EMERGING TECHNOLOGY REPORT:
DEMONSTRATION OF
AMBERSORB® 563 ADSORBENT TECHNOLOGY
bY
ROY F. WESTON, INC.
West Chester, PA 19380-1499
Project Officer
Ronald J. Turner
Water and Hazardous Waste Treatment Research Division
National Risk Management Research Laboratory
Cincinnati, OH 45268
This study was conducted in cooperation
with the U.S. Environmental Protection Agency
Under Cooperative Agreement No. CR-821352-01-0
NATIONAL RISK MANAGEMENT RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI. OH 45268
MK01\RPT:02659051.001\epuile.txt
-------
CONTACT
Ronald Turner is the EPA contact for this report. He is presently with the newly organized
National Risk Management Research Laboratory's new Land Remediation and Pollution
Control Division in Cincinnati, OH (formerly the Risk Reduction Engineering Laboratory).
The National Risk Management Research Laboratory is headquartered in Cincinnati, OH, and
is now responsible for research conducted by the Land Remediation and Pollution Control
Division in Cincinnati.
-------
DISCLAIMER
The information in this document has been funded in part by the United States
Environmental Protection Agency under Cooperative Agreement No. CR-8213 52-01-0 to Roy F.
Weston, Inc. It has been subjected to the Agency's peer and administrative review, and it has
been approved for publication as an EPA document. Mention of trade names or commercial
products does not constitute endorsement or recommendation for use.
MKOI\RPT:026J9051.001Vp«ite.ttt 11 07/07/95
-------
FOREWORD
The U.S. Environmental Protection Agency is charged by Congress with protecting the
Nation's land, air, and water resources. Under a mandate of national environmental laws, the Agency
strives to formulate and implement actions leading to a compatible balance between human activities
and the ability of natural systems to support and nurture life. To meet this mandate, EPA's research
program is providing data and technical support for solving environmental problems today and
building a science knowledge base necessary to manage our ecological resources wisely, understand
how pollutants affect our health, and prevent or reduce environmental risks in the future.
The National Risk Management Research Laboratory is the Agency's center for investigation
of technological and management approaches for reducing risks from threats to human health and the
environment. The focus of the Laboratory's research program is on methods for the prevention and
control of pollution to air, land, water and subsurface resources; protection of water quality in public
water systems; remediation of contaminated sites and ground water; and prevention and control of
indoor air pollution. The goal of this research effort is to catalyze development and implementation
of innovative, cost-effective environmental technologies; develop scientific and engineering
information needed by EPA to support regulatory and policy decisions; and provide technical support
and information transfer to ensure effective implementation of environmental regulations and
strategies.
This publication has been produced as part of the Laboratory's strategic long-term research
plan. It is published and made available by EPA's Office of Research and Development to assist the
user community and to link researchers with their clients.
E. Timothy Oppelt, Director
National Risk Management Research Laboratory
in
-------
ABSTRACT
Roy F. Weston, Inc., in conjunction with Rohm and Haas Company, conducted a field
pilot study to demonstrate the technical feasibility and cost-effectiveness of Ambersorb' 563
carbonaceous adsorbent for the remediation of groundwater contaminated with volatile organic
compounds (VOCs). The project was conducted under the Emerging Technology Program of the
EPA Super-fund Innovative Technology Evaluation (SITE) program.
The Ambersorb adsorbent technology demonstration was conducted over a 12-week period
during the period from 2 May to 20 July 1994 at Site 32/36 of the Pease Air Force Base (AFB)
in Newington, New Hampshire. The groundwater in this area is contaminated with a number of
chlorinated organics, including vinyl chloride, 1,1-dichloroethene, cis- 1,2-dichloroethene, trans-
1,2dichloroethene, and trichloroethene.
The Ambersorb adsorbent technology demonstration included four service cycles, three
steam regenerations, and one superloading cycle. The study was conducted using a 1-gallon-per-
minute (gpm) continuous pilot system, consisting of two adsorbent columns that can be operated
in parallel or series.
The demonstration study showed that Ambersorb 563 adsorbent is an effective technology
for the treatment of groundwater contaminated with chlorinated organics. The effluent
groundwater from the Ambersorb 563 adsorbent system consistently met drinking water
standards.
Direct comparison of the performance of Ambersorb 563 adsorbent with Filtrasorb* 400
granular activated carbon (GAC) showed that Ambersorb 563 adsorbent treated to the drinking
water standard approximately two to five times the number of bed volumes of water as GAC
while operating at five times the flow rate loading.
