
0.53
FIGURE 20 - IODINE NUMBERS OF REGENERATED DARCO, F 400 AND
WVP&P CARBONS AS FUNCTION OF BULK DENSITIES
102
-------
ber. Densities in the 0.460 to 0.484 g/cc range were at-
tained but the high iodine number of the virgin carbon was
never reached. A straight line through the nonleached car-
bon iodine number points indicates that the virgin carbon
iodine number cannot be attained by extrapolation of den-
sities in either direction. It is felt that the parameters;
temperature, residence time, gas composition and input rate,
both during baking and activating have been adequately cov-
ered. It is beginning to appear as an impossibility to re-
gain the original iodine number simultaneously with all
other initial properties of the carbon when the metallic
elements, brought in with the adsorbate, are in the carbon
during baking and activating. When these were removed by
HC1 leach, the iodine number increased from the 930 to the
1030 mg/g level, which is still below the virgin carbon
iodine number.
103
-------
SECTION VI
ACKNOWLEDGMENTS
This program was sponsored by the Environmental Protection Agency,
Water Quality Office, Washington, D.C.; Project Officer was Mr. Arthur
N. Masse, Chief, Municipal Treatment Research Program, Advanced Waste
Treatment Research Laboratory, Cincinnati, Ohio. Negotiator was Mr.
John H. Blake, Water Quality Office, Washington, D.C.
The cooperative phase of the program at Pomona, was performed
by Messrs John English and Jay Pitkin.
The major part of the investigation was performed by MSA Re-
search Corporation, Evans City, Pennsylvania, where Dr. J.W.
Mausteller, Associate Director of Research, is responsible
for all research activities at the Evans City laboratories.
Mr. Frederick Tepper was Head of Physical Section (to Septem-
ber 1, 1969) in which this program was carried out. The
principal investigator on the program was Dr. A.J. Juhola.
Performance of the direct experimental work was done by Mr.
Edward Krieger.
105
-------
APPENDIX A
INPUT AND OUTPUT GAS COMPOSITIONS OF ACTIVATION RUNS
Run
32
34
36
39
40
Gas
N2
CO
C02
H2
H20
N2
cO
C02
"2
H20
N2
CO
i C02
H2
H20
£•
No
C§
1 C02
H2
H20
c.
No
cS
) C02
t.
H2
£
H20
N2
1 C02
H2
H20
Input
ft3/hr (stp)
4.00
0.51
1.11
5.62
7.16
0.81
1.84
978T
8.30
1.10
2.41
11.81
8.26
1.00
2.14
11.40
5.36
0.64
_ — -
0.95
679T
8.3
1.10
2.63
12.03
% Comp
71.2
9.1
19.7
100.0
73.0
8.2
18.8
100.0
70.3
9.3
20.4
100.0
72.5
8.8
18.8
100.1
77.1
9.2
13.7
100.0
68.9
_ - -
9.1
21.9
99T9
Output
ft3/hr (stp)
4.00
1.36
0.29
0.89
0.22
6776
7.16
0.78
0.81
0.73
1.11
10.59
8.30
0.99
1.23
1 .03
1.38
12.93
8.26
1.39
1.07
1.35
0.79
12.86
5.36
1.03
0.50
0.70
0.25
7.84
8.3
1.53
1.22
1.55
1.08
13.68
% Comp
59.2
20.1
4.3
13.2
3.3
100.1
67.6
7.4
7.6
6.9
10.5
100.0
64.1
7.6
9.5
8.0
10.7
99.9
64.2
10.8
8.3
10.5
6.1
99.9
68.3
13.1
6.4
8.9
3.2
99.9
60.7
11.2
8.9
11.3
7.9
100.0
107
-------
APPENDIX A (Continued)
Input
Run
47
51
52
53
66
67
Gas
*
CO
COo
H2
H20
N-
Cl
C02
H2
H20
N2
CO
COo
Ho
H20
C
COo
H2
H20
?
