Predicting the performance of fixed-bed granular - Michigan ...

PREDICTING THE PERFORMANCE OF
FIXED-BED GRANULAR
ACTIVATED CARBON ADSORBERS
D. W. Hand, J. C. Crittenden, D. R. Hokanson, and J. L. Bulloch
Dept. of Civil and Environmental Engineering
Michigan Technological University
Houghton, Michigan 49931
USA
Presented at First IAWQ Specialized Conference on Adsorption in Water
Environment and Treatment Processes; Shirahama, Wakayama Japan
November 1996
Copyright © 1996-2002. Michigan Technological University. All Rights Reserved.
1

OUTLINE
Competitive Interactions
Between Known Components
• Correlation of Single Solute Adsorption Equilibria
• Prediction of Multicomponent Equilibria of Known
Components
• Model Predictions Using the Equilibrium Column
Model
• Model Predictions Using the Pore Surface Diffusion
Model
2

OUTLINE (continued)
Synthetic Organic Compound (SOC) and Unknown
Background Organic Matter (BOM) Interactions
• Impact of BOM on Fixed Bed Adsorption Capacity of
SOCs
– Influence of Exposure Time
– Influence of Compound Type
– Influence of Water Type
• Impact of BOM on Adsorption Rate
• Model Simulations for a Variety of Waters
3

Volume Adsorbed (ml/g)
1
0.1
0.01
0.001
0.0001
ISOTHERM CORRELATION
Data from 26 chemicals
Correlation
Adsorbent: Calgon F-400 GAC
Data Source: Speth and Miltner (1990)
Wi = Woexp[-β(ε i/Vi) σ ]
Wo = 0.24 ml/g
β = 0.0025 (ml/cal) σ
σ = 1.51
0 50 100 150 200 250
Adsorption Potential / Intrinsic Molar Volume (cal/ml)
4

Volume Adsorbed (ml/g)
0.01
0.001
0.0001
MTU DATA FOR SIX COMPONENTS
10
1
0.1
Naphthalene
1,2,4-Trichlorobenzene
Trichloroethylene
Toluene
m-Xylene
Correlation
Adsorbent: Calgon APA GAC
Particle Size: 200 x 400 USTM
Data Source: Bulloch et al. (1995)
Wi = Woexp[-β(ε i/Vi) σ ]
Wo = 1.03 ml/g
β = 0.045 (ml/cal) σ
σ = 1.00
0 20 40 60 80 100 120 140 160
Adsorption Potential / Intrinsic Molar Volume (cal/ml)
5

IDEAL ADSORBED SOLUTION THEORY
C
i
=
g q
i i
N
Â
j=
1
• Wilson Activity Coefficients
q
j
L
M
NM
N
Â
j=
1
nq
nK
i i
ln = 10 . -ln xA-
g i j ij
j=
1
n
Â
j j
( 1-zi
)
i i
g = a
n
Â
k = 1
O
P
QP
L
M
NM
n
Â
j=
1
x A
k ki
xA
j kj
• Empirical Activity Coefficients (less than 1.0)
O
P
QP
6

Solid Phase Concentration
(mMol/g)
10
1
0.1
0.01
IAST CALCULATIONS SHOWING
IMPROVEMENT WITH γ
Adsorbent: Calgon F-400 GAC
Particle Size: 200 x 400 USTM
Temperature: 13 o C
Data Source: Arora (1989)
0.00001 0.0001 0.001 0.01 0.1
Liquid Phase Concentration (mMol/L)
Multicomponent DBCM Data Original DBCM IAST Prediction
DBCM IAST Fit DBCM Single Solute
Multicomponent PCE Data Original PCE IAST Prediction
PCE IAST Fit PCE Single Solute
7