On-site steam regeneration was successfully demonstrated. The steam regenerations
yielded a separate organic phase that contained approximately 73% to 87% of the total VOC
mass loaded onto the adsorbent. The majority of VOC recovery was shown to occur within the
first 3 bed volumes of steam as condensate.
The principle of superloading was demonstrated as an effective treatment method for the
aqueous condensate layer generated during the steam regeneration of the Ambersorb adsorbent.
A condensate stream containing approximately 700,000 pg/L VOCs was treated to below the
drinking water standards using the superloading column of Ambersorb 563 adsorbent.
MK01WT:02659051.001>«p«j«e.l» IV 07/06/95
-------
Based on the results of the Ambersorb adsorbent demonstration study, conceptual designs
and cost estimates for full-scale groundwater treatment systems (100 gpm) using Ambersorb 563
adsorbent and GAC were developed. The installed costs for the 100-gpm treatment systems
using Ambersorb 563 adsorbent ($526,100) were significantly greater than those using GAC
($336,800). The total present worth cost analysis, however, showed that after approximately 2
years, the Ambersorb 563 adsorbent system would be less expensive due to its lower operating
costs. The annual operating costs of the Ambersorb 563 adsorbent system were approximately
$32,500/yr for the first 5 years, while the annual operating costs for the GAC system were
approximately $125,800/yr for the first five years.
This report was submitted in fulfillment of Cooperative Agreement CR-821352-01-0 by
Roy F. Weston, Inc. under the partial sponsorship of the U.S. Environmental Protection Agency.
This report covers a period of performance from 1 October 1993 to 28 February 1995, and work
was completed as of 28 February 1995.
MKOr*PT:02659051.001>ep«iile.at V 07/0695
-------
CONTENTS
Disclaimer
Foreword
Abstract
Figures
Tables
List of Acronyms
Acknowledgements
1. Introduction .................................................. 1
ProgramOverview ........................................ 1
Treatment Technology Description ............................. 2
Project Objectives ......................................... 3
2. Conclusions .................................................. 5
3. Recommendations ............................................. 7
4. Experimental Design and Procedures ................................. 8
Experimental Design ...................................... 8
Overview .......................................... 8
Breakthrough Capacity Model .......................... 8
Operating Conditions ....................................... 9
Service Cycles ...................................... 9
Steam Regenerations ................................ 12
Superloading ..................................... 15
Sample Collection and Analysis ............................. 15
Service Cycles .................................... 15
Steam Regenerations ................................ 18
Superloading ..................................... 18
Analytical Procedures ............................... 21
Data Collection and Analysis ............................... 21
Data Collection .................................... 21
Data Summary ..................................... 21
Data Analysis ..................................... 23
5. Equipment and Materials ....................................... 24
PilotUnit .............................................. 24
Site Requirements and Utilities .............................. 24
On-Site Field Laboratory .................................. 27
6. Results and Discussion .......................................... 28
Service Cycle Results .................................... 28
h4KOIW:oz659OSI.OOIkpsite.txt VJ 071071p5
-------
CONTENTS
(continued)
Cycle 1 28
Cycle 2 30
Cycle 3 .... 37
Cycle4 43
Steam Regeneration Results 54
Steam Regeneration 1 54
Steam Regeneration 2 58
Steam Regeneration 3 58
Summary of Steam Regeneration Results 58
Superloading Results 65
Data Quality Review 70
