 

v

.11.. mart.
)3. I) I 3..)
1...... 1......" ..
721...? 9&4... .
. 1... :55: . .
15.3145: huh-vb. . Para.“ 0...
‘39-‘91) ‘ﬁ I’k‘ ‘I
3h 5...... «in 15.1.
“what ..
1.113843...

.. nut, 1...... . . . .1.
.33... 59......mwsmmﬁ 03.. 1w... .1.
. . . .rmmu’uaunﬂ:myh .ﬂﬁkm.

- . A I.
. at;

”UN! . r 1 . s4.
.. . . inﬂux... 5.... 3...»... 5......
link ﬁx . . .

.1. Z. . ull:
.xﬂ .11.. .ng... 1a.. a. 51 .1
iii. .§'# . ”1¢N. V n I,
1 :13... .31... 1.. . . 5....
.58....0rreo 21.11%:-
9 t. 11 3.2 . .
3:1 5!... v
1.3.11 :3...»
It“! . ban»: a...
Ix. . at...) P1.

J
. .1 4 .4“ .
.. SWIHW. 1..... 1
. HBPMMM.
3
n.
L

.l. .s.
:13. Du“. .41
1.1.11.5 ,i‘til
2.40... Razaymn
1.... ...

1...... ..

.. . 5...... 1..:
6) 1.. . 1......
1.1.1.3. , .1.
5.5%... 1. 3.2.4.4.." as“
. . ... .. .
. .3“. .3 .....
, . INS-WW. é.

.1 ..
E... .3!
. i .. . 2.5.51... akin»
)1}... 0 I. ..I....).§u.‘)2)
.. 1..! 3? a... 1 .
. . . . a...
via... a). z
1.315.336.19
:1 .........axis 42.3.! a...
1......5. 1.... 3...... :. 1.3.5:...
.. . .:. .a a... 1.... y.
1 .

I!
i

.2 .. .13.: s
v. :31
D

3.0». D
”(rel-.3. I: . .. 5.5
$1.01 a.) El! 2:).- .xi... ;! 11..
. ‘1 3.5415 lira... .3‘3’AID‘NM. .1. .‘c
‘3'...) . d" alt}.|)>“.|)..r‘lq.
0i." :1“. “churn..-

« L...-ani..hlh.1v\£\l.ﬂ h

1 cl lei.“ \ .
513?. x?

ivnswukuba... .1

1 xi? . 1..
.m 311...:

.\ . r. I}; vs
\SIQA 1..... 1525.! .t.
I13 I ’31.. 1.14 x...‘..v.....
Eli-$5 0.51.9. ID"!-
.n. I: .xQunrv 5!) «Ill. 9...
3!}.Uavk.

‘v
5113?

Lu.”
w

v. .«\..t 1...: .
~53.¥I\Jr
. 1.. 3‘91}...
.‘xl’v‘ﬂa. .1 .1
1.. I).

(lift. v!!!..n|.l.|..
1!... n» 3 .|.| \.! .1

§Lu$.?.t\. :. .v $1.... Mn... :34 ..
1:1.‘ﬁ'w J‘ In 1“. \ g¥uv. . .

 

linuwm wogqyu 19“.":

 

THESlS

MCHIGAN

ll!!!

NIH W ..

3 mini/umumrmumi‘illm

3 01046 1634

“I

This is to certify that the

dissertation entitled

Biological Activated Carbon In Fluidized Bed Reactors
For The Treatment Of Groundwater Contaminated
With Volatile Aromatic Hydrocarbons

presented by

Xianda Zhao

has been accepted towards fulﬁllment
of the requirements for

Ph.D. degree in Environmental Engineering

Wm ()Vu—LZL

Major professor

Date 11-17-94

MS U i: an Afﬁrmative Action/Equal Opportunity Institution 0-12771

 

LIBRARY
Mlchigan State
Unlverslty

 

 

 

PLACE II RETURN BOXtoromwothb WM ywrrocord.
TO AVOID FINES Mum on or Mon date duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU IoAn Namath WW Oppottunlty Imon
Wt

 

BIOLOGICAL ACTIVATED CARBON IN FLUIDIZED BED REACT ORS FOR THE
TREATMENT OF GROUNDWATER CONTAMINATED WITH VOLATILE
AROMATIC HYDROCARB ON S

By

Xianda Zhao

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Civil and Environmental Engineering

1994

ABSTRACT

BIOLOGICAL ACTIVATED CARBON IN FLUIDIZED BED REACT ORS FOR THE
TREATMENT OF GROUNDWATER CONTAMNATED WITH VOLATILE
AROMATIC HYDROCARBONS

By

Xianda Zhao

Laboratory and pilot-scale ﬂuidized bed systems employing granular activated
carbon (GAC) as a biomass carrier were used for the remediation of groundwater
contaminated with the gasoline constituents, benzene, toluene and xylene (BTX). These
systems (designated GAC-FBR) were evaluated in start-up, pseudo-steady-state and step-
load rate increase modes of operation. The role of adsorption and the change in adsorption
capacity of GAC in the systems were investigated. The formation of partial oxidation
products was also investigated under steady-state and extremely high organic loading

conditions.

The role of adsorption was investigated by comparing identical systems with
adsorptive (GAC) and non-adsorptive (non-activated carbon) bioﬁlm carriers. The GAC-
FBR system, which had both adsorption and biodegradation capacities, showed higher
organic removal and more stable performance over the non-adsorptive bioreactor (FBR)
during transient conditions such as start-up and step-load increases. However, under
steady-state organic loading conditions (3 and 6 kg COD/m3-day) BTX removals were
comparable for the GAC-FBR and biological only (FBR) systems and dominated by
biodegradation. More than 90 percent of BTX was removed in both systems.

At steady-state, no intermediate decomposition products were detected in the efﬂuents
of either system. These materials were formed, however, during single substrate
concentration step increases (twenty, twelve and seven-fold step increases in the organic
loading rates of benzene, toluene and p-xylene, respectively). The concentration of

byproducts was found to be lower in the GAC-FBR efﬂuent than in the FBR system.

The adsorption capacity of biocoated GAC was maintained at more than 70% of its
initial level during the ﬁrst two months of system operation. After six months of operation,
the remaining capacity was approximately 50% of the initial value. It was determined that
this was not due to adsorption of the primary substrate, but may result from the adsorption
of other materials produced by the bioﬁlm. No direct relationship was found between the

amount of biomass on the carbon and the remaining adsorption capacity.

ACKNOWLEDGMENTS

I would like to give my greatest appreciation to my advisor Dr. Thomas Voice of
the Department of Civil and Environmental Engineering for his encouragement,

understanding, and assistance in my research.

A special thank you is given to Dr. Robert Hickey of the Michigan Biotechnology
Institute for introducing me to ﬂuidized bed technology and teaching me how to write a
research paper (let the data speak for itself and never exaggerate). Thank you also to Mr.
Dan Wagon at the Michigan Biotechnology Institute for teaching me the tricks of the

laboratory and making my job much easier.

I would also like to thank Dr. Craig Criddle of the Department of Civil and
Environmental Engineering for his fresh ideas which opened my eyes to new view points
and helped me think beyond my research. I would like to thank Dr. Stephen Boyd of the
Department of Crop and Soil Sciences for his understanding and help to insure my study
program ran smoothly, and for being the ﬁrst person to introduce the soil environment into
my study. I would also like to thank Dr. William Saul, chairmen of the Department of Civil
and Environmental Engineering, for serving on my committee and rescuing me from the

last minute regulation change.

I have been at MSU for seven years and have so many fellow students and friends
who gave me considerable assistance and made my Spartan life more enjoyable. I can't
write every name in this very small space but my heart will always be with them. I would

like to mention a few names to remember their speciﬁc contributions. Dr. Daewon Pak

iv

introduced me to biological reactors and helped me to establish my experimental protocol.
Dr. Myung-Keun Chang, as my fellow graduate student, worked very closely with me and
discussed many important ideas. Mr. Yanlyang Pan, who was once my fellow graduate
student, helped me to understand and operate my ﬁrst piece of analytical equipment (the IC)
in my research and continuously helped me on many different analytical methods since
then. Mr. Kwunhwan Doh conducted a series of experiments with me and kindly allowed
me to present the results in chapter ﬁve of this dissertation. I would like to thank Mr.
Chien-Chun Shih and Mr. I-yuang Hsu for their valuable comments on many aspects of the
environmental engineering ﬁeld. I would also thank Ms. Mara Hollinbeck for her help

correcting my not so good English.

I give a very special thank you to my grandfather for his encouragement and effort
which made it possible for me to study in the US. I would also like to thank my
granduncles, grandaunts, aunts and uncles for the encouragement and help in making us

life more enjoyable.

I have a very special thank you for my parents. They taught me how to work hard
and value life.

My greatest thank you goes to my lovely wife Jing and our daughter Aisling. They
make my life as happy as it can be. Jing, who was my classmate in college and my fellow
graduate student at MSU, not only makes my home life very comfortable and enjoyable,
but also helped me to conduct many experiments in my research. She also is the co—author
of chapter six in this dissertation. Aisling, my lovely daughter, you came into my life just
over a year ago, but you make my life so rich and bring happiness to my heart. My greatest

love to Jing and Aisling.

TABLE OF CONTENTS

List of Tables ...................................................................................... xi
List of Figures ..................................................................................... xiii
Nomenclature ...................................................................................... xvii
CHAPTER
I. INTRODUCTION AND RESEARCH OBJECTIVES .............................. 1
1.1 Introduction ........................................................................ 1
1.2 Research Objectives ............................................................... 2
1.3 Overview of the Chapters ......................................................... 3
II. BACKGROUND ......................................................................... 4
2.1 Deﬁnition and development of biological activated carbon ................... 4
2.2 Activated carbon ................................................................... 6
2.3 Adsorption, Desorption and Regeneration ...................................... 6
4 2.3.1 Adsorption ............................................................ 6
2.3.2 Multi-Component Adsorption ...................................... 8
2.3.3 Desorption ............................................................ 9
2.3.4 Regeneration .......................................................... 10
2.4 Biofilm Development on Activated Carbon ..................................... 10
2.5 Microbial Degradation of Benzene, Toluene and p-Xylene (BTX) .......... 13
2.6 Microbial Degradation Kinetics of Benzene, Toluene and p-Xylene
(BTX) ............................................................................... 18
2.7 Bioregeneration of BAC .......................................................... 22
2.8 Interaction of Adsorption and Biodegradation in the BAC Process ......... 24

vi

III.

IV.

2.9 Application of BAC processes ................................................... 30
2.9.1 Drinking Water Treatment ........................................... 30
2.9.2 Industry Wastewater Treatment .................................... 31

COMPARISON OF BIOFILM CARRIERS IN FLUIDIZED BED

REACTOR SYSTEMS .................................................................. 33
3.1 Introduction ........................................................................ 33
3 .2 Materials and Methods ............................................................ 35
3.2.1 Media and Chemicals ................................................ 35
3.2.2 Determination of Adsorptive Capacity of BTX ................... 36
3.2.3 Fluidized Bed Reactors .............................................. 36
3.2.4 Analytical Methods ................................................... 39
3.2.5 Scanning Electron Microscopy ..................................... 40
3.3 Results .............................................................................. 40
3.3.1 Adsorption Characteristics Of BTX ................................ 40
3.3.2 Breakthrough Proﬁles ............................................... 41
3.3.3 Comparison of Steady State BTX Removal Between
GAC-FBR and FBR Systems ...................................... 48
3.3.4 Response of FBR Reactors to Step Load Increases in BTX
Concentration ......................................................... 49
3.3.5 Extent of Surface Coverage of the Media by the Bioﬁlms ...... 53
3.4 Discussion .......................................................................... 55
3.5 Conclusions ........................................................................ 58
SINGLE SUBSTRATE STEP-LOADING RATE INCREASES .................. 60
4.1 Introduction ........................................................................ 60
4.2 Materials and Methods ............................................................ 61
4.2.1 Media and Chemicals ................................................ 61
4.2.2 Biodegradation Kinetic Assays ..................................... 62
4.2.3 Fluidized-Bed Reactors .............................................. 63
4.2.4 Step-Loading Rate Increase Experiments ......................... 65
4.2.5 Extraction of BTX from Seeded GAC and Baker Product ...... 65
4.2.6 Analytical Procedures ................................................ 65
4.3 Results .............................................................................. 66
4.3.1 Steady-State Conditions ............................................. 66
4.3.2 Biodegradation Kinetics ............................................. 68
4.3.3 Step-Load Increase Experiments ................................... 72
4.4 Discussion .......................................................................... 83

viii

4.5 Conclusions ........................................................................ 85

FORMATION OF PARTIAL OXIDIZED PRODUCTS IN THE FBR

SYSTEMS AT SINGLE STEP LOADING INCREASES .......................... 86
5.1 Introduction ........................................................................ 86
5.2 Materials and Methods ............................................................ 90
5.2.1 Media and Chemicals ................................................ 90
5.2.2 Fluidized Bed Reactors .............................................. 90
5.2.3 Step-Load Increase Experiments ................................... 93
5.2.4 Analytical Procedure ................................................. 94
5.2.5 Stoichiometry of Biological Reactions ............................. 96
5.3 Results .............................................................................. 100
5.3.1 Steady-State Conditions ................................................. 100
5.3.2 Step-Load Increases ..................................................... 101
5.4 Discussion .......................................................................... 113
5.5 Conclusions ........................................................................ 115

VI. ROLE OF ADSORPTION IN GRANULAR ACTIVATED CARBON

FLUIDIZED BED REACTORS ........................................................ 1 17
6.1 Introduction ........................................................................ 117
6.2 Materials and Methods ............................................................ 119
6 2 l Fluidized-Bed Reactor ............................................... 119
6.2.2 Materials ............................................................... 122
6.2.3 Step-OLR Increase Procedure ...................................... 123
6.2.4 Extraction Procedure ................................................ 123
6 2 5 Analytical Measurements ............................................ 124
6.3 Results .............................................................................. 125
6 3 1 Phase 1: Start-up ..................................................... 125
6.3.2 Phase 2: Pseudo Steady-State ...................................... 130
6.3.3 Phase 3: Step-OLR Increase ........................................ 133
6.3.4 Phase 4: Recovery ................................................... 137
6.4 Discussion .......................................................................... 137
6.5 Conclusions ........................................................................ 143

VII. ESTIMATION OF KINETIC PARAMETERS FOR BIODEGRADATION
OF BENZENE, TOLUENE AND XYLENE ......................................... 144

7.1 Introduction ........................................................................ 144
7.2 Materials and Methods ............................................................ 146
7.2.1 Biodegradation Kinetic Assays ..................................... 146

7.2.2 Yield Coefﬁcient Assay ............................................. 147

7.2.3 Decay Coefﬁcient Assay ............................................ 148

7.2.4 Theory ................................................................. 148

7.2.5 Estimation of Parameters ............................................ 149

7.2.6 Chemicals and Analytical Procedure ............................... 150

7.3 Results .............................................................................. 151
7.3.1 Estimation Method Verification ..................................... 151

7.3.2 Decay Coefﬁcient .................................................... 152

7.3.3 Yield Coefﬁcient ..................................................... 154

7.3.4 Estimation of Biodegradation Kinetic Parameters ................ 155

7.3.4.1 One Substrate .............................................. 156

7.3.4.2 Two Substrates ............................................ 159

7.3.4.3 Three Substrates .......................................... 164

7.4 Discussion .......................................................................... 169
7.5 Conclusions ........................................................................ 171

VIII. ADSORPTION CAPACITY OF BIOFILM COATED ACTIVATED

CARBON IN A BIOLOGICAL FLUIDIZED BED REACTOR SYSTEM ....... 173
8.1 Introduction ........................................................................ 173
8.2 Materials and Methods ............................................................ 174
8.2.1 Media and Chemicals ................................................ 174
8.2.2 Fluidized Bed Reactors .............................................. 175
8.2.3 Collection and Sterilization of Bioﬁlm Coated Activated
Carbon ................................................................. 176
8.2.4 Measurement of Bioﬁlm Thickness ................................ 177
8.2.5 Extraction Procedure ................................................. 179
8.2.6 Adsorption Isotherrn Assay ......................................... 179
8.2.7 Sodium Hydroxide Wash ........................................... 180
8.2.8 Estimation of Adsorption Parameters .............................. 181
8 .3 Results .............................................................................. 18 1
8.3.1 Fluidized Bed Performance and Collection of Carbon
Samples ............................................................... 181
8.3.2 Extraction of Adsorbed Toluene on Carbon ...................... 185
8.3.3 Adsorption Capacity ................................................. 187
8.3.4 Effect of Biomass on Adsorption .................................. 190
8.3.5 Sodium Hydroxide Wash ........................................... 192

8.4 Discussion .......................................................................... 193

8.5 Conclusions ........................................................................ 195

IX. DISSERTATION SUMMARY AND ENGINEERING DESIGN

IMPLICATIONS ......................................................................... 197

9.1 Summary ........................................................................... 195
9.1.1 Importance of Using GAC as a Carrier Medium in FBR

Systems ............................................................... 197

9.1.2 Biodegradation in FBR Systems ................................... 199

9.1.3 Interaction of Multiple Substrates .................................. 200

9.1.4 Formation of Byproducts ........................................... 200

9.1.5 Adsorption Capacity in a GAC-FBR .............................. 201

9.2 Engineering Design Implications ................................................ 202

9.3 Recommendations for Future Studies ........................................... 204

LIST OF REFERENCES ...................................................................... 205

TABLE

2-2.
3-1.
3-2.

3-3.
3-4.

3-5.

4-1.

4-2.
5-1.

5-3.
6-1.
7-1.
7-2

LIST OF TABLES

Biodegradation kinetic coefﬁcients for aerobic biodegradation of

BTX (modiﬁed from Alvarez et a1. 1991) .................................... 21
BAC processes for treatment of industrial wastewaters .................... 32
Adsorption characteristics of BTX ............................................ 41
Substantial breakthrough time for each BTX compound in the GAC

system ............................................................................ 42
Summary of performance of FBRs during the start-up period ............. 46
Comparison of steady state BTX removals in GAC-FBR and FBR

systems ........................................................................... 48
Summary of FBR responses to step-load increases from base

loading level of 6.0 kg COD/m3-day ......................................... 52
Summary of FBR performance during step load increases ................. 57
Summary of ﬂuidized-bed system performance at steady-state and

during a step-load increase ..................................................... 67
Biodegradation rate parameters for BTX by a mixed culture ............... 69
Electron balances at steady-state .......... .................................... 101
BTX removal during step-load increases ..................................... 106
Carbon balances during the step load increases .............................. 108
Summary of overall performance of the reactor ............................. 138
Procedure veriﬁcation for parameter estimation ............................. 151
Evaluated data sets .............................................................. 152
Decay coefficient determination ................................................ 152

7-4.
7-5.
7-6.

7-7.

7-9.

8-1.

8-3.

Net yield coefﬁcients for single substrates ................................... 155
Net yield coefﬁcients for multiple substrates ................................. 155
Kinetic parameters for biodegradation of benzene by a mixed culture

with and without other substrates present .................................... 158
Kinetic parameters for biodegradation of toluene by a mixed culture

with and without other substrates present .................................... 159
Kinetic parameters for biodegradation of xylene by a mixed culture

with and without other substrates present .................................... 159
Kinetic parameters for biodegradation of BTX by a mixed culture

with varying initial concentrations ............................................. 164
Effect of biomass and base washing on isotherm parameters .............. 178

Adsorption isotherm parameters for GAC collected from mid-height
in a GAC-FBR (150 cm) ....................................................... 186

Adsorption isotherm parameters for GAC collected from the bottom
portion of a GAC—FBR (50 cm) ............................................... 187

FIGURE

2-1.

2-2.

2-3.

3-2.

3-3.

3-5.

3- 6.

3-7.
4-1.
4-2.
4-3.

LIST OF FIGURES

Pathways for benzene biotransformation (modiﬁed from Atlas and

Bartha 1998). .................................................................... 14
Pathways for toluene biotransformation (modiﬁed from Zylstra and

Gibson 1991). ................................................................... l6
Pathways for p-xylene biotransformation. ................................... 17
Schematic of ﬂuidized bed reactor systems. ................................. 38

Comparison of breakthrough proﬁles for GAC and GAC-FBR
systems ........................................................................... 43

Comparison of breakthrough proﬁles for GAC-FBR and FBR
systems ........................................................................... 44

SEM of a particle without biological growth. (a) activated carbon, x
66 and (b) non-activated carbon (baker product), x 66. .................... 47

The response of FBR system to a 96 hour, 3-fold step load increase
from steady state loading level of 2.9 kg COD/m3 -day. The load
increase started at hour number 929 and ended at hour number 1025 .... 50

The response of GAC-FBR system to a 96 hour, 3-fold step load
increase from steady state loading level of 3.8 kg COD/m3-day. The
load increase started at hour number 929 and ended at hour number

1 02 5 ............................................................................... 51
SEM of a GAC with biological growth. (a) x 48 and (b) x 1200 .......... 54
Schematic of ﬂuidized-bed reactor systems .................................. 64
The biodegradation rate patterns for BTX in a mixed culture .............. 70

Proﬁle of BTX concentrations and degradation rates in a FBR under
steady-state conditions .......................................................... 71

4-10.

4-11.

5-1.
5—2.
5-3.

5-4.

5-5.

5-6.

5-7.

6-1.

6-2.

6-4.
6-5.

xiv

Substrate concentrations in the efﬂuent of ﬂuidized-bed systems

during a benzene step-load increase experiment ............................. 73

Proﬁle of BTX concentrations and degradation rates in a FBR at end

of benzene step-load increase .................................................. 74

Substrate concentrations in the efﬂuent of ﬂuidized-bed systems

during a toluene step-load increase experiment .............................. 77

Substrate concentrations in the efﬂuent of ﬂuidized-bed systems

during a p-xylene step-load increase experiment ............................ 78

The responses of ﬂuidized-bed systems to a p-xylene step-load

increase ........................................................................... 79

Concentration proﬁles of substrates and D0 in ﬂuidized-bed

systems at the end of a benzene step-load increase experiment ............ 80
Concentration proﬁles of substrates and D0 in ﬂuidized-bed

systems at the end of a p—xylene step-load increase experiment ........... 81
Proﬁles of adsorbed substrates on GAC-FBR before and p-xylene
step-load increase ................................................................ 82
Schematic of ﬂuidized-bed reactor systems .................................. 92
Experimental conditions for a xylene step load increase .................... 94

Spectra for the efﬂuent of FBR before, during (HI-1H), and after a

xylene step load increase ....................................................... 103

Removal of BTX, consumption of DO, and production of
byproducts in the GAC-FBR and FBR during a xylene step load

increase with different oxygen and nutrient supply ......................... 104

Performance of the GAC-FBR and FBR during a benzene step load

increase ........................................................................... 107

Performance of the GAC-FBR and FBR during a toluene step load

increase ........................................................................... 111

Performance of the GAC-FBR and FBR during a xylene step load

increase ........................................................................... 112
Schematic of the experimental GAC-FBR system ........................... 121
D0 consumption during the startup phase .................................... 127
The proﬁle of adsorbed of toluene during the startup phase ............... 128
Total mass of adsorbed toluene in the GAC-FBR system .................. 129

Height of the carbon bed in the GAC-FBR system ......................... 130

6-6.

6-8.

6-9.

6-10.
6-11.

7-10.

7-11.

XV

Concentration proﬁle aqueous-phase toluene, extracted toluene and
DO consumption before and after biomass removal ......................... 132

Toluene concentration in the inﬂuent and efﬂuent and DO
consumption during the pseudo-steady-state, step-OLR increase and
recovery phases .................................................................. 134

Concentration proﬁles of aqueous-phase toluene, adsorbed toluene
and DO consumption ............................................................ 135

The proﬁles of aqueous-phase toluene, adsorbed toluene and DO
consumption during the pseudo-steady-state and step increase

periods ............................................................................ 136
The oxygen utilization quotient (Q) for the overall experiment ............ 139

Cumulative toluene removed from the aqueous phase and that
removed biologically during the step-OLR increase and recovery

phases ............................................................................. 142
Veriﬁcation of parameter estimation protocol ................................ 153
Decay of biomass harvested from the FBR .................................. 154

Typical biodegradation trend of benzene as a single substrate in a
mixed culture ..................................................................... 157

Typical biodegradation trend of toluene as a single substrate in a
mixed culture ..................................................................... 157

Typical biodegradation trend of p-xylene as a single substrate in a
mixed culture ..................................................................... 158

Typical biodegradation patterns of benzene and toluene as dual
substrates in a mixed culture ................................................... 161

Typical biodegradation patterns of benzene and p-xylene as dual
substrates in a mixed culture ................................................... 162

Typical biodegradation patterns of toluene and p—xylene as dual
substrates in a mixed culture ................................................... 163

Typical biodegradation patterns of benzene, toluene and p-xylene as
multiple substrates in a mixed culture ......................................... 165

Biodegradation patterns of benzene, toluene and p-xylene as
multiple substrates with a high initial benzene concentration in a
mixed culture ..................................................................... 166

Biodegradation patterns of benzene, toluene and p-xylene as
multiple substrates with a high initial toluene concentration in a
mixed culture ..................................................................... 167

8-7.

8-9.

xvi

Biodegradation patterns of benzene, toluene and p-xylene as
multiple substrates with a high initial p-xylene concentration in a

mixed culture ..................................................................... 168
Schematic of the experimental GAC-FBR system ........................... 176
Inﬂuent and efﬂuent toluene concentrations and DO consumption in

a GAC-FBR during the ﬁrst phase ............................................ 183
Inﬂuent and efﬂuent toluene concentrations and DO consumption in

a GAC-FBR during the second phase. ....................................... 184
Adsorbed toluene on biocoated GAC in the middle (150 cm) and

bottom portion of the GAC-FBR .............................................. 185
Isotherm parameters for biocoated GAC. .................................... 188

Remaining capacity at three hypothetical equilibrium concentrations
(0.1, 3, and 10 mg/L representing the concentrations at efﬂuent,
inﬂuent, and a high inﬂuent shock load, respectively.) ..................... 189

Adsorption isotherms for carbon with and without biomass
removal. .......................................................................... 191

The effect of N aOH wash on the adsorption isotherm ...................... 193

NOMENCLATURE

English Symbols

b decay coefﬁcient (l/day)

Ce equilibrium concentration of toluene in water (mg/L)
dts empty bed hydraulic retention time of the section.

fb fraction of electrons diverted for byproduct.

fe fraction of electrons diverted for energy.

fs fraction of electrons diverted for synthesis.

f 3 fraction of electron donor originating from benzene
f “j fraction of electron donor originating from toluene

f 3,‘ fraction of electron donor originating from xylene

g electron equivalent mass of the electron donor (g)

h electron equivalent mass of biomass

k maximum speciﬁc utilization rate of substrate (mg substrate/mg VSS—day)
Kf a constant for adsorption isotherm

Ks half-saturation coefﬁcient (mg/L)X concentration of biomass (mg VSS/L)
Mb amount of toluene removed biologically

Moor) difference between the inﬂuent and efﬂuent COD
Mm mass of oxygen consumed in the FBR

1/n constant for adsorption isotherm

Q oxygen utilization quotient

Q5 S removal index at pseudo steady-state

xvii

xviii

qe additional amount of toluene adsorbed (mg/ g)

qo initial amount of adsorbed toluene (mg/g)

R average degradation rate in the section (mg/L—min)

Ra half reaction for the reduction of an electron acceptor for energy
Rb half reaction for the reduction of an electron acceptor for byproduct
Rd half reaction for the oxidation of an electron donor

Re half reaction for the reduction of an electron acceptor for synthesis
S concentration of substrate (mg/L)

S H concentration of substrate at higher point of the section (mg/L)
SL concentration of substrate at lower point of the section (mg/L)
So initial concentration of substrate (mg/L)

t time (day)

X concentration of biomass (mg VSS/L)

X0 initial concentration of biomass (mg VSS/L)

Y net yield coefﬁcient of biomass (mg VSS/mg substrate)

Mb production of byproduct (mg, or mg/L)

M0 consumption of oxygen (mg, or mg/L)

Mf consumption of benzene (mg, or mg/L)

M: consumption of toluene (mg, or mg/L)

M? consumption of p-xylene (mg, or mg/L)

Greek Symbols

11 speciﬁc growth rate constant (1/hr)

um maximum speciﬁc growth rate constant (1/hr)

CHAPTER 1

INTRODUCTION AND RESEARCH OBJECTIVES

1.1 Introduction

The contamination of groundwater by volatile organic compounds (VOCs) is being
reported with increasing frequency. More than 40% of US. population uses groundwater
as a drinking water supply, often without any treatment other than disinfection.
Groundwater contamination is thus a serious public health concern. Among the most
common types of groundwater contamination are those resulting from petroleum products.
In 1989, the department of natural resources in theEState of Michigan reported that 41.3 %
of their registered sites contained benzene, toluene or xylene (BTX) related contamination
(DNR 1989). The presence of these materials frequently results in elevated levels of BTX
in groundwater due to the fact that these compounds are relatively water soluble and hence

are poorly retarded by the soil matrix.

The conventional approaches to remediation of BTX contaminated groundwater
involve either liquid-phase adsorption by granular activated carbon (GAC) or air stripping.
In these processes, the VOCs are simply transferred from one phase to another (McCarty
1983; Voice 1989). Further treatment or disposal of the receiving phase is often required.
Biological treatment of contaminated groundwater would appear to be a desirable alternative
to such techniques. This approach has the potential to destroy the pollutant and is often less

expensive than physical-chemical treatment. Biological processes are frequently perceived

as being less stable than physical-chemical approaches, however, and thus are not always

considered for groundwater treatment.

The beneﬁcial aspects of integrated biodegradation/adsorption systems were
reported in early work on the use of GAC for secondary and tertiary treatment of municipal
wastewater (Weber et al. 1970). In systems designed as adsorbers, it has been shown that
biodegradation increases the period between GAC regeneration cycles (Bouwer and
McCarty 1982; Chudyk and Snoeyink 1984; De Laat et al. 1985; Kim et al. 1986; Gardner
et al. 1988; Speitel et al. 1989a; 1989b). In bioﬁlm systems, the use of GAC as a biomass
carrier has been shown to provide for removal of compounds resistant to biodegradation
(Suidan et al. 1983; 1987). In a treatment process for groundwater contaminated with
relatively low levels of volatile organic compounds, a biological granular activated carbon
ﬂuidized-bed reactor (GAC-FBR) system has been demonstrated to provide both the
efﬁciency of biological removal and the positive efﬂuent protection capability of activated
carbon adsorption (Hickey et al. 1990b; 1991a).

1.2 Research Objectives

This project was designed to investigate and exploit the advantages of the biological
activated carbon ﬂuidized bed system (GAC-FBR) for the treatment of gasoline
contaminated groundwater. The primary objective was to elucidate the roles of the two
removal mechanisms, adsorption and biodegradation, under a variety of operating
conditions. Secondary objectives included using this information to better deﬁne the
conditions under which GAC-FBR systems might be used and to explore the implications

to system design and operation.

1.3 Overview of the Chapters

This project was designed in six phases. Each phase constitutes one chapter in the
dissertation. The investigation and discovery of the advantages of GAC-FBR systems were
the major focus in the ﬁrst phase. In the second phase, the interaction of BTX and the role
of adsorption during a single-substrate step-load increase were, investigated. Formation of
byproducts in FBRs during transient conditions was explored in the third phase. In the
fourth phase, adsorption in a GAC-FBR was monitored by measurement of adsorbed
toluene on the GAC carriers. Because of the importance of biodegradation in the GAC-
FBR, the biodegradation kinetics of BTX were determined quantitatively in the ﬁfth phase.
In the sixth phase, the change in adsorption capacity of the GAC-FBR was measured in
terms of adsorption isotherms. The dissertation concludes with a summary of the most

important results from this project along with several recommendations for the design and

operation of GAC-FBR systems.

CHAPTER 2

BACKGROUND

2.1 Deﬁnition and development of biological activated carbon

It has been understood that activated carbon can remov' undesirable contaminants
from water since some time in the nineteenth century (Komegay 1978). The production of
synthetic organic chemicals has grown exponentially since 1930 (Schwarzenbach et al.
1993). Many of these compounds have entered the aqueous environment and contaminated
drinking water sources. The use of activated carbon has therefore grown widely in recent
years as a means of removing such pollutants. In addition to adsorbing aqueous
contaminants, activated carbon has also been shown to provide a very favorable
environment for the growth of microorganisms (Pirbazari et all. 1990). Development of a
microbial community on the carbon surface has generally been regarded as an undesirable,

but unavoidable aspect of activated carbon usage.

In the early 1970's, scientists and engineers discovered that bacterial activity
appeared to extend the life of carbon adsorbers used in drinking water treatment processes
in Western Europe. They investigated the effect of pretreatment and the requirement of
post-treatment for the process. A report written by AWWA Research and Technical Practice
Committee on Organic Contaminants summarized and discussed these studies and
suggested further research in this area (Committee Report 1981). Rice and Robson have

also presented a very comprehensive review of this topic (Rice and Robson 1982).

During this same period, researchers and engineers working on physical-chemical
treatment (PCT) and tertiary treatment of municipal wastewater had also identiﬁed the
beneﬁcial effect of biological activity in the GAC adsorption process. In these processes,
the activated carbon was designed to remove recalcitrant pollutants through adsorption. It is
not surprising that microorganisms grew in these systems. In what can be viewed as one of
the ﬁrst integrated biological activated carbon systems, Weber et al.. demonstrated the
beneﬁts of biological activity in expanded-bed adsorbers for treatment of municipal

wastewater (Weber et al. 1970; 1972a).

Following these initial discoveries, Rice et al. deﬁned the biological activated
carbon (BAC) process as a system for water or wastewater treatment "in which aerobic
microbial activity is deliberately promoted in granular activated carbon (GAC) adsorber
systems" (Rice et al. 1980). In 1981 DiGiano summarized the results of previous studies
and identiﬁed the following key features of biological activated carbon systems (DiGiano
1981): ‘

—-""‘"_'

a). The adsorption capability of a BAC system serves "to remove those difﬁcult to

biodegrade, but adsorbable, compounds."

b). The adsorbed compounds may create a high concentration region in the

macropore to hasten the acclimation of bacteria and the resulting biodegradation.

c). The GAC could prevent escape of compounds during the adoption and

acclimation period of microbial growth.

d). The adsorption capability is able to smooth out periods of high inﬂuent

concentration, even when little adsorptive capacity remains.

e). The adsorption capability will also protect the microbial system from potentially

toxic compounds.

2.2 Activated carbon

Activated carbon is manufactured from many carbon-based materials including
bituminous coal, coconut shells, lignite, wood and pulp mill residues (Rice and Robson
1982). As a result of the activation process, activated carbon exhibits a high degree of
porosity and an extensive associated surface area (Weber and Vliet 1980). The carbon
particles can be thought of as stacks of graphite planes. The channels and interstices which
pass through or between the planes are called macropores and have diameters larger than
1000 Angstroms. The micropores are within and parallel to the planes with diameters
between 10 and 1000 Angstroms (Weber 1972b). Commonly used activated carbon has a
speciﬁc surface area of 800 to 1200 m2/g (Weber and Vliet 1980). About 99% of the
available surface for adsorption consists of micropores while only 1% of the surface sites

are due to the macropores (Rice and Robson 1982).

2.3 Adsorption, Desorption and Regeneration
2.3.1 Adsorption

Adsorption is a phenomenon of accumulation and concentration of substances at a
surface or interface (Weber 1972b). Adsorption phenomena are commonly sub-divided into
three categories: exchange adsorption, physical adsorption and chemical adsorption.
Exchange adsorption results from the electrostatic attraction of ions to the charged sites at
the surface. Exchange adsorption can be a reversible process by changing the direction of
electrostatic attraction. Physical adsorption, which is also generally considered reversible,
involves weak van der Waals forces. In this process, the adsorbed molecule is not afﬁxed

to a speciﬁc site but remains free to undergo transitional movement within the interface

(Weber 1972b). Chemical adsorption results when there is a speciﬁc reaction between the
adsorbed substances and the surface sites. Chemical adsorption is usually considered an
irreversible process. The primary adsorption processes for GAC in water treatment are

physical adsorption and chemical adsorption (Rice and Robson 1982).

Activated carbon can adsorb a variety of compounds. Weber and Gould
investigated the adsorption of several organic pesticides from aqueous solution and
concluded that adsorption on activated carbon is an effective way to remove those
pollutants from water. Giusti et al. presented results demonstrating the adsorption ability of
93 petrochemicals on GAC and generalized that the higher molecular weight compounds
are favored for adsorption by activated carbon (Giusti et al. 1974). Yousseﬁ and Faust
reported that GAC is capable of removing low molecular-weight halogenated hydrocarbons
from water under conditions simulating treatment plant operation (Yousseﬁ and Faust
1980). The adsorption of polycyclic aromatic hydrocarbons (PAHs) on activated carbon
was studied by Borneff (1980). He concluded that 99 - 99.9% of soluble PAHs could be
eliminated from water by adsorption on activated carbon . The US. Environmental
Protection Agency conducted adsorption studies on 140 toxic organic and concluded that
pesticides, PAHs, phthalates, phenolics, and substituted benzenes are readily adsorbed by
activated carbon (Dobbs and Cohen 1980). Certain low molecular weight compounds
with high polarity, which include low molecular weight amines, nitrosamines, glycols, and
certain ethers, were found not to be amendable to treatment using activated carbon

adsorption.

The adsorption capacity of GAC can be evaluated by an adsorption isotherm
measurement. One of the accepted procedures used to generate isotherms is termed the
bottle point method (Randtke and Snoeyink 1983). In this technique, a series of bottles,

each containing activated carbon, are equilibrated with an aqueous solution of the target

compound. After equilibrium is achieved, the concentration of the compound in the liquid

phase is measured and the amount of solute in the solid phase is estimated by mass balance.

There are several models that are commonly used to describe adsorption equilibria.
The Langmuir model describes a single-layer adsorption model. The Brunauer, Emmett,
Teller (BET) model represents multilayer adsorption. Both models assume uniform
energies of adsorption on the surface. Another widely used equation for adsorption by
activated carbon is the Freundlich equation which is essentially empirical, but can be shown
to be a special case of Langmuir model which assumes heterogeneous surface energies vary

as a function of surface coverage (Weber 1972b).

2.3.2 Multi-Component Adsorption

In most practical applications, more than one adsorbable compound is present in the
water to be treated. Interactive and competitive effects can occur in such multi-component
systems. Jain and Snoeyink (l973)reported the results obtained from a study of several
bisolute systems which included neutral and anionic p-nitrophenol, benzenesulfonate, and
p-bromophenol. They concluded that competitive adsorption could occur when the
substances tend to occupy the same adsorption, sites while only minor competitive effects
occurred when the two species were adsorbed on different kinds of sites. Fritz et al.
studied competitive adsorption in bisolute water systems (Fritz et al. 1980). They
investigated the adsorption equilibrium and kinetics in a batch reactor and a ﬁxed-bed
adsorber. A p-nitrophenol and phenol bisolute system showed a higher reduction in
adsorptive capacity compared to a single-solute system. The displacement of more weakly
adsorbed compounds by stronger adsorbing compounds was found in both the batch

reactor and the ﬁxed-bed adsorber.