Solid PhaseConcentration
(mMol/g)
0.1
0.01
0.001
IAST CALCULATIONS SHOWING NO
IMPROVEMENT WITH γ
1
Adsorbent: Calgon F-400 GAC
Particle Size: 200 x 400 USTM
Temperature: 13 o C
Data Source: Arora (1989)
0.00001 0.0001 0.001 0.01 0.1
Liquid Phase Concentration (mMol/L)
Multicomponent Chloroform Data Original Chloroform IAST Prediction
Chloroform IAST Fit Chloroform Single Solute
Multicomponent TCE Data Original TCE IAST Prediction
TCE IAST Fit TCE Single Solute
8

Solid-phase Concentration, (mg/g)
ADSORPTION EQUILIBRIUM ISOTHERM
FOR TRICHLOROETHENE
100
10
1
TRICHLOROETHENE ISOTHERMS
PRELOADING TIME IN WEEKS
CALGON F-100 GAC (12x40 mesh)
v = 10 m/hr
Source: Zimmer et al. (1988)
Single Solute (q=3.27 C 0.458 )
34
70
100
2
7
25
50
1 10 100 1000
Liquid-Phase Concentration, (mg/m 3 )
9

Relative Freundlich Parameter K %
120
100
80
60
40
20
0
FREUNDLICH K REDUCTION
FOR VARIOUS WATERS
Houghton GW
(F-400 GAC)
Portage Lake
(F-400 GAC)
Rhine River
(F-100 GAC)
TRICHLOROETHENE
Source: Zimmer et al. (1988)
Hand et al. (1989)
El- Behlil (1990)
Wausau GW
(F-400 GAC)
Karlsruhe GW
(F-100 GAC)
0 10 20 30 40 50 60 70 80
Preloading Time (weeks)
10

C b
Bulk
Solution
r
C s
Linear
Driving
Force
Boundary
Layer
PSDM MECHANISMS
C p,r
Local Equilibrium
Between Fluid Phase
and Adsorbent Phase
R
Dr
q r
r+Dr
Fluid Phase Adsorbent Phase
Mass
Flux
= k fbCb -Csg
Mass
Flux
q
=-Dsra r
-
∂
∂
r
τ p = 1
De∂C t ∂r
r l p
p
pr ,
Pore
Diffusion
Surface
Diffusion
11

MASS TRANSFER COEFFICIENTS
• External Mass Transfer Coefficient (Gnielinski, 1978)
b g �
k f = 1+1.5 1- e fD
2R
2+0.644 Re Sc
1/ 2 1/ 3
12

MASS TRANSFER COEFFICIENTS (continued)
• Intraparticle Surface Diffusion Coefficient
D = SPDFR
s
L
M
N
D C
�e
p
*
0
KC
1/
n
t p 0
r
– Single Components:
• SPDFR between 4 and 8 (mean = 6.58)
– Multiple Components:
• SPDFR = 16.27 EBCT(min) -0.843
a
O
P
Q
13

MASS TRANSFER COEFFICIENTS (continued)
• Intraparticle Pore Diffusion Coefficient
D =
p
e
p
t
D �
– SOCs Alone (maximum pore diffusion flux):
• τp = 1.0
– SOCs in the Presence of BOM:
• τp = 1.0 when Time < 70 days
• τp = 0.334 + 6.61(10-6 ) * t when Time > 70 days
p
14

ADSORPTION DESIGN SOFTWARE
(AdDesignS TM )
• Gas and Liquid phase
• Visual Basic Front-End with FORTRAN DLLs
• Up To 6 Components (12 PDEs), Solved by
Orthogonal Collocation (Up To 18 Axial and 6 Radial
Points, Up To 126 ODEs for Each Component, 756
ODEs total)
• Structured Heuristics Based on Experience for Model
Parameter Estimation
• Data Base for Isotherms and Adsorbents
• SI and English Units
• 15,000 lines of code
15