CO
COo
H2
H20
N
C
CO
H20
ft3/hr (stp)
5.80
0.77
1.70
8.27
3.33
0.41
0.74
4.48
3.24
0.44
0.57
4.25
3.12
0.41
0.93
4.46
4.00
0.55
1.49
6.04
4.00
2.81
6.81
% Comp
70.1
9.3
20.6
100.6
74.3
9.2
16.5
100.0
76.2
10.3
13.4
99.9
70.0
9.2
20.8
100.0
66.2
9.1
24.7
100.0
58.7
41.3
100.0
Output
ft3/hr (stp)
5.80
0.76
0.84
0.74
0.96
9TTO"
3.33
0.90
0.26
0.54
0.20
5.23
3.24
1 .08
0.20
0.47
0.10
5.09
3.12
0.73
0.43
0.61
0.32
5.21
4.00
1.18
0.44
1.04
0.45
7.11
4.00
1.00
0.53
2.05*
0.76
8.34
% Comp
63.7
8.3
9.2
8.1
10.5
9~978
63.7
17.2
5.0
10.3
3.8
100.0
63.7
21.4
3.9
9.2
2.0
100.2
59.9
14.0
8.2
11.7
6.1
99.9
56.3
16.6
6.2
14.6
6.3
100.0
48.0
12.0
6.3
24.6
9.1
100.0
*Estimated from CO and C02 analyses, H2 concentration off
scale on analyses
108
-------
APPENDIX A (Continued)
Input Output
ft3/hr (stp) % Comp ft3/hr (stja) % Comp
9.22 73.3 9.22 67.2
--- --- 1.07 7.8
69 C02 1.10 8.8 1.19 8.7
H2 -" --- 1.03 7.5
H2° 2.24 17.8 1.21 8.8
12.56 99.9 13.72 100.0
No 10.14 72.2 10.14 67.7
CO --- --- 0.84 5.6
70 C02 1.21 8.6 1.31 8.7
H2 --- --- 0.95 6.3
H2° 2.69 19.1 1.74 11.6
14.04 99.9 14.98 99.9
No 3.3 73.5 3.3 64.4
CO --- --- 0.71 13.9
76 C02 0.41 9.1 0.33 6.4
H2 --- --- 0.49 9.6
H20 0.78 17.4 0.29 5.7
' 9 100.0 5TT2 100.0
N2 7.6 68.2 7.6 62.1
CO --- --- 1.16 9-5
77 C02 1.01 9.1 0.94 7.7
H2 --- --- 1.09 8.9
H20 2.53 22.7 1.44 11.8
11.14 100.0 12.23 100.0
No 6.00 68.7 6.00 62.5
CO --- --- 0.78 8.1
79 C02 0.80 9.1 0.88 9.2
H2 --- --- 0.84 8.7
1.94 22.2 1.10 11.5
8.74 100.0 9.60 100.0
No 3.50 69.1 3.50 62.4
CO --- --- 0.54 9.6
88 C02 0.43 8.5 0.44 7.7
WVP&P Ho --- --- 0.48 8.6
HoO 1.13 22.3 0.65 11.6
* 5.06 99.9 576T 99.9
109
-------
Run
89
WVP&P
90
98
Task
2
104
Task
2A
105
Task
2B
Gas
?
CO
C02
H2
H20
CO
C02
H2
H20
N2
CO
C02
H2
H20
C
C02
H2
H20
C
C02
H2
H20
APPENDIX A (Continued)
Input
ft3/hr (stp)
3.50
0.43
1.14
T70T
3.50
0.43
1.25
5.18
1.27
0.17
0.12
1.56
2.19
0.37
1.12
3.68
2.19
0.37
0.81
3.37
% Comp
69.0
8.5
22.5
100.0
67.5
8.3
24.1
99.9
81.4
10.9
7.8
100.1
59.5
10.0
30.4
99.9
65.0
11.0
24.0
100.0
Output
ft^/hr (stp)
3.50
0.55
0.44
0.54
0.60
5.63
3.50
0.55
0.54
0.63
0.62
5784
1.27
0.28
0.17
0.08
0.04
1.84
2.19
0.55
0.51
0.62
0.50
4.37
2.19
0.56
0.43
0.41
0.40
3.99
% Comp
62.2
9.8
7.8
9.6
10.7
100.1
59.9
9.4
9.2
10.8
10.6
99.9
69.0
15.2
9.2
4.3
2.2
99.9
50.1
12.6
11.7
14.2
11.4
100.0
54.9
14.0
10.8
10.3
10.0
100.0
110
-------
Run Gas
106
WVP&P
N2
CO
C02
H2
H20
108 C(
WVP&P C02
H.