An ideal adsorbed solution theory (IAST) has been used to describe the adsorption
equilibrium of multi-components (Crittenden et al. 1985; Annesini et al. 1987). This theory
is based upon the assumption that the single-solute surface loading and the mixture loading
cause the same spreading pressure. Crittenden et al. (1985) presented a prediction of
multicomponent adsorption of seven mixtures that contained various combinations of two,
three and six solutes, which include chloroform, bromoform, trichloroethene,
tetrachloroethene, 1,2-dibromform, and chlorodibromomethane, using the IAST and
Freundlich isotherm equations. The results show an average percent error of 29% and 16%
for the liquid phase and solid phase concentrations, respectively. Annesini et al. (1987)
reported satisfactory agreement between the IAST coupled with the Langmuir isotherm
equation and experimental results for four bisolute mixtures included methylisobutylketone,

acetone, propionaldehyde, and sucrose.

2.3.3 Desorption

Desorption can occur when there are changes in the conditions under which
equilibrium was established. If the concentration of the substance at the liquid-solid
interface decreases, the adsorbed substance will have a tendency to return to the liquid
phase. Depending upon the type of adsorption, some or all adsorbed substances may not be
released. Desorption can also occurs when a mixture of substances is fed into a ﬁxed—bed
absorber (Fritz et al. 1980). As the mixture enters the absorber, the different substances
will be adsorbed at different depths in the unit. The less favored compounds for GAC
adsorption will be adsorbed at greater depths in the absorber while the more strongly
favored compounds will be adsorbed closer to the inlet. Since the substances enter the
column continuously, the more strongly favored compounds will compete and replace the
less favored compounds in the deeper region of the column. This desorptive release of the

less strongly adsorbed compounds, is known as the chromatographic effect.

10

2.3.4 Regeneration

After the adsorption capacity of activated carbon is exhausted, the carbon must
either be disposed of regenerated to recover its capacity. Three potential reactivating
techniques are chemical, steam, and thermal (Clark and Benjamin W. Lykins 1989).
Chemical reactivating involves using inorganic or organic regenerates to recover the
capacity. The recovery efﬁciency is not very high, especially for low molecular weight
compounds (Martin and Ng 1985). Steam reactivation is used primarily when adsorbate
recovery is desired as it often causes the carbon to lose its ability to remove low
concentrations of organic (Clark and Benjamin W. Lykins 1989). Thermal regeneration
systems are the most widely used. Detailed design considerations have been reviewed by
others (Zanitsch and Lynch 1978; Clark and Benjamin W. Lykins 1989). According to a
study of the cost for GAC ﬁlters and contactors in the Cincinnati Water Works from 1979
to 1980, more then 70% of the operating and maintenance costs were due to regeneration
processes (Clark and Benjamin W. Lykins 1989). Neukrug et al. (1984) reported similar

results.

2.4 Bioﬁlm Development on Activated Carbon

Microorganisms can survive and grow in an environment which provides sufﬁcient
water, substrate, and nutrients at suitable pH and temperature conditions. These
requirements are not difﬁcult to accomplish in an engineered reactor. The growth of a
bioﬁlm involves several physical, chemical and biological processes (Characklis 1984):

a). Transfer of organic molecules and microbial cells to the wetted surface.

b). Adsorption of organic molecules to the wetted surface creating a "conditioned"

surface.

c). Attachment of microbial cells to the "conditioned" surface.

11

d). Attached microbial cell metabolism resulting in more attached cells and
associated material.

e). Detachment of portions of the bioﬁlm.

The studies of microbial growth on GAC mainly rely upon direct microscopic
observation and viable cell counts. Klotz et al. (1976) found that the number of
microorganism colonies reached 106 to 107 per gram of wet GAC for a ﬁlter treating river
water. Van der Kooij reported that cell counts were 10 time higher on a GAC sample than a
sand sample when the columns were fed with non-chlorinated tap water. Latoszek and
Benedek (1979) studied the microorganisms on GAC for the treatment of domestic
wastewater and found 109 bacteria per gram of wet carbon at 25 °C and 2x108 at 5 0C.
Brewer and Carmichael (1979) reported that the GAC ﬁlter could retain and accumulate
bacteria and fungi based on the difference between inﬂuent and efﬂuent cell counts in a
drinking water system. Bancroft et al. (1983) studied the overall growth rate of bacteria on
GAC contactors treating river water by measuring the cell count in the inﬂuent and efﬂuent
and on the GAC. They found an exponential decrease in speciﬁc growth rate of the

population for the ﬁrst 40 days after which time it remained constant.

The scanning electron microscope (SEM) is able to magnify the details of a solid
surface and has therefore been used to investigate the colonization of microorganisms on
GAC. Weber et al. (1978a) examined GAC exposed to coagulated-settled sewage and a 5
mg/L solution of humic acid using this approach. After seven days of exposure in a
completely mixed batch reactor, a diverse community of microorganisms was found on the
surface of the GAC. An expanded-bed charged with clean GAC was fed a 5 mg/L solution
of humic acid for 10 days and the GAC particles were also examined by SEM. It was
reported that the bioﬁlm was non-homogeneous and full surface coverage was not

observed. However, this may result from the relatively short growth period use in this

study.

12

An investigation of microorganism coverage on GAC in a two-stage process was
conducted by Petrovic et al. (1989) The two-stage process was structured with an activated
sludge system and a ﬁxed bed GAC adsorber system for the puriﬁcation of mixed oil
reﬁnery and municipal wastewater. An average of 107 cells per gram was counted on the
GAC. Examination of scanning electron microscopy results showed heterogeneity of the
microﬂora on the GAC. During the colonization period, the cells appeared as individuals or
in small groups. After one year of operation, a rich extracellular matrix on the GAC was
observed. A homogeneous covering was found on some of the GAC while a non-

homogeneous cobweb-like covering was found on the remainder.

Five different bioﬁlm support media including GAC, carbonaceous adsorbent resin,
anion-exchange resin, silica sand and glass beads were used to compare initial cell
attachment and cultivation on the media surface (Pirbazari et al. 1990). Each of the media
was added in each of several individual completely-mixed-batch reactors and seeded with
activated-sludge from the secondary stage of a municipal wastewater treatment plant. The
reactor was ﬁlled with the efﬂuent from the primary stage of the municipal wastewater
treatment plant. After 24 hours of stirring at 100 rpm without aeration, air was provide to
each reactor for 40 hours. The particles were then removed and processed for examination
by SEM. The results showed activated carbon provided a more favorable surface for the
attachment of microorganisms compared to the other media. The sand contained less
growth than the GAC. The authors concluded that adsorption on GAC partially helped the

growth of microorganisms.

Based upon this research, it has been concluded that GAC provides a surface that promotes

colonization and growth by microorganisms, making it a desirable bioﬁlm support material.

13

2.5 Microbial Degradation of Benzene, Toluene and p—Xylene (BTX)

Since the early part of this century, it has been known that microorganisms are able
to metabolize benzene, toluene and p-xylene (BTX) (Stormer 1908; Sohngen 1913).
Numerous investigations have been conducted on the degradation pathways, type of
organisms, and the degradation rates of BTX. There are several articles which review the
efforts in these areas (Evans 1963; Gibson 1968a; Ribbons and Eaton 1982; Gibson and
Subramanian 1984).

The accepted degradative pathways (Figure 2—1) for benzene involve transformation
to cis-benzene glycol by a dioxygenase and subsequent oxidation of cis-benzene glycol to
catechol (Marr and Stone 1961; Gibson et al. 1968b; Hou 1982). The catechol is degraded
by a ring cleavage reaction involving either the B-ketoadipate pathway (ortho cleavage) or
the meta ﬁssion pathway (meta cleavage) (Gibson and Subramanian 1984). In ortho
cleavage, the dihydroxylated aromatic ring is opened to produce cis, cis-muconic acid,
which is further metabolized to B-ketoadipic acid and then oxidized to succinic acid and
acetyl-CoA by the tricarboxylic cycle. In meta cleavage, the dihydroxylated aromatic ring is
opened to produce 2-hydroxy-cis, cis-muconic semialdehyde. Formic acid, pyruvic acid,
and acetaldehyde are further metabolic products. All of the products from both cleavage
mechanisms are used further for either biosynthesis or energy generation through the citric
acid cycle, which produces C02 and ATP by electron transport and oxidative

phosphorylation (Lehninger 1982).

Benzene

9510 :
oﬂ ”'0

OH

cis-Benzene Glycol OH

2-H dr ', '-M ' _ .
y $331313;de uconrc l cis, cis-Muconrc Acrd
CHO
OOH / OH
| E / coon
OOH
\ OH \OH \

H203 a... ..
Cleava_e Cleava e r02

HCOOH
COOH O
I 2-Keto-4- Beta.- . COOH
V§ Pentenoic Acid KCtOﬁfllplC OCH
0 Acrd
l rCoA
0
ll Acetaldehyde HOOC-CHz-CHz-COOH
CH3-CH . . .
Succmrc Acrd +
.1- A tyl-CoA fl)
Pyruvic Acid ‘36
CH3-(l!-C00H CH3-C-SCoA

Figure 2-1. Pathways for benzene biotransformation (modiﬁed from Atlas and Bartha
1998).
i

15

As a result of the methyl group on the aromatic ring, the initial steps in the
biodegradation of toluene can occur either by oxidation of the methyl group to form benzoic
acid or hydroxylation of the ring (Figure 2-2) (Zylstra and Gibson 1991). Depending upon
the hydroxylated position on the ring, the products are p-cresol, o-cresol, m-cresol, or cis-
toluene dihydrodiol. A catechol will be produced as a result of the transformation of the
benzoic acid which can be further degraded through meta or ortho cleavage of the aromatic
ring. p-cresol can be hydroxylated again to form protocatechuate and this can be broken
down by meta ring cleavage. A further product of o-cresol, m-cresol, or cis-toluene

degradation is dihydrodiol 3-methylcatechol which is a substrate for meta ring cleavage.

The biotransformation of p-xylene can also occur through one of two pathways
(Figure 2-3): aromatic ring oxidation or oxidation of the methyl substituent. Gibson et al.
(1974) studied ring oxidation of p-xylene by Pseudomonas putida 39/D. The p-xylene was
oxidized ﬁrst to cis-p-xylene dihydrodiol and further to 3,6-dimethylprocatechol, which
can be degraded further by the ortho ring ﬁssion pathway. The cis-p-xylene dihydrodiol
can also be transformed to 2,5-dimethyl phenol by a non-enzymatic reaction (Gibson et al.
1974). Methyl substituent oxidation of p-xylene has been reported by several researchers
(Davis et al. 1968; Omori and Yamada 1970a; 1970b; Davey and Gibson 1974). The
proposed pathways involve two variations. In the ﬁrst, one of the methyl groups is
oxidized to form a carboxylic acid, which is then transformed to 4-methylcatechol. The
aromatic ring of 4-methylcatechol can be broken by meta ring ﬁssion. The second pathway

involves oxidation of both methyl groups before ring ﬁssion.

OO- $H3 3

 

OH
OH
Benzoic cis—Toluene
Acid p—Cresol o-Cresol m—Cresol Dihydrodiol
OO' 3
OH OH
OH OH OH
OH
Catechol Protocatechuate 3-Methylcatechol
Ring Cleavage

Figure 2-2. Pathways for toluene biotransformation (modiﬁed from Zylstra and Gibson
1991).

17

p-Xylene

Ring Oxidation Methyl Substituent Oxidation

(EH3.H

CH 3
$3” 3
C OH p-Methyl-
benzyl
'1." OH Alcohol
H
r12

CH3 OH
cis-Xylene
Dihydrodiol (EH3
3 p-Tolualdehyde
OH
p-Hydroxy- p—Hydroxy-
CH0 -C res 01 benzyl benzal-
OH + P Alcohol dehyde
CH 3 (EH 3 (EH 3 YH 20H (EH0

3,6-Dimethyl-
Pr t h l
miec o P-Toluic acid —> O O

OH OH OH
Ring Fission 3001-1 /
OOH
$1.13 COOH
4-Methylcatechol '
OH OH OH

Protocatechuic Acid p-Hydroxy-
benzoic
Acid
Ring FiSSiOD Ring Fission

Figure 2-3. Pathways for p-xylene biotransformation.

18

In a study of cometabolic degradation of p-xylene by Pseudomonas sp. strain B1
(pure culture system) in a ﬁxed-bed biological activated carbon system, Chang (1994)
noted the formation of several partial oxidation products. These products were identiﬁed as
3,6-dimethyl procatechol and p—xylene dihydrodiol from the degradation of p-xylene when
toluene was utilized as both carbon and energy sources. They were found in the efﬂuent of
a biological activated carbon ﬁxed-bed and a biological ﬁxed-bed (without adsorption)

during the entire experimental period (about 30 days).

It is clear that various intermediate products (byproducts) will be generated as part
of the degradation reactions. It is also clear that the BTX can be utilized as sole source of
energy and carbon by various microorganisms. With a fully developed mixed-culture
community, the byproducts will not normally accumulate because of the enzyme regulation
systems in the microorganisms. In a continuous ﬂow bioreactor, a community of
microorganisms will develop to utilize the substrate resource to obtain the maximum
beneﬁt. The community will not produce more byproducts than are needed because such
production costs, rather than yields, energy. Under transient conditions, however,
byproducts may be generated at a rate higher than they are consumed, causing these

compounds to accumulate in the reactor or be washed out in the efﬂuent.

2.6 Microbial Degradation Kinetics of Benzene, Toluene and p-Xylene (BTX)

One of the most widely accepted rate equations is the Monod equation (Monod

1949) shown here.

—— (2-1)

19

S is the concentration of substrate (mg/L), Liﬂis the hspeciﬁc growth rate constant

 

(1/hr), um is the maximum speciﬁc growth rate constant (Mn), and K8 is the half-
saturation coefﬁcient (mg/L). The Monod equation is strictly an empirical equation (Grady

and Lirn 1980). "The kinetic meters can be evaluated by conducting a batch experiment.

 

In this approach, the compound (as a sole organic constituent) is introduced into a reactor

containing the required nutrients, electron acceptor, and gowth factors. The reactor is then

 

 

inoculated with a known amount of microorganisms. Subsequent samples are taken to
monitor degradation of the compound and growth of the microorganisms. The Monod

coefﬁcients can then be estimated from Eqn. 2-1.

Several studies have been conducted to estimate the Monod coefﬁcients for the
biodegradation of BTX (Table 2-1). There are tremendous differences in the values
reported, however. This is not unexpected given that different organisms and widely varied

environments were used in these studies.

Interactions among BTX during biodegradation has been evaluated by several
researchers. Goldsmith (1988) showed a slightly higher maximum utilization rate for
toluene in a BTX mixture by an enriched mixed culture. Arvin et al. (1989) studied
substrate interactions during aerobic biodegradation of benzene. They found that the
degradation rates of benzene were higher when toluene or xylene was also present in the
environment. Alvarez and Vogel (1991) investigated the interactions of BTX during
biodegradation by two pure cultures and a mixed culture from an aquifer. The enhanced
degradation patterns of benzene and p-xylene in the presence of toluene was revealed. The
degradation of benzene was inhibited by the presence of p-xylene for a toluene degrader. In
their studies, toluene could not be degraded by a benzene degrader unless benzene was also
presence. Recent studies by Chang et al. (1993) indicated competitive inhibition and
cometabolism during biodegradation of BTX by two Pseudomonas isolates. Strain Bl,

which grew on benzene and toluene as sole sources of carbon and energy, has the ability to

20

transform p-xylene by consumption of the growth substrates (benzene and toluene) or
consumption of biomass during cometabolic degradation. The degradation of benzene,
however, was inhibited by toluene even though the degradation rate of toluene was not
affected. For strain X1, which grew on toluene and p-xylene but not benzene, the
degradation of p-xylene was inhibited by the presence of toluene, but the degradation rate
of toluene was not affected. Quantiﬁcation of competitive inhibition and cometabolism was
performed in their studies. For a system treating a mixture of BTX, complete degradation
patterns should be expected. Such patterns may depend upon environmental conditions,

composition of the waste, and the maturity of the microbial population

Table 2-1. Biodegradation kinetic coefﬁcients for aerobic biodegradation of BTX (modiﬁed

from Alvarez et al. 1991)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 Substrate k Ks Y k/Ks Culture References \
(Grady et al.
4.7 10.8 0.39 0.44 mixed 1989) l
(Alvarez et al. ;
i Benzene 8.3 12.2 0.68 aquifer _1991) l
' Pseudomonas (Chang et al. .
7.7 3.17 1.04 2.43 sp. B1 1993) '
" Pseudomonas
i 0.004 0.33 0.01 0.012 sp. T2 (Button 1985) 1
Pseudomonas
11 0.43 0.28 25.5 sp. T2 1
Pseudomonas (Robertson and ;
i 0.013 0.034 0.1 0.38 sp. T2 Button 1987) g
' Pseudomonas l
Toluene 0.33 0.044 0.1 7.7 sp. T2
(MacQuarrie et
0.49 0.65 0.43 0.75 aquifer al. 1990)
denitrifying (Jorgensen et i
4.32 0.15 28.8 sewage sludge al. 1990) -
Pseudomonas
10.7 1.96 1.22 5.45 sp. B1 1
Pseudomonas (Chang et al. 3
I 10.9 1.88 0.99 5.82 Sp. x1 1993) ;
Pseudomonas
: -X lene 51.4 4.55 0.25 11.3 St. X1

K5 is the half—saturation coefﬁcient (mg/L).

Y is the yield coefﬁcient for the biomass (mg VSS/mg substrate).

Where: k is the maximum speciﬁc utilization rate of substrate (mg substrate/mg VSS-day).

22

2.7 Bioregeneration of BAC

Following the discovery of the beneﬁts of bioactivity on operating adsorption
systems, some researchers attempted to use biological methods to regenerate exhausted
GAC (Rodman and Skunney 1971). Rodrnan et al. (1978) summarized their studies in a
1978 article. A textile dye wastewater was adsorbed by a GAC column for 10 hours
followed by back ﬂushing with the liquor from a completely mixed activated sludge reactor
for 13 hours. The process accomplished a 99% removal efﬁciency of color for a period of
4 months. The removed COD represented eight times the single adsorption capacity of the
GAC. Sigurson and Robinson (1978) developed a mathematical model to estimate the time
required for bioregeneration of the GAC. In their study, a phenol solution was treated
using a GAC column. After the carbon equilibrated with the liquid-phase phenol, the
carbon was regenerated with a mixed liquid from a fennentor in a back ﬂush mode. They
found that the bioregenerated carbon had recovered only 47 percent of the fresh carbon
phenol adsorption capacity. They concluded that the low recovery may be due to
occupation of certain adsorption sites by excreted products or products of the intermediary
metabolism. Wallis and Bolton (1982) studied a similar regeneration system loaded with
phenol. They demonstrated that the regeneration efﬁciency increased with an increase in
regeneration contact time to a certain point (about 29 hours), then the efﬁciency would
decrease. A maximum of 87% of the original adsorptive capacity was recovered after a 29
hour contact time. In 1987 Goeddertz et al. (1988) named this type of bioregeneration
process "Off-line Bioregeneration". They used a similar system which contained a GAC
expanded bed and an aeration tank. After the GAC was saturated with phenol, a desired
amount of biomass was added in the aeration tank and the mixed liquid was recirculated in
the GAC column and the aeration tank. After the regeneration period, the GAC was
saturated again with phenol to measure the regeneration efﬁciency. The experiments were

conducted with different initial biomass levels and duration of regeneration. The results

23

showed that 64% to 75% regeneration efﬁciencies can be achieved for the maximum
regeneration cycle of 4 days. The higher the concentration of initial biomass, the faster the

system approached the maximum GAC regeneration efﬁciency for the GAC.

In situ bioregeneration phenomenon, which is described as the adsorption,
desorption, and biodegradation which occurs in an integrated reactor, was also studied by
several researchers. Chudyk and Snoeyink (1984) studied columns ﬁlled with phenol-
saturated GAC. The columns were inoculated and fed with phenol at two different
concentrations of dissolved oxygen (4 mg/L and 9 mg/L). The results from a mass balance
calculation (using inﬂuent and efﬂuent phenol concentrations and the DO consumption)
showed that bioregeneration of GAC occurred at a DO concentration of 9 mg/L while no
bioregeneration was found when the inﬂuent DO was 4 mg/L. They also demonstrated the
effect of bioregeneration on GAC columns in response to transient loading. There is a more
detailed discussion of this transient loading effect in a later section. Kim et al. (1986) also
provided evidence for in situ bioregeneration in an anaerobic granular activated carbon
reactor for the removal of phenol. They found that the total carbon in biogas production
was higher than the total carbon removed from the inﬂuent with a relatively low

concentration of phenol in the inﬂuent. Therefore, bioregeneration of the GAC occurred.

Speital and DiGiano (1987) studied the bioregeneration of GAC used to treat very
low concentrations of pollutants (less than 100 ug/L). They selected two model organic
contaminants, phenol and paranitrophenol (PNP). Both chemicals have similar
biodegradation characteristics, but PNP has a higher adsorptive capacity for GAC. Using a
radiochemical method, they demonstrated that 8% to 15% of pre-sorbed phenol and 5% to
22% of PNP were bioregenerated within 10 days. Mathematical modeling suggested that
the bioregeneration rate was controlled by diffusion of sorbed substrate to the outer surface
of the GAC particle. Suidan and his colleagues (Khan et al. 1981; Suidan et al. 1981; 1983;
Wang et al. 1984; 1986; Suidan et al. 1987) extensively studied anaerobic treatment of coal

24

gasiﬁcation wastewater using BAC expanded-beds. They found that the BAC processes
were able to remove pollutants from the inﬂuent during a long initial start-up phase (about

100 days) and consequently part of adsorbed pollutants were removed by bioregeneration.

The effect of chemical characteristics of an organic compound on bioregeneration
has also been studied. Speitel et al. (Speitel et al. 1988; 1989a; 1989c) conducted a series
of studies to further understand this effect. Four chemicals (phenol, dichlorophenol (DCP),
pentachlorophenol (PCP), and paranitrophenol (PNP)) were used in their studies. The
GAC in the ﬁxed-bed reactor was saturated with radiolabled substrates, then a stream of the
same non-radiolabled substrates was fed to the reactor at equilibrium concentrations. The
portion of substrate removed by bioregeneration could be monitored by the production of
radiolabled C02. The effect of bioregeneration is highly dependent on the adsorption and
desorption characteristics of the substrate. The moderately adsorbable chemicals such as
phenol and PNP could be removed from the surface of carbon signiﬁcantly by
bioregeneration. The absence of bioregeneration of highly adsorbable chemicals such as

DCP and PCP was linked to irreversible adsorption.

2.8 Interaction of Adsorption and Biodegradation in the BAC Process

Bouwer and McCarty (1982) conducted a study on the removal of trace chlorinated
organic compounds by BAC systems. A mixture of chlorobenzene (CB), 1,2-
dichlorobenzene (1,2-DCB), 1,3-dichlorobenzene (1,3-DCB), 1,4-dichlorobenzene (1,4-
DCB), 1,2,4-trichlorobenzene (1,2,4-TCB), chloroform (CLF), 1,1,1-trichloroethane
(1,1,1-TCLE), and tetrachloroethylene (TECE) with sodium acetate was continuously
applied to two upﬂow columns for 2 years. One of the columns (BAC) was ﬁlled with a
homogeneous mixture of GAC and glass beads and another (BC) contained only glass

beads. The empty-bed detention time was 60 min and the hydraulic loading was 6.0 mlday

25

for both columns. The concentration of each chlorinated compound was between 10 and 30
jig/L and the dissolved oxygen concentration was maintained above 5 mg/L in the efﬂuent
to ensure aerobic conditions existed throughout the columns. Removal efﬁciencies of 95%-
98% of the chlorinated organic compounds were achieved in the BAC column initially.
Complete breakthrough of CLF, 1,1,1-TCLE, and TECE, however, occurred after 4, 6,
and 18 months, respectively. The high removal efficiency of chlorinated benzenes
continued during the two years of operation. The chlorinated benzenes were also removed
in the BC column after a period of bio-acclimation. The removal of CLF, 1,1,1-TCLE, and
TECE, however, was not observed in the BC column. In order to investigate the role of
adsorption in the columns, carbon-14 labeled CB and 1,4-DCB were fed to the columns as
tracer compounds and the column was operated in a "shut off" mode by maintaining anoxic
conditions in the feed. Nearly complete removal of both compounds continued to occur in
the BAC column, but removal of CB almost completely stopped and only 39% of 1,4-DCB
was removed in the BC column. Adsorption became the dominant removal mechanism in
the BAC column during the "shut off'. After restoring aerobic conditions and changing to
nonlabeled substrates, a measurable amount of 1“C activity was found in both columns in
the ﬁrst 6 hours. Then, the 14C02 production in the BC column dropped rapidly to
background levels. The 14C02 production in the BC column, however, was found for
more than two weeks leading to the conclusion that bioregeneration occurred in the BAC

column.

The BAC systems are expected to remove more non-biodegradable but adsorptive
compounds, in a mixture of non-biodegradable and biodegradable chemicals than a strictly
adsorptive system. This effect is expected because biodegradation of some compounds can
reduce competition on the GAC active sites for adsorption. De Laat et al. (1985) studied the
treatment of two mixtures of salicylic acid with 4-chlorobenzoic acid and 4-nitrophenol
with 2-methyl 4,6-dinitrophenol in GAC adsorbers and BAC systems. Salicylic acid and 4-

nitrophenol were presented as the biodegradable compounds, while 4-chlorobenzoic acid

26

and 2-methyl 4,6—dinitrophenol were used as non-biodegradable compounds. The two
chemicals from each of the mixtures showed competitive adsorption in the sterile GAC
adsorbers. When the mixture of salicylic acid and 4-chlorobenzoic acid was introduced into
a BAC ﬁlter, salicylic acid did not breakthrough and breakthrough of 4-chlorobenzoic acid
was signiﬁcantly delayed compared with the breakthrough curve in a GAC adsorber. The
amount of adsorbed 4-chlorobenzoic acid was close to the adsorptive value in a single
solute system. In the case of treatment of a mixture of 4-nitrophenol and 2—methyl 4,6-
dinitrophenol in the BAC reactor, the system performed as an adsorber initially for almost
45 days. When significant biodegradation of 4-nitrophenol occurred, the efﬂuent
concentration of 4-nitrophenol decreased dramatically to almost zero and the adsorption of
2-methyl 4,6-dinitrophenol increased. As a result, the concentration of 2-methyl 4,6-
dinitrophenol decreased in the efﬂuent. The concentration of 2-methyl 4,6—dinitrophenol in
the efﬂuent, however, rebounded to the higher level after 20 days. This can be explained
due to exhaustion of the adsorption capacity. The total amount of adsorbed 2-methyl 4,6-
dinitrophenol was signiﬁcantly higher than the value in the GAC adsorber with the mixture
and equated to the adsorbed amount in a single solute GAC adsorber.

DeWaters and DiGiano (1990) studied the inﬂuence of ozonated and non-ozonated
natural organic matter (N OM) on the removal of phenol in a BAC bed. The ozonated NOM
was less adsorptive and more biodegradable than non-ozonated N OM. The radio-labeled
phenol was fed to the reactors under three conditions. First, the phenol and ozonated N OM
were fed from start-up of the reactors (t = 0 hour). Second, the phenol and non-ozonated
NOM were fed after bioﬁlm development (t = 210 hours). Third, the phenol without NOM
was fed after start-up of the reactors (t = 0 hour). High 1“C02 production was found in the
presence of N OM because the N OM was believed to compete with phenol for the
adsorption sites on the activated carbon and more phenol was available for biodegradation.

There was less 1‘lCOz production when no NOM was introduced because the adsorbed

27

phenol was not directly available to the microorganisms. When the phenol was introduced

into the reactor with a developed bioﬁlm, l4C02 production was observed immediately.

Li and DiGiano (1983) determined the availability of sorbed substrate for
biodegradation on GAC using an inﬁnite batch recycle reactor. The four organic substrates,
o-cresol, acetophenone, phenol, and benzoic acid, were selected in their studies. Ottawa
silica sand, anthracite coal, and three different sizes of GAC (Filtrasorb 400) were used to
compare the effect of bioﬁlm carrier. The GAC particles were loaded with substrate and
inoculated with a tertiary culture before use in the reactor. Constant substrates
concentrations were maintained in the reactor by adding the appropriate amount of
substrates with time. They found the surface characteristics of the particle had little effect
on bioactivity, but higher biodegradation rates and speciﬁc growth rates were observed on
GAC compared to sand or coal. Therefore, the utilization of internally sorbed substrate
rather than better attachment on GAC may contribute to this enhancement. They also found
that the enhanced speciﬁc growth rate was increased further with an increase in sorbed

substrate concentration and decreased with an increase in particle size.

Several researchers found that there is, however, no evidence to prove that
bioenhancement occurs under steady-state conditions or when the adsorption capacity is
exhausted. Lowry and Burkhead (1980) compared ﬁxed-bed reactors charged with sand,
coal, and GAC for the treatment of hexanol and sodium acetate. There was no signiﬁcant
difference in the performance of the reactors under steady-state conditions. Peel and
Benedek (1983) reported the results from a comparison of efﬂuent data from adsorption
systems and data from a model prediction. They suggest that activated carbon does not
have an inherent bioenhancement capability. Zhang et al. (1991) reported that there were no
difference in terms of removal efﬁciency of phenol between BAC and biological aerated

ﬁlters after the start-up period.

28

Even though the bioenhancement claim seems unlikely for BAC systems, activated
carbon adsorption provides a very important mechanism for removal of non-biodegradable
or inhibited compounds. Wang et al. (1984) reported that adsorption was an essential
mechanism for removal of methquuinoline from a polycyclic N-aromatic compounds
mixture in an anaerobic activated carbon expanded-bed reactor. Suidan et al. (I 987) studied
the treatment of synthetically prepared coal gasiﬁcation wastewater in anaerobic activated
carbon expanded-bed systems. The phenols, primarily phenol and the Cz-phenols, account
for 60 to 80 percent of the COD in coal gasiﬁcation wastewater (Singer and Yen 1980).
The 0- and m-cresol were not degraded anaerobically and the high concentrations of those
compounds inhibited the biodegradation of the other aromatic compounds (Suidan et al.
1987). Methane production in the reactors decreased when breakthrough of o- and m-cresol
occurred. GAC replacement became necessary and an increase in methane production
followed immediately after replacement. Further studies to determine the relationship
between carbon replacement rate and reactor performance were conducted by Nakhla et al.
(1988) . They found that the soluble COD in the effluent decreased linearly with an increase
in GAC replacement rate. After a replacement rate was selected, the minimum required

volumetric COD loading rate was determined.

When loading transients occur in a BAC system, adsorption may serve to dampen
concentration changes. Chudyk and Snoeyink (1984) demonstrated that a BAC system
was able to smooth out the ISO-fold phenol pulses if bioregeneration had removed enough
adsorbed phenol. Lowry and Burkhead (1980), however, reported that the GAC carrier
was not able to provide better COD removal than other biological ﬁlters charged with sand
and coal when the systems were subjected to a two-fold sodium acetate (less favorable for

adsorption) step increase.

The adsorption capacity of GAC can decrease with an increase in usage of the BAC

system. This reduction could be due to adsorption of some compounds from the inﬂuent

29

but may also be caused by microorganism growth on the surface of the carbon. Suidan et
al. (1983) reported that the adsorption capacity of GAC decreased after the carbon was
exposed to a phenol mixture, which contained some compounds resistant to
biodegradation, in an anaerobic expanded-bed reactor for a period of time. The capacity
was, however, slightly regained after a period of bioregeneration. Zhang et al. (1991)
found that a similar phenomena occurred in an aerobic BAC ﬁlter to treat phenol. They did

not, however, observe the effect bioregeneration had on the reactor.

Weber and Ying (1978b) assert that breakthrough of toluene sulfonate in an
expanded-bed GAC adsorber with sucrose pre-saturated-biocoated activated carbon was
earlier than an adsorber with sucrose pre-saturated activated carbon without the biocoat.
When Schultz and Keinath (1984) conducted a study of powdered activated carbon
treatment (PACT) process mechanisms in an activated sludge system, they found that the
biomass in the PACT culture impedes the rate of mass transfer of phenol to the PAC

surface.

Olmstead (1989) studied the effect of microbial interference on GAC adsorption . In
his experiment, a column containing GAC particles was fed glucose for 1 -2 weeks. After
the bioﬁlm developed, the carbon was removed from the column in sections. Because of
the plug-ﬂow pattern in the column, more bioactivity and biomass were expected on the
inﬂuent end of the column. Biocoated GAC with different amount of biomass could be
obtained. For trichloroethylene (TCE), he found that the multiplier term in the Freundlich
isotherm equation, Kf, decreased with increasing biomass on the carbon while the slope
term of the equation, 11, remained constant. The adsorption of para-toluene sulfonate
(PT S), which has a much lower adsorption capacity than TCE, was not decreased
signiﬁcantly. The author was not able to isolate the effect of a sample protein (BAS) on the
adsorption of PT S, but, discovered that preloading soluble microbial products (SMP) on

the GAC reduced the amount of PTS adsorbed. Results from column experiments

30

investigating adsorption of dodecylbenzene sulfonate (DBS) conﬁrmed that a smaller

amount of DBS was adsorbed by the biocoated GAC.

2.9 Application of BAC processes
2.9.1 Drinking Water Treatment

The early stages of understanding BAC processes started in the drinking water
treatment ﬁeld. Researchers in Western Europe discovered and studied the beneﬁcial use of
the bacterial activity to extend the service life of activated carbon processes (Rice and
Robson 1982). The BAC process was combined with pretreatment (including clariﬁcation,
ﬁltration, and oxidation by ozonation) and post-treatment (mainly disinfection) to achieve

high quality drinking water.

Ozonation can oxidize organic matter to what is termed assembled organic carbon
(AOC). The adsorption capacity of GAC for AOC is less than the capacity for the natural
organic carbon (Glaze et al. 1986), but the biodegradability of AOC is higher (Somiya et al.
1986). Janssens et al. (1985) examined a GAC ﬁltration system in a water production
center. They compared four different processes using different ozone dosages (0, 2, 4, 8
mg/L). They found that ozonation could increase ﬁlter service time by about 60% to 65%
due to the signiﬁcant increase in AOC. Larger increases in AOC correspond to higher
dosages of ozone. They also observed that a dosage higher than 2 mg/L provided no

additional effect on total organic carbon removal.

The BAC process is often followed by a disinfection procedure to remove
microorganisms. Tobin et al. (1981) found two secondary pathogenic organisms
(Pseudomonas aeruginosa and F lavobacterium) in the efﬂuent of a point-of-use GAC

ﬁlter. Camper et al. (1985; 1986; 1987) studied the release of colonized granular activated

31

carbon particles from the GAC bed to the treated drinking water. They found that three
enteric pathogens Yersinia enterocolitica 0:8, Salmonella trphimurium, and enterotoxigenic
Escherichia coli were able to colonize sterile GAC. The heterotrophic plate count bacteria
were found in 41.4% of 201 GAC-treated drinking water samples. They also found that the
GAC supported greater numbers of the colifonn Kebsiella oxytoca than sand or anthracite.
Suffet (1980) suggested the use of backwashing to control the detachment of
microorganisms from the BAC process. LeChevallier et al. (1984), however, reported that
the microorganisms attached on GAC, which can wash out with the treated water, resisted
disinfection by chlorination. Treatment plant operators need to be aware that the

microorganisms could penetrate disinfection barriers and enter drinking water supplies.

2.9.2 Industry Wastewater Treatment

The complexity of the components in industrial wastewater is well known. Some of
the constituents may be biologically degradable, but others may be non-biodegradable and
may even be toxic to the microbial process. BAC systems which can provide adsorption
and biodegradation in an integrated reactor are appropriate candidates to treat such
complicated waste. BAC systems have demonstrated high removal efﬁciencies and stable
performance for the treatment of many wastes such as phenolic wastewater, slaughterhouse
wastewater, musty odor compounds, geosmin, textile wastewater, and metal-cutting-ﬂuids

wastewater. A summary of the remarks related to those systems is presented in Table 2-2.

32

Table 2-2. BAC processes for treatment of industrial wastewaters

 

 

Vjor '... .g'.- emov
‘ astewater Components Reactors (kg COD/m3-d) Efﬁciency
Catechol 200- two anaerobic expanded- 1 - 5 Catechol > ® "4'
1000 mg/L reactors in series (35 0C) (in the ﬁrst COD 93%,
reactor) 94%, 96% i

Phenols

 

 

 

 

Eemark: Adsorption removed majority of C01") in the ﬁrst 140 days with a
ioregeneration period thereafter (Suidan et al. 1981).

 

 
 
    
   
 
  

o-Cresol two anaerobic expanded-bed 1.9 - 2.2 i
256 mg/L reactors in series (35 0C) (in the ﬁrst Breakthrough _
reactor

 

 

 

emark: 50 % of o—Cresol breakthrough after200 days in the first reactor and l
40 days in the second reactor. Reactor performance was similar as adsorbers i
Suidan et al. 1981).

Phenol 360 - an anaerobic expanded-bed phenol 99.9 ,

 

 

 

 
 
  

3000 mg/L reactor (35 0C) 0.9 -7.2 COD 98%
emark: Bioregeneration was evident (Wang et al. 198g. 1
P no] 200- three anaerobic expanded- 1 - 5 phenol > ' t ;
1000 mg/L reactors in series (35 0C (in the fnst '
reactor) |

 

 

emark: Adsorption removed the most of COD in the ﬁrst 135 days and a
ioregeneration period was occurred thereafter (Khan et al. 1981).

 

Phenol fberl-saddle packed anaerobic 0.8 - 1.7 .
200-1000 mg/L; ﬁlter and an anaerobic GAC (in the ﬁrst COD > 90% 5

 

 

 

 

 

 

 

Coal Cresols expanded-bed reactor in reactor)
= iﬁcatio 170-850 mg/L series (35 °C)
emark: No signiﬁcant removal occurred in the berl-saddle reactor and GAC ‘
replacement was necessary to prevent microbial toxic levels of some componen --
(Suidan et al. 1983; Suidan et al. 1987). l
Indole 50-300 COD > 95% 1
mg/L; anaerobic expanded-bed No target .
Polycylic Quinoline 50- reactor (35 °C) 0.3 - 2.2 compounds
N—aromatic 300 mg/L ere detected ' ‘
fompounds Methquuinoline the efﬂuent.
50-300 mg/L
IRemark: Bioregeneration was evident. Methy quinoline was removed strictIy
byadsorption (Wang et al. 1984). ’
2-methyl-
isobomeol
(MIB) 1.6.mg/L batch ﬁlter 2.7 99 %

 

.I usty odor geosrrnn

 

 

 

 

1.5 mg/L
I ' emark: More than 50% of substrates were biodegraded in the seeded GAC

nrlters, but there was no signiﬁcant removal in the seed sand ﬁlter after 43-54 .
. - . volumes. The seed time wasone da Ya 1 et 1988) ’

 

CHAPTER 3

COMPARISON OF BIOFILM CARRIERS IN FLUIDIZED BED REACTOR SYSTEMS

3.1 Introduction

The ever increasing presence of volatile organic compounds (VOCs) in water
supply wells has led to the realization that groundwater can no longer be considered as an
inherently pristine source of drinking water. Contamination by VOCs can be caused by
spills of hazardous materials and leaks from underground storage tanks and pipelines.
Because many people depend on groundwater as the sole source of their potable water

supply, the degradation of these water reserves has serious public health ramiﬁcations.