Liquid Phase Concentration ( µg/L)
2500
2250
2000
1750
1500
1250
1000
750
500
250
0
CHLOROFORM: 6 Component ECM and
PSDM Simulations for Organic Free Water
EBCT =
2.4 min.
EBCT =
9.56 min.
EBCT =
4.9 min.
ECM
EFFLUENT : EBCT = 2.4 min.
EFFLUENT : EBCT = 4.9 min.
EFFLUENT : EBCT = 9.56 min.
INFLUENT : PILOT SCALE
PSDM : EBCT = 2.4 min.
PSDM : EBCT = 4.90 min.
PSDM : EBCT = 9.56 min.
CHLOROFORM F-400 GAC (12 x 40 mesh)
C0,AVG. = 1020.9 mg/L v = 5.12 m/hr T = 13 o C
Data Source: Arora (1989)
0 5 10 15 20 25 30 35 40 45 50
Liters of Water Treated Per Gram of GAC
16

Liquid Phase Concentration ( µg/L)
5000
4500
4000
3500
3000
2500
2000
1500
1000
500
1,2-DBE RESULT: 6 Component ECM and
PSDM Simulations for Organic Free Water
0
ECM CALCULATION
EFFLUENT : EBCT = 2.4 min.
EFFLUENT : EBCT = 4.9 min.
EFFLUENT : EBCT = 9.56 min.
INFLUENT : PILOT SCALE
PSDM : EBCT = 2.40 min.
PSDM : EBCT = 4.90 min.
PSDM : EBCT = 9.56 min.
EBCT =
2.4 min.
EBCT =
4.9 min.
1,2-DIBROMOETHANE
F-400 GAC (12 x 40 mesh)
C0,AVG. = 1576.8 µg/L
v = 5.12 m/hr
T = 13 o C
Data Source: Arora (1989)
EBCT =
9.56 min.
0 5 10 15 20 25 30 35 40 45 50
Liters of Water Treated Per Gram of GAC
17

Liquid Phase Concentration ( µg/L)
TCE RESULT: 6 Component ECM and PSDM
Simulations for Organic Free Water
1500
1350
1200
1050
900
750
600
450
300
150
0
EBCT =
9.56 min.
EBCT =
4.9 min.
EBCT =
2.4 min.
ECM CALCULATION
EFFLUENT : EBCT = 2.4 min.
EFFLUENT : EBCT = 4.9 min.
EFFLUENT : EBCT = 9.56 min.
INFLUENT : PILOT SCALE
PSDM : EBCT = 2.40 min.
PSDM : EBCT = 4.90 min.
PSDM : EBCT = 9.56 min.
TRICHLOROETHENE F-400 GAC (12 x 40 mesh)
C0,AVG. = 1062.6 µg/L v = 5.12 m/hr T = 13 o C
Data Source: Arora (1989)
0 20 40 60 80 100 120 140 160
Liters of Water Treated Per Gram of GAC
18

MODEL VERIFICATION EFFORT
• 11 Case Studies
•
– 9 Pilot Plant Experiments
– 2 Full-Scale Plants
11 Water Sources (USA, Germany, Netherlands)
– 8 Groundwaters
– 3 Surface Waters
• 4 Adsorbents
• 50 Empty Bed Contact Times
• 15 Synthetic Organic Chemicals
19

Liquid-Phase Concentration ( µg/L)
TCE: PSDM PREDICTION - KARLSRUHE
TAP WATER CORRELATIONS
200
180
160
140
120
100
80
60
40
20
0
GROUNDWATER : MANNHEIM, GERMANY
TRICHLOROETHENE (C0 = 90 µg/L)
SURROGATE GROUNDWATER : KARLSRUHE, GERMANY
EBCT : 15.6 min. v = 10 m/hr CALGON F-100 GAC
DATA SOURCE: ZIMMER (1988)
Influent Data
Effluent Data
PDM Model Prediction
0 10 20 30 40 50 60 70 80 90
Liters of Water Treated Per Gram of GAC
20