APPENDIX A (Continued)
Input
ft3/hr (stp)
4.00
0.67
1.49
67T16
7.60
1.00
2.68
11 .28
% Comp
65.0
10.9
24.2
100.1
67.4
8.9
23.7
100.0
Output
ft-Vhr (stp)
4.00
0.57
0.70
0.54
0.95
6.76
7.60
0.79
1.15
0.90
1.78
12.22
% Comp
59.1
8.4
10.4
8.0
14.0
9979
62.1
6.5
9.4
7.4
14.6
100.0
111
-------
APPENDIX B
GAS ANALYSES WITH NO CARBON IN REGENERATOR
TEST 1 - APPROXIMATE FLUE GAS MIXTURE
THROUGH HEATED REGENERATOR
Ijiput Output
Gas ft3/hr (stp)% Comp ft3/hr (stp) % Comp
No 7.0 73.0 7.0 73.0
CO --- --- 0.01 0.1
C02 0.93 9.7 0.92 9.6
H2 --- --- 0.03 0.3
H?0 1.66 17.3 1.63 17.0
9.59 100.0 9.59 100.0
TEST 2 - APPROXIMATE FLUE GAS MIXTURE THROUGH
HEATED REGENERATOR WITH HYDROGEN ADDED
No 7.0 65.7 7.00
CO --- --- 0.20
COo 0.93 8.7 0.72
Ho 0.79 7.4 0.58
HoO 1.93 18.1 2.14
* 10.65 99.9 10.64
TEST 3 - APPROXIMATE FLUE GAS MIXTURE THROUGH
HEATED REGENERATOR WITH 0.90 ft3/hr CO ADDED
No 7.0 65.7 7.0 65.7
CO 0.90 8.4 0.49 4.6
COo 0.93 8.7 1.35 12.7
H * .-- --- 0.38 3.6
HoO 1.83 17.2 1.45 13.6
2 10.66 100.0 TOT T0672
113
-------
APPENDIX B (Continued)
TEST 4 - APPROXIMATE FLUE GAS MIXTURE THROUGH
HEATED REGENERATOR WITH 0.50 ft3/hr CO ADDED
Input Output
Gas ft3/hr (stp)% Comp ft3/hr (stp) % Comp
N2 7.0 69.7 7.0 69.7
CO 0.50 5.0 0.23 2.3
CO? 0.93 9.3 1.21 12.0
Ho --- --- 0.25 2.5
HoO 1.61 16.0 1.35 13.4
10.04 100.0 10.04 99.9
TEST 5 - APPROXIMATE FLUE GAS MIXTURE THROUGH
HEATED REGENERATOR WITH 0.20 ft3/hr CO ADDED
N2 7.0 70.8 7.0 70.9
CO 0.20 2.0 0.10 1.0
C02 0.93 9.4 1.00 10.1
Hp --- --- 0.17 1.7
H20 1.77 17.9 1.60 16.2
" " 100.1 9.87 99.9
114
-------
REFERENCES
1. Juhola, A.J. and Tepper, F., "Regeneration of Spent Gran-
ular Activated Carbon," Robert A. Taft Water Research
Center Report No. TWRC-7, February, 1969.
2. Juhola, A.J., Matz, W.M., and labor, J.W., paper pre-
sented at the American Chemical Society meeting, Division
of Sugar Chemistry, April 1, 1951, "Adsorptive Properties
of Activated Carbons."
3. Grant, R.J., "Basic Concept of Adsorption on Activated
Carbon", Pittsburgh Activated Carbon Company.
4. Wiig, E.O., and Juhola, A.J., JACS 71., 2069, 2078 (1949).
5. Measurements made by American Instrument Company, Incor-
porated, Silver Springs, Maryland.
115
-------
1
5
Accession Number
2
Subject Field & Group
10 A
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Organization
Evans City, Pennsylvania
Title
OPTIMIZATION OF THE REGENERATION PROCEDURE
FOR GRANULAR ACTIVATED CARBON
1 Q Authors)
Juhola, A, J0
16
21
Project Designation
Program #17020
DAO
Note
Report NOo TWRC-7
Contract No. 14-12-107
22
Citation
Contract Report, 106 pages, 41 tables, 20 figures
23
Descriptors (Starred First)
*Wastewater purification, Tertiary treatment
25
Identifiers (Starred First)
*Thermal regeneration, *Carbon regeneration losses, Carbon regeneration, Steam
activation, C09 activation, Carbon pore structure8 Chemical regeneration, Iodine
number,.Molasses numbera Rotary tube regenerator^Caustic regeneration, Carbonization,
Activation-gas reactions, Cyclic regenerations, HCI treatment, Spent carbon drying
27
Abstract
Spent granular activated carbons from tertiary water treatment, on multiple-
hearth furnace regenerations, suffer a volume loss of 5% to 107, per regeneration. On
the first regeneration, activity loss is as high as 137o but diminishes on subsequent
cycles. Laboratory studies to improve regeneration have demonstrated that on regenera-
tion of wet spent carbon three steps occur; (1) drying at about 220°F, (2) pyrolysis of
adsorbed pollutants at 500° to 1500°F(baking step) and (3) activation with flue gas and
steam at 1600° to 1700°F<, Alkaline and iron oxides accumulate in the carbon and catalyze
Q o
oxidation of pores in the ISA to 28 A diameter range. When metallic elements are leached
from the carbon, prior to regeneration, less carbon and activity loss occurs. Steam
regeneration is more effective than that with CC>2. Carbon volume decreases during
laboratory baking and activation average 2% and I087o, respectively. The apparent
volume decrease during baking is due to pyrolysis of colloidal pollutants on the
particle surfaces; true carbon loss is then Io87>o The latter proceeds as submicron
fines formation from particle surfaces. Regeneration attempts by leaching with
solutions of NaOH, H_0_ and CC1. were ineffective and/or uneconomical. Report contains
41 tables and 20 figures.
Abstractor
Juhola_j_A» Jo
Institution
Mine Safety Appliances
Research Corporation
WR:102 (REV. JULY 1969)
WRSIC
SEND TO: WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERIOR
WASHINGTON, D. C, 20240
* CPO: 1969-359-339
------- |