One of the most widespread contamination problems is the release of gasoline and
other petroleum fuels into the subsurface. The constituents of principle concern are the
aromatic components, benzene, toluene, and xylenes (BTX), which are less strongly
adsorbed to the soil matrix than the aliphatic components in fuel. The BTX components
are, therefore, more mobile in the soil and therefore more likely to contaminant water

supplies.

The most widely used remediation techniques, liquid-phase adsorption using
granular activated carbon (GAC) and air stripping, simply transfer contaminants from one
phase to another (McCarty 1983; Voice 1989). Further treatment or disposal of the
receiving phase is required. Biological treatment would appear to be a desirable alternative
to such concentration techniques because it has the potential to completely destroy the

contaminant compounds and it is generally less expensive than physical-chemical treatment

33

34

processes. Biological treatment has not been widely accepted for groundwater treatment,
however, due to the widespread impression that biological systems are not sufﬁciently
stable to consistently meet the stringent discharge limitations that are often required. In
cases where biological treatment has been employed, it is typically followed by GAC
adsorption for efﬂuent polishing and to provide back-up treatment in the event of failure of

the biological system.

A promising new approach, termed biological activated carbon (BAC), integrates
biological removal and granular activated carbon adsorption into a single unit process. The
beneﬁcial aspects of integrated biodegradation/adsorption systems were reported in the
early work on the use of GAC for secondary and tertiary treatment of municipal wastewater
(Weber et al. 1970). Researchers observed that adsorption columns continued to effectively
remove organic material far beyond the point at which adsorption capacity would normally
be exhausted (Bouwer and McCarty 1982; Kim et al. 1986; Rittmann 1987; Gardner et al.
1988; Speitel et al. 1989a; 1989b; 1989c). This removal has been shown to be the result of
growth of microorganisms on the surface of the GAC particles, and the subsequent
biodegradation of the waste constituents by these communities. In some systems this
phenomenon has been regarded as a fortuitous beneﬁt because less frequent carbon
regeneration is required. In others, however, it has proven to be problematic because the
biomass can grow to the point where it interferes with the hydraulic operation of the system
and it may impede the adsorption of non-degradable compounds. Little attention has been
given in the literature to understanding how adsorption and biode gradation work together in
biological activated carbon systems to affect both efﬂuent quality and system stability. In
addition, BAC has only recently been considered as a potential treatment process for
groundwater contaminated with relatively low levels of toxic materials, such as petroleum

hydrocarbons or chlorinated solvents (Hickey et al. 1990b).

35

In this research an integrated BAC system consisting of a biological ﬂuidized bed
reactor employing granular activated carbon as a bioﬁlm support (BAC) was evaluated and
compared to identical systems with adsorptive removal only (GAC) and biological removal
only (FBR) for the remediation of groundwater contaminated with the gasoline
constituents, benzene, toluene, and xylene. It has been hypothesized that BAC systems
provide for faster start-up times and better protection against shock-loads than similar
systems with only biological removal, while preserving the beneﬁts of low operating costs
and pollutant destruction (Hickey 1990a). These hypotheses were evaluated by comparing
the breakthrough proﬁles, steady-state removals, and system responses to step increases in

applied organic loading rates for the three systems.

3.2 Materials and Methods

3.2.1 Media and Chemicals

Two different support media, GAC and a non—adsorbent carbon material (non-
activated carbon) were used in this study. The GAC used was Calgon Filtrasorb 400
(Calgon Co., Pittsburgh, PA). Non-activated carbon, termed "baker product" by the
manufacturer, is the same material as the GAC that it has not undergone the activation step
in the manufacturing process and has therefore little adsorption capacity. Both media were
sieved to obtain a 20 x 30 mesh fraction, resulting in an average particle diameter of 0.75
mm. After sieving, the GAC and baker product were rinsed with distilled water to remove
oil and ﬁnes and dried overnight at 100 0C. The media were then stored in sealed

containers until needed.

Benzene, toluene, and p-xylene (BTX) were obtained from except Chemical Co.,
Milwaukee, WI. Ammonium chloride and potassium phosphate dibasic were obtained from
J. T. Baker Chemical Co., Phillipsburg, NJ. All chemicals were reagent grade. The water

36

supplied to Michigan State University was used directly as a source of groundwater. This
relatively hard water (450 mg/L as CaCOg) is pumped from a deep aquifer underlying the

campus and distributed without further treatment.

3.2.2 Determination of Adsorptive Capacity of BTX

Isotherms for each BTX compound were performed using serum bottles (160 ml)
sealed with Teﬂon-coated septa. Activated carbon (10 mg, 20 mg, and 30 mg for benzene,
toluene, and xylene, respectively) was added to oven dried serum bottles, which were then
capped and sterilized. The serum bottles were subsequently ﬁlled, without leaving a head
space, with water (pH = 7.0) containing the desired concentration of BTX. Ten bottles,
each with different initial concentrations between 1 and 20 mg/L, were prepared for each
compound. The bottles were tumbled at 6' rpm and 15 0C for 7 days to ensure that
equilibrium was obtained. The liquid-phase concentration was then analyzed to determine
the residual amount of BTX. The solid-phase concentration was calculated by difference.
An isotherm was also performed for xylene with non-activated carbon using the same

procedure.

3.2.3 Fluidized Bed Reactors

The laboratory-scale ﬂuidized bed reactors used in this study were glass columns
with a 2.5 cm diameter and 92 cm height (Figure 3-1). All reactors were operated as one-
pass systems with no recycle. The GAC-FBR and FBR (biological only) systems were
charged with activated carbon and non-activated carbon, respectively, and _ W
ng'xed culture. This culture was taken from a wpilot-scale ﬂuidized bed system that was
originally seeded with activated sludge and was subsequently supplied with BTX as the _

sole carbonsource. The reactors were provided with sufﬁcient dissolved oxygen and

37

essential nutrients )9 encomage_bioﬂrlr_n__forr_njrgpg and permit complete biglogicaloxjdatior}
of the BTX. The feed water was oxygenated to a concentration sufﬁcient to maintain an
efﬂuent D0 of more than 2.0 mg/L using pure oxygen and was supplemented with nitrogen
and phosphorous at a stoichiometric ratio of 100/5/1:COD/N/P. A mixture of benzene,
toluene, and xylene, in 1:1:1 volumetric ratio, was injected into this stream via a syringe
pump and dissolved using an mixer. A recycle loop around the mixer was employed to
create turbulerrtjlgy inside the mixer and help ensure complete BTX dissolution. This
BTX contaminated water was then pumped to the bottom of the reactors using_peristaltic
pumps with variable speed drives. To limit microbial growth, the GAC system (adsorption
additional nitrogen or phosphate. During the course of this experiment the in-line water

temperature averaged 15 0C :1: 1.0 °C.

GAC

H

-—> Efﬂuent

Efﬂuent Sampling
Point

 
   
 
    
  
 

Granular
Activated
Carbon

Inﬂuent
Sampling
Point

 

 

Groundwater

BTX Injection
Syringe Pump

Inﬂuent Injection Unit

Mixing Pump

Figure 3-1. Schematic of ﬂuidized bed reactor systems.

I I

 

38

GAIL-EBB

——> Efﬂuent

Efﬂuent Sampling Point

 

  

’
Inline Mixer

Granular
Activated
Carbon

Granular
Baker
Product

Inﬂuent Sampling Point

)’\

 

 

_/ \—

—‘—> Efﬂuent

 

 

 
 
   
   

Tap Water

Inﬂuent
Reservoir

 

'1

Nutrient Addition
—>

__>=

Overﬂow to

Waste
—>

02

 

 

 

 

ﬁ-

BTX Injection

          
   
  

  

39

Breakthrough proﬁles were obtained for the three reactor types (GAC, GAC-FBR ,
and FBR) using a 500 ml working volume e(92 cm (working height) and a 180 ml initial
charge of media, which produced a settled b:d(lheightqofar3proximately 33 cm. BTX
contaminated water was fed at a ﬂow rate ofl140;@ to each reactor during 2111: period,
resulting in a hydraulic ﬂux rate of 0.29 mil/min-m2 and an empty bed hydraulic retention

QTM/pur,

time 0503.6 min. Thginitialﬂuidized bed height for the three reactors was Mn. After
steady-stale; soliditions were reached in {BEETLE with biological activity, higher loading
rates were applied to these reactors by increasing BTX concentration 1.5-fold and
increasing the ﬂow rate to 300 ml/min. The working volume of each reactor was \. f7
concurrently increased to 1000 ml by adding another column unit and subsequently x! c
increasing the w9r_king_height§ tom This resulted in a hydraulic ﬂux rate of 0.61 /
m3/min-m2 and hydraulic retention :time of 3.3 min. Step load increases in BTX

concentration (ca. 3- and 5-fold) were applied to the two biological reactors (GAC-FBR

and FBR) operating at two different steady-state loading rates.

3.2.4 Analytical Methods

Aqueous BTX concentrations were determined using an grtomated headspaceﬁ
sampler (Perkin-Elmer, model HS-101) coupled to a gas chromatograph (Perkin-Elmer
model 8700) equipped with a ﬂame ionization detector (FID) using helium as a carrier gas.
Samples collected from the ﬂuidized bed reactors (5 ml or 10 ml) were transferred to 203$
glass vials with Teﬂon-coated septa_and aluminum seals. The samples were equilibrated for

1 hour at 80 09 and an aliquot of the headspace gas was injected onto a 30 m x 0.53 mm
W4 column (J & W Scientiﬁc). The detection limit of this method was 1 u for each

..... EV
BTX component. The—accuracy for measurements of a sample containing a concentration of

1.5 ug/l was 10.1 jig/l. q.

4O

Inﬂuent and efﬂuent dissolved oxygon concentrations were analyzed using an
glarographic electrode ooopleito a digital pI-l/millivolt meter. If the DO concentration
exceeded the limit of the instrument (15 mg/L), the sample was diluted to the measurable
range using water of a known (low) DO concentration. DO consumption was calculated as

the difference between inlet and efﬂuent DO.

3.2.5 Scanning Electron Microscopy

Samples of both virgin and bioﬁlm coated activated carbon and non-activated
carbon were ﬁxed at 4 0C for 1-2 hours in 4% Jghrtaraldehyde buffered with 0.1 m sodium
phosphate (pH 7.4). Following a brief rinse in the buffer, samples were dehydrated in a
ethanol sorjos (25%, 50%, 75%, 95%) for 10-15 min at each gradation followed by three

 

10 min changes in 100% ethanol. ABalzersgjtical MM, using liquid carbon dioxide
as the transitional ﬂuid, was used to dry the samples. Samples were sputter coated with
gold to a thickness of 20 nm (Emscope Sputter Coater model SC 500), purged with argon
gas and examined via scanning electron microscopy (Japan Electron Optics Ltd., model

JSM-35CF).

3.3 Results
3.3.1 Adsorption Characteristics Of BTX

The single solute adsorption isotherm of each BTX compound can be related
through the Freundlich isotherm.
q. =Kfo <3-1)
where q, = mass solute adsorbed/mass adsorbent

C6 = concentration of solute in solution, mass/volume

41

Kr, 1 = empirical parameters representing the sorption capacity and non-linearity,

respectively.

The experimental data was ﬁt to the Freundlich equation using non-linear regression
analysis, resulting in the parameter estimates shown in Table 3-1. The KP values of BTX
were observed to follow the expected order: xylene > toluene > benzene. This means that
the amount of xylene, adsorbed per unit mass of activated carbon is greater than that of
toluene which in turn is greater than that of benzene for a given aqueous concentration. The
KF value for xyl/ene with the non-activated carbon at 15 0C was (2,223“ of that found for
GAC, thereby substantiating the assumption that this material had negligible adsorption
capacity. The adsorption capacity of baker product for benzene and toluene was not

measured, but it can be assumed that it is also low relative to GAC.

Ma) 4'; €79; //’"”*‘/ .3
Table 3-1. Adsorption characteristics of BTX

emperature
oc , Coefﬁcient

 

* (mg/g)(L/rn8)“n
Note: BTX: Benzene, Toluene and p-Xylene

3.3.2 Breakthrough Proﬁles

Start-up and breakthrough proﬁles for the three systems are shown in Figures 3-2
and 3-3. The initial ﬂuidized bed height for the all reactors was 43 cm. The concentration of

42

each BTX component in the inlet was maintained at approximately 1 mg/L during this

period.

In the GAC system, where adsorption was the only removal mechanism,
substantial breakthrough of benzene and toluene occurred after 160 and 460 hours,
respectively. Xylene concentration in the efﬂuent increased slowly up to 0.3 mg/L by 800
hours. These breakthrough proﬁles indicate that the GAC unit can be treated as upﬂow
ﬁxed bed reactor with essentially plug-ﬂow liquid transport conditions. This is because
liquid-solid, two-phase ﬂuidized beds become stratiﬁed according to particle size and
particle motion in the GAC unit can be assumed to be conﬁned to reasonably deﬁned
patterns of localized movement. To predict the time required for substantial breakthrough of
BTX, an equilibrium column model (ECM), which neglects mass transfer resistance and
uses ideal adsorbed solution theory to include the competitive effects in multicomponent
mixtures, was used. The ECM has been shown to be able to calculate the elutEn/gder of
the adsorbates in multicomponent mixture using single solute isotherms (Crittenden et al.
1987). The predicted and actual breakthrough times for BTX are presented in Table 3-2.

Differences between measured and predicted breakthrough times are likely due to

competitive interactions among adsorbates and non-ideal adsorption conditions.

Table 3-2. Substantial breakthrough time for each BTX compound in the GAC system

Breakthrough Trme

Compounds GAC system Predicted using Predicted based on
(hrs) ECM (hrs) single solute

_ isotherm (hrs)
Benzene 218 286
Toluene 585 790

. X lene

 

 

 

 

 

 

 

 

Note: GAC: Granular Activated Carbon;
ECM: Equilibrium Column Model

43

 

 

 

 

 

3
‘ +oAc
‘GAC-FBR
2-
o
S
{- ‘I'
+ T 4 4|.
0 + f +
m1- *‘ 4 ‘¢++"' a": ‘9 *1-6-4-
‘ + I -r I ”*4. 1-
‘ ﬁ“;
04144:. MW
in 3
a I GAC
z " GAC-FBR
o
n o ‘
E
5 "o' "
U I—
g 1-1 * + '1' +9".
U 4‘.’ -Il-++‘_ {Iii-F"
8 +2“
:3 r 4
E O‘Wd " '
m 3
+ GAC
a ‘ GAC-FBR
8 2'
2
>‘ d
if"
Q11-
‘0'?
0 lillll II‘Illltl Ilhﬁtﬂ’t M‘ﬂttttﬂ “Ti". . ‘

Time (hours)

Figure 3-2. Comparison of breakthrough proﬁles for GAC and GAC-FBR systems.
(hydraulic ﬂux rate: 0.29 m3 /rr1in-m2 ; empty bed hydraulic retention time: 3.6 min;
inﬂuent concentration of each BTX component in the inﬂuent was maintained at ca. 1

mg/l.)

 

 

 

 

 

A GAC-FBR
- I FBR
a ' . - ' -
1. j I - I
I. ' I. ' . I
A
ll ‘ “ I. -
‘ ‘ A I -- I
o irﬁl—r—H—q—H—W
g 3
E A GAC-FBR
v - I FBR
8
.3 g 2 ..
S “:1
5 E " -
a I ' I I
o 1- - ' - - _
U . ' . I I
E? . ' - -
g I
{WWW
m ‘3’
A GAC-FBR
- - FBR
8 2 ‘
2
>5
X 4
5- I I. '
1 " I l I . _
I ' ' _ I
_ _ . I I I
o {WM—WW ‘
0 100 200 300 400 500 600 700 800
Time(hours)

Figure 3-3. Comparison of breakthrough proﬁles for GAC-FBR and FBR systems.
(hydraulic ﬂux rate: 0.29 m3 lmin-m2 ; empty bed hydraulic retention time: 3.6 min; .
inﬂuent concentration of each BTX component in the inﬂuent was maintained at ca. 1

Ing/1-)

 

45

In the FBR system, where biodegradation was the only signiﬁcant mechanism,
breakthrough of BTX occurred soon after introduction of the inﬂuent (Figure 3-3). After
approximately 400 hours, sufﬁcient biomass had developed on the surface of the non-
activated carbon so that biodegradation became signiﬁcant and the concentrations of BTX
in the efﬂuent began to decrease. At the same time, the bed height and DO consumption

began to increase.

The GAC-FBR system, where both adsorption and biodegradation are possible,
initially performed much like the GAC system. Efﬂuent benzene concentrations increased
up to 50% of the in-line levels by 140 hours, in the same manner as the GAC system. After
this initial period, however, the GAC-FBR system performed quite differently: the
benzene concentration began to decrease, while the DO consumption and bed height started
to increase. Low concentrations of toluene and xylene were also detected in the efﬂuent
during this start-up period, but these also decreased with the onset of biological activity.
These compounds were detected in the efﬂuent slightly earlier than in the GAC system.
This may be a result of particle motion changing from well-deﬁned patterns of localized
movement at the beginning of operation to more random patterns as surface growth
developed and the seeded particles became more buoyant and started migrating in the
reactor (Shieh and Keenan 1986). This type of movement would serve to increase the
amount of mixing in the reactor which would produce faster breakthrough. An alternative
explanation is that the developing bioﬁlm interferes with adsorption, again resulting in

earlier breakthrough.

Even though the same amount of inoculum was added to both the GAC-FBR and
FBR systems, the time required until noticeable biodegradation commenced was
signiﬁcantly less in the GAC—FBR reactor. Previous reports have attributed this to the fact
that the activated carbon surface has a large number of crevasses and ridges that provide a

sheltered environment for colonization. It can be seen from Figure 3-4, however, that the

46

non-activated carbon used in this study has a very similar surface texture. This suggests
that some other property of the activated carbon, such as the adsorption capacity, is

responsible for faster colonization rates.

A comparison of the biological systems during start-up, that is until a constant
biomass develops and steady-state operating conditions are reached, reveals that the GAC-
FBR system produced a higher quality efﬂuent and allowed considerably less BTX to leave
the reactor untreated than did the FBR system. The mass of each BTX compound which
left the system untreated was determined by calculating the area under the efﬂuent
concentration curves using the trapezoidal rule (Burden et al. 1981). These results, which
are presented in Table 3-3, show that the FBR system allowed approximately 15 times

more BTX to leave the system untreated than was observed for the GAC-FBR system.

Table 3-3. Summary of performance of FBRs during the start-up period

I Aumcc *— ated N vgeera co.
Compounds in efﬂuent in efﬂuent

(mg) tug/1)
Benzene 349 83
Toluene 61 15
Xylene 75 18
Benzene 2925 696
Toluene 2388 569
X lene 5 12

 

 

 

 

 

 

 

 

 

 

 

 

47

(b) 4 C L. 1 l l_‘.l .- "

 

Figure 3-4. SEM of a particle without biological growth. (a) activated carbon, x 66 and (b)
non-activated carbon (baker product), x 66.

48
3.3.3 Comparison of Steady State BTX Removal Between GAC-FBR and FBR Systems

The reactors were assumed to be operating under steady- state conditions when
constant DO consumption, consistent with that expected to be required for degradation of
the added BTX, was found in the efﬂuent. In this initial test, the average total COD in the
in-line to FBR was calculated to be 7.25 mg/L, or an applied organic loading rate of 2.9 kg
COD/m3-day. More than 90 percent of the BTX (94% for benzene and toluene and 90%
for xylene) was removed (Table 3-4). The average DO consumed was 4.6 mg/L. The
average total COD in the inlet to GAC-FBR system was somewhat higher, 9.6 mg COD/1
or 3.8 kg COD/m3-day. Despite the higher loading rate to the GAC-FBR system, greater
substrate removal was observed: 99% for benzene and toluene and 92% for xylene. The

average DO consumed was 6.1 mg/L.

Performance of the two systems at a higher loading rate was also examined. With a
total COD loading to both reactors of 6.0 kg COD/m3-day, the two systems performed
comparably. As shown in Table 3-5, removal rates averaged 94%, 90% and 81% for
benzene, toluene, and xylene. The GAC-FBR system was found to consume slightly more

dissolved oxygen.

Table 34. Comparison of steady state BTX removals in GAC-FBR and FBR systems

Compound Inﬂuent Efﬂuent Removal
(pg/L) (ug/L) (7°)

Benzene 1066 (21303) 3 (E) 99.7
Toluene 1023 (21340) 5 (5) 99.6

Xylene 996 (:t340) 81 (21:53) 92.0
Benzene 834 (21:151) 45 (:40) 94.6
Toluene 758 (1:380) 40 (i40) 94.7
X lene 732 ($340) 68 ($42) 90.7

 

 

 

 

 

 

 

 

 

 

 

 

 

49
3.3.4 Response of FBR Reactors to Step Load Increases in BTX Concentration

The GAC-FBR and FBR systems were subjected to several step-load increases to
determine whether the use of an adsorptive biomass carrier contributed to system stability.
A 96-hour, three-fold step-load increase from a base organic loading rate of approximately
3.0 kg COD/m3-day was introduced to the two FBRs by increasing the BTX concentration
in the in-line, while keeping the ﬂow constant. BTX concentrations in the in-line and
efﬂuent are presented in Figures 3-5 and 3-6 for the FBR and GAC-FBR systems,
respectively. The FBR system showed an immediate response to the step load increase
(Figure 3-5). Benzene and toluene concentrations increased from 40 to 200 jig/l and xylene
increased from 70 to 1000 jig/1. DO consumed increased from 4.6 to 13 mg/L within one
hour. After initiating a similar increase to the GAC-FBR system, efﬂuent concentrations of
the BTX remained at nearly identical levels to those observed prior to the increase. DO
consumption increased from 6 to 13 mg/L within one hour. After the loading rate was
returned to the base loading level, DO consumption remained at a level higher than that
observed during the period prior to the increase. Consumption of DO thereafter decreased

slowly to the previous steady-state value.

 

a InfBenzene I *‘ I
O EffBenzene '-

U)
r
I
I
I
I

I
'11
_'_..__1

c

p

r

>

r

>

r

p

P
f ﬁa—
Q

9

o

W
>

 

 

 

 

 

 

 

 

 

 

 

 

A 4 I InfToluene I I _ I h
3 3i ‘ EffToluene I ' L
g o l I
a r: I J! I
,o 8 2' I I;
g "0‘ II
I- I
§ 1- .. I .. ' w u
t: I 0 ‘ ’
5 0 - --,-4a¢- -, Jr,
4 . Inf p—Xylene 1" .I a a _ I
3 _ ‘ Effp-Xylene | 9 " f
g I I r
.9.
>. 2- I
5‘: 1.. I
n 4
1 d l -I I. I: O o . F.
JL—‘hlwr—K—PA A . A
0 j ‘ f I
800 850 900 950 1000 1050 1 100
Time(hours)

Figure 3-5. The response of FBR system to a 96 hour, 3-fold step load increase from
steady state loading level of 2.9 kg COD/m3 -day. The load increase started at hour number

929 and ended at hour number 1025. (hydraulic ﬂux rate: 0.29 m3/min-m2; hydraulic
retention time: 3.6 min)

Benzene

Toluene

Concentration (mg/L)

p-Xylene

 

4 IE I
I InfBenzene " I I
3‘ o EffBenzene % F I 'J.
EI- II I I
2'1
_ I I
I I .

1‘

I
L AA AA AA A A
I I I

 

 

 

 

 

 

 

0
4 F
I InfToluene I a I a a .l
3 . ‘ Eff Toluene -" 5 1
I a I
2‘ r a I
I -. I
1' ' ' a 1 to I
o-pu—Aﬁrhmﬁ—rA—b-M—ﬂ—y—hL—rﬁ—
4 I I
I Infp-Xylene " I
3‘ ‘ Effp-Xylene I. " I I. t
I
4 £51 D'- T
2 I
. - mi" . I
. I
1- I 3 T :- .
' n
0 A '. .W A
800 850 900 950 1000 1050 1100

 

Time (hours)

Figure 3- 6. The response of GAC-FBR system to a 96 hour, 3-fold step load increase
from steady state loading level of 3.8 kg COD/m3-day. The load increase started at hour

number 929 and ended at hour number 1025. (hydraulic flux rate: 0.29 m3/min-m2;
hydraulic retention time: 3.6 min)

52

Because the performance of neither biological system was seriously perturbed by
the 3-fold step-load increase, the reactors were subsequently subjected to a ﬁve-fold step
load increase from a higher base loading level (6.0 kg COD/m3-day). Step-load increases
of this magnitude were conducted for two different durations, 4 and 8 hours. These results,
shown in Table 3-5, show that during the increases, BTX removal and DO consumption in
the GAC-FBR system were higher than in the FBR system. The higher DO consumption
indicates enhanced bioactivity, which may result from a greater amount of biomass or more
active cells in the GAC-FBR system. While more dissolved oxygen was consumed, the
ratio of DO consumed to COD removed for the GAC-FBR system was lower than for the
FBR system (Table 3-5). This indicates that some of COD (BTX) was being removed

concurrently via adsorption.

Table 3-5. Summary of FBR responses to step-load increases from base loading level of
6.0 kg COD/m3-day

State Inﬂuent . Efﬂuent Subst. IX) 99mm
(pg/L) (pg/L) C0D

(pg/L) (mg/L) removed

 

53

3.3.5 Extent of Surface Coverage of the Media by the Bioﬁlms

Microbial attachment to the external surface of GAC has been shown using
scanning electron microscopy (Weber et al. 1978a; Pirbazari et al. 1990). These studies
employed completely mixed batch reactors to contact carbon with chemically coagulated
and settled sewage for up to 17 days. Under these conditions, a contiguous, uniform
bioﬁlm over the entire surface area was not formed. Localized concentrations of bacteria
were observed to occur in crevices and areas which are sheltered from ﬂuid shear forces.
During this current study, scanning electron microscopy (SEM) was used to examine the
extent of bioﬁlm coverage on the GAC and baker product. Samples of GAC that had been
taken from the ﬂuidized bed reactors while they were operated under steady state conditions
are shown in Figures 3-7. Contiguous bioﬁlms were observed on the support media in
both biological systems. These ﬁlms were quite thick (100 to 200 um) on both the activated
and non-activated carbons. The organisms observed on the surface of both carriers had

similar morphologies.

54

 

BU [£095

Figure 3—7. SEM of a GAC with biological growth. (a) x 48 and (b) x 1200.

55
3.4 Discussion

A comparison of the performance of the three reactor systems during start-up
reveals one of the most distinctive characteristics of GAC-FBR systems. The adsorption
capacity of the biomass support serves to protect efﬂuent quality until the biodegradation
capability of the system is established. Only benzene was released at significant
concentrations, and this proved to be short-lived. Since activated carbon has a relatively
low capacity for benzene, the amount of carbon in these systems was not sufﬁcient to
adsorb the total benzene loading during the time required to develop a bioﬁlm. GAC has a
higher adsorption capacity for toluene and xylene, so the system had the ability to remove
these compounds during the start-up period. The total mass of toluene and xylene released
during start-up was 30-40 times greater in the FBR system than in the GAC-FBR system.
It thus appears that efﬂuent protection during start-up is a straightforward design
consideration involving the amount of carbon in the reactor, the loading rate, the time
required to develop a bioﬁlm, and the adsorption capacity of the carbon for the compounds
of interest. A second start-up characteristic of GAC-FBR is that these systems developed a
biofilm and began degrading BTX faster than the biological system employing non-
adsorbing support media. This is supported by visual observation of the bioﬁlm, and data
collected on oxygen consumption and BTX in the efﬂuent. This is consistent with previous
findings showing that GAC particles support richer microbial communities than other
materials (Pirbazari et al. 1990). It has been hypothesized that this results from the ability
of the carbon to concentrate chemicals necessary for microbial growth while providing an
environment that is well protected from ﬂuid shear forces for cell attachment. Since the
activated and non-activated carbons were observed to have similar surface textures, it is

unlikely that the differences can be explained by shear-force protection alone.

Under steady-state conditions at an organic loading rate of 3.8 kg COD/m3-day and

a hydraulic retention time of 3.6 minutes, the GAC-FBR system produced efﬂuent benzene

56

and toluene concentrations of 3 and 5 ug/l, respectively. The FBR system, operating at a
somewhat lower loading rate of 2.9 kg COD/m3-day had efﬂuent benzene and toluene
levels of 45 and 40 ug/l. Karlson et al. (Karlson and Frankenberger 1989) conducted a
batch assay using a mixed bacterial culture capable of utilizing gasoline as a sole carbon
source. They reported that 22-35 ug/l of benzene did not provide enough carbon source to
sustain an active bacterial population. Higher BTX concentrations were not degraded to
levels below the 20-40 ug/l concentration range. One possible explanation for the low
efﬂuent concentrations in the GAC-FBR system is that the adsorption capability of the
activated carbon may serve to enrich substrate and oxygen levels, thereby allowing the
bacterial population to utilize BTX despite the low concentrations in the bulk solution.
Alternatively, the low efﬂuent levels reached in the GAC-FBR systems may result from
continued efﬂuent polishing via adsorption. At the higher loading rate of 6.0 of kg
COD/m3-day performance of the GAC-FBR and FBR systems were comparable. Efﬂuent
concentrations in both systems were above the critical minimum range of 20-40 ug/l
discussed above. At this loading rate it appears that adsorptivity of the biomass support has
little effect on steady-state performance and the systems are dominated by bioﬁlm

pl'OCCSSCS .

The abilities of the GAC-FBR and FBR systems to respond to in-line shock loads,
as simulated by the step concentration increase, demonstrates another distinctive feature of
systems with adsorptive biomass supports. Efﬂuent quality during the step increase was
better for the GAC-FBR system in all cases. During the highest loading rate applied in the
step increase experiments, the GAC-FBR system was found to remove 82, 74 and 64

percent of the benzene, toluene and xylene applied, compared to 65, 56, and 46 percent in
the FBR system.

It was previously suggested that the superior performance of the GAC-FBR

system during the step increase could be attributed to adsorption. This conclusion is

57

supported by the lower ratios of DO consumed to COD removed in the GAC and the
implied assumption that the observed differences could not be explained entirely by
changes in metabolic efﬁciency. If we assume that efﬁciency remains constant at the value
found under steady-state conditions (where it is assumed that only biodegradation is
active), we can calculate the amounts of COD removed by biodegradation and by
adsorption during the step increase. This analysis is presented in Table 3-6. From this data
it appears that biodegradation is essentially the same in both systems and the lower efﬂuent

levels in the GAC-FBR system are attributable to additional removal by adsorption.

Table 3-6. Summary of FBR performance during step load increases

State

4 hrs
load

 

* Amount of COD removed by biodegradation equals the DO consumed divided by 0.68.
1' Amount of COD removed by adsorption equals the total COD removal minus the amount
of COD removed by biodegradation.

1 Amount of COD removed by biodegradation equals the total COD removal.

Elevated levels of DO consumption following a step increase and subsequent return
to original loading levels were observed in the GAC-FBR but not in the FBR systems.
This is strong circumstantial evidence for bioregeneration of adsorption sites on the GAC.

The BTX adsorbed during the load increase were apparently desorbed from the carbon to

58

the bioﬁlm when bulk aqueous phase concentration decreased. Such a conclusion is also
consistent with the observation that there was slightly more DO consumption under the
initial steady-state conditions, since this may have resulted from desorption and degradation

of the BTX adsorbed during start-up.

3.5 Conclusions

This study demonstrates that the use of activated carbon as a biomass carrier in
ﬂuidized bed reactors produces a system in which both adsorption and biodegradation
affect substrate removal. To a large extent, the performance of such biological activated
carbon systems is controlled by the additive contributions of these two removal
mechanisms. During the start-up period, before a fully functional biomass has developed,
the substrate is removed primarily by adsorption. After the biomass is established and
steady-state conditions are reached, system performance is dominated by biodegradation.
The system retains adsorption capacity under steady state conditions, however, as
evidenced by the response of the system to in-line concentration shocks. Lower efﬂuent
concentration levels were found in GAC-FBR systems during a step concentration increase
than in similar systems using a non-adsorbing biomass carrier. The data suggest that this
results from adsorption of a portion of the increase, and that this material can be
subsequently desorbed and biodegraded when the inlet concentration returns to pre-shock

levels.

The combination of adsorption and biodegradation in a single reactor system also
produced two synergistic effects. First, biomass developed more quickly on the surface of
activated carbon than it did on non-activated carbon with a similar surface texture. Second,
higher steady-state substrate removal was observed in the GAC-FBR system at lower
loading levels (3.8 kg COD/m3-day). A potential explanation for both of these observations

is that the adsorptivity of the activated carbon surface serves to concentrate substances

59

including the substrates, nutrients, and oxygen on the carrier surface. This concentration
effect may promote more rapid colonization and may allow degradation to occur when the

substrate concentrations in the bulk solution are too low to support growth.

CHAPTER 4

SINGLE SUBSTRATE STEP-LOADING RATE INCREASES

4.1 Introduction

The beneﬁcial aspects of integrated biodegradation/adsorption systems were
reported in the early work on the use of GAC for secondary and tertiary treatment of
municipal wastewater (Weber et al. 1970). In systems designed as adsorbers, it has been
shown that biodegradation increases the period between GAC regeneration cycles
(Bouwer and McCarty 1982; Chudyk and Snoeyink 1984; De Laat et al. 1985; Kim et al.
1986; Gardner et al. 1988; Speitel et al. 1989a; 1989b; 1989c). In bioﬁlm systems, the
use of GAC as a biomass carrier has been shown to provide for removal of compounds
resistant to biodegradation (Suidan et al. 1983; Suidan et al. 1987). In a treatment process
for groundwater contaminated with relatively low levels of volatile organic compounds, a
biological activated carbon ﬂuidized-bed reactor (GAC-FBR) system has been
demonstrated to provide both the efﬁciency of biological removal and the positive

efﬂuent protection capability of activated carbon adsorption (Hickey et al. 1990b; 1991a;
Voice etal. 1992),

In the previous chapter, the performance of adsorptive and non-adsorptive bioﬁlm
carriers in biological ﬂuidized-bed reactors (FBR) was compared for the treatment of
groundwater contaminated with benzene, toluene and p-xylene (Voice et al. 1992). The
results demonstrated that the GAC-FBR system functioned primarily as a biological

reactors under steady-state conditions, but that the GAC carrier contributed signiﬁcantly

60

61

to overall removal during transient conditions. This study extends the previous work by
characterizing the interactions (such as inhibition) among benzene, toluene and p-xylene
before and during a step-load increase in the influent concentration of one of these

substrates while the other two were held constant.

4.2 Materials and Methods

4.2.1 Media and Chemicals

GAC (Calgon Filtrasorb 400, Calgon Co., Pittsburgh, PA) and non-activated
carbon, termed "baker product" by the manufacturer, were used as carrier media in two
ﬂuidized-bed reactor systems. The baker product is the same material as GAC except that
it has not undergone the activation step in the manufacturing process and has little
adsorptive capacity (Voice et al. 1992). Both media were sieved to obtain a 20 x 30 mesh
fraction (average particle diameter 0.75 mm). Prior to addition to the reactors, the GAC
and baker product were rinsed with distilled water to remove ﬁnes and dried at 100 0C for

24 hours.

Benzene, toluene, p-xylene (BTX), potassium phosphate dibasic and ammonia
chloride were obtained from J.T. Baker Chemical Co., Phillipsburg, N]. All chemicals
were reagent grade. The water supplied to Michigan State University was used directly as
a source of groundwater. This relatively hard water (450 mg/L as CaCO3) is pumped

from a deep aquifer underlying the campus and distributed without further treatment.

62

4.2.2 Biodegradation Kinetic Assays

After a mature bioﬁlm developed on the baker product, a small amount of
biomass coated material was removed from the FBR. The biomass was stripped from the
carrier particles, purged with oxygen to remove residual BTX, and homogenized by
shaking the biomass using glass beads in a 300 mL glass bottle. Kinetic assays were
performed using a 50 mL glass syringe equipped with a Lure-lock valve. A stirring bar (6
mm in diameter and 15 mm long) was placed in the syringe. The syringe was placed on a
magnetic stirrer to provide mixing. The growth media was groundwater amended with
nitrogen and phosphorous at a weight ratio of 100/5/1:COD/N/P. Prior to addition of the
BTX this water was oxygenated with pure oxygen to obtain an initial dissolved oxygen
concentration of greater than 30 mg/L. Stock substrates of benzene, toluene and p-xylene
were dissolved in tap water and then measured amounts of these solutions were
transferred to the syringes. The initial concentration of BTX was varied from 1200 to
3900 ug/L for each substrate. A constant ratio of B:T:X of 1:1:1 was used for all assays.
The initial concentration of biomass was varied from 20 to 80 mg VSS/L. After biomass
was introduced into the syringe, a one-mL liquid sample was taken. Subsequent one-mL
samples were taken at regular intervals to track degradation. Assays were conducted in

triplicate.

The net yield coefﬁcient (Y) was measured using a similar procedure as used in
the kinetic assay except only one substrate was added to the syringes at a time. Liquid
samples were taken at the beginning and end of experiment to measure the concentration
of substrate and biomass. Biomass concentrations (VSS) were measured according to

Standard Methods (Greenberg et al. 1985). All assays were conducted at 20 0C.

63

4.2.3 Fluidized-Bed Reactors

The laboratory-scale ﬂuidized-bed reactors were constructed using a 2.5 cm
diameter by 184 cm height glass column (Figure 4-1) to produce a system with a 1000-
mL working volume. Two reactors were used in this study to allow comparison of
adsorptive and non-adsorptive bioﬁlm carriers. The GAC-FBR (combined adsorption and
biodegradation) and FBR (biodegradation only) were charged with activated carbon and
baker product, respectively. The reactors were operated as single pass systems with no
recycle. Both were inoculated with a mixed culture obtained from a pilot-scale FBR that
was originally seeded with activated sludge and was subsequently supplied with BTX as
the sole carbon source for more than two years (Hickey et al. 1990b). Dissolved oxygen
(DO) was supplied to the reactors by oxygenating the reactor feed water with pure
oxygen to a concentration sufﬁcient to maintain an efﬂuent D0 of more than 4.0 mg/L.
Nitrogen and phosphorus were supplemented at a weight ratio of 100/5/1:COD/N/P. A
mixture of benzene, toluene and p-xylene, at a 1:1:1: volumetric ratio, was injected into
feed water using a syringe pump (Harvard Apparatus 22) and dissolved using an in-line
mixer. A recycle loop around the mixer was employed to create turbulent ﬂow inside the
mixer and help ensure complete BTX dissolution. This BTX contaminated groundwater
was then pumped to the bottom of the reactors using peristaltic pumps (W atson-Marrow
502E with 501R head). The bed height was controlled by mixing the bed twice a week
with a long thin wire brush that sheared some biomass from the carrier medium. The
reactors were charged with 180 mL media initially, which produced a settled bed height
of 33 cm. BTX contaminated water was fed at flow rate of 200 mL/min to each reactor,
resulting in a hydraulic ﬂux rate of 0.41 m/min, an empty bed hydraulic retention time of
5 min and organic loading rate about 3 kg COD/m3-day. The temperature of the feed

water was 15 :l: 1 °C.