Overview 70
Accuracy 70
Precision 74
Other Data Quality Measures 74
summary 77
Comparison of Ambersorb Adsorbent and Fiiltrasorb GAC Performance - - 77
Summary of Ambersorb Adsorbent Performance 77
Comparison of Predicted and Measured Performance Results 81
Scale-Up Parameters 81
Process Configuration 83
EBCT' or Flow Rate Loading 84
Vessel Configuration 84
Steam Regeneration Conditions 85
7. Conceptual Design and Preliminary Cost Estimate 86
Conceptual Design 86
Preliminary Cost Estimate 92
References 99
vn
-------
FIGURES
Number Page
1 Photograph of Condensate Separation Following Steam Regeneration ... 14
2 Schematic of Ambersorb Adsorbent Pilot Unit 25
3 Photograph of Ambersorb Adsorbent Pilot Unit 26
4 Cycle 1 Ambersorb 563 Adsorbent VOC Breakthrough Curves 32
5 Cycle 1 Filtrasorb 400 GAC VOC Breakthrough Curves 33
6 Cycle 1 Ambersorb 563 Adsorbent VOC Leakage Curves
(Expanded Ordinate) 34
7 Cycle 1 Filtrasorb 400 GAC VOC Leakage Curves
(Expanded Ordinate) 35
8 Cycle 2 Ambersorb 563 Adsorbent Lead Column VOC
Breakthrough Curves 39
9 Cycle 2 Ambersorb 563 Adsorbent Lead Column VOC
Leakage Curves 40
10 Cycle 2 Ambersorb 563 Adsorbent Lag Column VOC
LeakageCurves 41
11 Cycle 3 Amber-sorb 563 Adsorbent Lead Column VOC
Breakthrough Curves 45
12 Cycle 3 Ambersorb 563 Adsorbent Lead Column VOC
Leakage Curves 46
13 Cycle 3 Ambersorb 563 Adsorbent Lag Column VOC
Leakage Curves 47
14 Cycle 4 Amber-sorb 563 Adsorbent Lead Column VOC
Breakthrough Curves 51
15 Cycle 4 Ambersorb 563 Adsorbent Lead Column VOC
Leakage Curves 52
16 Cycle 4 Ambersorb 563 Adsorbent Lag Column VOC
Leakage Curves 53
17 Steam Regeneration 1 Total VOC Mass Recovery Profile 57
18 Steam Regeneration 2 Total VOC Mass Recovery Profile 60
19 Steam Regeneration 3 Total VOC Mass Recovery Profile 62
20 Steam Regenerations Total VOC Mass Recovery Profiles 63
21 Steam Regenerations Condensate Aqueous Phase pH Profiles
22 Ambersorb 563 Adsorbent Superioading Column VOC
Leakage Curves 69
via
-------
FIGURES
(continued)
Number Page
23 Comparison of Ambersorb 563 Adsorbent and Filtrasorb 400 GAC VC
and TCE Leakage Curves 79
24 Process Flow Diagram of 100-gpm Ambersorb 563 Adsorbent
Treatment System 88
25 Process Flow Diagram for 100-gpm GAC Treatment System 90
26 Total Present Worth Cost Profiles for 100-gpm Treatment Systems 98
IX
-------
TABLES
Number Page
Operating Conditions for Service Cycles 10
2 Operating Conditions for Steam Regenerations 13
3 Operating Conditions for Superloading 16
4 VOC Sampling and Analysis Program for Service Cycles 17
5 VOC Sampling and Analysis Program for Steam Regenerations 19
6 VOC Sampling and Analysis Program for Superloading 20
7 Analytical Laboratories and Methods 22
8 Cycle 1 Process Operations Data 29
9 Cycle 1 Performance Results 31
10 Cycle 2 Process Operations Data 36
11 Cycle 2 Performance Results 38
12 Cycle 3 Process Operations Data 42
13 Cycle 3 Performance Results 44
14 Cycle 4 Process Operations Data 49
15 Cycle 4 Performance Results 50
16 Steam Regenerations Process Operations Data 55
17 Steam Regeneration 1 VOC Mass Recovery Results 56
18 Steam Regeneration 2 VOC Mass Recovery Results 59
19 Steam Regeneration 3 VOC Mass Recovery Results 61
20 Summary of Steam Regenerations Total VOC Mass Recovery Results ... 66
21 Superloading Process Operations Data 67
22 Superloading Performance Results 68
23 Typical Detection Limits for Target VOCs 71
24 Summary of Sample Analysis Program 72
25 Summary of Accuracy Data for Matrix Spike/Matrix Spike Duplicate
Samples 73
26 Summary of Precision Data for Duplicate Samples 75
27 QA Objectives for Precision, Accuracy, and MDL for Target VOCs 78
28 Summary of Ambersorb 563 Adsorbent Performance Results 80
29 Comparison of Predicted and Measured Performance Results 82
30 Design Parameters for 100-gpm Treatment Systems 87
31 Major Equipment List for 100~gpm Amber-sorb 563 Adsorbent
Treatment System 89
32 Major Equipment List for 100-gpm GAC Treatment System 91
33 Installed Costs for 100-gpm Ambersorb 563 Adsorbent Treatment System . 93