Liquid-Phase Concentration ( µg/L)
275
250
225
200
175
150
125
100
75
50
25
0
TCE: PSDM PREDICTION - RHINE
RIVER WATER CORRELATIONS
SURFACE WATER : HUDSON RIVER (WATERFORD, NY)
TRICHLOROETHENE (C0 = 121 µg/L)
SURROGATE SURFACE WATER : RHINE RIVER (GERMANY)
EBCT : 16 min. v = 4 m/hr CALGON F-300 GAC
DATA SOURCE: ALBEN ET AL. (1992)
0 10 20 30 40 50 60 70 80 90 100 110
Liters of Water Treated Per Gram of GAC
Influent Data
Effluent Data
PDM Model Prediction
21

1,2-DCP: PSDM PREDICTION - KARLSRUHE
TAP WATER CORRELATIONS
Liquid-Phase Concentration ( µg/L)
1.5
1.35
1.2
1.05
0.9
0.75
0.6
0.45
0.3
0.15
0
GROUNDWATER : NETHERLANDS
1,2-DICHLOROPROPANE (C0 = 0.5 µg/L)
SURROGATE GROUNDWATER : KARLSRUHE, GERMANY
EBCT : 14.3 min. v = 7.23 m/hr CALGON F-400 GAC
DATA SOURCE: KRUITHOFF ET AL. (1989)
Influent Data
Effluent Data
PDM Model Prediction
0 10 20 30 40 50 60 70 80 90 100
Liters of Water Treated Per Gram of GAC
22

BENTAZONE: PSDM PREDICTION - KARLSRUHE
TAP WATER CORRELATIONS
Liquid-Phase Concentration ( µg/L)
2
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
GROUNDWATER : NETHERLANDS
BENTAZONE (C0 = 0.9 µg/L)
SURROGATE GROUNDWATER : KARLSRUHE, GERMANY
EBCT : 10.2 min. v = 9.0 m/hr CALGON F-400 GAC
DATA SOURCE: KRUITHOFF ET AL. (1992)
0 10 20 30 40 50 60 70 80 90 100
Liters of Water Treated Per Gram of GAC
Influent Data
Effluent Data
PDM Model Prediction
23

ATRAZINE: PSDM PREDICTION - KARLSRUHE
TAP WATER CORRELATIONS
Liquid-Phase Concentration ( µg/L)
22
20
18
16
14
12
10
8
6
4
2
0
SURFACE WATER : HUDSON RIVER (WATERFORD, NY)
ATRAZINE (C0 = 8.4 µg/L)
SURROGATE SURFACE WATER : RHINE RIVER (GERMANY)
EBCT : 16 min. v = 4 m/hr CALGON F-300 GAC
DATA SOURCE: ALBEN ET AL. (1992)
0 5 10 15 20 25 30 35 40 45 50 55 60
Liters of Water Treated Per Gram of GAC
Influent Data
Effluent Data
PDM Model Prediction
24

CHLOROFORM: PSDM PREDICTION USING THE
RHINE RIVER WATER FOULING CORRELATION
Liquid Phase Concentration ( µg/L)
2000
1750
1500
1250
1000
750
500
250
0
CHLOROFORM
Portage Lake Water, Houghton, MI USA
F-400 GAC (12 x 40 mesh) Co,AVG. = 1021 µg/L
EBCT = 9.78 min. v = 5.18 m/hr T = 16 o C
Data Source: Arora (1989)
PSDM PREDICTION
INFLUENT DATA
EFFLUENT DATA
0 1 2 3 4 5 6 7 8 9 10 11
Liters of Water Treated per Gram of GAC
25

Liquid Phase Concentration ( µg/L)
TCE: PSDM PREDICTION USING THE RHINE
RIVER WATER FOULING CORRELATION
2000
1750
1500
1250
1000
750
500
250
0
TRICHLOROETHENE
Portage Lake Water, Houghton, MI USA
F-400 GAC (12 x 40 mesh) Co,AVG. = 881 µg/L
EBCT = 9.78 min. v = 5.18 m/hr T = 16 o C
Data Source: Arora (1989)
PSDM PREDICTION
INFLUENT DATA
EFFLUENT DATA
0 5 10 15 20 25 30
Liters of Water Treated per Gram of GAC
26