 

Efﬂuent ‘—‘

 

  

Efﬂuent Sampling Point

 

—> Efﬂuen

J |

 

 

t

 

Overﬂow to
Tap Water I = was”
ular Granular Influent
Activated Baker Reservou'
Carbon Product L oxygen
..... Oxygenator

 

7:. a 29.03“

 

 

  

Feed Pump

 

 

Additional
Substrate

Inﬂuent

Injection
Unit

 

Inﬂuent Sampling Point

  

 

Mixing Pump b‘

 

 

......
..;.q .........

..........

Feed Pump

Nutrient Addition
—.-

 

 

 

BTX Injection
Syringe Pump
é
Inline Mixer

Figure 4-1. Schematic of ﬂuidized-bed reactor systems

65
4.2.4 Step-Loading Rate Increase Experiments

After steady-state conditions reached, as determined by biomass growth and BTX
removal, the organic loading rates were increased seven-fold by raising the concentration
of one of the BTX substrates in the feed water. The organic loading rates during the step-
load increases were 9 to 11 kg COD/m3 -day. Step-loading increases were maintained for
four hours, after which the concentration was returned to the steady-state loading level.
During the increased loading period, the amount of DO and essential nutrients supplies

were concurrently increased to ensure that these factors did not limit performance.

4.2.5 Extraction of BTX from Seeded GAC and Baker Product

Biomass carrier (GAC and baker-product) samples were collected from each
reactor, dried with adsorbent tissues and then transferred to 17 mL glass vials containing
15 mL of methanol. The vials were capped with Teﬂon®-coated septa and aluminum
seals and tumbled at 6 rpm for three days. Analysis involved transferring 100 11L of the
extract to 20 mL headspace vials containing 10 mL NaCl saturated water and 120 pL 6 M
HCl. The vials were sealed with Teﬂon®-coated septa and aluminum crimps seal and
analyzed by headspace chromatography. After samples were withdrawn for analysis, the
carbon was dried at 90 0C for 24 hours and weighed on an analytical balance. The
extraction efﬁciencies for a clean GAC with a known amount of BTX were 93%, 79%

and 66% for benzene, toluene and p-xylene, respectively.

4.2.6 Analytical Procedures

BTX concentrations were determined by using a gas chromatograph equipped

with a ﬂame ionization detector (FID) using helium as a carrier gas (Perkin-Elmer, model

- -—
_._

F

66

8700) coupled with an automated headspace sampler (Perkin-Elmer, model HS-lOl).
Separation was accomplished using a 30 m x 0.53 mm DB-624 column (J & W
Scientiﬁc). Water samples, collected from the ﬂuidized-bed reactors (5 mL or 10 mL),
were transferred to 20 mL glass vials with Teﬂon® -coated septa and aluminum crimps
seals. The samples were equilibrated for 1 hour at 80 °C and an aliquot of the headspace
gas was injected into the gas chromatograph. The detection limit of this procedure was 1
ug/L for each BTX component. The accuracy for measurements of split samples

containing a concentration of 1.5 ug/L was 21:01 ug/L

DO concentrations in the inﬂuent and efﬂuent were measured using a
polarographic electrode (Orion 97-08-00) coupled to a digital pH meter (Orion 611). If
the DO concentration exceeded the limit of the instrument (15 mg/L), the sample was

diluted 1:1 or 1:2 with tap water of a known low DO concentration.

4.3 Results
4.3.1 Steady-State Conditions

During steady-state operation, both ﬂuidized-bed systems performed similarly;
the removal rates were approximate 99 percent for all substrates. Inﬂuent and efﬂuent
substrate concentrations and DO consumption are shown in Table 4-1. Efﬂuent
concentrations were typically less than 4 ug/L, 8 ug/L and 15 [Lg/L for benzene, toluene
and p-xylene, respectively. Total DO consumption was 9.36 mg/L for GAC-FBR and
7.71 mg/L for the FBR. The slightly higher DO consumption in the GAC-FBR was
attributed to slightly higher loading rate for GAC-FBR (3.21 kg COD/m3 -day vs. 3.07 kg
COD/m3-day).

    
 
   

     
   

67

Table 4-1. Summary of ﬂuidized-bed system performance at steady-state and during a
step-load increase

 

 

 

 

 

 

 

 

 

 

 

  
 
  
  
 

 

 

 

 

 

 

 

 

  

 

 

  
 

 

 

 

 

 

 

 

 

 

ompound In uent E uent C I 0 d 1‘
onsume =
vﬂw n... we»
Benzene 1234 i141 4 i3
GAC FBR Toluene 1211 i146 8 i7 9.36 21:0.22 !
Steady-State p-Xylene 1125 i 109 15 i8 .
Benzene 1188 $130 1 21:0 1
FBR Toluene 1157 21:125 3 3:2 7.71 21:0.65 1
p-Xylene 1070 3:65 8 i4 1
— —Benzene 9756 i384 3 i4 1
GAC-FBR Toluene 1367 $57 2 i4 19.90 1110.72 ‘
Benzene p-Xylene 1291 :64 5 i 3 I
loading Benzene 9312 i871 5 i6
increase FBR Toluene 1298 $127 10 1:6 20.35 11.03 1
p-Xylene 1224 1127 36 21:8 ’
Benzene 1456 i 187 4 to “I T _ '
GAC-FBR Toluene 9734 i1153 16 1:18 17. 53 :|:2. 19 '
Toluene p-Xylene 1198 :1: 143 32 1:13
loading Benzene 1535 :76 2 21:3
increase FBR Toluene 10566 $482 18 i16 18.28 $1.88 '
1253 21:57 32 21:33 I
1246 i243 2 i4
GAC-FBR Toluene 1185 i234 2 i3 17.05 i057 ,
p-Xylene p-Xylene 8421 i 1704 19 i8
loading Benzene 1214 i204 4 :t4
increase FBR Toluene 1151 i197 5 :13 15.96 321.59 1
p-Xylene 8357 21:1488 680 1:234 f

 

CC
CC

10

68

4.3.2 Biodegradation Kinetics

The degradation kinetics for benzene, toluene and p-xylene by a mixed culture
obtained from the FBR were studied in batch experiments. Degradation rate parameters

were estimated using a kinetic equation and a mass balance on biomass growth:

 

d§ = _ kSX (44)
dt Ks +S
X = X0 + Y(S0 - S) (42)

Where: S is the concentration of substrate (mg/L).
So = initial concentration of substrate (mg/L).
X = concentration of biomass (mg VSS/L).
X0 = initial concentration of biomass (mg VSS/L).
k = maximum speciﬁc utilization rate of substrate (mg substrate/mg VSS-day).
Ks = half-saturation coefﬁcient (mg/L).
Y = net yield coefﬁcient of biomass (mg VSS/mg substrate).

t = time (day).
Combining Eqn. 4-1 and Eqn. 4-2, integrating and rearranging results in:

,=_L[__Is_ln(§n,_[_1<_._,1)m( x. H
k (X0+YSO) s (X0+YSO) Y X0+Y(So-S)

(4-3)

 

The net yield coefﬁcient (Y) was calculated using Eqn. 4-2. The values of k and
Ks were determined using a nonlinear regression analysis (Systate 5.1, SYSTAT, Inc.) of
Eqn. 4-3. These results are presented in Table 4-2. The biodegradation pattern for these
compounds is shown in Figure 4-2. When the concentration of each of the BTX
constituents was greater than 530 pig/L, the degradation rate for toluene was highest and

followed by benzene degradation. If the concentrations were lower than 530 jig/L,

69

however, the rate of benzene degradation was fastest. The degradation rate for p—xylene

was the lowest for the concentration range tested.

Table 4-2. Biodegradation rate parameters for BTX by a mixed culture

       

 

 

 

 

 

Suostrate . Ks
(mg/mg/day) (mg/L)
Benzene 0.99 i0.12 0.19 :t0.16
Toluene 1.48 $0.21 0.47 120.23
p-Xylene 0.73 10.16 0.58 i058

 

 

To understand the effect of biodegradation rate on performance of the biological
ﬂuidized-bed at steady-state, a concentration proﬁle for the bottom half of FBR is shown
in Figure 4-3. The average degradation rates were calculated using Eqn. 4-4 for each
compound in ﬁrst three sections of a FBR.

81. —SH
4.4
at, < )

 

R:

Where: R = average degradation rate in the section (mg/L-min).
SL = concentration of substrate at lower point of the section (mg/L).

SH = concentration of substrate at higher point of the section (mg/L).

dts = empty bed hydraulic retention time of the section.

The degradation rate for toluene was highest in ﬁrst section. After the ﬁrst

section, the degradation rate for benzene became highest because the concentrations of

BTX were lower than 530 ug/L.

 

 

70

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.6
rToluene I
1.4 '- F__'—a—__"_'_q
aF'_’F'-'
e; I Benzene I
a
.31
g I p-Xylene I
g ____,_ ______ _,_-
m __‘_,__,__---
'U
i i
0 2 4 6 8
S(mg/L)

Figure 4-2. The biodegradation rate patterns for BTX in a mixed culture

 

10

71

50

 

 

I Steady-State

 

 

 

 

Carbon Bed Height (cm)

 

 

 

0 i i
0 0.2 0.4 0.6 0.8 1 1.2 1.4
Concentration (mg/L)
- - - 'l - ' - Benzene -—X— Toluene - - 'lr - - - p-Xylene

Figure 4-3. Proﬁle of BTX concentrations and degradation rates in a FBR under steady-
state conditions

72

4.3.3 Step-Load Increase Experiments

After steady-state conditions were achieved, the concentration of benzene in the
inﬂuent was increased seven-fold while the concentrations of toluene and p-xylene were
keep constant. The concentrations of nutrients and DO were also increased corresponding
to the loading increase. The step-load increase was maintained for four hours; the loading
rate was subsequently decreased to the initial steady-state level. The step-load increase

experiments were repeated for both toluene and p-xylene.

The substrate concentration and the DO consumption during the benzene step-
load increase are shown in Table 4-1. The concentration of BTX in efﬂuent of the
ﬂuidized-bed systems is shown in Figure 4-4. The DO consumption and efﬂuent quality
of the two systems were similar. The removal of BTX was greater than 97 percent. p-
Xylene concentrations in the efﬂuent of the FBR were slightly higher than in GAC-FBR
and higher than the value at steady-state. The average concentration of benzene and

toluene in the efﬂuents were similar to the steady-state levels.

During the benzene step-load increase, the increase in the concentration of
benzene could effect the degradation rate of toluene and p-xylene (Figure 4-5). The
degradation rate for benzene increased from the steady-state value of 1.70 mg/L-min to
9.35 mg/L-min in the bottom section, from 1.25 mg/L-min to 4.95 mg/L-min in second
section and from 0.16 mg/L-min to 4.87 mg/L-min in third section. The degradation rates
for toluene and p-xylene did not change in the bottom section compared to steady-state
value. The degradation rates of toluene and p-xylene did, however, decrease in the second
section despite the fact that the average concentration for toluene was higher than the
value at steady-state (410 vs. 260 11 g/L). Increases in the toluene and p-xylene

degradation rates were observed in third section.

73

 

0.04 -- l period of elevated benzene loading |

 

Benzene
o
o
N
l
I

 

 

 

 

 

 

 

 

 

g
E
V
5..
an:
'30
I-E
H
G
81—
9
E
G
U
H
E
0
E
m
0
ﬂ
2
>5
’5
O.

 

 

 

 

Figure 4-4. Substrate concentrations in the efﬂuent of ﬂuidized-bed systems during a
benzene step-load increase experiment

74

 

 

 

 

 

 

 

 

 

50 I
T I
40 __ Rate \‘ Rate
T: 0.44 “ B: 4.87 ‘
x: 0.39 '
a 1.
e I ‘~
3 30 qr- ! ‘
0g .
a Rate ~. Rate
T: 0.65 ‘ B: 4.95
'3 x: 0.65 ‘. ‘
a:
E 20 --
8 in
U “‘ I Rate
Rate .‘5 . B: 9.35
T: 1.70 ‘ ‘
10 ‘_ X: 1.65 ‘ .
o -__. i 1 g i e-—
0 2 4 6 8 10
Concentration (ppm)
- - - -Il - - - Benzene —-x— Toluene " ' 'k ' ‘ ' p-Xylene

Figure 4-5. Proﬁle of BTX concentrations and degradation rates in a FBR at end of

benzene step-load increase

75

The two ﬂuidized-bed systems performed almost identically during the toluene
step-load increase (Table 4-1, Figure 4-6). The concentration of toluene and p-xylene in
the efﬂuent streams were slightly higher than the concentrations observed at steady-state
(from 3 - 8 ug/L compared to 16 - 18 ug/L for toluene and 8 - 15 ug/L compared to 32

[lg/L for p-xylene). The concentration of benzene in the efﬂuent did not change when

compared to the steady-state levels.

During the p-xylene step increase, the efﬂuent concentration of benzene and
toluene were the same for the GAC-FBR and FBR. These values were essentially the
same as the values observed at steady state (about 3 ug/L and 4 [lg/L for benzene and
toluene, respectively). The concentration of p-xylene in efﬂuent of GAC-FBR was,
however, signiﬁcantly lower than in the FBR (Table 4-1, and Figure 4-7) The total mass
of p-xylene in the efﬂuent of the non-adsorption system (FBR) was 34 times higher (19
[lg/L compared to 680 ug/L) than in the GAC-FBR (Figure 4-8).

The concentration proﬁle of substrate was measured in both systems at the end of
each step increase. The proﬁles of BTX at the end of the benzene step increase are plotted
in Figure 4-9. There were no signiﬁcant differences in terms of the concentrations of
BTX and oxygen consumption in the two ﬂuidized-bed systems. This indicates the same
biodegradation rate in the two ﬂuidized-bed systems. Likewise, similar concentration
proﬁle patterns were observed for both reactors during the toluene step-load increase.
During the p-xylene step-load increase experiment, the concentration proﬁles of benzene
and toluene were similar for both the adsorptive and non-adsorptive systems (Figure 4-
10). The nearly identical DO proﬁles indicates that similar level of biological activity
occurred in two ﬂuidized-bed systems. The p-xylene concentration proﬁles were,
however, very different (Figure 4-10). Lower concentrations of p-xylene were observed

at all heights in the GAC-FBR compared to the FBR.

76

A small amount of carrier media were removed from the GAC-FBR and subjected
to solvent extraction prior to and at the conclusion of each step-load increase to determine
changes in the amount of BTX adsorbed. During the benzene step-load increase, no
change in the concentrations of adsorbed BTX were observed. However, there was a
slight but signiﬁcant change in adsorbed BTX during the toluene step-load increase based
on 90% conﬁdence levels. Likewise a small mass of p-xylene and toluene were adsorbed

during the p-xylene step-load increase (Figure 4-11).

77

 

 

 

 

 

 

 

 

 

0.05
0.04--
0
5 0.03--
N
5 0.02..
an
0.01.. /\ x
x“ _ , .. u
_*F\._._eem—
a -50 0 50 100 150 200 250 300
E 0.08
v I period of elevated toluene loading |
g 0 0.06 4- | I
33 :1 35
b E 0.04-m :0. x“
s l: T
g 0.02~- x/OIX‘ ﬁ—ﬁG—ax
O _(ra’x'100x“’?$
‘E
0
:1
E
H

 

 

 

 

Figure 4-6. Substrate concentrations in the efﬂuent of ﬂuidized-bed systems during a
toluene step«load increase experiment

78

 

 

 

 

 

 

 

 

 

 

 

 

 

0.05
0.04--
0
s 0.03.-
N
g 0.02...
0.01-Il- j x k .,ox"o~x
0 .—5—x-xx x-- x-l—x??/ \.“x—>xx-Xr-'—'-i:i
a -50 0 50 100 150 200 250 300
0.05
E
1: 0.04--
'3 3 003
8 i '
l: '5 0.02..
3 i-
: 0.01.. .x.
e ,. ..
D X x. XX-x. -.X' X . W“
.. 0- "" xii—"ﬁxn—M/ max—.1
§ -50 0 50 100 150 200 250 300
1.2
E 1 I period of elevated p-xylene loading
" | ,.x‘ Kid
§08.. ,x" x" t
t. 0.6-- ’,.X ‘x " z.
E 0.4 -- ,X x‘ 1
x I
0.2.. ; g
o ._x-x.§m_u.qa_aq_n_q_a_Fn—:2Sx.x—x
-50 0 50 100 150 200 250 300

Figure 4-7. Substrate concentrations in the efﬂuent of ﬂuidized-bed systems during a p-
xylene step-load increase experiment

79

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

500 " Influent
399 389
a 4m ‘- -.
E“
E 300 --
a
e
A
a 200 w
2
57
Benzene Toluene Xylene
40.00
4.
Effluent I 3 6
”a? ..
E 30.00
‘5
O
:l
E 2000 ..
Lil
.E
2 10.00 ..
0.12 0.19 0.13 0.22 1-01
0.00 ; :
Benzene Toluene Xylene
Substrate

El GAC-FBR FBR

Figure 4-8. The responses of ﬂuidized-bed systems to a p-xylene step-load increase

80

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

150a
1204. I Benzene J
90k
60 .“x
30I N‘xmx_
0. : : e Tn—xn—
6 8 10
150; 2 4
120.; Toluene
’é‘ 90%
O
V 60 '\
E 30Ih.‘
.29 M%_
a 0 r ‘r """ ‘:‘"'=-x-l: .L
. 1.5 2
g 150991 05 1
2 1204. rip-Xylene
‘E 90
6 60 '~
”fo x;
0. :'_"":"—-xl:+
0 0.5 1 1.5 2
150 i
120.. $1 | DO
90“ .'\j\
60“ xu.“x°\
30T- .‘-X_INX
0 : : Nx
5 10 15 20 25 30
Concentration (mg/L)
—-D—GAC-FBR ----X--- FBR

Figure 4-9. Concentration proﬁles of substrates and D0 in ﬂuidized-bed systems at the

end of a benzene step-load increase experiment

81

 

 

 

120; I Benzene I

.0 e
X
1.7
l/
11

 

 

0 0.5 1 1.5 2

 

 

120 :- Toluene

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

i?
3
E
.2."
0
:1
E 2
g J
.n
In
a
U
10
150 X
120 .. ‘5‘ | DO |
90 .. “x
60 " \‘h;x
30 .. ‘Dtxux
0 : : NXH—
5 10 15 20 25 30
Concentration (mg/L)

—u—GAc-FBR -------x FBR

Figure 4-10. Concentration proﬁles of substrates and DO in ﬂuidized-bed systems at the
end of a p-xylene step-load increase experiment

82

 

 

 

 

 

 

 

150
120 ..
90 ..
60 ..
30 .P I Benzene I
0 : 1L i i 5
0 0.05 0.1 0.15 0.2 0.25 0.3

150

 

120 u

 

30A I Toluene I

 

 

 

 

Carbon Bed Height (cm)
8

 

150 0

 

120 v-
90 u.

60--

 

30 ..

 

 

 

 

 

Adsorbed Substrate on Bioﬁlm Coated GAC (mg/g)

Figure 4—1 1. Proﬁles of adsorbed substrates on GAC-FBR before and p-xylene step-load
increase; Squares represents the average value before the step load increase; Crosses
represents the average value at end of the step load increase; Bands represent the 90 %
conﬁdence intervals.

83
4.4 Discussion

There are two mechanisms for removal of organic contaminants in GAC-FBR '
systems: biodegradation and adsorption. Under dynamic loading conditions, the removal
of an organic chemical depends upon the rate of biodegradation and the adsorption
capacity of activated carbon for the compound in addition to other factors such as bioﬁlm
thickness. Three substrates, benzene, toluene and p-xylene, were used in this study. The
ranking of adsorption capacities for these compounds from high to low is p-xylene,
toluene and benzene (Dobbs and Cohen 1980). The maximum rate of biodegradation for
this bioﬁlm community, however, is fastest for toluene followed by benzene and then p-
xylene. The maximum rate of biodegradation of p-xylene is approximately half of the
value of toluene (Table 4-2). As indicated by the low half-saturation coefﬁcient (Ks)
value, the rate of biodegradation of benzene at low concentrations (less than 530 ug/L) is
the greatest. The ability of the ﬂuidized-bed systems to remove BTX is highly related to

these characteristics.

Under steady-state conditions at an organic loading rate of 3 kg COD/m3/day and
an empty bed retention time of 5 min, the GAC-FBR performed as a biological reactor.
The high concentration of biomass that fully coated the GAC was able to remove 99
percent of the inﬂuent BTX. The concentrations of BTX in efﬂuents from two systems

were essentially the same as was observed in a previous study (Voice et al. 1992).

The low concentrations of BTX in the efﬂuent are due to the mixed culture which
has the relatively low Ks values and the high biomass concentrations attainable in the
ﬂuidized-bed systems. Higher Ks values were reported for pure and mixed cultures
isolated from aquifer material (Goldsmith and Balderson 1988; Alvarez et al. 1991). The
continuous feed of low concentrations of BTX in ﬂuidized-bed systems apparently

enhanced the growth of microorganisms with low KS values.

84

During the step-load increase of benzene, the two systems performed similarly
and the efﬂuent concentrations of all three compounds remained at essentially the steady-
state level (Table 4-1). Because of the high k value for benzene and the high biomass
levels in ﬂuidized-bed systems, the reactors are able to maintain high quality efﬂuents
due to biodegradation alone. The increase in p-xylene concentration over that found at
steady state indicates that competitive inhibition due to the high concentration of benzene

may have occurred.

The toluene step increase did not effect benzene removal in either ﬂuidized-bed
system. Even though the a high k value was found for toluene, the high value of Ks
resulted in low degradation rates at low concentrations (Table 4-1). The concentrations of
the toluene in the efﬂuent were higher than the values at steady state. Although the mass
of toluene on the GAC carrier increased slightly during toluene step increase, the two
systems performed essentially as biological reactors. The concentrations of p-xylene
increased by a factor of two from the steady-state value. There was no signiﬁcant change
in the concentration of p-xylene in the inﬂuent and no evidence of desorption during the
toluene step increase. The increase in p-xylene concentration in the efﬂuents is probably

due to substrate competition between toluene and p—xylene.

Our previous results demonstrated that adsorption is an important removal
mechanism for BTX during a step-load increase experiment where the loading rate
increased from 6 kg COD/m3 -day to 16 kg COD/m3 -day. The amount of BTX removed
by biological oxidation during this step-load increase was about 9 kg COD/m3-day
(Voice et al. 1992). In the present study, the organic loading rates were increase from 3
kg COD/m3-day to 11 kg COD/m3-day during the benzene and toluene step-load
increases. Apparently, biodegradation was sufﬁcient in the current study to remove all of

the benzene and toluene added during the step-load increase. It thus appears that

85

adsorption serves to remove the portion of a shock-load that can not be degraded by the
bioﬁlm.

The p-xylene step-load increase provided the most convincing evidence of the
role of adsorption in GAC-FBR systems. Due to the relatively low biodegradation rate of
p-xylene (compared to toluene and benzene), high concentrations were found in the
efﬂuent of the FBR system but not in GAC-FBR system (Table 4-1 and Figure 4-10). The
similar DO proﬁles in two systems indicated similar biodegradation rates (Figure 4-10).
Apparently the adsorption capacity present in the GAC-FBR prevented the escape of the
p-xylene in the efﬂuent (Figures 8 and 9) Extraction the GAC carrier veriﬁed that the
adsorption was responsible for maintaining in high treatment efﬁciencies during the step

increase (Figure 4-11).

4.5 Conclusions

The results from this study demonstrate the roles of the biodegradation and
adsorption in the ﬂuidized-bed systems treating BTX. Under steady-state conditions with
an organic loading of 3 kg-COD/m3-day, both systems removed more than 99 percent of
the applied BTX and the GAC-FBR functioned primarily as a biological reactor. When
the two systems were subjected to the seven-fold step-load increases in benzene and
toluene, biodegradation in the ﬂuidized-bed system was sufﬁcient to maintain a high
efﬂuent quality. For a substrate that is biodegraded more slowly, such as p-xylene,
adsorption onto the GAC carrier contributed signiﬁcantly to removal and helped to
provide more robust performance in the GAC-FBR. Because of the complicated substrate
interactions, the concentration of some substrates (e. g. benzene and toluene) can effect
the degradation rate of other substrates in a ﬂuidized-bed systems and other biological

treatment systems.

CHAPTER 5

FORMATION OF PARTIAL OXIDIZED PRODUCTS IN THE FBR SYSTEMS AT
SINGLE STEP LOADING INCREASES

5.1 Introduction

The contamination of soil and groundwater by petroleum products is a common
occurrence in many parts of the country. Of the numerous gasoline components,
benzene, toluene and xylenes (BTX) are the most mobile, and thus the most likely to
contaminate water supplies. One of most common approaches to remediate the
contaminated groundwater involves a "pump and treat" procedure, in which groundwater
is removed from the subsurface, treated in an above-ground system and discharged to
either a publicly owned treatment works or a surface water, or is returned to the aquifer.
Increasing attention is being given to biological treatment because these systems can
destroy the pollutant and are often less expensive than physical-chemical treatment.
Recent work has shown that biological activated carbon in ﬂuidized bed reactor (GAC-
FBR) systems provide a high level of removal efﬁciency and stability in the treatment of
BTX contaminated water (Hickey et al. 1990b; 1991a; Voice et al. 1992; Zhao et al.
1993).

Since the early part of this century, it has been known that microorganisms are
able to metabolize hydrocarbons, including BTX (Stormer 1908; Sohngen 1913).
Numerous investigations have been conducted on the degradation pathways, type of

organisms, and the degradation rates of BTX. The degradative pathway for benzene

86

87

involves the transformation to cis-benzene glycol by a dioxygenase and subsequent
oxidation of cis-benzene glycol to catechol (Marr and Stone 1961; Gibson et al. 1968b;
Hou 1982). The catechol is degraded by a ring cleavage reaction involving either the B-
ketoadipate pathway (ortho cleavage) or the meta ﬁssion pathway (meta cleavage)
(Gibson and Subramanian 1984). In ortho cleavage, the dihydroxylated aromatic ring is
opened to produce cis, cis -muconic acid, which is further metabolized to B-ketoadipic
acid and then oxidized to succinic acid and acetyl-CoA by the tricarboxylic cycle. In
meta cleavage, the dihydroxylated aromatic ring is opened to produce 2-hydroxy-cis,
cis-muconic semialdehyde. Formic acid, pyruvic acid, and acetaldehyde are further
metabolic products. All of the products from both cleavage mechanisms are used further
for either biosynthesis or energy generation through the citric acid cycle, which produces

002 and ATP by electron transport and oxidative phosphorylation (Lehninger 1982).

As a result of the methyl group on the aromatic ring, the initial steps in the
biodegradation of toluene can occur either by the oxidation of the methyl group to form
benzoic acid or hydroxylation of the ring (Zylstra and Gibson 1991). Depending upon
the hydroxylated position on the ring, the products are p—cresol, o-cresol, m-cresol, or
cis-toluene dihydrodiol. The catechol will be produced as a result of the transformation
of the benzoic acid and this can be further degraded through meta or ortho cleavage of
the aromatic ring. p-Cresol can be hydroxylated again to form protocatechuate and this
can be broken by meta ring cleavage. A further product of o-cresol, m-cresol, or cis-
toluene degradation is dihydrodiol 3-methylcatechol which is a substrate for meta ring

cleavage.

The biotransformation of p-xylene can also be characterized in two categories:
aromatic ring oxidation and oxidation of the methyl substituent. Gibson et al. (Gibson et
al. 1974) studied ring oxidation of p-xylene by Pseudomonas putida 39/D. The p-xylene

was oxidized ﬁrst to cis-p-xylene dihydrodiol and further to 3,6-dimethylprocatechol,

88

which can be degraded further by the ortho ring ﬁssion pathway. The cis-p-xylene
dihydrodiol can also be transformed to 2,5-dimethy1 phenol by a non-enzymatic reaction
(Gibson et al. 1974). Methyl substituent oxidation of p-xylene has been reported by
several researchers (Davis et al. 1968; Omori and Yamada 1970a; 1970b; Davey and
Gibson 1974). The proposed pathways involve two variations. In the ﬁrst, one of the
methyl groups is oxidized to from a carboxylic acid, which is then transformed to 4-
methylcatechol. The aromatic ring of 4-methylcatechol can be broken by meta ring
ﬁssion. The second pathway involves the oxidation of both methyl group before ring

ﬁssion.

In a study of cometabolic degradation of p-xylene by Pseudomonas Sp. strain Bl
in ﬁxed-bed biological activated carbon systems, Chang (1994) demonstrated the
formation of partial oxidized products in this pure culture system when toluene was
utilized as carbon and energy source. The partial oxidized products were identiﬁed as
3,6—dimethyl pyrocatechol and p-xylene dihydrodiol. These partial oxidized products
were found in the efﬂuents of biological activated carbon ﬁx-bed and a biological ﬁx-

bed (without adsorption) for the duration of the experiment (about 30 days).

It is clear that various partial oxidized products will be generated as a part of
BTX degradation reactions. It is also clear that the BTX can be utilized as a sole energy
and carbon source by various microorganisms. With a fully developed mixed culture
community, the partial oxidized products will not normally accumulate because of the
enzyme regulation systems in the microorganisms. In a continuous ﬂow bioreactor, a
community of microorganisms will develop to utilize the substrate resource for
maximum beneﬁt. The community will not produce more partial oxidized products than
are needed because such production costs, rather than yields, energy. Under transient

conditions, however, partial oxidized products may be generated at a rate higher than

89

they are consumed, and these compounds may accumulate in the reactor or be washed

out in the efﬂuent.

BTX are listed as priority pollutants by the US. Environmental Protection
Agency (EPA). Benzene is an acute and chronic toxicant, and is listed as a suspected
human carcinogen (Patnaik 1992). Toluene and xylene have similar acute toxicity to that
of benzene, but less severe chronic effects. A limited amount of information is available
about most of the partial oxidized products of BTX degradation. Pyrocatechol, for
example, has higher acute oral and percutaneous toxicities than that of phenol, which is
an EPA priority pollutant (Patnaik 1992). Cresols exhibit similar toxicity to those of
phenol, and p-cresol has been found to be hepatotoxic and nephrotoxic (Patnaik 1992).
Even on the basis of this limited information, it is clear that the potential public health
impacts of BTX degradation products should be considered. The disappearance of BTX
in a treatment process does not ensure that the environmental impacts of the

contamination are eliminated.

The performance of adsorptive and non-adsorptive bioﬁlm carriers in biological
ﬂuidized-bed reactors for the treatment of groundwater contaminated with BTX has
been evaluated in this laboratory (Voice et al. 1992; Zhao et al. 1993). The GAC-FBR
system demonstrated the ability to remove more BTX from the inﬂuent than a traditional
FBR (non-adsorptive bioﬁlm carrier) during step-load increases. The present work
focuses on the effect of adsorption on the extent of mineralization of BTX during
transient step loading in the FBRs. The formation of partial oxidized products in both
GAC-FBR and FBR systems was compared during a step-load increase in inﬂuent

concentration of one substrate while the other two were held constant.

90

5.2 Materials and Methods
5.2.1 Media and Chemicals

Granular activated carbon (GAC Calgon Filtrasorb 400, Calgon Co., Pittsburgh,
PA) and non-activated carbon, termed "baker product" by the manufacturer, were used
as carrier media in two FBR systems. Baker product is the same material as GAC except
that it has not undergone the activation step in the manufacturing process and has little
adsorptive capacity (Voice et al. 1992). Both media were sieved to obtain a 20 x 30
mesh fraction (average particle diameter 0.7 5 mm). Prior to addition to the reactors, the
GAC and baker product were rinsed with distilled water to remove ﬁnes and dried at 100
°C for 24 hours.

Benzene, toluene, p-xylene (BTX), potassium phosphate dibasic and ammonia
chloride were obtained from J .T. Baker Chemical Co., Phillipsburg, NJ. All chemicals
were reagent grade. Michigan State University tap water was used directly as a source of
groundwater. This relatively hard water (450 mg/L as CaCO3) is pumped from a deep

aquifer underlying the campus and distributed without further treatment.

5.2.2 Fluidized Bed Reactors

Laboratory-scale ﬂuidized bed reactors were constructed using a 2.5 cm diameter
by 184 cm height glass column to produce a system with a 1000-mL working volume
(Figure 5-1). Two reactors were used in this study to allow comparison of adsorptive
and non-adsorptive biofilm carriers. The GAC-FBR (combined adsorption and
biodegradation) and FBR (biodegradation only) were charged with activated carbon and
baker product, respectively. The reactors were operated as single-pass systems with no

recycle. Both were inoculated with a mixed culture obtained from a pilot-scale FBR that

91

was originally seeded with activated sludge and was subsequently supplied with BTX as
the sole carbon source for more than two years (Hickey et al. 1990b). Dissolved oxygen
(DO) was supplied to the reactors by oxygenating the reactor feed water with pure
oxygen to a concentration sufﬁcient to maintain an efﬂuent DO of more than 4.0 mg/L.
Nitrogen and phosphorus were supplemented at a weight ratio of 100/5/1:COD/N/P. A
mixture of benzene, toluene and p-xylene, at a 1:1:1: volumetric ratio, was injected into
feed water using a syringe pump (Harvard Apparatus 22) and dissolved using a mixing
tank (2 liters) and an in-line mixer. A recycle loop around the mixing tank was
employed to create complete mixing conditions inside the tank and to help complete
BTX dissolution. This BTX contaminated groundwater was then pumped to the bottom
of the reactors using peristaltic pumps (W atson-Marrow 502E with a 501R head). The
bed height was controlled by mixing the bed bi-weekly with a long thin wire brush to
shear some biomass from the carrier medium. The reactors were charged with 180 mL
media initially, which produced a settled bed height of 33 cm. BTX contaminated water
was fed at ﬂow rate of 200 mL/min to each reactor, resulting in a hydraulic ﬂux rate of
0.41 m/min, an empty bed hydraulic retention time of 5 min and organic loading rate

about 2.2 kg COD/m3-day. The temperature of the feed water was 15:1:1 oC.

92

 

Efﬂuent Efﬂuent

Efﬂuent Sampling Point

_/ \_

   

Overﬂow to
Iz‘i‘i‘i’231’11i3‘1é25 Waste

Granular 5‘2: Granular Tap Water =___>
Activated Baker
Carbon 3715.? Product Influent

 

Oxygen

Oxygenator

 

 

 

 

 

 

Inﬂuent Sampling Point

)’ \ i*€?i22§§2é2:5

 

A A
Feed Pump Feed Pump

Inline Mixer ‘ i, Mixing Pump

Additional g
Substrate LA

Nutrient CHE—l

Addition ,
BTK Injection
‘li- Syringe Pump

 

 

 

 

Mixing Tank
Figure 5-1. Schematic of ﬂuidized-bed reactor systems

93
5.2.3 Step-Load Increase Experiments

In the ﬁrst phase of the experiment, the presence of partial oxidized products in
the reactor efﬂuents was investigated during a step-load increase of p-xylene under
different concentrations of dissolved oxygen and nutrients in the inﬂuents. After steady-
state conditions were reached, as determined by biomass growth and BTX removal, the
organic loading rates (Figure 5-2) were increased seven-fold by raising the concentration
of p-xylene in the feed water. The increase was maintained for ﬁve hours. In the ﬁrst
hour of increase (referred to as HLL: high organic, low oxygen, and low nutrients), the
oxygen and nutrients supplies were maintained at steady-state level. During the next
hour of loading period (HI-IL: high organic, high oxygen, and low nutrients), the amount
of DO was increased 2.5-fold. Essential nutrients supplies were then increased (HHH:
high organic, high oxygen, and high nutrients ). Finally, the loading sequences were
reversed: beginning with HLH, proceeding to HLL and returning to the initial condition
(ILL)-

In phase two of the experiment, after steady-state conditions were established
again, the organic loading rates were increased twenty, twelve and seven-fold for
benzene, toluene, and p-xylene, respectively. The increases were achieved by raising the
concentration of one of the BTX substrates in the feed water. Step-load increases were
maintained for four hours. During the increased loading period, the feed concentrations

of DO and essential nutrients were increased to ensure that these factors did not limit

degradation.

 

94

 

 

 

 

 

 

 

30
-I- -I I
g2“ Hi I933 :I i !
e ! ! ! !
€20" ! ! ! !
8 ! I ! I ! !
Eu.- ! | s- ! !
5' I ’ I ’
ear-""1 : ----:-----
g I L_m __I
I | I
U 5" ! l ! l
I I I I
or===.="""'rr':"|. 1. Imta
0 50 100 150 200 250 300 350 400

Time (min)

Figure 5-2. Experimental conditions for a xylene step load increase

5.2.4 Analytical Procedure

BTX concentrations were determined using a gas chromatograph equipped with
a ﬂame ionization detector (FID) using helium as a carrier gas (Perkin-Elmer, model
8700), coupled with an automatic headspace sampler (Perkin-Elmer, model HS-101).
Separation was accomplished using a 30 m x 0.53 mm DB-624 column (J & W
Scientiﬁc). Water samples collected from the ﬂuidized bed reactors (5 mL or 10 mL)
were transferred to 20 mL glass vials with Teﬂon ®-coated septa and aluminum crimp-
seals. The samples were equilibrated for 1 hour at 80 °C and an aliquot of the headspace

gas was injected into the gas chromatograph. The detection limit of this procedure was 1

 

95

jig/L for each BTX component. The accuracy for measurements of split samples

containing a concentration of 1.5 jig/L was :1: 0.1 ug/L

DO concentrations in the inﬂuent and efﬂuent were measured using a
polarographic electrode (Orion 97-08-00) coupled to a digital pH meter (Orion 611). If
the DO concentration exceeded the limit of the instrument (15 mg/L), the sample was

diluted 1:1 or 1:2 with tap water of a known low DO concentration.

The amount of partial oxidized products in the efﬂuent of FBRs can be measured
by several methods. Two analytical procedures, non-volatile organic carbon (non-
volatile TOC) measurement and the measurement of adsorption on the ultraviolet
spectrophotometer, are easily conducted in a general analytical laboratory. Oxidation on
the organic molecule, results in partial oxidized products that are less volatile than the
parent compounds. A simple gas stripping process can separate the partial oxidized
products from the parent compounds. The carbon in the partial oxidized products can be
measured by a total organic carbon analyzer. Before ring cleavage, the partial oxidized
products contain the conjugated aromatic ring structure which will show ultraviolet
adsorption in the wavelength of 200 nm to 300 nm (McMurry 1988; Willard et al. 1988).
Formation of partial oxidized products was monitored by measurement of adsorption on
the ultraviolet spectrophotometer. Chang (1994) reported that the 3,6-dimethyl
pyrocatechol, 2,5-dimethylphenol, and p-xylene dihydrodiol, which are the partial
oxidized products of degradation of p-xylene, can be measured by a high pressure liquid
chromatograph equipped with a C13 reverse-phase column and a UV detector at

wavelength of 280 nm.

The concentration of non-volatile total organic carbon (non-volatile TOC) was
measured on a total organic carbon analyzer (Simazdu, Model TOC-500) equipped with
an auto-sample injector (Simazdu, ASI—502). Before the measurement of non-volatile

TOC, the water sample was acidiﬁed with hydrochloric acid and purged with oxygen for

$0 r;iTz‘-__l-§.J L

96

4 minutes to remove inorganic carbon and volatile organic carbon. A water sample
spiked with BTX and catechol (one of the expected intermediates) was tested to ensure
the efﬁcacy of the procedure. The results showed that the BTX and inorganic carbon
were removed completely and there was no signiﬁcant removal of the non-volatile
organic carbon (catechol). The TOC detection limit was 0.6 mg/L and the accuracy for

measurements of standard solution of 1 mg/L was :1: 0.2 mg/L.