-------
TABLES
(continued)
Number Page
34 Installed Costs for 100-gpm GAC Treatment System 95
35 Present Worth Costs for 100-gpm Ambersorb 563 Adsorbent Treatment
System 96
36 Present Worth Costs for 100-gpm GAC Treatment System 97
XI
-------
LIST OF ACRONYMS
1,1-DCE
A563
AEL
AFB
BVs
cis- 1,2-DCE
DA
EBCT
EPA
ETL
F400
GAC
GFIC
gpm
HC1
IRP
MCLs
MDLs
MS
MSD
NPL
RPD
SITE
SOC's
TCE
trans- 1,2-DCE
umhos/cm
UV
V C
VOG,
1,1 -dichloroethene
Ambersorb 563 adsorbent
Analytics Environmental Laboratory
Air Force Base
bed volumes
cis- 1,2-dichloroethene
Dubinin-Astakov
empty bed contact time
U.S. Environmental Protection Agency
Environmental Technology Laboratory
Filtrasorb 400 GAC
granular activated carbon
ground fault interrupted circuits
gallon-per-minute
hydrochloric acid
Installation Restoration Program
maximum contaminant levels
minimum detectable levels
matrix spike
matrix spike duplicate
National Priorities List
part per million
quality assurance
quality assurance project plan
Remedial Investigation
relative percent difference
Super-fund Innovative Technology Evaluation
synthetic organic chemicals
trkhloroethene
trans- 1,2dichloroethene
micromhos per centimeter
ultraviolet
vinyl chloride
volatile organic compounds
xn
-------
ACKNOWLEDGEMENTS
This project was conducted as part of the U.S. EPA SITE program under Cooperative
Agreement No. CR-821352-01-0. Ronald J. Turner of the U.S. EPA National Risk Management
Research Laboratory, Cincinnati, Ohio, served as the Project Officer.
The cooperation and assistance of the U.S. Air Force and personnel at Pease AFB,
Newington, New Hampshire, in providing the site for this technology demonstration project is
greatly appreciated.
The key project participants included Russell E. Turner, Joseph F. Martino, Russell W.
Frye, and Anthony G. Bove of WESTON and Deborah A. Plantz, Eric G. Isacoff, and Richard
D. Link of Rohm and Haas. Chemical analysis to support the technology demonstration was
provided by Analytics Environmental Laboratory, Inc. (AEL) in Portsmouth, New Hampshire.
xlll
-------
SECTION 1
INTRODUCTION
PROGRAM OVERVIEW
Roy F. Weston, Inc. (WESTON,), in conjunction with Rohm and Haas Company (Rohm
and Haas), conducted a field pilot study to demonstrate the technical feasibility and cost-
effectiveness of Ambersorb* 563 carbonaceous adsorbent for the remediation of groundwater
contaminated with volatile organic compounds (VOCs). (Ambersorb is a registered trademark
of Rohm and Haas Company, Philadelphia, Pennsylvania.) The Ambersorb 563 adsorbent
technology is currently commercially available. The WESTON/Rohm and Haas team conducted
the Ambersorb adsorbent technology demonstration under the Emerging Technology Program of
the U.S. Environmental Protection Agency (EPA) Superfund Innovative Technology Evaluation
(SITE) Program.
The Ambersorb carbonaceous adsorbent system can remove organic contaminants so that
they can be isolated and disposed of or reclaimed. Ambersorb adsorbents are targeted for
applications on long-term remediation projects where the advantages of on site regeneration will
provide a cost-effective water treatment alternative to granular activated carbon (GAC).
The Ambersorb adsorbent technology demonstration was conducted at Pease Air Force
Base (AFB) in Newington, New Hampshire. The base is included on the National Priorities List
(NPL), and WESTON has been conducting an Installation Restoration Program (IRP) Stage 3
Remedial Investigation (RI) at Pease AFB over the past several years. Based on a review of
groundwater data for various sites at Pease AFB, Site 32/36 was selected for the field trial to
demonstrate the use of Ambersorb 563 adsorbent for the treatment of contaminated groundwater.
The groundwater in this area is contaminated with a number of chlorinated organics, including
vinyl chloride (VC), 1, Idichloroethene (1,1-DCE), cis- l,2dichloroethene (cis- 1,2-DCE), trans-
1 ,2-dichloroethene (trans- 1,2-DCE), and trichloroethene (TCE).