CONCLUSIONS
• AdDesignSTM interfaces several fixed bed adsorption
models, model parameter estimation methods, and
isotherm and adsorbent data bases.
• AdDesignSTM allows the user to make adsorber
•
performance predictions with greater ease and
archive the results.
The heuristics for determining the most appropriate
models and parameters should be considered work in
progress and the Adsorption Design Software can be
greatly improved as more information is gathered on
the practical application of the models.
27

CONCLUSIONS (continued)
• Single component isotherm data can be correlated
using the intrinsic molar volume and the Polanyi
Potential Theory
• Multicomponent equilibria can be predicted from
single solute isotherms using IAST for similar sized
molecules
• The Equilibrium Column Model can predict the
longest bed life and the highest overshoot
concentrations for multicomponent mixtures of known
components.
28

CONCLUSIONS (continued)
• The PSDM can predict the effluent concentration
history profiles for multicomponent mixtures of known
components.
• The PSDM can simulate the effluent concentration
history profiles for SOCs in the presence of BOM.
– Reductions in capacity and diffusivities which
were estimated from a ground water and surface
water span the range of fixed bed data from 11
different studies.
– Additional comparisons are needed to develop
more general guidelines.
29

FURTHER READING
Crittenden, J.C., N.J. Hutzler, D.G. Geyer, J.L. Oravitz, and G. Friedman, "Transport of
Organic Compounds with Saturated Groundwater Flow: Model Development and
Parameter Sensitivity," Water Resources Research, 22 (3), 271-284 (1986).
Crittenden, J.C., T.F. Speth, D.W. Hand, P.J. Luft, and B. Lykins, "Evaluating Multicomponent
Competition in Fixed Beds," Journal of Environmental Engineering, 113 (6),
1363-1375 (1987a).
Crittenden, J.C., D.W. Hand, H. Arora, and B.W. Lykins Jr., "Design Considerations for
GAC Treatment of Organic Chemicals," Jour. of AWWA, 79 (1), 74-82 (1987b).
Crittenden, J.C., R.D. Cortright, B. Rick, S.R. Tang, and D. Perram, "Using Granular
Activated Carbon to Remove Volatile Organic Chemicals from Air Stripping Off-Gas,"
Jour. of AWWA, 80 (5), 73-84 (1988).
Hand, D. W., J. C. Crittenden, and W. E. Thacker, "Simplified Models for Design of Fixed-
Bed Adsorbers," Jour. of Env. Eng. Div., Proceedings of ASCE, 110 (EE2), 1984.
Hand, D. W., J. C. Crittenden, H. Arora, J. Miller, and B.W.Lykins Jr., "Design of Fixed-
Beds to Remove Multi-component Mixtures of Volatile and Synthetic Organic
Chemicals," Jour. of AWWA, 81(1) 1989.
Hand, D. W., J.C. Crittenden, D.R. Hokanson, and J.L. Bulloch, “Predicting the performance
of fixed-bed granular activated carbon adsorbers,” Water Science and Technology,
35(7), 235-241 (1997).
Sontheimer, H., J.C. Crittenden, and R.S. Summers, “Activated Carbon for Water
Treatment,” DVGW-Forschungsstelle, Engler-Bunte-Institut, Univ. Karlsruhe, Fed. Rep.
of Germany (1988).
30

ACKNOWLEDGMENTS
• National Center for Clean Industrial and Treatment
Technologies (CenCITT)
• U.S. Environmental Protection Agency
• National Science Foundation (No. ECE 8603615)
• Environmental Engineering Center (MTU)
• Some laboratory and field data were analyzed and
provided by G. Baldauf, Gerhard Zimmer, and the
late Professor Heinrich Sontheimer at Engler-Bunte-
Institut, University of Karlsruhe
31