The absorbance of ultraviolet and visible light by water samples was measured
by a spectrophotometer (Perkin-Elmer, model Lambda 6 UV/VIS). Before measurement,
samples were passed through a 0.22 pm ﬁlter to remove suspended solids and purged
with oxygen for two minutes. A test showed that the purging procedure was able to
remove all BTX from the sample. The samples were scanned from a wavelength of 600
nm to a wavelength of 200 nm with a slit of 4 and scan speed of 200 nm per min. The
absorbance was also determined at a wavelength of 265 nm. Deionized water was used

as reference sample.

5.2.5 Stoichiometry of Biological Reactions

The removal of substrates in a GAC-FBR system can result from either
adsorption or biodegradation. Biodegradation can further be divided into that which is
used 1) as an electron donor, 2) for the growth for biomass, and 3) to form incomplete

oxidation products. The overall reaction (R) can be written as:

R=Rd+f¢Ra+fSRc+beb (5-1)

where: Rd is the half reaction for the oxidation of an electron donor.

"v ‘t 5 .. j
_.d

~ 1 stir—it

97

Ra is the half reaction for the reduction of an electron acceptor for
energy.

Re is the half reaction for the reduction of an electron acceptor for
synthesis.

Rb is the half reaction for the reduction of an electron acceptor for
byproduct.

fe is the fraction of electrons diverted for energy.

fs is the fraction of electrons diverted for synthesis.

fb is the fraction of electrons diverted for byproduct.

The half reactions for Rd, Ra, RC, and Rb can be written as:

Electron donor half reaction (Rd):

Benzene: 0.0333C6H6 + 0.4HZO = 0. 2CO2 + H+ + e’
Toluene: 0.0278C7H8 + 0.389HZO = 0.194CO2 + H+ + e”
p-Xylene: 0.0238C8H10 + 0.381HZO = 0.19CO2 + H+ + e"

Electron acceptor half reaction for energy (Ra):
0.2502 + H+ + e" = 0.5H20
Electron acceptor half reaction for cell synthesis (RC):
0.25CO2 + 0.05NH3 + H+ + e" = 0.05C5H7OZN + 0.4HZO

Electron acceptor half reaction for byproduct (Rb):

Organic + 0.502 + H+ + e“ = Byproduct + H20

In this system there are three electron donors in the inﬂuent. The terms f 3, f (I, and f if

denote the fraction of electron donor originating from benzene, toluene, and xylene,
respectively. The half reaction for the production of byproduct is presented in a general

form since speciﬁc partial oxidized products were not identiﬁed in this study.
The overall reactions (R)are: R = f ER: + f :R: + f fRI‘ + feR,I + stc + f I,Rb
f3 + f I + f ff =1

C” Ti'-L‘ R 2'71.
1' _.—-1 '5

98

fe+fs+fb=l

An expression for the overall reactions for BTX biodegradation under steady-state

conditions (without the production of byproduct, f b=0) is:

 

0. 0333f§ic,H,5 + 0. 027in,T (1,11, + 0. 0238f§fc8Hlo + 0. 251,02 + 0.051,NH3
= 0.051,C,H,o,N + (0.213 + 0.1946,T + 0.1913: — 0.251, )co2
+(0.5£f + 0.4ff - 0.41“: — 0. 389i,T — 0.381ff)HZO

 

 

Because the partial oxidized products were not explicitly identiﬁed, the overall reactions

under the step-load increase conditions can not be written.

If yield (Y) data is available, the values of f5, can be estimated by the

relationship (Criddle et al. 1991):

8
f = Y — 5-2
where: g is the electron equivalent mass of the electron donor, g.

h is the electron equivalent mass of biomass (5.65 g for ammonia as the

nitrogen source).

The values of fe and fs may be calculated from the consumption of substrates
and oxygen under steady-sate conditions as the following:
The calculation of the fraction of electrons at steady-state:

MB MT
(0.0333)(78)ff ' (0.0278)(92)de

 

MT M"
(0.0278)(92)de — (0.0238)(106)f;‘

 

MI _ M.
(0.0278)(92)de - (0.25)(32)l;

Where: Mo = consumption of oxygen (mg, or mg/L).

99

M? = consumption of benzene (mg, or mg/L).
M? = consumption of toluene (mg, or mg/L).
M? = consumption of p-xylene (mg, or mg/L).
Mb = production of byproduct (mg, or mg/L).
Because there is no production of partial oxidized products at steady-state condition,
f b = 0
and

f =1-(fc+fb)

An estimate of the fraction of BTX removed by the two mechanisms,
biodegradation and adsorption, was performed. Two assumptions were made in order to
perform this calculation. The ﬁrst was that the removal of BTX in the FBR results only
from biodegradation. The second was that the microbial communities in both reactors
were similar, allowing the use of the same values for fe, f s, and fb. The amount of
biologically removed BTX could be estimated using dissolved oxygen consumption

data:

 

(Milne... =1M 1.91:0“;

 

(Mam-.. =1“; 1 (Mann...

 

(Mane... =11”, I...(M°’m-m

 

M
(Mb)BAc-FBR = (Mb ] (Mo )BAC—FBR
FBR

100

The fraction removed by adsorption was calculated from the difference of the total
removed BTX and the biologically removed BTX. The adsorbed byproduct in the GAC-
FBR was estimated from the difference between the calculated amount of byproduct

from biodegradation and the amount in the efﬂuent.

5.3 Results
5.3.1 Steady-State Conditions

Under steady-state conditions, BTX removal rates were greater than 98% in both
reactors. The consumption rate of D0 in both reactors were similar. It is believed that
the primary removal mechanism is biodegradation in both systems. The same conclusion
was presented in our previous studies (Voice et al. 1992; Zhao et al. 1993). No
byproduct was found in the efﬂuents under these conditions. The values of fe and f3,
were calculated from this and previous studies (Voice et al. 1992) and the results are
summarized in the Table 5-1. The average value of fe and f5, were 0.70 and 0.30,
respectively. A yield coefﬁcient was determined in a batch experiment which was
described in our previous study (Zhao et al. 1993). Assuming that fb is approximately
zero, the yield coefﬁcient can be used to calculate fe and fs values of 0.72 and 0.28,
respectively. The observation that similar values were found in the batch and both types

of FBR systems.

101

Table 5-1. Electron balances at steady-state

COD/m3

this study

Voice et al. 1992

 

5.3.2 Step-Load Increases

In the ﬁrst phase of the experiment, the organic loading rates were increased
from 2.2 to 14.9 kg COD/m3-day by raising the concentration of p-xylene in the feed
water from 0.5 to 14 mg/L. Step-loading increases were maintained for ﬁve hours.
During the increased loading period, the supply of DO was increased from 11 to 27
mg/L in the second and third hours of the increase, and the amount of essential nutrients

supplies were increased in the third and fourth hours.

The spectra of absorbance for the efﬂuent of FBR before, during (HHH), and
after a xylene step load increase is shown in Figure 5-3. The absorbance 0e < 300 nm)
increased signiﬁcantly during the step-load increase. A wavelength of 265 nm was
selected as an indicator of the presence of partial oxidized products. Removal
percentages for benzene, toluene, and p-xylene, DO consumption, and efﬂuent

absorbance at 265 nm are shown in Figure 5-4. Benzene and toluene removal were

102

similar in both reactors with greater than 90% of the two compounds being removed
under most loading conditions (Figure 5-4a, b). Removal dropped from steady-state
values of greater than 99% to 96% in the initial hour of shock loading. Removal
increased to 99% in the second hour with an increase in DO concentration. Efﬂuent DO
levels increased from 2 mg/L to 8 mg/L in this interval and DO consumption increased
from 10 mg/L to 17 mg/L (Figure 4d). Increasing the nutrient levels in the third hour did
not appear to effect performance. Similar results were found in the fourth and ﬁfth hours

as ﬁrst the DO and then nutrient levels were decreased.

The removal of p-xylene (Figure 5-4c) in the GAC—FBR was consistently higher
than in the FBR (75% vs. 55%). The average absorbance at 265 nm in the efﬂuent of the
GAC-FBR was also lower than that from the efﬂuent of the FBR (0.064 vs. 0.072 ,
Figure 5-4e.). The higher removal of xylene was observed under high DO conditions in
both reactors (80% vs. 72% for GAC-FBR and 62% vs. 50% for FBR) and the
absorbances were also higher (0.074 vs. 0.058 for GAC-FBR and 0.082 vs. 0.067 for
FBR).

103

 

 

 

 

 

0.3 ,
!
I
i 265 nm
0.2 i
1 During Step Load
3: I
a 0 1 a. i \
8 ' \ Before Step Load
1 ' \,
0 1 =-
1 After Step Load
-0-1 I i : i :
200 300 400 500 600 700
Wavelength (nm)

Figure 5-3. Spectra for the efﬂuent of FBR before, during (HHH), and after a xylene
step load increase

104

 

 

 

 

 

 

 

 

 

 

18.8881

 

 

3380M
ocuuaom

E
onus—oh

E

REES—
one—xx

. .
0 5 O
2

 

 

- I - d
5 09 6 3
0.

1 1 0. 0.
O 0 0

SEE on 8: new a
n nogmnou a 85353<

8:1
:1

Condition

GAC-FBR U FBR
Figure 5-4. Removal of BTX, consumption of DO, and production of byproducts in the

GAC-FBR and FBR during a xylene step load increase with different oxygen and

nutrient supply

105

In phase two of the experiment, three step-load increases were applied to both
reactors. The organic loading rates were increased twenty, twelve and seven-fold for
benzene, toluene, and p-xylene, respectively. The increases were achieved by raising the
concentration of one of the BTX substrates in the feed water. Step-load increases were
maintained for four hours. During the loading increase period, the amount of DO and
essential nutrients supplied were also increased to ensure that these factors did not limit

performance.

During the benzene step-load increase, the concentration of benzene in the feed
was increased from the steady-state value of 1 mg/L to 45 mg/L. The DO concentration
in the efﬂuents of both reactors were maintained at greater than 10 mg/L. The
concentrations of BTX in the inﬂuents and efﬂuents, DO consumption and non-volatile
TOC production, are shown in Figure 5-5. The removal rates (Figure 5-5 a, b, c; Table
5-2) for applied BTX in the GAC-FBR were 63%, 85%, and 93% for BTX, respectively.
The removal rates in the FBR were 24%, 55%, and 89% for BTX, respectively. DO
consumption was similar in the two systems during the step-load increase, but it did not
drop to pre-increase levels in the GAC-FBR system as it did in the FBR system (Figure
5-5d). Non-volatile TOC was considerably higher in the FBR system throughout the
step-load increase compared to the GAC-FBR (Figure 5-5e).

106

Table 5-2. BTX removal during step-load increases

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Increase eactor ompound Tot Mass (mg) Remov
Inﬂuent Efﬂuent Remov (%)
Benzene 2 7 723 1 8

GAC-FBR Toluene 41 7 T4 85
Benzene p-Xylene 29 2 27 93
Benzene 2154 1617 526 25
FBR Toluene 40 18 22 55
-X lene 28 3 25 89

Benzene l 7 8
GAC-FBR Toluene 1340 2_03 1137 85
Toluene p-Xylene 1 16 31 85 73

Benzene 95 66 29

FBR Toluene 1333 883 450 34

-X lene 115 101 14 1
Benzene 51 0.2 50.8 99
GAC-FBR Toluene 59 l 58 98
p-Xylene p-XLlene 64? 56 591 91
Benzene 50 2 48 96

FBR Toluene 57 4 53 9
-X lene 638 270 368 58

 

 

 

 

 

 

107

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a 50 x’xpx-5_xﬁ=§ﬁil GAC-FBR Inf I
A 40-- -+"+"+
.9. |
1:: 30-- J FBRFJT + -----
85 +
3 g 20..
8 10.
0
a 1
£03 X.

x’
$506 an?“
3 c50.4 +.'+"+"+"+'“~~-+--~ +
t— : ".

8 0.2 : GAC-FBREff b
0+-—-+-+ : i 4—4-
El 1
£08} [GAC-FBRInf
D 0.6XN x\x=xv‘§m~.‘
G a '--xf§-x-_f/ "" A "--x_-- X~
330.4 x m '--><““*-x
L8) 02 (GAC-FBRHT] FBREff
- -'+' -U-Iw-u. ‘u-‘II-ﬁ'," +
0+_...+._+_$d+ +£t_+___F—..a+_+
2° *
3% 15" X‘s; | GAC-FBRDO |
.. ﬁ-W
ﬁg“ 10 X 'x _____ x---- x.~
FBRDO -
U 5 Tﬂx-‘x‘x 'x ' -x ----- X
0 r : i 4. 1
6 +
8% 5.. FBR'IOC , " :
:5 3:: ."+'-+§-+.-+-o-"+' E
38 2.»- ‘i i
O :1 '
HE 1" I | GACFBR'IOC
am 0 ----- +- ‘ '
205 -1.” =4]..- : :
-120 0 120
Time (min)

Figure 5-5. Performance of the GAC-FBR and FBR during a benzene step load increase

108

Of the total BTX fed in the GAC-FBR, 64% was removed. From the mass
balance calculation of GAC-FBR, 35% was estimated to result from adsorption and 28%
due to biodegradation (Table 5-3). The production of byproduct accounted for 10% of
the BTX that was removed in the GAC-FBR (Table 5-3). However, only 3% of the
removed BTX was detected in the efﬂuent of GAC-FBR as the partial oxidized products
In the FBR removal of 26% of the applied BTX was biological, with 9% of the total
showing up in the efﬂuent as partial oxidized products (Table 5-3). Three times as much
non-volatile TOC (partial oxidized products) was found in the efﬂuent of FBR than in

the GAC-FBR.

Table 5-3. Carbon balances during the step load increases (unit: % of inﬂuent carbon)

 

 

 

 

 

 

 

Removal by

Load Reactor Total Adsorp- Biodegradation

Increase Removal tion Bio-mass L Byproduct | Total
of BTX and C02 in Eff. Adsorbed

Benzene GA -FBR 64.2% 35.6% 18.6% 2.9% 7.1% 28.6%
FBR 26.1% N/A 17.0% 9.2% N/A 26.1%
ouene GAC-FBR 83.9% 44.8% 20.9% 3.4% 14.8% 39.1%
FBR 32.1% N/A 17.2% 14.9% N/A 32.1%
yene AC-FBR 93.0% 32. % 43.2% 11.4% .1% .7%
FBR 63.4% N/A 45.2% 18.2% N/A 63.4%

 

 

 

 

 

 

After the reactors were returned to steady—state conditions again, the
concentration of toluene in the inﬂuents was increased from 0.9 mg/L to 27 mg/L for
four hours. DO concentrations in the efﬂuents were maintained at greater than 10 mg/L.
The concentrations of BTX in the inﬂuents and efﬂuents of both reactors, DO
consumption and non-volatile TOC production are shown in Figure 5-6. Due to a failure
of the substrate feeding system, high concentrations of all three compounds were
introduced into both reactor systems for the ﬁrst 30 rrrinutes. BTX removal rates, after

recovery from this situation, were 80%, 85%, and 73% respectively in the GAC-FBR.

109

Removals in the FBR were 31%, 34%, and 12% for BTX, respectively. DO
consumption was higher and non-BTX production lower, in the GAC-FBR system
during the increase period. As in the case of benzene, DO consumption continued at
elevated levels in the GAC—FBR system after the OLR increase period for more than 48

hours.

A carbon balance for the toluene increase experiment is presented in Table 5-3.
Approximately 84% of the total BTX fed to the GAC-FBR was removed. As with the
benzene step increase, 45% was estimated to result from adsorption and 39% due to
biodegradation in GAC-FBR (Table 5-3). About 18% of removed BTX was transformed
to the partial oxidized products in GAC-FBR (Table 5-3). However, only 3% of
removed BTX was found in the efﬂuent of the GAC-FBR as partial oxidized products.
In the FBR 32% of the applied BTX removed biologically, with 15% of the total
showing up in the efﬂuent as partial oxidized products (Table 5-3). More than four times
as much non-volatile TOC (partial oxidized products) were found in the efﬂuent of FBR

compared to the GAC-FBR.

The last step load increase experiment was a xylene increase (Figure 5-7). The
concentrations of xylene in the inﬂuents were increased from 0.6 mg/L to 13 mg/L and,
as with previous increases, the concentrations of DO in the efﬂuents were maintained
greater than 10 mg/L. The concentrations of BTX in the inﬂuents and efﬂuents, DO
consumption and non-BTX TOC production, for both reactors are shown in Figure 5-7.
The removal rates (Table 5-2) of applied BTX in the GAC-FBR were 99%, 98%, and
91% for BTX, respectively compared to 96%, 93%, and 57% in the FBR, respectively.

The carbon balance for both reactors is shown in Table 5-3. Because the xylene
step load increase was less than the benzene and toluene increases (the organic loading
rates were increased twenty, twelve and seven-fold for benzene, toluene, and xylene,

respectively), The removal of BTX in the GAC-FBR remained at 93%, but removal

110

dropped to 63% in the FBR. About 32% of loaded BTX was removed by the activated
carbon in the GAC-FBR and the 60% of the removed BTX was degraded biologically.
In the FBR, 18% of the BTX was removed as products, all of which were present in the
efﬂuent. Similar to the other two step increases, the carbon in the GAC-FBR was able to
reduce the amount of partial oxidized products in the efﬂuent from 17% of applied BTX
to 11%.

111

 

 

 

  
      

 

 

 

 

 

 

        

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a 10 2
g 8..[GAc-FBR1nﬁ FBRInf
8 a 6-- -.
83’ 4--
co 5 2 IGAC-FBREff]
7" s; ﬁXg-Vm"
x-Iu—Irxﬁx— “+:¥-'+ """ + """ +8 “x x
a x
A 30" \ WAC-Fulani] PERM
o g 20 31“.... FBREff
8 o " ‘+~-+--+ ----- + ------
[-1 a
C3 10.. W
E FBRInf
A 12.. LGAC-FBRIan
,g.
$31 6.. ‘.
><2:
U 3-- \- GAC-FBREff

tufafﬁaﬁ¥ﬁf=¥=¥—n_¥

 

N
GO

 

  

 

  

I GAC-FBR DO I

 

Consumed
D0 (mg/L) E
2’5 17.

 

 

  

 

 

 

    

 

 

ST-r—‘x-«ax ‘- .x ----- X
X!
gl 8 + l I I
g? i.- '0' ‘+',+~‘+~- FBR-1m
v "" ' ~-‘.+""'
ﬁe 3" I I
'50 2._ l I
>: 1.. I
éé 0+ ' fGAc-Falzrocl +""+--.g:-
' -1‘ : : :
-120 0 120 240 360
Time(min)

Figure 5-6. Performance of the GAC-FBR and FBR during a toluene step load increase

112

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. ’x- 0“ _ -'-'x.
g 09x ----- x " X’:\"§_n \
ﬁg ° [GAC-FBR Inf] X—x
a g 0.6
U 0'3 [GAC-FBREff] FBREff
.1 2 +_+"+“+“‘5'-‘?"?=+~—' - t ------ +~+
. x — FBRInf
a 1.2 kX--x._x“x_!ﬂx
§ 8 0.9 ',x--- [GAC-FBRInf] I
3" wa X‘X
[3 8 0.6
O
O 0.31 [GAC-FBREffI .+. +m+ -------
IE] 0+ +-+—+io;v~+ ' ,_, - +,__.. --- +_=‘_4__ + H

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

15
A 12 xa—X’X'V- -X%~x.___.x
0% 9 IGACFBRInf] FBRInf
£5 6“ m
X 8 H ‘+--+-~+~.+_----+-_,--+
8 3*" *-' GAC-FBREff
°“=¥=¥ 1 x=x====¢
I20
A .. --_ - [GAC-FBRDO]
“g? 15 X‘XW‘Z' X _____ “
8" 10" / s
o FBRDO i
U 8 5 x—V'x'w-x-ux x"x ----- T
o g = = .
6
8g 5
ES 4 _R'POC
' 7: 3 -+-
'+' ~+ ------ Q .....
8§ 2 +
2% 1 IGACFBR'IOCI
2.9 31*" +:..+—1+2L__+ + ..

Time (min)

Figure 5-7. Performance of the GAC-FBR and FBR during a xylene step load increase

1 13
5.4 Discussion

The GAC-FBR performs as a biological reactor under steady-state conditions
(Voice et al. 1992; Zhao et al. 1993). The calculated yields of biomass from mass
balances at different loading conditions produce similar values to that obtained in batch
experiments. This suggests that the production of partial oxidized products is not
signiﬁcant under steady-state conditions. This is consistent with the hypothesis that
microorganisms obtain maximum energy from complete mineralization and will

naturally form a community capable of mineralization under constant feed conditions.

Under transient conditions, such as those that occur during a step-load increase,
the community is presented with a feed condition different from that around which it
was organized. Changes in community structure are slow, and can not occur to any
signiﬁcant level during short-term perturbations. As a result, the community will not
operate at full efficiency, and complete mineralization may not occur. This was
demonstrated in this study. In the all of the step load increase experiments, partial

oxidized products were found in the efﬂuents of both reactor types.

The formation of partial oxidized products is a result of incomplete oxidation and
consumes oxygen as a reactant. When the amount of available DO was varied under
step-loading conditions, higher DO consumption corresponded to higher byproduct
concentrations in the efﬂuent (Figure 5-4). Total BTX removal also increased at higher

DO levels, however.

The levels of essential nutrients (N and P) in the inﬂuent feeds were not a
signiﬁcant factor in BTX removal and the production of partial oxidized products over
the short time intervals studied. A related study of nutrient pulse feeding conducted on a
GAC-FBR system showed that a nutrient supply cycle of 30 minutes on and 30 minutes

off did not effect BTX removal (Xing and Hickey 1994) . When nutrients were withheld

114

for two days, BTX removal decreased approximately 10% from the steady-state level.
The present study conﬁrms and extends this observation by demonstrating that short
term variations in nutrient supply do not effect the removal of BTX or the formation of

partial oxidized products.

The mass balance calculations in this study were based upon the assumption that,
in the FBR, BTX is removed only by biodegradation, and that both reactors have similar
levels of bio-activity. The ﬁrst assumption may slightly overestimate biodegradation
because the non-activated carbon has a small amount of adsorption capacity (Voice et al.
1992). The second assumption is only supported by similarity of performance of the two
systems under steady-state conditions. As a result, these calculations must be viewed as
estimates. Nevertheless, the results do indicate the relative magnitudes of adsorption and

biodegradation in these systems.

During the benzene and toluene step-load increases, about half of the removed
BTX was adsorbed by the GAC in the GAC-FBR. Consequently, the total removals of
BTX were much higher in the GAC-FBR than that in the FBR (64% vs. 26% for
benzene step load increase and 84% vs. 32% for toluene step load increase). There were
also lower concentrations of partial oxidized products in the efﬂuent of the GAC-FBR.
This system also had slightly higher levels of D0 consumption. In the ﬁrst experiment it
was found that increased DO consumption corresponded to increased BTX removal and
increased byproduct formation. We would thus expect the slightly higher DO
consumption in the GAC-FBR to indicate slightly higher byproduct formation, and thus
the lower byproduct levels in the efﬂuent must result from adsorption of the byproduct

compounds by the activated carbon.

During the xylene step-load increase, 35% of the removed BTX was adsorbed by
the carbon in the GAC-FBR and 35% of the partial oxidized products were removed by

adsorption. Because the amount of the xylene increase was lower than the benzene and

115

toluene increases, biodegradation could remove larger portion of inﬂuent BTX and

adsorption constituted a smaller portion of the total removal.

During the step-load increases, a signiﬁcant amount of the biologically removed
BTX resulted in the production of partial oxidized products (approximately 35%, 46%,
and 29% for benzene, toluene, and xylene increases, respectively). Since the
disappearance of BTX is normally used as the only indicator of treatment system
performance, the relatively high levels of byproduct compounds in the efﬂuent
corresponding to these percentages would not be observed. This may pose a potential
health problem. The activated carbon bioﬁlm carrier in the GAC-FBR prevents these
partial oxidized products from leaving the reactor. In concept, such systems could be
designed to insure protection against certain types of shock loads. As shown in previous
studies, this efﬂuent protection capability applies equally well to BTX. During the
toluene step load increase experiment, high concentrations of BTX inadvertently entered
both reactors during the ﬁrst 30 minutes due to an operational error. The GAC-FBR
demonstrated the ability to maintain low levels of BTX in the efﬂuent. The FBR system,

however, was not able to handle the increase (Figure 5-6).

5.5 Conclusions

The production of partial oxidized products was not observed under steady-state
conditions, at a organic loading rate of 2.2 kg COD/m3-day. Under application of a step-
load increase, however, signiﬁcant concentrations of partial oxidized products were
produced. The production of partial oxidized products was not effected by the nutrient
levels but increased with an increase in D0 in the inﬂuent. A signiﬁcant amount of the
biologically removed BTX resulted in the production of partial oxidized products (35%,

46%, and 29% for benzene, toluene, and xylene increases, respectively). use of activated

116

carbon as a bioﬁlm carrier in the GAC-FBR dramatically increased the total removal of

BTX and reduced the amount of partial oxidized products in the efﬂuent.

CHAPTER 6

ROLE OF ADSORPTION IN GRANULAR ACTIVATED CARBON
FLUIDIZED BED REACTORS

6.1 Introduction

Groundwater contamination by volatile aromatic hydrocarbons is widespread as a
result of leaks and spills from petroleum storage tanks, pipelines, and production facilities.
Benzene, toluene, and xylenes (BTX) are frequently observed far from the contaminant
source because of the relatively high water solubility and low soil/water partition
coefﬁcients of the these petroleum constituents. Remediation of groundwater contaminated
with BTX is often accomplished using either air stripping or activated carbon adsorption.
Recently, a technique referred to as biological activated carbon (BAC) has been used treat
BTX contaminated groundwater. In the Granular Activated Carbon-Fluidized Bed Reactor
(GAC-FBR) system, BAC has been shown to provide both the efﬁciency of a biological
removal system and the positive efﬂuent protection capability of activated carbon

adsorption (Hickey et al. 1991a; Voice et al. 1992).

The precise roles of the two removal mechanisms in GAC-FBR systems --
biodegradation and adsorption -- have not been well deﬁned. However, several
hypotheses have been formulated based on the overall performance of operating systems. It
has been observed that microorganisms colonize granular activated carbon (GAC) surfaces

faster than surfaces that are either smoother or less adsorptive (Pirbazari et al. 1990). In

117

118

addition, BAC systems are observed to perform well at low substrate concentrations. It is
thought that granular activated carbon provides a favorable environment for microbial
growth by locally enriching the concentrations of substrates, oxygen and nutrients, and by
providing a surface with sufﬁcient texture to protect a sparse bioﬁlm from ﬂuid shear
forces (Kim and Pirbazari 1989; Speitel et al. 1989a). In addition to promoting bioﬁlm
development, activated carbon provides a removal mechanism during initial the reactor

start-up period (DiGiano 1981).

In previous BAC studies, considerable attention has been given to the question of
whether substrate which adsorbs can later be biodegraded, a phenomenon referred to as
bioregeneration (Khan et al. 1981; Suidan et al. 1981; 1983; Wang et al. 1984; Kim et al.
1986; Wang et al. 1986; Speitel and DiGiano 1987; Suidan et al. 1987). Speitel and co-
workers (Speitel and DiGiano 1987; Speitel et al. 1988; 1989) conducted a series
experiments designed to help determine how bioregeneration varied among four chemicals:
phenol, dichlorophenol (DCP), pentachlorophenol (PCP), and paranitrophenol (PNP). The
bioregeneration rate, as measured by the production of radio-labeled C02, which was
produced by the biodegradation of pre-adsorbed radio-labeled substrates, was found to be
highly dependent on the adsorption and desorption characteristics of these compounds.
Moderately absorbable chemicals such as phenol and PNP could be signiﬁcantly removed
from carbon by bioregeneration. In contrast, compounds with a great afﬁnity for activated
carbon, such as DCP and PCP, were observed to exhibit a high degree of irreversibility of
adsorption. These authors also observed that the diffusive transport resistance in GAC

may control the rate of bioregeneration.

Under the types of operating conditions that can be expected in BAC treatment
systems, it is hypothesized that activated carbon will protect the microorganisms by

removing materials that may be toxic or inhibitory (Suidan et al. 1987). Adsorption will

119

serve to hold these compounds in the reactor for longer than the hydraulic retention time,
thereby allowing time for the organisms to acclimate, and in some cases to degrade, these
potentially toxic compounds (Andrews and Tien 1982). The ability of GAC to adsorb and
desorb compounds has led researchers to suggest that BAC systems may be able to better
respond to changes in inﬂuent concentrations than systems employing non-adsorbing
carrier particles. During an inﬂuent concentration spike, excess substrate could be
adsorbed and then, following the spike, the substrate would be desorbed and biodegraded.
It is thought that the bioﬁlm can thereby serve to bioregenerate the activated carbon (Kim et

al. 1986).

Previous work in this laboratory investigated the performance of GAC-FBR
systems designed to remove BTX from groundwater under both constant inﬂuent and
spike-loading conditions. By comparing the GAC-FBR to an identical system using a non-
adsorbing bioﬁlm carrier, it was observed that the GAC-FBR functioned primarily as a
biological reactor under steady-state conditions, but that the GAC carrier contributed
signiﬁcantly to overall system removal capacity during transient conditions. It was
hypothesized and demonstrated numerically that this resulted from adsorption/desorption of
the substrate by the GAC. However this conclusion was not experimentally veriﬁed (Voice
er al. 1992). The purpose of the present study was to verify this hypothesis by monitoring
the amount of adsorbed substrate before, during and after a step-increase in inﬂuent toluene

concentration to a laboratory scale GAC-FBR reactor.

6.2 Materials and Methods
6.2.1 Fluidized-Bed Reactor

The experimental set-up used for the GAC-FBR studies is illustrated in Figure 6-1.

The reactor was constructed using two sections of 5 cm id. x 92 cm long glass pipe,

120

containing side sampling ports. The working volume was 4 liters. The reactor was
operated as a one-pass system without recycle, at a ﬂow rate of approximately 900
mL/min. This provided an empty bed contact time of approximately 4 minutes. A peristaltic
pump (W atson-Marlow 503S with a 501R head) was used to introduce groundwater into
the column after it had been pre-oxygenated using pure oxygen. Greater than 25 mg/L of
dissolved oxygen could be achieved in the inﬂuent. Toluene was introduced into the
inﬂuent line using a syringe pump (Harvard Apparatus 22). An in-line mixer combined
with a recirculation pump (Watson-Marrow 502E with a 501R head) was employed to
increase the efﬁciency of mixing and dissolution of the toluene. A nutrient solution was
added to the inﬂuent just prior to the feed stream entering the column to minimize any

bacterial growth in the tubing.

In full-scale GAC-FBRs the bed height is maintained at a set control level using one
of several proprietary systems (Hickey et al. 1991b). In this laboratory pilot-scale system,
control of the bed was maintained by mixing the bed once daily with a long thin wire that

sheared some of the biomass from the GAC carrier.

l
(I

121

 

Sampling
Point

\

l'
Effluent g
I
I
I

 

   

—' Treated Ground water

GranularActivated Carbon

 

, -. -. .
. 9 (6‘
,. .
I’IS'?‘ /: a
I a ."
7 .~ CCUOII
\-,._ I
A. " ‘.._
\

 

 

  
 
 

 

 

 

 

 

Sampling Point .7. __.> Overﬂow to Waste
- -Dl> Ground water
I
I ‘.. x 3: O
: Section 3 ﬁnﬂuent. 2
11'
Sampling Point : eservo _ _;
_ .H, "' " -:
I
I GAC Filter :
I if '
| Section 2 +4 :-
- - I " "* N
Sampling Prim-L:> P . X
I . Filter
I =-’ x 5.3": _ -
: .3..‘ Section 1 I.
I i f '
11°93'91" I .t 0.45 11 Filter '
r ' " ‘1 : <— Inﬂuent Sampling Point
: g" " ‘
I ,
I
I l" """
| I
I I 1m M' I- - z
: n e Ixer Toluene Injection
_-----1 SyringePump
submte Additio 1 Inline Mixer
Nutrient Addition -I '
‘1 ‘ Inﬂuent Injection Unit
I- - Stiring bar
a
—N
Mixing Pump

Figure 6-1. Schematic of the experimental GAC-FBR system

L———-I

122

In order to concurrently investigate the ability of inoculated bacterial strains to
compete with an indigenous microﬁlm (results reported elsewhere) the system was initially
heat sterilized and the inﬂuent feed water was subsequently sterilized by 0.22 mm ﬁlter .
Prior to start-up, the glass column was ﬁlled with a 20% chlorine bleach solution which
was recirculated through the system for two days. This was followed by a 0.2 M HCl
solution for which was recirculated for one day. The reactor was subsequently ﬂushed
with deionized water to remove the acid. Initially, the reactor was inoculated with three
bacterial cultures: Pseudomonas pickettii (PKUl), Psueodmonas putida (Pan) and
Pseudomonas cepacia (G4) obtained from the Center for Microbial Ecology at Michigan
State University. Approximately 450 mL of cell suspension containing 3.37 x 107, 1.58 x
108 and 9.50 x 107 cells/mL respectively were added to the reactor (direct counts were
obtaining using acridine orange stained samples). The reactor was operated in a recycle
mode for the ﬁrst two hours to ensure sufﬁcient contact between the inoculum and the
GAC carrier. During the start-up period (initial seventeen days of operation) a ﬁlter system
was placed in the inﬂuent water supply line to prevent organisms in the raw water feed
from entering the reactor. This allowed the inoculated organisms the opportunity to
preemptively colonize the GAC. The ﬁlter system contained activated carbon, a glass-ﬁber
ﬁlter and a 0.45 mm ﬁlter. After day 17, when the bioﬁlm was well established on the

carrier particles, the ﬁlter system was removed.

6.2.2 Materials

Activated carbon (Calgon Filtrasorb 400) was sieved to select only particles passing
a No. 20 sieve (0.85 mm) and retained on a No. 30 sieve (0.65 mm). The geometric mean
diameter of the media was 0.70 mm. The carbon was washed several times with deionized
water to remove carbon ﬁnes, dried in an oven at 105 °C, and sterilized in an autoclave.

Five hundred grams of carbon were added to the reactor. Toluene was used as the sole

123

source of carbon for the bioﬁlm. The nutrient solution was formulated to maintain a ratio

of 100/5/1: COD/NIP using NH4C1 and KH2P04 and boiled deionized water.

6.2.3 Step-OLR Increase Procedure

After a stable bioﬁlm had developed and the reactor reached pseudo-steady-state
conditions at a COD loading rate of 5.4 kg COD/m3-day, the OLR was increased by a
factor of 4 without changing the hydraulic loading rate. An additional oxygen diffuser was
added to the inﬂuent line to insure that dissolved oxygen availability did not limit
biodegradation. The nutrient concentration was increased proportionally to the increase in
OLR. The step-OLR increase was maintained for 77 hours, after which time the inﬂuent
was reduced to the previous loading rate. The inﬂuent dissolved oxygen and nutrient
concentrations were not decreased for an additional 50 hours, in order to ensure that these
factors did not limit biological removal of the adsorbed toluene during this period. When
oxygen consumption decreased to the pre-step—OLR increase level, the nutrient and oxygen

feeds were reduced to the initial conditions.

6.2.4 Extraction Procedure

To measure the amount of toluene adsorbed onto the GAC, carbon samples (0.03 to
0.1 g) were collected from reactor sampling points located at 0%, 25%, 50%, 75%, and
100% of the carbon bed height, placed in aluminum dishes, separated from the liquid using
adsorbent tissue and transferred to 15 mL glass vials. The vials were capped with Teﬂon-
faced septa and aluminum seals. Fifteen mL of methanol was added to each vial as the
extraction solvent, and the vials were tumbled at 6 rpm for 3 to 4 days. This methanol
extract was stored for no more than 3 days at 4 °C prior to analysis. Analysis involved
transferring 0.1 mL of the extract to a 20 mL headspace vial containing 10 mL NaCl
saturated water and 0.12 mL 6 M HCl, sealing the vials with Teﬂon-faced septa, and

124

analyzing this solution by headspace gas chromatography. The mass of carbon extracted
was determined by drying at 90 0C overnight, desiccating, and weighing on an analytical

balance.

In order to determine the effect of adsorption time on extraction efﬁciency, a carbon
sample was collected from a GAC-FBR which was. operated in a manner similar to the
system used in this study. Two adsorption isotherm assays were conducted by the bottle-
point method, (Voice et al. 1992) using different equilibrium contact times (7 days and 35
days). Adsorbed toluene was determined using ﬁve successive extractions with methanol.
No differences in recovery were found between the two isotherms using either the
concentration in the ﬁrst extract or the total mass recovered in all ﬁve. The total mass

recovered was greater than 90% of that lost from solution in all cases.

In this study, only one extraction was performed for each carbon sample and thus,
the extraction efﬁciency is dependent upon the ratio of the volume of extraction solvent to
mass of carbon used in the extraction procedure. A standard multiple extraction calculation
was used to estimate the total amount of toluene adsorbed using the amount recovered in
the ﬁrst extraction and the results of ﬁve successive extraction tests conducted at different

solvent to carbon ratios (Schwarzenbach et al. 1993).

6.2.5 Analytical Measurements

Ten-mL inﬂuent and efﬂuent samples were transferred directly to 20 mL headspace
vials for analysis and sterilized with 0.12 mL 6 M HCl. Both these samples and the
methanol extracts described above were analyzed using an automatic headspace sampler
(Hewlett-Packard 19395A) coupled to a gas chromatograph (Perkin-Elmer Autosystem)

equipped with a ﬂame ionization detector using helium as a carrier gas. Samples were

125

equilibrated at 90 °C for 1 hour and an aliquot of the headspace gas was injected onto a 30
m x 0.53 mm DB-624 column (J & W Scientiﬁc). An overall relative standard deviation of

4% was found for toluene sample collection and analysis. The method detection limit was

determined to be 0.2 jig/L.

Inﬂuent and efﬂuent dissolved oxygen (DO) concentrations were determined using
a polarographic electrode (Orion 97-08-00) and pH meter (Orion 611). Samples were

diluted 1:1 or 1:2 with tap water of a known low DO concentration.

6.3 Results

The operation and performance of the GAC-FBR reactor system can be
conveniently divided into four phases: 1) start-up, the period from inoculation to
development of a mature bioﬁlm; 2) pseudo-steady-state, the period after development of a
mature bioﬁlm, as evidenced by stable inﬂuent, efﬂuent toluene concentrations and DO
consumption levels, 3) step-OLR increase, the period of elevated inﬂuent concentration,
and 4) recovery, the period following the step-OLR increase during which the inﬂuent

concentration was restored to the original level.

6.3.1 Phase 1: Start-up

Immediately following inoculation, the ﬂuidized bed height was 96.5 cm.
Scanning electron microscopy of carbon particles removed from the middle of the bed
revealed that organisms were attached to the carbon surfaces. There was no appreciable
DO consumption during the ﬁrst 4 to 5 days (Figure 6-2); toluene concentrations averaged

2700 [lg/L in the inﬂuent and 18 jig/L in the efﬂuent. This indicates that adsorption was

the primary removal mechanism during this period. This is conﬁrmed by measurements of

126

adsorbed toluene, which is shown both as a function of time and bed height in Figure 6-3.
Toluene concentrations increased sharply at all bed positions during the ﬁrst 10 days of

operation, with higher toluene levels found at lower bed positions.

DO consumption increased sharply between days 4 and 8 and continued to increase
until day 10 as the toluene degradation capability of the bioﬁlm developed. After day 10,
DO consumption decreased slowly from the maximum value of 8 mg/L to a pseudo-steady-
state average value of 5.6 mg/L. During this period, the levels of adsorbed toluene
decreased and the distribution of adsorbed toluene became more uniform throughout the
bed. This suggests that as the degradation capability of the bioﬁlm developed, not only
was the toluene in solution degraded but the toluene adsorbed on the GAC was desorbed
and degraded as well. This is graphically illustrated in Figure 6-4 which shows that the
total amount of adsorbed toluene decreased signiﬁcantly after reaching a peak value

observed at day 10.

Although the bed height of the reactor did not increase signiﬁcantly during the ﬁrst
20 days (Figure 6-5), the level of DO consumption conﬁrmed that biodegradation was
occurring. After day 20, an observable amount of biomass began to accumulate on the

surface of the GAC particles, and the bed height began to increase rapidly.

127

 

 

 

 

10.00
Steady state DO
consumption 5.6 mg/l
8.00 .- I
A I I I
ﬁ, I
3 ._ I I I I
5 6.00 I -
E I I
§ 4.00 T
2.00 ~-
I
I ' I
0-00 . 'r l l i i
0 4 8 12 16 20 24
Time (days)

Figure 6-2. DO consumption during the startup phase (hydraulic ﬂux rate: 0.44 m/min ;
empty bed hydraulic retention time: 4.4 min; inﬂuent concentration of toluene was
maintained at 2.75 mg/L)

'7’ Tan‘S

 

Extracted Toluene Conc. (mg/g carbon)

 

 

 

Figure 6-3. The proﬁle of adsorbed of toluene during the startup phase (hydraulic ﬂux rate
= 0.44 m/min; empty bed hydraulic retention time = 4.4 min; inﬂuent toluene concentration
=2.75 mg/l.)

129

 

 

 

 

 

 

 

14
Startup I 8:33 I l Recovery I
| | | |
12 .. Step
Increase
4
10 --
‘ nd 0
Step
3.9 Increase
a s -. .
.15
o
E J’. .
g 6 -- l
<
4 __ I
2 -- tart ’
2
p Removal Increase
0 l i i l
0 20 40 60 80
Time (days)

Figure 6-4. Total mass of adsorbed toluene in the GAC-FBR system

 

100

130

 

 

 

 

 

 

 

70 IF
- F
60 ~+ I '- I
I I
i
A I
.5. 50 di- ... J
E f I—-
E 40 I.“ .- Biomass Removed I...
I M I
E II
E 30 --
'5 20 ..
E Pseudo
,_ Startup steady Recovery
10 ._ I state
Step Increase
0 l I l i I I i r

 

 

 

0 10 20 30 40 50 60 70 80
Time (days)

Figure 6-5. Height of the carbon bed in the GAC-FBR system

6.3.2 Phase 2: Pseudo Steady-State

The reactor was assumed to be operating under pseudo-steady-state conditions
when constant DO consumption, commensurate with that required for degradation of the
added toluene, was observed. The system performance stabilized with average efﬂuent
toluene concentrations and DO consumption of 14 [lg/L and 5.6 mg/L, respectively. This

corresponded to an overall toluene removal rate of 99.4%.

In an attempt to assess the inﬂuence of the bioﬁlm thickness on adsorption and

biological removal within the proﬁle of the bed, all but a thin ﬁlm of biomass was removed

131

from the GAC particles in the bed on day 39 by vigorous mixing. Prior to biomass
removal, 81% of the DO consumption in the system occurred in the ﬁrst section of the
reactor (15 cm). After the removal of a signiﬁcant amount of the total biomass, the amount
of biodegradation in the ﬁrst section of the bed decreased and a greater bed height was
required to obtain the same level of toluene removal. This resulted in an increase in the
aqueous phase toluene concentration throughout the proﬁle of the bed and a concomitant
increase in adsorbed toluene level (Figure 6-6). With the thinner bioﬁlm, adsorption
became, temporarily, an important removal mechanism for toluene in the lower portion of
the bed. This demonstrates that the GAC had additional adsorptive capacity available. As is
illustrated in Figure 6-4, the mass of adsorbed toluene subsequently leveled-off and then

decreased over the next several weeks as the bioﬁlm regrew.

132

 

   
  
  
 
   
   

2. 5 . Substrate Concentration Proﬁle

N
I

 

Aqueous-phase toluene
concentration (mg/l)
3

 

 

 

 

 

 

 

 

 

 

 

       

 

 

 

 

 

 

0.5 .
: : "'71—
75 100
8
g Adsorbed Concentration Proﬁle
c:
g ’a
8i
O
c:
0 N)
o
8
<1
50 75 100
5 ' ' DO Consumption Proﬁle
3 4 ..
m ‘8 ..
8 a 3
0 v
8 2 ..
1 -l- :3: E
o _ 555 = 35:3. 7/4... 97/4
0-25 25-50 50-75 75-
100
% ofbed height

Before E] After

Figure 6-6. Concentration proﬁle aqueous-phase toluene, extracted toluene and DO
consumption before and after biomass removal

133

6.3.3 Phase 3: Step-OLR Increase

The reactor system was subjected to a step-OLR increase starting on day 58. The
OLR was increased by approximately 400% by increasing the inﬂuent concentration to
11,000 ug/L for a 77 hour period. The DO consumption increased substantially and
rapidly, with most of the increase occurring in the ﬁrst section of the bed (Figures 6-7 and
6-8). In response to the increased aqueous-phase concentration, the concentration of
adsorbed toluene also increased rapidly (Figure 6-4). The increase in bioactivity and
adsorption did not completely mitigate the effects of the increase in inﬂuent toluene at this
elevated OLR; the efﬂuent concentration increased to an average value of 520 rig/L during
this period in which the system was challenged with an applied organic loading of 22.6 kg
COD/m3-d (removal was 95.1 % or 21.5 kg COD/m3-d).

A comparison of aqueous-phase toluene concentrations, adsorbed toluene
concentrations, and DO consumption before and during the step-OLR increase is shown in
Figure 6-9. It is clear that the increased inﬂuent toluene concentration served to increase
aqueous toluene concentrations throughout the proﬁle of the bed. These higher levels
drove an increase in adsorption, which served to dampen the changes in the aqueous phase.
As expected, the higher concentrations also resulted in increasing biodegradation and a

corresponding increase in oxygen consumption.

134

 

 

 

 

 
 

 

 

 

 

 

 

 

 

16
14 . _ Steady State I Recovery I
I
g 12 " Step-OLR I I
E: 10 ﬂ Increaser
2 I Inﬂuent I
5 8-- I
g 6.. OEfﬂuent
.2
E9, 4..
2 I I II I I I °
0
30 35 40 45 50 55 60 65 7O 75 80
30
25L Steady State | | Recovery |
”$3 I
g 20 Step-OLR. I
5 " Increase-l.
E- 15 ..
m I
8 10 -- I '
O 5.
8 5 ~ I I u .- I 'I
T I —-I v I I I I T
0 . I I : : I : I I : I
30 35 40 4 50 55 60 65 70 75 80
Running Time (day)

Figure 6-7. Toluene concentration in the inﬂuent and efﬂuent and DO consumption during
the pseudo-steady-state, step-OLR increase and recovery phases

135

 

 

 

A3588

onus—oh 3853‘

.88 35%/I.

 

 

 

 

Figure 6—8. Concentration proﬁles of aqueous-phase toluene, adsorbed toluene and DO

consumption

136

 

‘:'; : E: Substrate Concentration Proﬁle

 

Loquid-phase toluene
concentration (mg/l)
m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

25 50 75 100
’3 Adsorbed Concentration Proﬁle
I
DO
I I
< c:
I.
50 75 100
D0 Consumption Proﬁle
g g
I 5
0-25 25-50 50—75 75-
100
% of bed height

Steady-state (52 day) [3 Step increase (60.3 day)

Figure 6-9. The proﬁles of aqueous-phase toluene, adsorbed toluene and DO consumption
during the pseudo-steady-state and step increase periods.

137

6.3.4 Phase 4: Recovery

After 77 hours of the step-OLR increase, the organic loading rate was returned to
the original level. DO consumption decreased from 24.3 to 11.3 mg/L within 4 hours, but
did not return to the pre-step OLR increase level of 5.6 mg/L for another 50 hours (Figure
6-7). The total amount of adsorbed toluene, which had increased signiﬁcantly during the
step OLR increase, began to slowly decrease, but did not decline to the pre-OLR step
increase level until 17 days after the OLR was reduced.

6.4 Discussion

Average performance of the system during each of the four phases is shown in
Table 6-1. For the purpose of comparing reactor performance during different phases of
operation, an oxygen utilization quotient (Q) can be deﬁned as the ratio of the mass of DO

consumed to COD removed,

 

(6-1)

where Q: the oxygen utilization quotient (when all COD is removed biologically, the Q
will equals to fe which is the fraction of electrons diverted for energy.), MDO = the mass

of oxygen consumed in the FBR, and MCOD = the difference between the inﬂuent and

efﬂuent COD. In the absence of adsorption, Q indicates the fraction of COD consumed that
was used for energy by the biomass, while the remainder was used for cell synthesis (i.e.
biomass production). During the pseudo-steady-state phase, the average Q was observed
to be 0.74, which is consistent with previous reports for toluene degradation in FBR

systems (Hickey et al. 1991a; Voice et al. 1992). Since it has been demonstrated that

138

biodegradation is the primary removal mechanism and the production of incomplete
oxidation products (byproducts) was negligible during steady-state, this value can be
considered a property of the biological (bioﬁlm) removal mechanism. Assuming that the
fraction of energy diverted to cell synthesis was not signiﬁcantly different during the other
phases of operation, changes in the value Q from the pseudo-steady-state value will reﬂect
the effect of adsorption and desorption on substrate removal and the production and

degradation of byproducts in the system.

Table 6-1. Summary of overall performance of the reactor

 

 

 

 

 

 

 

 

 

 

 

 

In uent Euent GD 0.: g Q I

Phase Time Conc. Conc. Rate Removal DO ‘
(daY) (Hg/1) (Hg/1) kg/m3'DaY % (COD) .

Startup 155" _'2700 18(ilW 5.8 99.3 0110.74
(1250) '

. SteacV 25-58 5500 l4(:l:5) 5.3 99.4 0.7 l
l state (i200) ,
1 tep-OLR 58-61 11000 520(1270) 22.6 95.1 0.58 l
increase (:I: 1400) _ ‘

1 Recovery 61-82 2300 34(125) 4.9 98.5 1510.77
(1150) 1

As shown in Figure 6-10, Q increased from essentially zero to the pseudo-state-
state value of 0.74 over the initial part of the start-up phase. This indicates that more COD
was removed initially than can be explained by biodegradation (as indicated by oxygen
consumption), and thus adsorption must be responsible. This was conﬁrmed by the
concentration of toluene extracted from the GAC (a steady increase in the amount of
adsorbed toluene, Figure 64). After establishment of a bioﬁlm capable of degrading
toluene, a continual decrease in the amount of adsorbed toluene was observed (Figure 6-4).
The desorbed toluene was subsequently degraded by microorganisms in the bioﬁlm.
During the pseudo-steady-state period the value of Q remained essentially constant as

would be anticipated if biodegradation was the predominate removal mechanism.

139

 

 

 

 

 

 

 

 

 

1.6
A 1-4 " Steady State F. Recovery
E Startup Q=0.74 Q: 1.51-0.77
O 1.2 .. I
8
2 I
Q 1 -II- -
8 I I I I
I I I
EO.81:- ----- .l-L----.-----------.‘--
I I. I I I
0.6 -- I
g - q-
8 0~4 -- Step-Load
v Increase
O 0.2 I- Q=0-59
II
0 I I I I I I I I I I I I I I I I
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80

Running Time (day)

Figure 6-10. The oxygen utilization quotient (Q) for the overall experiment

During the step OLR increase, the value of Q decreased to an average value of 0.59.
As was found during the start-up period, more COD was removed than could be explained
by biodegradation; this was accompanied by a measured increase in the amount of adsorbed
toluene (Figure 6-4). This appears to conﬁrm the hypothesis of previous work (Hickey et
al. 1991b; Voice et al. 1992) that adsorption serves to reduce the effect of concentration

increases.

During the recovery phase, the value of Q increased to a maximum value of 1.51
immediately after the loading was reduced. Approximately 20 days was required before a
return to the pseudo-steady-state value of 0.74 was observed. This indicates that during
this period more dissolved oxygen was consumed than necessary for removal of the

inﬂuent toluene. Based on Figure 6—4 it can also be seen that amount of adsorbed toluene

140

decreased, rapidly at ﬁrst and then somewhat more slowly, during the recovery period.
Since the efﬂuent concentration did not increase, it must be concluded from these two
observations that the excess toluene adsorbed during the step increase was both desorbed
and biodegraded after the OLR was returned to 5 kg COD/m3-day. This indicates that
adsorption serves not only to protect the system against concentration increases but that, at
least for toluene, adsorption is reversible and dampens the systems against increases or
decreases in concentration. Furthermore, the elevated DO consumption levels observed
after the OLR was returned to the steady-state level conﬁrms that the bioﬁlm can regenerate
the adsorptive capacity of the carbon. This process of bioregeneration should provide a

means by which GAC-FBR systems can handle repeated transient loading increases.

Speitel and DiGiano (1987) demonstrated that diffusive transport resistance is the
rate-limiting step in bioregeneration of GAC using a homogeneous surface diffusion
model. While this study was not designed to quantitatively evaluate the desorption rate, the
results showed that the bioregeneration was a slow process. It can be inferred from
previous work that this probably results from such a limitation. More information on the
kinetics of desorption, degradation, and time varying concentration gradients would be

needed to fully understand this issue.

The large increase in Q during the recovery period can't be explained by toluene
desorption alone. In a separate study conducted in this laboratory, it was found that during
an extremely high shock-OLR period not all of the toluene was completely oxidized; some
intermediate products were detected (Zhao et al. 1994). These intermediate oxidation
products may have also adsorbed during the step-OLR period and subsequently desorbed
and were degraded during the recovery period. This would be consistent with the

observation that Q was larger than could be accounted for by desorption and degradation of

141

inlet toluene only. Additional experimentation would be necessary to substantiate this

explanation.

The oxygen utilization quotient index can be used to estimate the amount of toluene

removed biologically,

_ MDO
' (3.13)Q“

where: Mb = the amount of toluene removed biologically, MDO = the mass of DO

Mb (6-2)

consumed through the proﬁle of the reactor, QSS = the removal index at pseudo steady-

state (0.74), and (3.13) = conversion factor for mass of COD to mass of toluene. The total
amount of toluene removed from the liquid phase, as determined by inﬂuent and efﬂuent
concentrations, and the amount removed biologically, as determined by equation 6-2, are
shown in Figure 6-11. It can be seen that during the step-OLR increase, total removal
exceeded the amount that could be accounted for by complete oxidation. The difference
between these two curves is the amount of toluene adsorbed from the start of the step-OLR
increase. After the OLR was returned to normal, the biological removal rate exceeded the
total removal rate, as evidenced by the greater slope of the lower line in Figure 6-11, and
the amount toluene adsorbed onto the GAC is reduced. At approximately day 69, 11 days
after the end of the step-OLR increase, the two curves meet, indicating that the mass of
toluene adsorbed during the increase has been removed by bioregeneration. However,
because of the potential production of the byproducts, the results from Eqn. 2 could under-
estimate the biological removal of toluene during the step-OLR increase and over-estimate
biological removal during recovery. As such, this analysis must be treated as a

simpliﬁcation of a presumably more complex set of processes.

142

 

 

 

 

 

 

 

 

 

 

80 .
Step increase ‘ '.
phase Recovery /X/
3 60 ..
"a
S
E
2
I
T; 40 «-
D
.>
g /
I —0—Removed
x from liquid
g 20 “i I phase
ix — —x- - -Removed
biologically
x1
0 9:. I ‘r ‘r
58 62 66 70
Time (day)

Figure 6-11. Cumulative toluene removed from the aqueous phase and that removed

biologically during the step-OLR increase and recovery phases

 

 

143
6.5 Conclusions

This study provides conclusive evidence that a major role of adsorption in
integrated biological activated carbon (BAC) systems is to dampen concentration changes
during loading transients. During the start-up period in a GAC-FBR before a bioﬁlm is
established, the substrate (toluene) was adsorbed rather being released, as would occur in a
biological-only system. Upon establishment of a bioﬁlm capable of degrading the
substrate, inﬂuent toluene was biologically oxidized and a portion of the previously
adsorbed toluene was desorbed and degraded. Under continued constant organic loading
rate conditions, the system stabilized and functioned as a biological reactor. During a 77
hour, 400% step-OLR increase, adsorption again became important; excess substrate was
observed to accumulate on the GAC carrier particles. This material was desorbed and
degraded over an 11 day period following resumption the initial organic loading rate of 5.3
kg coo Im3-d. Thus it was established that the dampening effect of activated carbon is not

exhausted, but is recovered via bioregeneration.

It is clear from these results that GAC-FBR systems are advantageous for treating
BTX-contaminated waters when variations in loading rates are expected or when positive
removal is required during the start-up period. Furthermore, this concept provides a
foundation upon which future studies can be conducted in order to develop a rational

design procedure which incorporates criteria for transient loading performance.

CHAPTER 7

ESTIMATION OF KINETIC PARAMETERS FOR BIODEGRADATION OF
BENZENE, TOLUENE AND XYLENE

7.1 Introduction

Biological ﬂuidized bed reactors (FBR) have demonstrated the ability to treat
groundwater contaminated with petroleum hydrocarbons (Hickey et al. 1990b; 1991a;
Voice et al. 1992). Even when granular activated carbon particles are used as the bioﬁlm
carrier, biodegradation remains the most important removal mechanism in the system,
especially under steady-state conditions (Voice et al. 1992). Biodegradation kinetics,
therefore, are very important to describe this system and assist engineers during process

design.

A kinetic experiment can be used to evaluate the rate of biodegradation. One of
the most widely accepted rate equations is the Monod equation (Monod 1949) shown

here.

=_um_s_ (7-1)

S is the concentration of substrate (mg/L), u is the speciﬁc growth rate constant (l/hr), pm
is the maximum speciﬁc growth rate constant (1/hr), and Ks is the half-saturation
coefﬁcient (mg/L). It should be noted that the Monod equation is strictly an empirical

equation (Grady and Lim 1980). Lawrence and McCarty (1970) modiﬁed the equation to:

144

\\

vs;

 

145 J ....a
dS = _ kSX (7_2)
dt K, + S
Where: u = _YlS_ and um = kY
th

X is the concentration of biomass (mg VSS/L), k is the maximum speciﬁc utilization rate
of substrate (mg substrate/mg VSS-day), K8 is the half-saturation coefﬁcient (mg/L), and

t is time (day). Y is the yield coefﬁcient of biomass (mg VSS/mg substrate).

The _kineﬁqparamete§_ can be evaluated_by {bitch experimejt. In this approach,

 

 

a compound, as the sole organic constituent, is introduced into a reactor containing the
required Ms, electron acceptor, and growth factors. The reactor is then inoculated
with a/Qown amount oLQc_roorgmisms. Samples are collected periodically to monitor
degradation of the compound and growth of the microorganisms. The kinetic coefﬁcients

can then be estimated from Eqn. 7-2.

Several studies have been conducted to estimate the kinetic coefﬁcients for the
biodegradation of BTX (Table 2-1). There is a tremendous difference in the reported
values, however, and different values for different environments should be expected. For

the mixed culture used during this work, the parameters had to be evaluated.

Interactions between BTX during biodegradation have been evaluated by several
researchers. Goldsmith (1988) showed a slightly higher maximum utilization rate for
toluene in a BTX mixture by an enriched mixed culture. Arvin et al. (1989) studied
substrate interaction during aerobic biodegradation of benzene. They found that the
degradation rates of benzene were higher when toluene or xylene was also present in the
environment. Alverez and Vogel (1991) investigated the interactions of BTX during
biodegradation by two pure cultures and a mixed culture from an aquifer. The enhanced
degradation patterns of benzene and p-xylene in the presence of toluene was revealed.

The degradation of benzene was inhibited by the presence of p-xylene for a toluene

146

degrader. In their studies, toluene could not be degraded by a benzene degrader unless
benzene was also present. Recent studies by Chang et al. (1993) revealed competitive
inhibition and cometabolism during biodegradation of BTX by two Pseudomonas
isolates. Strain B1, which grew on benzene and toluene as sole sources of carbon and
energy, has the ability to transform p-xylene by consumption of the growth substrates
(benzene and toluene) or consumption of biomass during cometabolic degradation. The
degradation of benzene, however, was inhibited by the toluene even though the
degradation rate of toluene was not effected. For strain X1, which grew on toluene and p-
xylene but not benzene, the degradation of p-xylene was inhibited by the presence of
toluene, but the degradation rate of toluene was not effected. Competitive inhibition and
cometabolism effects were quantiﬁed in their studies. For a system treating a mixture of
BTX, complete degradation patterns should be expected. Such patterns may depend upon
the environmental conditions, composition of the waste, and maturity of the microbial
population

In this chapter, the kinetic parameters for the degradation of BTX by a mixed
/' .

culture (gym a 12153195ng gaidiizedhed inactjo‘Itteating BTX.991_1t_aIr1in.a_ted aonndaatcr)

were estimated. ”,3

O

7.2 Materials and Methods
7 .2.1 Biodegradation Kinetic Assays

Biomass was collected from a biological ﬂuidized bed (FBR) reactor treating
groundwater contaminated with benzene, toluene, and p-xylene (Voice et al. 1992). Non-
activated carbon (without adsorption capacity) was used as the bioﬁlm carrier in the FBR.

After a mature bioﬁlm developed, a small amount of biomass coated material was

removed from the FBR. The biomasswavsﬂstijpped ﬁW purged with
I“ . ﬂ’

147

oxygen to remove residual BTX, and homogenized by; shaking the biomass using glass 7

beads in a 300 mL glass bottle. Kinetic assays were performedtugggﬁﬁggﬂml, 3145i
Wm a Lucr-lock valve A stirring bar (6 mm in diameter and 15 mm \‘ ,,...I

’I'.‘ ‘4“
- aﬂﬂt -. ﬂ
. .01' J.
1 f ‘

long) was placed in the syringe and the syringe was placed on a magnetic stirrer to I;

provide mmg.ﬂe growth media consisted of groundwater amended with nitrogen and

-._._._---\_’

WhLEQO of 100/5/ l:COD/NQJrior to addition of the BTX this Q , i,» t

 

water was oxygenated with pure oxyggi to obtain an initial dissolved oxygen ‘9

 

concentration of greater than 30 mg/L. Stock substratesof benzene, toluene and p-xylene
were dissolved in tap water and then measured amounts of these solutions were
transferred to the syringes. The initial concentration of BTX varied from 1.2 to 15 mg/L
for each substrate. Three different substrate compositions were used in the assays/Thea.
Wsngngimnly one substrate (benzene, toluene or p-xylene) as a single
substrate and seven experiments were conducted./ ﬁe seQQIIQQQmJDQSitIQQEQmQIDPSI two
substrates (benzene and toluene, benzene and p-xylene or toluene and p-xylene) as a dual
substrate system with a 1:1 substrate ratio and six experiments were conducted. The_thi_r_d___
com ' ' 'ned all three substrates (benzene, toluene and p-xylene) in a constant
ratio of B:T:X of 1:1:1 and ﬁve experiments were conducted he initial concentration of
biomass varied fromJ} to L20mgj§§££ for all of the assays. After biomass was
introduced into the syringe, a one-mL liquid sample was taken. Subsequent one-mL
samples were taken at regular intervals to track degradation. W
(VSS) were measured according to Standard Methods (greenberg et al. 1985). All assays

 

were conducted at 20 °C.

7.2.2 Yield Coefﬁcient Assay

The net yield coefﬁcient (Y) was measured using a procedure similar to that used

for the kinetic assay. One substrate was added to the syringes at a time for the single

148

substrate measurement. Benzene and toluene were used to estimate the yield coefﬁcients
for the dual substrate condition. All three substrates were added to the system to estimate
the coefﬁcient under the three substrate condition. Liquid samples were taken at the
beginning and end of each experiment (about 6 to 8 hours) to measure the substrate and

biomass concentrations.

7.2.3 Decay Coefﬁcient Assay

One hundred mL of biomass in growth media (100 mg VSS/L) was transferred to
a 250 mL glass, bottle with cap. The biomass was collected from thewEBB and treated
using a similar procedure as used in the kinetic assay. Subsequent one-mL samples were

taken at was to track the disappearance e biomass. The concentration of

 

 

 

biomass was determined by an optical density method using a spectrophotometer (PE
I K
Lambda 6 UV/VIS) at 500 nm. The optical density was converted to VSS using an

external calibration curve with the same mixed culture.

7.2.4 Theory

The modiﬁed Monod equation (Eqn. 7-2) was used to evaluate the degradation

capability of the microorganisms in the FBR. The growth of biomass can be described as:

X = X0 + Y(S0 - S) (7-3)
where: S = concentration of substrate (mg/L).
So = initial concentration of substrate (mg/L).
X = concentration of biomass (mg VSS/L).
X0 = initial concentration of biomass (mg VSS/L).

Y = net yield coefﬁcient of biomass (mg VSS/mg substrate).

Agonm

149

Combining Eqn. 7-2 and 7- 3, integrating and rearranging results in:

1 .3 W . ‘lt .. 11
t=— ——ln — - ——+— In
R (x,+Ys,) s (X0+YSO) Y X0+Y(So—S)

 

The decay of biomass without substrate can be described as follows:

EX. = _bx
dt
where: b = decay coefﬁcient (1/day)

The integrated form of Eqn. 7-5 is:

 

The net yield coefﬁcient (Y)can be expressed as follows:

_. X—XO
so-s

 

Y

7.2.5 Estimation of Parameters

(7-4)

(7-5)

(7-6)

(7-7)

(7-8)

A decay coefﬁcient was estimated using Eqn. 7-7 and a non-linear regression

computer software program (Solver, Excel 4.0). The net yield coefﬁcients were

calculated using Eqn. 7-8. The kinetic parameters, maximum speciﬁc utilization rate of

substrate (k) and half-saturation coefﬁcient (Ks), were estimated using Eqn. 74 and non-

linear regression. The following criteria were applied during selection of experimental

data and parameter estimation.

1). discard data sets with fewer than 6 usable data points;

 

150

2). discard data sets with an initial substrate concentration higher than 45 mg
COD/L;

3). discard data set with end substrate concentrations higher than the GC detection
limit (about 15 ug/L for each of BTX);

4). if a lag-phase was observed, discard the data points during the lag-phase, as
deﬁned by a rate of substrate disappearance signiﬁcantly (50%) lower than the
rate at later time period, and use a calculated initial substrate concentration
estimated from the intercept of an extension-line of the ﬁrst three points after
the lag-phase on a graph of remaining substrate concentration vs. time;

5). estimate k and Ks simultaneously.

7.2.6 Chemicals and Analytical Procedure

Benzene, toluene, p-xylene (BTX), potassium phosphate dibasic and ammonia
chloride were obtained from J.T. Baker Chemical Co., Phillipsburg, NJ. Michigan State

University tap water was used directly as a source of groundwater. This relatively hard

water 450 m as CaCO3) is pumped from a deep aquifer underlying the campus and

distributed without further treatment.

 

BTX concentrations were determined using a gas chromatograph equipped with a
ﬂame ionization detector (FID) using helium as the carrier gas (Perkin-Elmer, model
8700), coupled with an automatic headspace sampler (Perkin-Elmer, model HS-101).
Separation was accomplished using a 30 m x 0.53 mm DB-624 column (I & W
Scientiﬁc). Water samples collected from the kinetic assays (1 mL) were transferred to 20
mL glass vials with Teﬂon®-coated septa and aluminum crimp-seals. The samples were
equilibrated for 1 hour at 80 °C and an aliquot of the headspace gas was injected into the

gas chromatograph. The detection limit of this procedure was 10 Eg/L for eachﬁBT“

U V‘HIII _ 7,“
( HE \MMwVana, ‘0

151

component. The accuracy for measurements of split samples containing a concentration

of 15 ug/L was {lug/L.

7.3 Results

7.3.1 Estimation Method Veriﬁcation

In order to test the variability of the results from the estimation procedure, a 15
point degradation curve was generated based upon Eqn. 7-4 using the parameters in Table
7 -1. The substrate concentration (S) was then randomly varied between -10 % and +10 %
of the theoretical value and ten groups of generated data sets were generated (Table 7-2
and Figure 7-1). The parameters were estimated from the data sets using Eqn. 7-4 with a
non-linear parameter estimation procedure. The results are shown in Table 7-1. The
relative standard deviation of the each created point was approximately 6%. The results
showed that the estimation can be signiﬁcantly affected by the variance in data which
resulted in a large error being introduced to the estimation of Ks (the relative standard

deviation is 36%).

Table 7 -1. Procedure veriﬁcation for parameter estimation

 

 

 

 

 

 

Parameters Theoretical Value Surfrournarl ()3 eesis ts % SD
Xo 114.2 114.2
So 9.1 8.92 6.91%
Y 1.1 16 1.1 16
Ks 0.299 0.285 35.91%
R 2.248 2.194 11.61%
r2 N f 1.000 _ 7 >0.985

 

 

 

152

Table 7-2 Evaluated data sets

Time

7“.

 

7.3.2 Decay Coefﬁcient

Two decay curves are shown in Figure 7-2. Was concentration decreased
M without signiﬁcantly changing 3119339333. The decay coefﬁcients were
calculated based upon Eqn. 7-7 using a non-linear regression procedure and presented in
Table 7-3. The data from the ﬁrst 9 and 8 days was used. The average decay coefﬁcient
value was found to be 0.049 l/day.

Table 7-3. Decay coefﬁcient determination

Experiment Dates Usable Days

 

9/15/92-9/24/93 9.06
9/3/93-9/17/92 7.88

 

 

 

 

 

 

 

153

 

Substrate Concentration (mg/L)

 

 

 

 

0 I I 4. I I ~"x-x-
0 10 20 30 40 50 60
Time (min)

Figure 7-1. Veriﬁcation of parameter estimation protocol

154

 

 

 

 

 

110
x Experiment
No.1
100
Regression
90 '3 Experiment
g No.2
E .
v 30 -------- Regressron
m
‘9 x
70 x x
60 -.
X
50 I i i i i I i
0 2 4 6 8 10 12 14 16

Time (days)

Figure 7-2. Decay of biomass harvested from the FBR

7.3.3 Yield Coefﬁcient

The yield coefﬁcients were determined under different combinations of the
substrates, single substrate(the initial concentration of benzene, toluene, and xylene were
18 mg/L, 10 mg/L, and 17 mg/L, respectively), two substrates (the initial concentration of
benzene and toluene were 4.8 mg/L and 5.1 mg/L, respectively), and three substrates (the
initial concentration of benzene, toluene, and xylene were 4.8 mg/L, 5.1 mg/L, and 3.9
mg/L, respectively). The net yield coefﬁcient was calculated from Eqn. 7-8, which
includes the effect of biomass decay. The coefﬁcients are reported in Tables 7-4 and 7-5.
The net yield coefﬁcient in a biological ﬂuidized bed reactor treating groundwater

contaminated with BTX was also measured by tracking the waste biomass and the COD

155

consumption during a two week period. The net yield coefﬁcient was 0.23 mg VSS/mg

COD at a solids retention time of 20 days.

Table 7-4. Net ield coefﬁcients for sin 1e substrates

Net Yield Coefﬁcient
Y
(mg VSS/mg
substrate) substrate)
Benzene 3 1.12 0.15
Toluene 3 1. 13 0. 13

 

 

 

 

 

 

 

 

Table 7-5. Net ield coefﬁcients for multi 1e substrates

 

 

 

 

 

 

Net Yield Coefﬁcient
Substrate n Y SD
(mg VSS/mg (mg VSS/mg
COD) COD)
Benzene and Toluene 1 0.22 -
Benzene Toluene and 2 0.17 0.04
X lene

 

7.3.4 Estimation of Biodegradation Kinetic Parameters

The kinetic parameters, maximum speciﬁc utilization rate of substrate (1:) and

 

 

half-saturation coefﬁcient (Ks), were estimated under different substrate combinations:
one substrate (B, T, and X), two substrates (BT, BX, and TX), and three substrates
(BTX). The disappearance of each substrate was measured allowing the parameters to be

estimated using Eqn. 7—4 and non-linear regression.

In order to avoid large variation during. kinetic parameter cstimatiim. MEL
kinetic assays were conducted. Seve__r_1__assa¥s were conducted for eachsubstrate in“ the

156

W. Six assays were conducted for each combination of two
substrates. Five assays were conducted for the three substrate case. The initial biomass
and substrate concentrations were varied to determine the effect on parameter estimation.
Compared to the large variation in the parameters, however, the initial concentration
effect was not signiﬁcant. Therefore, the results reported in this chapter are a

combinations all estimations without considering the initial concentration effect.

7.3.4.1 One Substrate

In this phase of the study, one substrate was added to the syringe with the results
indicating the ability of the mixed culture to degrade one substrate without interaction

effects caused by other substrate. Six sets of data satisﬁed the criteria for kinetic

-..- p...— ‘—

benzene were 13.8 to 1W and 1.5 to 9.6 mg/L, respectively. Five sets of data

 

satisﬁed the criteria for kinetic estimation of the degradation of toluene. The initial
concentrations of biomass and toluene were 49.2 to 114.2 mg VSS/L and 8.5 to 11.1
mg/L, respectively. Six sets of data satisﬁed the criteria for kinetic estimation of the
degradation of xylene. The initial concentrations of biomass and xylene were 27.6 to
228.4 mg VSS/L and 1.1 to 12.1 mg/L, respectively. Typical biodegradation curves for
each single substrate system are shown in Figures 7-3, 7-4 and 7-5. The kinetic
parameters are reported in Tables 7-6, 7-7, and 7-8. The average maximum speciﬁc
utilization rates of substrate (k) are 2.09, 3.71, and 1.59 l/day for benzene, toluene, and
xylene, respectively. The average half-saturation coefﬁcient values (K3) are 0.25, 0.84,

and 0.63 mg/L for benzene, toluene, and xylene, respectively.

157

 

 

 

 

10
x Experiment
8 Regression
g xx x
X
5 6 Il- \x\
a x\
a 4 .. X\x
e \
U x“
‘X
2 " iXx
Xxx
0 : : : : xxxx-xx-x—x
0 30 60 90 120 150 180
Time (min)

Figure 7-3. Typical biodegradation trend of benzene as a single substrate in a mixed
culture; (estimation results: k=1.87 1/day; Ks=0.16 mg/L; R2=0.989; n=l9).

 

Conc. (mg/L)

 

X

 

Experiment

Regression

 

Figure 7-4. Typical biodegradation trend of toluene as a single substrate in a mixed
culture; (estimation results: k=3.70 l/day; Ks=0.9l mg/L; R2=0.970; n=13).

 

158

 

 

 

 

12 x
3‘ x Experiment
10 -- .
A X ——Regress10n
3 ..
X
E» \
v 6 .P X X\
‘3 x
I:
8 41- "\
X
2 «i- \x\
XXX
0 : : : xxxxxxxxx-xx-x—
0 30 60 90 120 150 180
Time (min)

Figure 7-5. Typical biodegradation trend of p-xylene as a single substrate in a mixed
culture; (estimation results: k=1.83 1/day; Ks=0.56 mg/L; R2=0.991; n=l9).

Table 7 -6. Kinetic parameters for biodegradation of benzene by a mixed culture with and
without other substrates present

verage n

Substrate Substrate

 

159

Table 7-7. Kinetic parameters for biodegradation of toluene by a mixed culture with and
without other substrates present

verage n

Substrate

Toluene

 

Table 7-8. Kinetic parameters for biodegradation of xylene by a mixed culture with and
without other substrates present

verage n
Substrate

Toluene

 

7.3.4.2 Two Substrates

Two substrates were added to the syringe simultaneously in the second phase of
this study to produce results indicating the ability of the mixed culture to degrade two
interacting substrates. A constant ratio of two substrates (B:T, B:X, and T:X) of 1:1 was

used in most of the assays.

160

Four sets of data satisﬁed the criteria for kinetic estimation of the degradation of
benzene and toluene. The initial concentrations of biomass, benzene, and toluene were
48.3 to 64.6 mg VSS/L, 3.5 to 6.5 mg/L, and 4.9 to 6.4 mg/L, respectively. Typical
biodegradation curves are shown in Figure 7-6. The kinetic parameters are reported in
Tables 7-6 and 7-7. The average values of maximum speciﬁc utilization rate of substrate
(k) are 1.32 and 3.15 1/day for benzene and toluene, respectively. The average values of
half-saturation coefﬁcient values (K5) are 0.10 and 0.90 mg/L for benzene and toluene,

respectively.

There were six sets of data which satisﬁed the criteria for kinetic estimation of the
degradation of benzene and xylene. The initial concentrations of biomass, benzene, and
xylene were 24.0 to 96.0 mg VSS/L, 1.5 to 7.0 mg/L, and 1.3 to 6.8 mg/L, respectively.
Typical biodegradation curves are shown in Figure 7-7. The parameters are reported in
Tables 7-6 and 7-8. The average maximum speciﬁc utilization rates of substrate (k) are
2.39 and 1.33 l/day for benzene and xylene , respectively. The average value of half-

saturation coefﬁcients (K5) are 0.43 and 0.17 mg/L for benzene and xylene, respectively.

Only two sets of data satisﬁed the criteria for kinetic estimation of toluene and
xylene degradation. The initial concentrations of biomass, toluene , and xylene were 48.3
and 64.6 mg VSS/L, 3.9 and 4.3 mg/L, and 2.4 and 2.6 mg/L, respectively. The initial
concentrations of toluene were higher than the concentrations of xylene. Typical
biodegradation curves are shown in Figure 7-8. The parameters are reported in Tables 7-7
and 7-8. The average value of maximum speciﬁc utilization rate of substrate (k) are 3.12
and 1.18 1/day for toluene and xylene , respectively. The average value of half-saturation

coefﬁcients (K5) are 0.53 and 0.13 mg/L for toluene and xylene , respectively.

 

161

 

 

 

 

 

 

6
x Benzene
Experiment
Benzene
5 Regression
Toluene
Experiment
4 Toluene
Regression
E
. 3
U
a
e
U
2
1
\ x. . .
0 g : $°ﬂuottuuuuunquuuuud
0 30 60 90 120 150 180
Time (min)

Figure 7 -6. Typical biodegradation patterns of benzene and toluene as dual substrates in a
mixed culture; (estimation results for benzene: k=1.12 llday; Ks=0.03 mg/L; R2=0.973;
n=16; estimation results for toluene: k=2.17 llday; Ks=0.58 mg/L; R2=0.973; n=16).

 

162

 

x Benzene
Experiment

Benzene
Regression

n p-Xylene
Experiment

-------- p-Xylene
Regression

Cone. (mg/L)

 

 

\
x
0 : : .L‘xx R-pnnuunﬁrnutru—u

0 30 60 90 120 150 180
Time (min)

 

 

Figure 7-7. Typical biodegradation patterns of benzene and p-xylene as dual substrates in
a mixed culture; (estimation results for benzene: k=2.15 llday; Ks=0.45 mg/L; R2=0.991;
n=12; estimation results for p—xylene: k=1.29 llday; Ks=0. 16 mg/L; R2=0.998; n=15).

 

163

 

 

 

 

 

 

6
x Toluene
Experiment
:1
Toluene
5 -. Regression
n p-Xylene
Experiment
4 -------- p-Xylene
Re '
gressron
g a
E
. 3
U
I:
e
U
2
‘\
q\
x
l \
line
X\ 419
O 'r : x+xxxxx>sxx nuuuuﬂnu-
0 30 60 90 120 150 180
Time (min)

Figure 7-8. Typical biodegradation patterns of toluene and p-xylene as dual substrates in
a mixed culture; (estimation results for toluene: k=2.32 llday; Ks=0.49 mg/L; R2 =0.988;
n=10; estimation results for p-xylene: k=0.96 llday; Ks=0. 14 mg/L; R2=0.979; n=18).

164
7.3.4.3 Three Substrates

All ﬁve sets of data satisﬁed the criteria for kinetic estimation under the three
substrate condition. The initial concentrations of biomass, benzene, toluene , and xylene
were between 22.8 and 63.7 mg VSS/L, 1.3 and 3.4 mg/L, 1.2 and 3.5 mg/L, and 1.1 and
3.1 mg/L, respectively. Typical biodegradation curves are shown in Figure 7-9. The
parameters are reported in Tables 7-6, 7-7, and 7-8. The average value of the maximum
speciﬁc utilization rate of substrate (k) are 1.08, 2.12, and 0.79 llday for benzene,
toluene, and xylene , respectively. The average value of half-saturation coefﬁcients (K5)

are 0.18, 0.71, and 0.13 mg/L for benzene, toluene, and xylene, respectively.

To investigate the substrate effect interaction has on the degradation rate, three
experiments were conducted for the BTX system. In the experiments, the initial
concentration of one of the three substrates was three times higher than the other two
compounds (Table 7-9). The initial biomass concentrations were 63.7 mg VSS/L. The

parameters were estimated and are reported in Table 7-9.

Table 7-9. Kinetic parameters for biodegradation of BTX by a mixed culture with varying
initial concentrations.

0 Benzene o uene D- y ene
(mg/L) S0 R Ks SO Ks S0 k 5
(mg/L) (llday) (mg/L) (mg/L) (llday) (mg/L) (mg/L) (lldaY) (mg/L)

3.7 5.74 1.67 0.37 1.34 1.08 0.72 1.34 0.36 0.05

3. 1.76 0.47 0.02 6.47 2.57 0.69 1.58 0.36 0.018

63.7 1.72 0.73 0.31 1.59 7.27 1.84 5.81 1.00 0.35

 

 

 

 

 

 

 

 

 

 

 

 

 

 

165

 

x Benzene
Experiment

Benzene
Regression

D Toluene
Experiment

-------- Toluene
Regression

A p.Xylene
Experiment

N

————— p-Xylene
Regression

Cone. (mg/L)

 

 

 

D. A\
0 g i ME'ﬁQuBB-BIBBBBBBUBB-K
0 30 6O 90 120 150 180
Time (min)

 

Figure 7-9. Typical biodegradation patterns of benzene, toluene and p—xylene as multiple
substrates in a mixed culture; (estimation results for benzene: k=1. 15 llday; Ks=0.24
mg/L; R2=0.992; n=1 1; estimation results for toluene: k=2.43 llday; Ks=1.33 mg/L;
R2=0.987; n=12; estimation results for p—xylene: k=0.92 llday; K,=0.05 mg/L; R2=0.982;
n=l4 ).

 

166

 

 

 

 

 

 

7
AX X Benzene
Experiment
6 _ —— Benzene
F Regression
X
X D Toluene
X Experiment
5 - _
-------- Toluene
X Regression
a 4 _ _ X A p-Xylene
E Experiment
T: ————— p-Xylene
‘5 X Regression
° 3
U — n-
X
2 . ..
X
1
X
‘A X
\ "x
o "Mun-3883335113534
90 120 150 180
Time (min)

Figure 7-10. Biodegradation patterns of benzene, toluene and p-xylene as multiple
substrates with a high initial benzene concentration in a mixed culture; (estimation results
for benzene: k=1.67 llday; Ks=0.37 mg/L; R2=0.986; n=16; estimation results for
toluene: k=1.08 llday; Ks=0.72 mg/L; R2 =0.978; n=11; estimation results for p-xylene:
k=0.36 llday; Ks=0.05 mg/L; R2=0.985; n=15).

167

 

X Benzene
Experiment

6 ‘_~‘ —Benzene
. Regression

\ a Toluene
‘ Experiment

\ -------- Toluene ,
‘\ Regression
‘ '.

4 u ‘ A p-Xylene
\ Expenment

 

. ----- p-Xylene
t Regression

Cone. (mg/L)

 

 

 

"mﬂ-TiuﬂuB-puulllﬂluBB-ﬁ
90 120 150 180
Time (min)

 

Figure 7-11. Biodegradation patterns of benzene, toluene and p-xylene as multiple
substrates with a high initial toluene concentration in a mixed culture; (estimation results
for benzene: k=0.47 llday; Ks=0.02 mg/L; R2 =0.976; n=11; estimation results for
toluene: k=2.57 llday; Ks=0.69 mg/L; R2=0.949; n=12; estimation results for p-xylene:
k=0.36 l/day; Ks=0.01 mg/L; R2=0.971; n=12).

168

 

*A A X Benzene
\ Experiment

\ -——-Benzene
5 -n— \ A Regression

\ A n Toluene
\ Experiment

4 _ _ \ -------- Toluene
\ A Regression

\ A p«Xylene
\ Experiment

3 - . \ ————— p-Xylene
\ Regression

Cone. (mg/L)
D
'6

 

 

\ x
bk} A
Rafxt
0 J, :°"§D5+unmu?gﬂumﬂDBIK-A——
0 30 60 90 120 150 180 210
Time (min)

 

 

 

Figure 7-12. Biodegradation patterns of benzene, toluene and p-xylene as multiple
substrates with a high initial p-xylene concentration in a mixed culture; (estimation
results for benzene: k=0.7 3 llday; Ks=0.31 mg/L; R2=0.994; n=11; estimation results for
toluene: k=2.27 llday; KS=1.84 mg/L; R2=0.995; n=11; estimation results for p-xylene:
k=l.00 llday; Ks=0.35 mg/L;R2=0.986;n=19).

 

169

7.4 Discussion

 

t._-~——' W -.——-

 

The net yield coefﬁcient for each BTX compound is about 0.36 mg VSS/mg

 

_COD, which is a typical value for geggpigmigroqrgamnismmso Each 0f the BTX

 

contaminants could be degraded by the organisms as a sole substrate. When both benzene
and toluene were present in the system, the net yield coefﬁcient decreased to 0.22 mg

VSS/mg COD. When all three substrates were available to the microorganisms, the

..—.—.

W by two Pseudomonas isolates isolated from the same FBR system
used in this study. He documented a similar yield coefﬁcient value but a much larger
value of decay coefﬁcient for toluene by a Pseudomonas strain Bl which could grow on
benzene or toluene. He also found an increase in decay rate with an increase in xylene
concentration, while the xylene was cometabolized by strain Bl. Therefore, a decrease in

the net yield coefﬁcient was expected in the multiple substrate system.

Estimation of kinetic parameters is delicate work due to the signiﬁcant variation
which can be associated with the procedure. The relative standard deviation in k for the
generated data sets was 11.6% based on the veriﬁcation test. The degradation ability of
the mixed culture for BTX was high as indicated by the high k and relatively low Ks. In
the veriﬁcation test, the relative standard deviation of Ks was 35.9% for the sets of
generated data which had a relative standard deviation of 6%. Therefore, the comparison
of the k values at Wﬁoyﬂs was possible, but it was very difﬁcult to draw any

conclusions regarding a trend in the Ks under those conditions.

There are several reasons for the variation in estimatmgﬁresﬁults. Because of the
low value of Ks, there are a limited number of points which occur during the transient
period of the degradation curve. This transient phase is located between the end of the
zero-order degradation period and the point at zero substrate concentration. Therefore, an

accurate estimate of Ks was very difﬁcult. A second reason for variation is due to

170

collection of biomass from the FBR. Because the bioﬁlm thickness was approximately
0.1 m, it is possible that different activity levels and even different populations of
microorganisms occur at different depths within the bioﬁlm. At the time of measurement,
because the FBR was operated without recycle which allowed a signiﬁcant substrate
concentration gradient to develop along. the height of the reactor, the activity was found
to decrease with an increase in reactor height. When each of the kinetic assays was
conducted, the bioﬁlm coated particles were collected from the same position in the
reactor and the biomass was stripped from the carrier particle. However, this did not

guarantee that the biomass obtained had the same activity.

The mixed culture in the ﬂuidizedﬂbedt reactor can degrade BTX. The rates,

 

 

 

however, were different for the individual substrates and affected by the presence of the
other substrates. Using a single substrate kinetic test allowed the potential ability of the
community to degrade each substrate to be found. This community had the highest
maximum speciﬁc utilization rate for toluene, followed by benzene, and ﬁnally by

xylene, which was the lowest.

When multiple substrates were present in the system, inhibited degradation of
some substrates occurred. The xylene degradation rates were decreased by the presence
of benzene and toluene. However, the benzene degradation rates were decreased by
toluene while remaining virtually unaffected by the presence of xylene. The presence of
benzene and xylene did not effect the toluene degradation rates. The maximum speciﬁc
utilization rates were also effected by the initial concentrations of substrate. When the
initial concentrations of benzene were three times higher than the other substrate
concentrations (Table 7-9), the value of k for benzene was higher and the values of k for
toluene and xylene were lower than the values for equal concentration cases. A similar
phenomena occurred for high concentrations of toluene. When the concentration of

xylene was higher, the k for xylene increased and the k for benzene decreased, while the

 

171

k value for toluene remained unchanged. Even though the maximum speciﬁc utilization
rates were changing based on initial concentrations, all of the substrates were utilized

simultaneously. Therefore, sequential utilization was not observed.

Kinetic parameters for the degradation of BTX have been estimated by several
researchers. There are tremendous differences in the values reported for BTX
degradation. Comparison of those results is difﬁcult and not very useful, because many
different microorganisms are able to degrade BTX and the rates are affected by
environmental conditions. The community structure in the FBR is very complicated.
Chang (1993) isolated ﬁve different strains with different degradation characteristics
from the same FBR. Those strains were not necessarily the dominant groups in the
community. Inhibition of the substrates was observed, however, due to the complex
nature of the community and the large variance in the estimation of Ks, the type of

inhibition was not determined.

Large variation in kinetic parameters was found in this study. This fact should not
limit use of the results. Because of the continuous change in activity and population of
the microorganisms in the bioreactor, the variation in substrate utilization rates are
expected. Engineers will need to consider such uncertainty in their designs (such as

increase the height of reactor).

7.5 Conclusions

The yield and WMor a mixed culture treating BTX were evaluated

 

 

and the kinetic parametersfgggcgradationgiﬁmere determined. The degradation rate
for toluene was the highest, while the rate for xylene was the lowest. There was inhibition

between all three substrates but £6 _type of the inhibition was not determined, The effect

 

of inhibition on toluene degradation was limited. A large variance during estimation of

. “new
I

172

Ks was found and may be due to its low value for this mixed culture. The parameters,

however, are useful for bioreactordesignv‘vithkaconsideration of uncertainty.

 

 

CHAPTER 8

ADSORPTION CAPACITY OF BIOFILM COATED ACTIVATED CARBON IN A
BIOLOGICAL FLUIDIZED BED REACTOR SYSTEM

8.1 Introduction

Biological granular activated carbon ﬂuidized bed reactors (GAC-FBR) have been
shown to have the ability to remove the gasoline contamination from groundwater and
provide both the efﬁciency of biological removal and the positive efﬂuent protection
capability of activated carbon adsorption (Hickey et al. 1990b; 1991a). In the previous
chapters, the role of the adsorption in a GAC-FBR was discussed. The adsorption capacity
was found to be a very important factor during transient conditions. However, the change

in adsorptive capacity during operation of the GAC-FBR has not been fully explored.

The adsorptive capacity of GAC can decrease with increased usage in BAC
systems. This reduction could be due to adsorption of some non-degradable compounds
from the inﬂuent but may also be caused by microorganism growth on the surface of the
carbon. Suidan et al. (1983) reported that the adsorption capacity of GAC decreased after
the carbon was exposed to a phenol mixture, which contained some compounds resistant to
biodegradation, in an anaerobic expanded-bed reactor. The capacity was, however, slightly
regained after a period of bioregeneration. Zhang et al. (1991) found that a similar
phenomena occurred in an aerobic BAC ﬁlter used to treat phenol. They did not, however,
investigate the effect bioregeneration had on the reactor. Weber and Ying (1978b) asserted

that breakthrough of toluene sulfate in an expanded-bed GAC adsorber with a bioﬁlm

173

 

174

grown on sucrose was earlier than an adsorber without the biocoat. When Schultz and
Keinath (1984) conducted a study of powdered activated carbon treatment (PACT) process
mechanisms in an activated sludge system, they found that the biomass in the PACT culture
impedes the rate of mass transfer of phenol to the PAC surface. Olmstead (1989) studied
the effect of microbial interference on GAC adsorption. In his experiment, a column
containing GAC particles was fed glucose for 1 -2 weeks. After the bioﬁlm developed, the
carbon was removed from the column in sections to obtain the biocoated GAC with
different amount of biomass. Because of the plug-ﬂow pattern in the column, more
bioactivity and biomass were expected at the inﬂuent end of the column. He found that the
multiplier term in the Freundlich isotherm equation, Kf, for trichloroethylene (TCE)
decreased with increasing biomass on the carbon while the slope term of the equation, n,
remained constant. The adsorption of para-toluene sulfate (PTS), which has a much lower
K0W than TCE, did not decrease signiﬁcantly. The author failed to determined the effect of
a sample protein (BAS) had on the adsorption of PTS, but, discovered that preloading
soluble microbial products (SMP) on the GAC reduced the amount of PTS adsorbed.

In this chapter, the adsorption capacity of bioﬁlm coated activated carbon from a
biological ﬂuidized bed reactor which treated toluene contaminated water was determined

periodically over an extended period of operation.

8.2 Materials and Methods

8.2.1 Media and Chemicals

Granular activated carbon (GAC Calgon Filtrasorb 400, Calgon Co., Pittsburgh, PA)
was used as the carrier media in the GAC-FBR. The media were sieved to obtain a 18 x 20
mesh fraction (average particle diameter 0.90 mm). Prior to addition to the reactors, the

GAC was rinsed with distilled water to remove ﬁnes and dried at 100 °C for 24 hours.

175

All chemicals were obtained from J .T. Baker Chemical Co., Phillipsburg, NJ and
were reagent grade. The water used to feed the reactor was from an industrial grade soft

water system which produces water with a hardness of 103 mg CaCOglL.

8.2.2 Fluidized Bed Reactors

Pilot-scale ﬂuidized bed reactor was constructed using a 7.6 cm diameter by 322 cm
high polyvinyl acetate (PVC) column to produce a system with a 14.7-L working volume
Oiigure 8-1). The reactor was charged with activated carbon and operated as single-pass
systems with no recycle. The reactor were inoculated with a mixed culture obtained from a
laboratory-scale FBR that was supplied with toluene as the sole carbon source. Dissolved
oxygen (DO) was supplied to the reactors by oxygenating the reactor feed water with pure
oxygen to a concentration sufﬁcient to maintain an efﬂuent DO exceeding 4.0 mg/L.
Nitrogen and phosphorus were supplemented at a weight ratio of 100/5/1:COD/N/P.
Toluene was injected into the feed water using a syringe pump (Harvard Apparatus 22). A
mixing tank (15 liters) and a recycle loop around the mixing tank were installed in the
second run of this study to improve dissolution of the pure toluene in the inﬂuent. This
toluene contaminated groundwater was then pumped to the bottom of the reactor using
peristaltic pumps (W atson-Marrow 6018). The bed height was controlled by a peristaltic
pump installed on the top of the carbon bed to shear some biomass from the carrier
medium. The reactor was charged with 2850 g of media initially to produce a settled bed
height of 156 cm. Toluene contaminated water was fed at a ﬂow rate of 2.2 IJmin resulting
in a hydraulic ﬂux rate of 0.48 m/min, an empty bed hydraulic retention time of 6.7 min,
and an organic loading rate of about 1.4 kg COD/m3-day. The temperature of the feed
water was 20:] 0C and the pH was about 8.

176

 

    
 

 

 

 

Treated
Water <— T
:- _> Efﬂuent
:.j Samplrng
Fluidized I31; Oxygen
Reactor 3-
; Groundwater OVCI'ﬂOW
._: —>
3: Inﬂuent ___
5.: Sampling Oxygenation Nutrient Tank
3 Tank
~:-: '
E:
Feed
Pump

Figure 8-1. Schematic of the experimental GAC-FBR system.

8.2.3 Collection and Sterilization of Bioﬁlm Coated Activated Carbon

Carbon samples were collected from two points in the reactor (50 cm and 150 cm
from the bottom of the reactor) using a sampling bottle (50 mL). The bottle was ﬁlled with
water, closed with a rubber stopper and lowered into the reactor to the desired height. The
stopper was then removed and the carbon allowed to ﬁll in the bottle. After one minute ﬁll,
the bottle was removed from the reactor. The carbon was transferred to 50 mL serum
bottles ﬁlled with the efﬂuent water (headspace free) and sealed with Teﬂon-coated septa

and aluminum seals.

The carbon and water ﬁlled bottles were sterilized by gamma radiation in a cobalt-
60 irradiator located at the Phoenix Memorial Laboratory (Ann Arbor, Michigan). The

samples were sterilized for one hour at a dose rate of 2 Mrad/hr. The effect of the

 

177

sterilization procedure was tested by plating the sterilized carbon samples on nutrient agar.

There was no colony growth after two weeks incubation at 25 0C.

There are two reasons to consider that the sterilization procedure may change the
adsorption characteristics of the GAC. The ﬁrst possibility is that gamma radiation may
change the carbon molecular structure and adsorption characteristics. The second
possibility is that the radiation may result in biomass alteration, creating adsorbable
substances which may reduce the adsorption capacity. Therefore, two tests were conducted
to investigate the effect of sterilization on the adsorption capacity of the GAC. In the ﬁrst
test, two clean GAC samples, one with and one without sterilization, were used for the
isotherm studies. The results indicate that sterilization does not effect the adsorption
capacity of the GAC. In the second test, a clean GAC sample (1 g) was mixed with 20 mL
of 500 mg VSS/L fresh biomass to create a similar concentration of biomass on the bioﬁlm
coated carbon. The mixture was sterilized by gamma radiation. After a 48 hour contact
time, this mixture was used to perform an isotherm study along with a clean GAC sample
for comparison purpose. The results are reported in Table 8-1 and show that the biomass

addition does not signiﬁcantly effect the adsorption capacity of the clear GAC.

8.2.4 Measurement of Bioﬁlm Thickness

Carbon particles were photographed after the sterilization using a stereo microscope
(Olympus SZH) equipped with a camera system (Olympus AD Exposure Control Unit).
Because the bioﬁlm color (light yellow) was signiﬁcantly different from the color of the
carbon (black), the bioﬁlm thickness could be directly measured from the photograph. The
thickness of bioﬁlm on the carbon was often not uniform, therefore, three particles were
selected randomly and ﬁve measurements were taken from each particle at equally

distributed locations. An average value was then calculated.

178

Table 8-1. Effect of biomass and base washing on isotherm parameters

  
   
   

Carbon
Description

 

   

Data
Pornt ‘10 (mg/ g)

E ect o Biomas

         

Isotherrn Parameters

Bioﬁlm

 

S

Kf (mg/9| 1/n

Thickness
(um)

 

   
 
  

Clean GAC

10

71.9:t3.4i

      
     

0.39i0.0 -
2

 

   

Clean GAC with Biomass

10

      

78.8:l:5.4

 

  

 

    
   

O.35:t0.0
4

 

     

Non-disturbed BAC

 

10

6.38

  

66.7:t2.8

  

   

  

0.37:|:0.0
2

     

 

     

   

Bioﬁlm Removed

10

 

5.90

62.5:t0.9

   

   

    

  

O.36:t0.0

 

 

1

 

        

   

Non-disturbed BAC

29

3.71

34.4:1:1.2

  

   

0.54:1:0.0

 

2

 

   

Bioﬁhn Removed

   

10

 

   

Base V ash ( .

3.36

  

37.5:t0.7

  

  

0.54:t0.0

 

M NaOH)

 

 

 

 

          

374

     

N on-disturbed BAC

       

 

10

0.88

27.9:t3.1

 

 
       

   

374

 

:1:: 95 % conﬁdent interval

Washed with NaOH

 

10

 

    

0.85

32.2:t1.7

 

 

 
   
   

 

179

8.2.5 Extraction Procedure

To measure the amount of adsorbed toluene, the carbon samples (0.03 to 0.1 g)
were placed in aluminum dishes, separated from the liquid using adsorbent tissue and
transferred to 15 mL glass vials. Fifteen mL of methanol was added to each vial as the
extraction solvent, then, the vials were capped with Teﬂon-faced septa and aluminum seals
and tumbled at 6 rpm for 3 to 4 days. Analysis involved transferring 0.1 mL of the extract
to a 20 mL headspace vial containing 5 mL DI. water and 0.12 mL 6 M HCl, sealing with
Teﬂon-faced septa, and analyzing by headspace gas chromatography. Three successive
extractions were performed. The mass of carbon was determined by drying at 90 0C
overnight, desiccating, and weighing on an analytical balance. The total mass of adsorbed
toluene was calculated based on the summary of three extractions. The efﬁciency of the

extraction procedure is higher than 90% (Shi et al. 1994).

8.2.6 Adsorption Isotherm Assay

The isotherm solution was prepared by adding calcium carbonate (5 mg as Ca/L),
magnesium carbonate (0.6 mg as Mg/L), and sodium bicarbonate (15 mg/L) to DI. water
and adjusting to pH 8.0 by adding sodium hydroxide. The toluene solution was prepared
by dissolving toluene in the isotherm solution followed by ﬁltration (0.22 um ﬁlter,
Whatrnan PolyCap 75 AS) for sterilization. 160 mL serum bottles were then ﬁlled, capped

and stored at 4 0C until use.

Isotherms for toluene were performed using serum bottles (160 mL) sealed with
Teﬂon-coated septa. The serum bottles contained the desired volume of isotherm solution
and bioﬁlm coated activated carbon (60 mg) and were then ﬁlled, without leaving a head

space, with toluene solution to achieve a desired concentration of toluene. Ten to ﬁfteen

180

bottles, each with different initial concentrations between 10 and 70 mg/L, were prepared
for each assay. The bottles were tumbled at 6 rpm and 20 0C for 7 days to ensure that
equilibrium was obtained. The mass of carbon was determined by drying at 90 °C
overnight, desiccating, and weighing on an analytical balance. After equilibration, ﬁve
milliliter liquid samples were transferred to headspace vials, sealed with Teﬂon-coated
septa, and analyzed by headspace gas chromatography to determine the residual amount of
toluene. The solid-phase concentration was calculated by difference. An isotherm assay
was also performed with new activated carbon. All of the serum bottles and septa were

sterilized by autoclave.

To evaluate the equilibration time, four isotherm studies were performed using
different equilibration times (4, 14, 21, and 28 days, respectively). Each isotherm curve
contained ﬁve points and the equilibrium concentrations in the water ranged from 0.5 mg/L
to 10 mg/L. There were no signiﬁcant differences in the adsorption results, therefore,

seven days was chosen as a convenient time to allow equilibrium to be occurred.

In order to study the effect the amount of biomass on the carbon adsorption, the
bioﬁlm on a carbon sample collected on day 122 was removed prior to sterilization by
shaking in the presence of a plastic tube to knock the bioﬁlm free. The carbon was then
used in an isotherm study. The same procedure was performed on a carbon sample
collected on day 64, and the results were compared to those found previously without

bioﬁlm removal.

8.2.7 Sodium Hydroxide Wash

The protein and other biological products can be digested and dissolved by a base
wash. To investigate the possibility of the base wash for the recovery of the adsorption

capacity, a carbon sample collected on day 374 was digested with 0.5 M sodium hydroxide

181

(NaOH) at 20 °C for 24 hours and washed with 0.5 M NaOH three times. Then, the
carbon rinsed with isotherm solution until the pH of the leachate reached 8. An isotherm

experiment was performed on this carbon and the adsorption parameters were estimated.

8.2.8 Estimation of Adsorption Parameters

The extended Freudlich isotherm equation was used to estimate the adsorption
characteristics of the carbon. This relationship follows.
q. +q. =K.CE (3-1)
where: qe is the additional amount of toluene adsorbed (mg/g).
q0 is the initial amount of adsorbed toluene (mg/g).
Kf is a constant.
CC is the equilibrium concentration of toluene in water (mg/L).

1/n is a constant.

The initial amount of adsorbed toluene was measured by the extraction procedure
described previously. The isotherm data was ﬁt to the equation using a non-linear
regression computer software program (Solver, Excel 4.0, Microsoft Inc.) and both
parameters were estimated simultaneously. The 95% conﬁdence intervals for each

parameter were also determined (Beck and Arnold 1977).

8.3 Results
8.3.1 Fluidized Bed Performance and Collection of Carbon Samples

The ﬂuidized bed reactor was operated in two phases. The ﬁrst phase lasted for 406

days. The long term effect of operation time on the adsorption capacity of biocoated GAC

182

was monitored and the recovery of capacity after stop of toluene feed was investigated. The
second phase lasted for 77 days and the short term changes of the adsorption capacity were

determined.

The inﬂuent and efﬂuent toluene concentrations and the DO consumption in the ﬁrst
phase are shown in Figure 8-2. The average inﬂuent and efﬂuent concentrations of toluene
were 1.71 mg/L and 0.10 mg/L, respectively. The average toluene removal rate was 94%.
The ﬁrst carbon sample was collected at day 94 and collected every 30 days thereafter until
day 202 while toluene feed was stopped. In order to understand the capacity recovery
potential, toluene feed was stopped and the nutrient supplies were reduced by half at day
202. To investigate the effect of supplied nutrient levels on the recovery of the adsorption
capacity, nutrients stopped being supplied at day 259 and resumed on day 304. Carbon
samples were collected on days 255, 304, and 374. On day 375, the toluene feed was
resumed to allow new biomass growth for the start of the second run. All of the carbon

samples were collected from mid-depth of the carbon bed (150 cm).

The FBR was operated for 77 days for the second phase. A modiﬁcation of the
inﬂuent mixing unit produced a more stable and higher inﬂuent toluene concentration than
that in the ﬁrst phase. The inﬂuent and efﬂuent toluene concentration and the DO
consumption are shown in Figure 8-3. The average inﬂuent and efﬂuent toluene
concentrations were 2.73 mg/L and 0.08 mg/L, respectively. The average toluene removal
rate was 97%. Signiﬁcant DO consumption started on day 5 and the carbon bed height
reached a stable level at day 50. The carbon samples were collected from the start of the
second run. In the ﬁrst week, the carbon samples were collected twice a week, then
collection was reduced to once a week for the next four weeks. For the ﬁnal six weeks,
carbon was sampled every two weeks. The carbon samples were collected from the bottom

portion of FBR at 50 cm and the middle portion of the FBR at 150 cm.

183

 

 

 

 

 

 

 

 

 

SEE 25.280 5»th
w 8 6 4 2 0
a
\da
ﬁlm;
|-1-..n_h4
Qing- lllll E 1
mm
mm
ml
6

SEAS .280 onus—oh

 

300

 

Time (days)

Consumed

_-.n-_- m

----A--- Eff

—X— Inf

Figure 8-2. Inﬂuent and efﬂuent toluene concentrations and DO consumption in a GAC-

FBR during the ﬁrst phase.

184

 

 

 

 

 

 

 

 

 

Shad 25.800 samba
H u 9 6 3 O
u u “ ﬂ .
\n x
\\\ \
bs X
n. s /X
[I]! K
a \
a x/
a, \.
\L a
an“... x
lUluIII-IIII1VI .. ......
XIﬂH R.I.PI.VI.VJIDA. ...........
ix .....
“I
01-1. m A.
wanrlnlJullnHuJul-Ju
albino x\
XI? Au
X uu-lx_ﬂ1ﬂIXu-lul.nT-l-
jH . -. . p
T u . - *4? . .
7 6 5 4 3 2 1 0

SEE .280 ones—eh.

60 80 100

Time (days)

20

Consume

_-.n--- m

----A--- Eff

—X— Inf

Figure 8-3. Inﬂuent and efﬂuent toluene concentrations and DO consumption in a GAC-

FBR during the second phase.

185

8.3.2 Extraction of Adsorbed Toluene on Carbon

The adsorbed toluene concentration reached its highest value at day 4 in the bottom
of the reactor followed by a sharp decrease after day 7 (Figure 84). Signiﬁcant DO
consumption started on day 5 which indicated that biodegradation became signiﬁcant. On
day 27, the adsorbed toluene concentration at the bottom of the FBR reached a level similar
to that on carbon in the middle of the reactor. The adsorbed toluene concentration reached
its highest value on day 21 in the middle of the reactor and decreased gradually thereafter
(Table 8-2 and Figure 8-4). Discontinuation of toluene feed did not increase the desorption

 

 

  
 

 

 

 

 

 

 

rate.

18
1,3 16 1 Toluene Feed
h) 14 ', Stopped
é '. 50 cm
0 12 t
t: I9
E 10 1,".
ﬂ 8 f; p
E 2 7%.

~X
0 3% : 4 ﬁx a:
0 100 200 300 400
Time (day)

Figure 8-4. Adsorbed toluene on biocoated GAC in the middle (150 cm) and bottom
portion of the GAC-FBR

186

Table 8-2. Adsorption isotherm parameters for GAC collected from mid-height in a GAC-
FBR (150 cm)

Bioﬁlm I
Thickness (um) I
I 79.3:334 I 0.37:0.03 -

Data Isotherm Parameters
Points qo (mg/g) I Kf (mg/g) I l/n

70 -

 

 

 

 

 

 

The Second Phase

 

10

79.315.9

0.4310.06

 

10

71.511.4

0.3810.02

 

10

69.611.4

0.3710.02

ND
4
5

 

20

62.611.7

0.4410.01

 

20

61.011.1

0.4210.03

 

10

67.511.7

0.4410.01

 

20

64.513.0

0.3510.04

 

10

56.713.9

0.4310.04

 

10

66.712.8

0.3710.02

160

 

 

10

 

 

46.3112

 

0.421002

 

76

 

The First Phase

 

42

40.912.5

0.481002

31

 

122

29

34.411.2

05410.02

76

 

158

44

35.710.9

05110.01

42

 

187

20

 

 

37.71l.2

 

05110.02

 

23

 

203

Stop Toluene Feed

 

255

15

34.411.5

0.461003

12

 

304

16

32.511.1

0.4910.02

 

374

10

 

 

i: 95 % conﬁdent interval

ND: non-detectable

 

f 27.93.1__

 

_ W5i05

 

 

ND

187

8.3.3 Adsorption Capacity

The isotherm parameters are presented in Table 8-2 and Figure 8-5. Adsorption
capacities throughout the experiment were calculated by comparing to the capacity at day 0.
The Kf value decreased slowly in the ﬁrst 60 days and quickly in the following 30 days to a
low level. The Kf value continued to decrease after the toluene feed stopped. Values of 1/n
increased slightly throughout the experimental period. Because of the non-linear
relationship between the equilibrium concentration and the adsorption capacity, the
capacities were also evaluated at three hypothetical concentrations of toluene (0.1, 3, and
10 mg/L to represent the concentrations at efﬂuent, inﬂuent, and a high inﬂuent shock
conditions, respectively) and the results are plotted in Figure 8-6. The adsorption capacities
remained between 70% to 100% of the initial value for all three concentrations in the ﬁrst
60 days and dropped to 40%, 50%, and 55% for 0.1, 3, and 10 mg/L, respectively, at later
times. The adsorption capacities decreased slightly and stopping the toluene feed did not
increase the capacities. The isotherm parameters for the carbon collected from the bottom of
the reactor were also estimated (Table 8-3) and found to be similar to those corresponding
to carbon in the middle portion of the reactor.

Table 8-3. Adsorption isotherm parameters for GAC collected from the bottom portion of a
GAC-FBR (50 cm)

Data Isotherm Parameters Bioﬁlm
Points qo (mg/g) Kf (mg/g) 1/n Thickness (11m)

10 0.20 77.812.91: 03910.03

 

 

 

10 17.87 77.0116 0.361002

 

10 17.84 76.3115 0.301002
20 8.45 65111.7 03910.01
20 11.54 63.9105 03710.01
10 7.45 __ f 47.613 0.52+004

1: 95 % conﬁdent interval
ND: non-detectable

 

 

 

 

 

 

 

 

 

 

20

10

188

 

 

 

Toluene Feed
Stopped

 

 

 

 

200 300

Time (day)

Figure 8-5. Isotherm parameters for biocoated GAC.

 

0.8

0.6

0.4

0.2

1/n

189

 

 

 

 

 

 

 

 

 

100% ..

30% TolueneFeed i
% Stopped
E
.13:
§ 60% A

‘3‘

.5 ~ °’_,A,“\ Ce = 10 mg/L
g" A-'-« ‘§.-\. ___D__,__..—-D
3 ----
<1 / Ce = 3 m
B /x \x\g/L
3 Ce = 0.1 mg/L X
‘3‘ 20% ..

0% : 11 :

0 100 200 300 400
Time (day)

Figure 8-6. Remaining capacity at three hypothetical equilibrium concentrations (0.1, 3,
and 10 mg/L representing the concentrations at efﬂuent, inﬂuent, and a high inﬂuent shock
load, respectively.)

190

8.3.4 Effect of Biomass on Adsorption

The bioﬁlm thicknesses on the carbon varied during this experiment. No consistent
relationship was found between adsorption capacity and bioﬁlm thickness. After toluene
feed stopped, the reactor operated in a decay mode , therefore, the bioﬁlm was decomposed
and was reduced to non-detectable levels. The adsorption capacity, however, did not

increase.

To investigate the effect of bioﬁlm thickness on adsorption, the bioﬁlm on two
carbon samples (day 64 and day 122) were removed before sterilization and the carbon was
used in the adsorption isotherm experiment. The average bioﬁlm thicknesses were reduced
from 160 to 76 um and 76 to 9 um for the day 64 and day 122 samples, respectively. The
isotherm parameters are similar to those for the non-disturbed carbon (Table 8-1 and Figure
8-7).

Clean GAC was mixed with biomass before sterilization and contacted with the
biomass for 48 hours. An adsorption isotherm was conducted using this carbon. The

biomass addition did not effect the isotherm results (Table 8-1 and Figure 8-8).

191

180
160 1

140 -

 

-. :1:”i ,4:
120 / /.o’
/* 9’ \

100 .. x .
/ IESS Bromass]

 

 

80-.
60.. 3"
40 j

20

\

 

 

-
q-
-

120 0 2 4 6 8 10 12

100 .-
. x :1:
[Less B1omassj - X
/

 

qe+qo (mg/g)

 

 

 

 

 

 

80 4
f"
60 .
40 -
20 .
Da 122
0 . : 1 1
0 2 4 6 8
Ce (mg/L)

Figure 8-7. Adsorption isotherms for carbon with and without biomass removal.

(A: carbon collected at day 64, bioﬁlm thicknesses were 160 and 76 mm for non-disturbed
carbon and biomass removed carbon, respectively;

B: carbon collected at day 122, bioﬁlm thicknesses were 76 and 9 mm for non-disturbed
carbon and biomass removed carbon, respectively.)

192

 

 

 

 

 

 

250

200 III- *,X
a FGAC with Biomass] /
E” 150 .. :x
to, x
g 100 -- D x
0"

50 ..

0 : :
0 10 15

Ce (mg/L)

Figure 8-8. Adsorption isotherm for GAC with and without contact with biomass.

8.3.5 Sodium Hydroxide Wash

A carbon sample collected at day 374 was digested and washed with 0.5 M sodium
hydroxide (NaOH). Some unknown brown substance was observed in the solution. The

adsorption pattern did not change signiﬁcantly although the Kf increased slightly (Table 8-1
and Figure 8—9).

193

 

120 :1:——

ﬂ
1001 I-
x’{ a

 

 

% 80 "' X 1"”0

E [washed with NaOH]
V 60 .-

o

55’

cr 40 -- ’

204

 

 

 

Cc (mg/L)

Figure 8-9. The effect of NaOH wash on the adsorption isotherm

8.4 Discussion

Adsorption is an important removal mechanism in biological activated carbon
ﬂuidized bed reactors. In the previous chapters, we have demonstrated that adsorption
reduced the concentrations of target compounds and byproducts in the efﬂuent during the

transient loading condition. Adsorption capacity, however, decreased with an increase in

usage of the GAC-FBR.

GAC-FBRs were operated at steady-state conditions without intentionally imposing
any perturbations. The adsorbed toluene on the carbon decreased after one week of
operation while DO consumption reached steady-state levels (ratio between consumed DO
and COD was 0.7 to 0.8). The adsorption capacity of the biocoated carbon decreased
slightly during the ﬁrst two months and was maintained at greater than 70% of the initial

level. The samples, which were collected from two portions of the reactor, indicate a

194

uniform distribution of the adsorption capacity in the GAC-FBR. After the bioﬁlm covered
the carbon particles, migration of the particles was expected in the reactor because of the
change in particle density and operation of the biomass control device. This migration could
increase particle mixing in the reactor, therefore, sampling carbon from any portion of the

reactor could represent the remaining capacity in the reactor.

After the biomass developed fully in GAC-FBR, which took about three weeks, the
adsorption capacity decreased gradually. The rate of decrease, however, was accelerated
after two months of operation to reach a new low level followed by a continued gradual
decrease. After six months of use, the remaining capacity was approximately 40%, 52%,
and 57% of it's initial value for the equilibrium toluene concentrations of 0.1, 3, and 10
mg/L, respectively. The remaining capacity may still provide a positive response when the

organic loading rate was increased.

Bioregeneration is a term used to describe the use of biodegradation to recover
adsorption capacity. In this study, the adsorbed toluene on biocoated GAC decreased,
however, the adsorption capacity did not recover. Therefore, bioregeneration did not occur.
After toluene feed stopped, the adsorption capacity decreased continuously even though the
adsorbed toluene level decreased as well. This indicates that the decrease in capacity is not
due to the adsorbed toluene but rather due to biomass growth. After toluene feed stopped, a
small amount of organic carbon was available in the inﬂuent and the DO consumption was
continuous. The DO consumption decreased after the nutrient supply was stopped and
increased immediately after the nutrient supply was resumed. The bioactivity never stopped
in this reactor. It is not clear what chemicals or factors cause the decrease in adsorption
capacity. However, it is known that microorganisms can produce soluble microbial
products (SMP) and adsorption of SMP can reduce the adsorption capacity (Olmstead
1989). When clean GAC was mixed with biomass for 48 hours, the adsorption capacity of
the GAC did not change. This indicates that if the SMP causes the reduction in capacity, the

195

carbon has to contact the SMP for a long period to allow a large amount of SMP to be
adsorbed by the carbon. The sodium hydroxide wash did not improve the recovery. If the
carbon is temporarily exhausted by toluene due to a transient loading condition,

bioregeneration can occur and the capacity can be recovered to the pre-transient level.

Olmstead (1989) reported that adsorption of TCE decreased with an increase in
biomass on the carbon. In this study, there is direct relationship between the amount of
biomass on the carbon and the remaining adsorption capacity. For example, at days 21 and
158, the carbon had a similar biocoat, but the capacity was much higher on day 21 than on
day 158 (Table 8-2). When some biomass was removed from the carbon, the adsorption
isotherm did not change signiﬁcantly (Figure 8-7). In Olmstead's study, the GAC particles
were fed with glucose for one to two weeks and coverage of the bioﬁlm on the GAC was
not reported. During this current study, a contiguous bioﬁlm was found using stereo
microscope examination. In a previous study, the contiguous bioﬁlm was also observed on
the carbon by scanning electron microscopy (SEM) examination of sample from a similar
GAC-FBR system (Voice et al. 1992). It is unlikely that bioﬁlm thickness could inﬂuence
the adsorption capacity.

The adsorption capacity found using an isotherm study indicates the potential of
biocoated GAC to adsorb the target chemical. The rate of adsorption, however, is another
very important factor which effects reactor performance during transient conditions. This

issue has not been addressed in this study therefore further experimentation is required.

8.5 Conclusions

This study provides conclusive evidence of the effect of bioﬁlm growth on the
adsorption capacity of GAC. The adsorption capacity of biocoated carbon was maintained

at more than 70% of initial levels during the ﬁrst two months. After six months of

196

operation, the remaining capacity was approximately 40%, 52%, and 57% of the initial
value for the equilibrium toluene concentrations of 0.1, 3, and 10 mg/L, respectively. If the
adsorption removal mechanism is used to provide a buffer during the transient conditions,
this remaining capacity may be sufﬁcient without carbon replacement. The adsorbed
toluene on biocoated GAC decreased continuously after the bioﬁlm developed, however,
the adsorption capacity did not recover and the bioregeneration did not occur. There is no
direct relation between the amount of biomass on the carbon and the remaining adsorption

capacity.

CHAPTER 9

DISSERTATION SUMMARY AND ENGINEERING DESIGN IMPLICATIONS

9.1 Summary

The objectives of this project were to investigate and exploit the advantages and
limitations of biological activated carbon ﬂuidized bed systems for treatment of gasoline
contaminated groundwater. The systems were tested under different conditions such as
start-up, steady-state with different organic loading rates, step loading rate increases, and
recovery. The research focused on several aspects including adsorption removal
mechanisms, biodegradation removal mechanisms, interaction between multiple substrates,

and adsorption capacity maintenance.

9.1.1 Importance of Using GAC as a Carrier Medium in FBR Systems

A comparison of ﬂuidized bed reactor systems with (1) adsorption removal capacity
only using granular activated carbon (GAC) without microbial growth, (2) combined
biological and adsorptive removal mechanisms using GAC with microbial growth and (3)
biological removal only using non-activated carbon with microbial growth was performed.
These three systems were fed groundwater contaminated with benzene, toluene and xylene
(BTX). The breakthrough proﬁles, steady-state removal of BTX, and system responses to
applied organic step loading rate increases were investigated. The use of activated carbon as

a biomass carrier in ﬂuidized bed reactors produces a system in which both adsorption and

197

198

biodegradation affect substrate removal. During start-up, even though the same amount of
inoculum was added to the two biological systems, the time (ca. 200 hrs) required until
effective biodegradation commenced in the system employing GAC was less than that (ca.
500 hrs) observed for the system employing the non-adsorptive biomass carrier (non-
activated carbon). Complete breakthrough of BTX did not occur in the system with
combined removal mechanisms and the development of a contiguous bioﬁlm was more
rapid. To a large extent, the performance of such biological activated carbon systems is
controlled by the additive contributions of these two removal mechanisms. During the start-
up period, before a fully functional biomass has developed, the substrate is removed
primarily by adsorption. The system retains adsorption capacity under steady state
conditions, however, as evidenced by the response of the system to in-line concentration
shocks. Lower efﬂuent concentration levels were found in BAC systems during a step
concentration increase than in similar systems using a non-adsorbing biomass carrier. The
data suggest that this results from adsorption of a portion of the increase, and that this
material can be subsequently desorbed and biodegraded when the in-line concentration

returns to pre-shock levels.

The biocoated GAC were collected from the granular activated carbon ﬂuidized bed
reactor (GAC-FBR) fed with toluene contaminated groundwater. The amount of adsorbed
toluene on the GAC was measured at various points in time by performing a solvent
extraction of a sample of the GAC during start-up, pseudo-steady-state, step-load-increase
and recovery phases of operation. This study provides conclusive evidence that a major
role of adsorption in integrated biological activated carbon (BAC) systems is to dampen
concentration changes during loading transients. During the start-up period in the GAC-
FBR before a bioﬁlm was established, the substrate (toluene) was adsorbed rather being
released, as would occur in a biological-only system. Upon establishment of a bioﬁlm
capable of degrading the substrate, inﬂuent toluene was biologically oxidized and a portion

of the previously adsorbed toluene was desorbed and degraded. Under continued constant

199

organic loading rate conditions, the system stabilized and functioned as a biological reactor.
During a 77 hour, 400% step-OLR increase, adsorption again became important; excess
substrate was observed to accumulate on the GAC carrier particles. This material was
desorbed and degraded over an 11 day period following resumption of the initial organic
loading rate of 5.3 kg COD lm3-d. Thus it was established that the dampening ability of

activated carbon is not exhausted, but is recovered via bioregeneration.

It is clear from these results that GAC-FBR systems are advantageous for treating
BTX-contaminated waters when variations in loading rates are expected or when positive
removal is required during the start-up period. Furthermore, this concept provides a
foundation upon which future studies can be conducted in order to develop a rational

design procedure which incorporates criteria for transient loading performance.

9.1.2 Biodegradation in FBR Systems

After the biomass is established and steady-state conditions are reached, system
performance is dominated by biodegradation. Scanning electron microscopy was used to
examine the extent of surface coverage of the GAC and non-activated carbon by the
bioﬁlm. Particles from both systems were observed to be completely covered by a
contiguous, thick bioﬁlm. Under constant, steady-state, organic loading conditions (3 and
6 kg COD/m3-day) BTX removals were comparable for the GAC-FBR and biological only
(FBR) systems. More than 90 percent of BTX was removed in both systems.

The yield and decay coefﬁcients for a mixed culture treating BTX were evaluated
and the kinetic parameters for degradation of BTX were determined. The degradation rate
for toluene was the highest, while the rate for xylene was the lowest. There was inhibition
between all three substrates but the type of the inhibition was not determined. The effect of

inhibition on toluene degradation was limited. A large variance during estimation of Ks was

200

found and may be due to its low value for this mixed culture. The parameters, however, are

useful for bioreactor design with a consideration of uncertainty.

9.1.3 Interaction of Multiple Substrates

Parallel ﬂuidized bed systems were operated to treat groundwater containing one
milligram per liter each of benzene, toluene and p-xylene (BTX). Granular activated carbon
(GAC) was used as the carrier media for microbial growth in one system and granular non-
activated carbon (with very low adsorption capacity) was used in another. The performance
of the systems during the single-substrate step-loading increases were investigated. When
the two systems were subjected to seven-fold step-load increases in benzene and toluene,
biodegradation in the ﬂuidized-bed system was sufﬁcient to maintain a high efﬂuent
quality. For a substrate that is biodegraded more slowly, such as p-xylene, adsorption onto
the GAC carrier contributed signiﬁcantly to removal and helped to provide more robust
performance in the BAC-FBR. Because of the complicated substrate interactions, the
concentration of some substrates (e. g. benzene and toluene) can effect the degradation rate
of other substrates in a ﬂuidized-bed systems and other biological treatment system.
Extraction of the GAC carrier and oxygen consumption in the GAC-FBR veriﬁed that both
adsorption and biodegradation were responsible for maintaining high treatment efﬁciencies

during the step increases.

9.1.4 Formation of Byproducts

Groundwater containing benzene, toluene and p-xylene (BTX) was treated in a
ﬂuidized bed reactor (FBR) using non-activated carbon as the bioﬁlm carrier and a
biological activated carbon ﬂuidized bed reactor (GAC-FBR) using activated carbon as the

bioﬁlm carrier. Formation of metabolic intermediate products at steady-state and during

201

single-substrate step-loading increases was determined by measuring non-volatile organic
carbon (non-volatile TOC) and UV spectra. At steady-state (organic loading rate of 2.2 kg-
COD/m3-day), no intermediate decomposition products were detected in the efﬂuent. Both
reactors were subjected to a seven-fold p-xylene step increase under different
concentrations of dissolved oxygen and nutrients in the inﬂuent. The production of
byproducts was not effected by the nutrient levels but increased with an increase in inﬂuent
DO. Both systems were subsequently subjected to twenty, twelve and seven-fold step
increases in the organic loading rates of benzene, toluene and p-xylene, respectively. A
significant amount of the biologically removed BTX resulted in the production of
byproducts (35%, 46%, and 29% for benzene, toluene, and xylene increases,
respectively). Efﬂuent TOC corresponded to the appearance of new peaks in the UV
spectra. The concentration of byproducts was less in the BAC-FBR efﬂuent than in the
FBR system. The combination of biological and adsorptive removal mechanisms resulted
in enhanced removal of BTX and byproducts and more stable operation of the BAC-FBR

system.

9.1.5 Adsorption Capacity in a GAC-FBR

Biocoated GAC was collected from a GAC-FBR treating groundwater contaminated
with toluene at different times. The adsorption capacity, as described by the adsorption
isotherm, was measured. The results provide conclusive evidence of the effect bioﬁlm
growth has on the adsorption capacity of GAC. The adsorption capacity of biocoated
carbon was maintained at more than 70% of its initial levels during the ﬁrst two months.
After six months of operation, the remaining capacity was approximately 40%, 52%, and
57% of its initial value for the equilibrium toluene concentrations of 0.1, 3, and 10 mg/L,
respectively. If the adsorption removal mechanism is used to provide a barrier for handling

the transient condition, this remaining capacity may be sufficient to avoid carbon

202

replacement. The amount of adsorbed toluene on biocoated GAC decreased continuously
after the bioﬁlm developed, however, the adsorption capacity did not recover. No direct
relationship was found between the amount of biomass on the carbon and the remaining

adsorption capacity.

9.2 Engineering Design Implications

9 Biological activated carbon ﬂuidized bed systems have very high treatment
efficiencies when fed groundwater contaminated with volatile aromatic

hydrocarbons and provide stable performance.

0 Using GAC as the bioﬁlm carrier allowed for more rapid biomass development
during the start-up period. The adsorption capacity of the GAC prevented
breakthrough of some strongly adsorptive compounds, thereby, avoiding use of
additional treatment components such as an adsorption column may not be

necessary.

- At a constant, steady-state organic loading rate of 3 kg-COD/m3-day, the GAC-
FBR is able to remove 98% of the applied BTX. At a constant, steady-state
organic loading rate of 6 kg-COD/m3—day, the GAC-FBR is able to remove 90%
of the applied BTX (B:T:X of 1:1:1 volume to volume).

- During steady-state conditions, the performance of the GAC-FBR is dominated by
biodegradation. Engineers, therefore, do not need to consider the adsorption

capacity in design of such conditions.

9 The biokinetic parameters reported in this dissertation can be used to design the
GAC-FBR systems with a help of modeling tool. Because of the continuous

change in activity and population of the microorganisms in the bioreactor,

203

however, variation in substrate utilization rates are expected. Engineers will need

to consider such uncertainty in their designs.

0 During a transient condition such as an organic loading rate increase, the
combination of biological and adsorption removal capacity results in enhanced

BTX removal and more stable operation.

0 70% of the initial adsorption capacity of biocoated carbon can be maintained after
the ﬁrst two months. After six months of operation, the remaining capacity was
approximately 40%, 52%, and 57% of its initial value for the equilibrium
concentrations of 0.1, 3, and 10 mg/L, respectively. If the adsorption removal
mechanism is used to provide a barrier for handling the transient condition, this

remaining capacity may be enough to make new carbon replacement unnecessary.

0 Formation of byproducts was not found during steady-state condition. At
extremely high organic loading rate increases (twenty, twelve and seven-fold step
increases in the organic loading rates of benzene, toluene and p-xylene,
respectively), a signiﬁcant amount of the biologically removed BTX, however,
resulted in the production of byproducts (35%, 46%, and 29% for benzene,
toluene, and xylene increases, respectively). The GAC-FBR can minimize the
concentration of byproducts in the efﬂuent, however, longer hydraulic retention

time and higher carbon bed height are required to compensate for this effects

- A two-step procedure should be considered for design of a GAC-FBR treating
biodegradable substrates. First, the reactor volume and height are calculated from
a steady-state organic loading rate without considering adsorption effects.
Second, the parameters are checked with a possible high organic loading rate to
verify that efﬂuent quality is maintained. The combined effect of the adsorption

and biodegradation removal mechanisms should be considered for the removal of

204

substrates. The production of byproducts should also be considered with
adsorption being the only removal mechanism for the byproducts. The reactor
volume and height can be adjusted from the steady-state value to satisfy the

efﬂuent requirements.

9.3 Recommendations for Future Studies

0 Investigate the effects of usage time of biocoated GAC on adsorption kinetics of

toluene.

0 Investigate the relationship between increased organic loading rates and
production of byproducts. Develop a method to identify the byproducts and

investigate their adsorption characteristics.

0 Develop a single substrate dynamics model to describe adsorption and
biodegradation including production and adsorption of byproducts. Use this
model to develop a rational design procedure for GAC-FBR incorporating steady-

state and shock loading conditions.

0 Understand the long-term effects of high organic loading rate on reactor

performance and the production of byproducts.
- Identify the cause of reduced adsorption capacity in GAC-FBR systems.
0 Study multiple substrate adsorption isotherms and kinetics for biocoated GAC.

0 Develop a dynamic model to describe multiple substrates in GAC-FBR system
including adsorption and biodegradation of the substrates and production and

adsorption of byproducts under high organic loading rate conditions.

LIST OF REFERENCES

Alvarez, P. J. J ., P. J. Anid and T. M. Vogel 1991. “Kinetics of aerobic biodegradation of
benzene and toluene in sandy aquifer material.” Biodegradation, 2: 43-51.

Andrews, G. F. and C. Tien 1982. “An analysis of bacterial growth in ﬂuidized-bed
adsorption column.” AlChE Journal, 28(2): 182-190.

Annesini, M. C., F. Gilroni, M. Ruzzi and C. ATomel 1987. “Adsorption of organic
compounds onto activated carbon.” Wat. Sci. Tech., 21(5): 567-571.

Arvin, E., B. K. Jensen and A. T. Gundersen 1989. “Substrate interactions during aerobic

biodegradation of benzene.” Applied and Environmental Microbiology, 55(12): 3221-
3225.

Atlas, R. M. and R. Bartha 1998. '
Menlo Park, California, The Benjanmin/Cummings Publishing Company, Inc.

Bancroft, K., S. W. Maloney, J. McElhaney, I. H. Suffet and W. O. Pipes 1983.
“Assessment of bacterial growth and total organic carbon removal on granular activated
carbon contactors.” Appl. Environ. Microbiol., 46(3): 683-8.

Beck. J. V. and K. J. Arnold 1977. W New
York, John Wiley & Sons.

Borneff, J. 1980. Elimination of carcinogens (excluding haloforms) by activated carbon.
'1 11111.1 ”1.111 0'11. 1111.1 ‘1-1-’1-.....IHSuffetandMJ
McGuire. Ann Arbor, Michigan, Ann Arbor Science. 1: 145- 56.

Bouwer, E. J. and P. L. McCarty 1982. “Removal of trace chlorinated organic compounds
by activated carbon and ﬁxed-ﬁlm bacteria.” Environmental Science and Technology,
16(12): 836-43.

Brewer , W. S. and W. W. Carmichael 1979. “Microbiological characterization of
granular activated carbon ﬁlter systems.” journal AWWA, 71(12): 738-40.

Burden, R. L., J. D. Faires and A. Reynolds 1981. W, PWS, Boston,
Mass.

Button, D. K. 1985. “Kinetics of nutrient- limited transport and microbial growth. ”
Microbiological Reviews, 49: 270-97.

205

206

Camper , A. K., S. C. Broadaway and M. W. LeChevallier 1987. “Operational variables
and the release of colonized granular activated carbon particles in drinking water.” (5): 70-
73.

Camper , A. K., M. W. LeChevallier , S. C. Broadaway and G. A. McFeters 1986.
“Bacteria associated with granular activated carbon particles in drinking water.” Applied
and Environmental Microbiology, 52: 434-438.

Camper, A. K., M. W. LeChevallier, S. C. Broadaway and G. A. McFeters 1985.
“Growth and persistence of pathogens on granular activated carbon ﬁlters.” Appl.
Environ. Microbiol., 50(6): 1378-82.

Chang, M.-K. 1994. Substrate Interactions during the degradation of benzene, toluene,
and p—xylene and their effects on sorption in a ﬁxed bed biological activated carbon reactor,
Michigan State University.

Chang, M.-K., T. C. Voice and C. S. Criddle 1993. “Kinetics of competitive inhibition
and cometabolis in the biodegradation of benzene, toluene, and p-xylene by two
pseudomonas isolates.” Biotechnology and Bioengineering, 41(11): 1057-65.

Characklis, W. G. 1984. “Bioﬁlm development: a process analysis.” Microbial Adhesion
and Aggregation,: 137-157.

Chudyk, W. A. and V. L. Snoeyink 1984. “Bioregeneration of activated carbon saturated
with phenol.” Environmental Science and Technology, 18(1): 1-5.

Clark, R. M. and J. Benjamin W. Lykins 1989. 'v
WM. Chelsea, Michigan, Lewis Publishers, Inc.

Committee Report 1981. “An assessment of microbial activity on GAC.” Journal of
American Water Work Association, 73(8): 447-454.

Criddle, C. S., L. A. Alvarez and P. L. McCarty 1991. Microbial processes in porous
media. W. J. Bear and M. Y. Corapcioglu.
Netherlands, Kluwer Academic Publishers: 639-91.

Crittenden, J. C., D. W. Hand, H. H. Arora and B. W. Lykins 1987. “Design
considerations for GAC treatment of organic chemicals.” J AWWA,: 74-82.

Crittenden, J. C., P. Luft, D. W. Hand, J. L. Oravitz, S. W. Loper and M. Ari 1985.
“Prediction of multicomponent adsorption equilibria using ideal adsorbed solution theory.”
Env. Sci. Tech., 19(11): 1037.

Davey, J. F. and D. T. Gibson 1974. “Bacterial metabolism of para- and meta-xylene:
oxidation of a methyl substituent.” Journal of Bacteriology, 119(3): 923-9.

Davis, R. S., H. F. E and R. W. Stone 1968. “Metabolism of p— and m-xylene by species
of Pseudomonas.” Canadian Journal of Microbiology, 14: 1005-9.

De Laat, J., F. Bouanga and M. Dore 1985. “Inﬂuence of microbiological activity in
granular activated carbon ﬁlters on the removal of organic compounds.” The Science of the
Total Environment, 47: 115-20.

207

DeWaters, J. E. and F. A. DiGiano 1990. “The inﬂuence of ozonated natural organic
matter on the biodegradation of a micropollutant in a GAC bed.” Journal of American

Water Work Association, 82(8): 69-75.

DiGiano, F. A. 1981.m1mrcentb1010gieal_a§tmm_QACrierfonnanee Conference on
Applications of Adsorption in Wastewater Treatment, Enviro Press, Nashville, Tenn.

Dobbs, R. A. and J. M. Cohen 1980. Carbon adsorption isotherms for toxic organics,
Municipal Environmental Research Laboratory, US. EPA.

Evans, W. C. 1963. “The microbiological degradation of aromatic compounds.” J. Gen.
Microbial, 32: 177-184.

Fritz, W., W. Merk, E. U. Schlunder and H. vSontheirner 1980. Competitive adsorption of
dissolved organics on activated carbon. . - _ 1 -_ f ._
..I H. Suffet and M. J. McGuire. Ann Arbor, Michigan, Ann Arbor

 

Science. 1.

Gardner, D. A., M. T. Suidan and H. A. Kobayashi 1988. “Role of GAC activity and
particle size during the ﬂuidized-bed anaerobic treatment of reﬁnery sour water stripper
bottoms.” Journal of Water Pollution Control Federal, 60(4): 505-13.

Gibson, D. T. 1968a. “Microbial degradation of aromatic compounds.” Science,
161(3846): 1093-7.

Gibson, D. T., J. R. Koch and R. E. Kallio 1968b. “Oxidative degradation of aromatic
hydrocarbons by microorganisms. I. enzymatic formation of catechol from benzene.”
Biochemistry, 7(7): 2653-62.

Gibson, D. T., V. Mahadevan and J. F. Davey 1974. “Bacterial metabolism of para- and
meta-xylene: oxidation of the aromatic ring.” Journal of Bacteriology, 119(3): 930-6.

Gibson, D. T. and V. Subramanian 1984. Microbial degradation of aromatic
hydrocarbons W D T Gidson New York
NY, Marcel Dekker: 181-254.

Giusti, D. M., R. A. Conway and C. T. Lawson 1974. “Activated carbon adsorption of
petrochemicals.” Journal of Water Pollution Control Federal, 46(5): 947-65.

Glaze, W. H., C. C. Lin, J. C. Crittenden and R. Cotton 1986. “Adsorption and
microbiological mechanisms for removal of natural organics in granular activated carbon
columns.” Ozone: Sci. Eng., 8(4): 299-319.

Goeddertz, J. G., M. R. Matsumoto and A. S. Weber 1988. “Ofﬂine bioregeneration of
granular activated carbon.” J. Environ. Eng. (N. Y.), 114(5): 1063-76.

Goldsmith, J ., C. D. and R. K. Balderson 1988. “Biodegradation and growth kinetics of
enrichment isolates on benzene, toluene and xylene.” Water Science and Technology,
20(11): 505-7.

VGrady.C P L J andH C Lim1980 Wham
W- New York, Marcel Dekker, Inc.

208

Grady, L. J., G. Aichinger, S. F. Cooper and M. Naziruddin 1989. '
.AWMA/EPA International
Symposium, Cincinnati, OH.

Greenberg, A. E., R. R. Trussell and L. S. Clesceri 1985. W
W- Washington, American Public Health Association.

Hickey,R...F1990a'111e 1 '1 -11111 -11 111111-31
Biological Fluidized Beds for Water and Wastewater Treatment- A User' s Forum 2425
May, Ann Arbor, Michigan.

Hickey, R. F., D. Wagner and G. Mazewski 1990b. -

2.111111 111- 11 '11rBI. 11 -111'1' 1 11W',1'1-'1"11.F0urthNatn.
Outdoor Action Conf. on Aquifer Restorating, Groundwater Monitoring and Geo physical
methods-NWWA, Las Vegas, Nev.,.

Hickey, R. F., D. Wagner and G. Mazewski 1991a. “Treating contaminated groundwater
using a ﬂuidized-bed reactor.” Remediation, 1(2): 447-60.

Hickey, R. F., W. M. Wu, M. C. Veiga and R. Jones 1991b. “Start-up, operation,
monitoring and control of high-rate anaerobic treatment systems.” Water Science and
Technology, 24(8): 207-55.

Hou, C. T. 1982. Microbial transformation of important industrial hydrocarbons.
' . J. P. Rosazza. Boca Raton, Florida,

 

CRC Press, Inc. 18.1107.

Jain, K. S. and V. L. Snoeyink 1973. “Adsorption from bisolute systems on activated
carbon.” Journal of Water Pollution Control Federal, 45(12): 2463-79.

Janssens, J. G., J. Meheus and J. Dirickx 1985. “Ozone enhanced biological activated
carbon ﬁltration and its effect on organic matter removal, and in particular on AOC
reduction.” Water Sci. Technol., 17(6-7): 1055-68.

Jorgensen, C., J. Flyvberg, B. K. Jensen, E. Arvin, S. K. Olsen and E. Mortensen 1990.

n'11'1111111-11 -'1' '10-,‘0 1-11111 11 -11-112‘ '1 1‘
9.011311119113- COST Seminar on Anaerobic Biodegradation of xenobiotic Compounds,
Copenhagen, November 1990.

Karlson, U. and J. Frankenberger, W. T. 1989. “Microbial degradation of benzene and
toluene in groundwater.” Bull. Environ. Contam. Toxicol., 43: 505-510.

Khan, K. A., M. T. Suidan and W. H. Cross 1981. “Anaerobic activated carbon ﬁlter for
the treatment of phenol-beating wastewater.” J WPCF, 53(10): 1519-32.

Kim, R., E. S. K. Chian, W. H. Cross and S. S. Cheng 1986. “Adsorption, desorption,
and bioregeneration in an anaerobic, granular activated carbon reactor for the removal of
phenol.” J WPCF, 58(1): 35-40.

Kim, S. H. and M. Pirbazari 1989. “Bioactive adsorber model for industrial wastewater
treatment.” Journal of Environmental Engineering, 115(6): 1235-65.

Klotz, M., P. Werner and R. Schweisfurth 1976. Investigations concerning the

microbiology of activated carbon ﬁlters. WM

209

W. W. Kuhn and H. Sontheimer, US. EPA Report
EPA-600/9-76-030. 9: 3 12-331.

Kornegay, B. H. 1978. ' ' ' 'v
WWW. Seminar on Control of Organic Chemical
Contaminants in Drinking Water, Los Angeles, CA, US. EPA, Ofﬁce of Drinking Water,
Washington, DC.

Latoszek, A. and A. Benedek 1979. “Some aspects of the microbiology of activated carbon
columns treating domestic wastewater.” Environ. Sci. Technol., 13(10): 1285-7.

Lawrence, A. W. and P. L. McCarty 1970. “Uniﬁed basis for biological treatment design
and operation.” ASCE,: 757-778.

LeChevallier , M. W., T. S. Hassenauer , A. K. Camper and G. A. McFeters 1984.
“Disinfection of bacteria attached to granular activated carbon.” Applied and Environmental
Microbiology, 48: 918-23.

Lehninger, A. L. 1982. W. New York, Worth Publishers, Inc.

Li, A. Y. L. and F. A. DiGiano 1983. “Availability of sorbed substrate for microbial
degradation on granular activated carbon.” Journal of Water Pollution Control Federal,
55(4): 392-9.

Lowry , J. D. and C. E. Burkhead 1980. “The role of adsorption in biologically extended
activated carbon columns.” journal WPCF, 52: 389-98.

MacQuarrie, K. T. B., E. A. Sudicky and E. O. Frinf 1990. “Simulation of biodegradable
organic contaminants in groundwater 1. numerical formulation in principal directions.”
Water Resources Research, 26: 207-22.

Marr, E. K. and R. W. Stone 1961. “Bacterial oxidation of benzene.” J. Bacterial, 81:
425-30.

Martin, R. J. and W. J. Ng 1985. “Chemical regeneration of exhausted activated carbon -
II.” Water Res., 19(12): 1527-35.

McCarty, P. L. 1983. Removal of organic substances from water by air stripping. Control
12W B Berger EPA- 600/8- 33-011

McMurry, J. 1988. W, Brooks/Cole Publishing Company, Belmont, CA.

Monod, J. 1949. “The growth of bacterial cultures.” MW, 3:
371-394

W

Nakhla, G. F., M. T. Suidan and J. T. Pfeffer 1988. “Operational control of an anaerobic
GAC reactor treating hazardous wastes.” Water Science and Technology, 21: 167-73.

Neukrug, H. M., M. G. Smith, S. W. Maloney andI. H. Suffet 1984. “Biological
activated carbon- at what cost?” J. Am. Water Works Assoc. ,:76(4) 158- 67.

Olmstead, K. P. 1989. Microbial interference with the adsorption of target organic
contaminants by granular activated carbon, University of Michigan.

210

Omori, T. and K. Yamada 1970a. “Studies on the utilization of hydrocarbons by
microorganisms part XVI. detection of metabolic intermediates of xylene and
pseudocumene.” Agr. Biol. Chem., 34(5): 659-63.

Omori, T. and K. Yamada 1970b. “Studies on the utilization of hydrocarbons by

microorganisms part XVII. metabolism p-xylene and related compounds.” Agr. Biol.
Chem., 34(664-9).

Patnaik, P. 1992. 1 1'1‘ 11 V 1_' 1 1' 1:11-11 ' 11'1' 1 1'11,-
Suhstms. New York, Van Nostrand Reinhold.

Peel, R. G. and A. Benedek 1983. “Biodegradation and adsorption within activated carbon
adsorbers.” Journal WPCF, 55(9): 1168-73.

Petrovic, O., M. Gantar and M. Bokorov 1989. “Scanning electron microscope
investigation of activated sludge and activated carbon from a wastewater puriﬁcation
process.” Biol. Wastes, 30(2): 81-8.

Pirbazari, M., T. C. Voice and J. Walter J. Weber 1990. “Evaluation of bioﬁlm
development on various natural and synthetic media.” Hazardous Waste & Hazardous
Materials, 7(3): 239-50.

Randtke, S. J. and V. L. Snoeyink 1983. “Evaluating GAC adsorptive capacity.” Journal
of American Water Work Association, 75(8): 406-13.

Ribbons, D. W. and R. W. Eaton 1982. Chemical transformations of aromatic

hydrocarbons that support the growth of microorganisms. W51
v ..A M. Chakrabarty. Boca Raton, Florida, CRC

Press, Inc.: 59-84.

Rice, R. G. and C. M. Robson 1982.
WWW Ann Arbor Mich” Ann Arbor Sci Publ

Rice, R. G., C. M. Robson, G. W. Miller, J. C. Clark and W. Kuhn 1980. Biological
processes in the treatment of municipal water supplies, US. EPA, Municipal
Environmental Research Laboratory.

Rittmann, B. E. 1987. “Aerobic biological treatment.” Environmental Science and
Technology., 21(2): 128-136.

Robertson, B. R. and D. K. Button 1987. “Toluene induction and uptake kinetics and their
inclusion in the speciﬁc-afﬁnity relationships describing rates of hydrocarbon metabolism.”
Applied and Environmental Microbiology, 53: 2193-205.

Rodrnan, C. A, E. L. Shunney and A. E. Perrotti 1978. Biological regeneration of
activated carbon. W. P. N. Cheremisinoff and F. Ellerbusch.
Fram Corp., East Providence, R. I, USA, Ann Arbor Sci.: Ann Arbor, Mich. 449- 83.

Rodrnan, C. A. and E. L. Skunney 1971. Bioregenerated activated carbon treatment of
textile dye wastewater, EPA, Washington, DC.

Schultz, J. R. and T. M. Keinath 1984. “Powdered activated carbon treatment process
mechanisms.” Journal of Water Pollution Control Federal, 56(2): 143- 51.

211

Schwarzenbach, R. P, P. M. Gschwend and D. M. Irnboden 1993. W
W. New York, John Wiley & Sons, Inc.

Shi, J., X. Zhao, R. F. Hickey and T. C. Voice 1994. “Role of adsorption in granular
activated carbon ﬂuidized bed reactors.” Submitted for publication on Water Environmental
Research,.

Shieh, W. K. and J. D. Keenan 1986. “Fluidized bed bioﬁlm reactors for wastewater
treatment.” Advances in Biochemical Engineering/Biotechnology, 33: 131-169.

Sigurdson, S. P. and C. W. Robinson 1978. “Substrate-inhibited microbiological
regeneration of granular activated carbon.” Can. J. Chem. Eng, 56(3): 330-9.

Singer, P. C. and C. -.Y Yen 1980. Adsorption of alkyl phenols by activated carbon.
- - .I...HSuffetandMJ

 

McGuire. Ann Arbor, Ann Arbor Science. 1: 167- 89. I ‘

Sohngen, N. L. 1913. “Benzin, Petroleum, parafﬁnol und parafﬁn als. kohlenstoff und
energiequelle fur rnikroben.” Centr. Bakteriol. Parasitenk., Abt.II, 37: 595-609.

Sorniya, I., H. Yamada, E. Nozawa and M. Mohri 1986. “Biodegradability and GAC
adsorbability of micropollutants by preozonization.” Ozone: Sci. Eng., 8(1): 11-26.

Speitel, G. E. and F. A. DiGiano 1987. “The bioregeneration of GAC used to treat
micropollutants.” journal AWWA, 79: 64-73.

Speitel, G. E., Jr., C. J. Lu, M. Turakhia and X. J. Zhu 1989a. “Biodegradation of trace
concentrations of substituted phenols in granular activated carbon columns.” Environ. Sci.
Technol., 23(1): 68-74.

Speitel, G. E., Jr., C. J. Lu and M. H. Turakhia 1988. “Biodegradation of synthetic
organic chemicals in GAC beds.” Water Science and Technology, 20(11-12): 463-5.

Speitel, G. E., Jr., C. J. Lu, X. J. Zhu and M. H. Turakhia 1989b. “Biodegradation and
adsorption of a bisolute mixture in GAC columns.” Journal of Water Pollution Control
Federal, 61(2): 221-9.

Speitel, G. E., Jr., M. H. Turakhia and C. J. Lu 1989c. “Initiation of micropollutant
biodegradation in virgin GAC columns.” Journal of American Water Work Association,
81(4): 168-76.

Storrner, K. 1908. “Ueber die wirkung des schwefel kohlenmstoffs und ahnlicher stoffe
auf den boden. II.” Zentralbl. Bakteriol. Parasitenkd. Infektionskr., 20: 282.

Suffet, I. H. 1980. “An evaluation of activated carbon for drinking water treatment: a
national academy of science report.” J. AWWA, 72(1): 41-50.

Suidan, M. T., W. H. Cross, M. Fong and J. W. Calvert 1981. “Anaerobic carbon ﬁlter
for degradation of phenols.” J. Environ. Eng. Div. (Am. Soc. Civ. Eng.), 107(EE3): 563-
79.

Suidan, M. T., P. Fox and J. T. Pfeffer 1987. “Anaerobic treatment of coal gasiﬁcation
wastewater.” Wat.Sci.Tech., 19: 229-36.

212

Suidan, M. T., G. L. Siekerka, S. W. Kao and J. T. Pfeffer 1983. “Anaerobic ﬁlters for
the treatment of coal gasiﬁcation wastewater.” Biotechnology and Bioengineering, 25:
1581-96.

Tobin, R. S., D. K. Smith and J. A. Lindsay 1981. “Effects of activated carbon and
bacteriostatic filters on microbiological quality of drinking water.” Appl. Environ.
Microbiol., 41(3): 646-51.

Voice, T. C. 1989. Activated Carbon Adsorption. 333W
W. H. Freeman. New York, McGraw-Hill.

Voice, T. C., D. Pak, X. Zhao, J. Shi and R. F. Hickey 1992. “Biological activated
carbon in ﬂuidized bed reactors for the treatment of groundwater contaminated with volatile
aromatic hydrocarbons.” Water Research, 26(10): 1389-401.

Wallis, D. A. and E. E. Bolton 1982. “Biological regeneration of activated carbon.” AlChE
Symp. Ser., 78(219): 64-70.

Wang, Y. T., M. T. Suidan and B. E. Rittman 1986. “Anaerobic treatment of phenol by
an expanded-bed reactor.” J WPCF, 58(3): 227-33.

Wang, Y. T., M. T. Suidan and J. T. Pfeffer 1984. “Anaerobic activated carbon ﬁlter for
the degradation of polycyclic N-aromatic compounds.” J WPCF, 56(12): 1247-53.

Weber, W. J., L. D. Friedman and R. Bloom 1972a. Biologically- -extended

physicochemical treatment. W. S. H. Jenkin, Pergamon
Press (Oxford): 641 -56.

Weber, W. J ., C. B. Hopkins and R. Bloom 1970. “Expanded-bed adsorption for
treatment of sewage efﬂuents.” J. Wat. Pollut. Control Fed., 42(1): 81-3.

Weber, W. J ., Jr., M. Pirbazari and G. L. Melson 1978a. “Biological growth on activated
carbon: an investigation by scanning electron microscopy.” Environ. Sci. Technol.,
12(7): 817-9.

Weber, W. J ., Jr. and W.-C. Ying 1978b. “Integrated biological and physicochemical
treatment for reclamation of wastewater.” Prog. Water Technol., 10(1-2): 217-33.

Weber. W. J. J. 1972b. Wand New York,

Wiley-Interscience.

Weber, W. J. J. and B. M. V. Vliet 1980. Fundamental concepts for application of
activated carbon in water and wastewater treatment. v

I. H. Suffet and M. J. McGuire. Ann Arbor, Michigan,
Ann Arbor Science. 1: 15—41.

Willard, H. H., J. Lynne L. Merritt, J. A. Dean and J. Frank A. Settle 1988. mm
W. Wadworth Publishing Company, Belmont, CA.

Xing, J. and R. F. Hickey 1994. “Response in performance of the GAC- ﬂuidized bed
reactor process for BTX removal to perturbation in oxygen and nutrient supply.”
International Biodeterioration & Biodegradation, 33. 23- 39.

213

Yagi, M., S. Nakashima and S. Muramoto 1988. “Biological degradation of musty odor
compounds, 2-Methylisobomeol and Geosmin, in a bio-activated carbon ﬁlted.” Wat. Sci.
Tech., 20(8/9): 255-60.

Yousseﬁ, M. and S. D. Faust 1980. Adsorption and formation of light halogenated
hydrocarbons and humic acid 1n water by granular activated carbon. W

.I. H. Suffet and M. J. McGuire. Ann
Arbor, Michigan, Ann Arbor Science. 1: 133-43.

Zanitsch, R., H. and R. T. Lynch 1978. “Selecting a thermal regeneration system for
activated carbon.” Chem. Eng. (N. Y.), 85(1): 95-100.

Zhang, X., Z. Wang and X. Gu 1991. “Simple combination of biodegradation and carbon
adsorption - the mechanism of the biological activated carbon process.” Water Research,
25(2): 165-72.

Zhao, X., K. Doh, C. S. Criddle and T. C. Voice 1994. BMW

OlLtI'lZ Ono-11‘s 0" 1'!l|!' 10010....
W. 49th Annual Purdue University Industrial Waste Conference, West
Lafayette, Indiana.

Zhao, X., J. Shi, R. F. Hickey andT. C. Voice 1993. Comparisonoﬂaosomﬁmrpon;

c. o . iv noiﬁlm carr r 'n nil-co i lﬂ ic-iz . uec r actrs fr e r .
W 48th Annual Purdue

University Industrial Waste Conference, Purdue University, West Lafayette, Indiana, May
10- l 2.

Zylstra, G. J. and D. T. Gibson 1991. Aromatic hydrocarbon degradation. a molecular
approach. Gonog’anginooﬁng. J. K. Setlow. New York, Plenum Press. 13: 183- 203.

            

            

     

 

MICHIGAN STRTE N V. L

U I IBRRRIES
1 IIWHIIIHIIHIIW llllll ll llllllllllllllﬂlll
93 1 6 634

312 (@1041