The Ambersorb adsorbent technology demonstration used a 1-gallon-per-minute (gpm)
continuous pilot-scale system to evaluate the treatment of groundwater from Site 32/36 at Pease
AFB. A slip stream from the influent line to the two air strippers currently operating at Site
32/36 was used as the groundwater source for the pilot-scale demonstration. The field study was
performed over a 12-week period from 2 May through 27 July 1994 and consisted of four service
cycles, three steam regenerations, and one superloading to obtain sufficient data to compare the
performance and economics of an Ambersorb 563 adsorbent system with the performance and
economics of a liquid-phase GAC system. Filtrasorb 400* GAC was selected for use during the
I
-------
first service cycle for direct comparison with Ambersorb 563 adsorbent performance. (Filtrasorb
is a registered trademark of Calgon Corporation, Pittsburgh, Pennsylvania.) Filtrasorb 400 GAC
represents a commonly used GAC for liquid-phase fixed-bed groundwater treatment systems.
The field trial was performed using staff personnel from WESTON's Environmental
Technology Laboratory (ETL) in Lionville, Pennsylvania; WESTON's on-site operations office
at Pease AFB; WESTON's office in Concord, New Hampshire; and Rohm and Haas Research
Laboratories, in Spring House, Pennsylvania. Chemical analyses to support the technology
demonstration were provided by a local analytical laboratory, Analytics Environmental
Laboratory, Inc. (AEL), in Portsmouth, New Hampshire.
This report describes the Ambersorb adsorbent technology study objectives and
experimental procedures and equipment used; summarizes the test results; and provides
discussions and recommendations on design parameters and treatment costs for full-scale
treatment systems. A work plan for the Ambersorb adsorbent demonstration project was
presented in a separate document.' A separate quality assurance project plan (QAPP) was also
prepared for the Ambersorb adsorbent technology demonstration study? The QAPP was prepared
in accordance with guidance for the development of a Category III project.
TREATMENT TECHNOLOGY DESCRIPTION
Current field-tested technologies available for the removal of VOCs from groundwater are
based on carbon adsorption and air stripping or aeration.3 Experimental technologies being
investigated include powdered activated carbon, biodegradation, reverse osmosis, ultraviolet (UV)
catalyzed oxidation, and ultrafiltration.3 Generally, the lower the level of the contaminant
concentration that is desired in the treated effluent, the more expensive the treatment technique.
Adsorption techniques using GAC are well-established for groundwater remediation: but
require either disposal or thermal regeneration of the spent carbon. In these adsorbent systems,
the GAC has to be removed from the remediation site and shipped as a hazardous material to the
disposal or regeneration facility. For large systems, on-site regeneration of spent GAC is
sometimes economically justified.
Ambersorb carbonaceous adsorbents are a family of synthetic, tailorable adsorbents that
were first developed in the 1970s for the remediation of contaminated groundwater.5'6 Rohm and
Haas has commercialized several Ambersorb carbonaceous adsorbents.7'*1''10 One particular grade,
Ambersorb 563 adsorbent, based on recently patented technology, has been found to be extremely
effective in the removal of low-level VOCs and synthetic organic chemicals (SOC's) from
contaminated water. The unique properties of Ambersorb 563 adsorbent result in several key
performance benefits:5'11'1113'14'15
Ambersorb 563 adsorbent can be regenerated on-site using steam, thus eliminating
the liability and cost of off-site regeneration or disposal associated with adsorbents
such as GAC. Condensed contaminants are recovered through phase separation.
-------
• Ambersorb 563 adsorbent has 5 to 10 times the capacity of GAC for adsorbing
VOC contaminants, such as chlorinated hydrocarbons, when the contaminants are
present at low concentrations (ppb to ppm levels). This higher adsorptive capacity
translates into significantly longer service cycle times before regeneration is
required
• Ambersorb 563 adsorbent can operate at much higher flow rates than G.AC, while
maintaining effluent water quality below drinking water standards. This advantage
results in a compact system with smaller, hence, less expensive components.
• Ambersorb adsorbents are comprised of hard, nondusting, spherical beads with
excellent physical integrity, thus reducing or eliminating handling problems and
attrition losses typically associated with GAC.
• Ambersorb adsorbent performance is not adversely affected by background levels
of heavy metals or other ionic species in groundwater. Changes in groundwater
pH, temperature, and alkalinity also have no deleterious effect on performance.
• Ambersorb 563 adsorbent is not prone to bacterial fouling.
• Ambersorb adsorbents can be manufactured with consistent reproducible
characteristics.
This combination of performance benefits can result in a more cost-effective alternative
to currently available technologies for the treatment of low-level VGC-contaminated groundwater.
Ambersorb adsorbent technology can be considered for wellhead treatment as well as for a
centralized treatment facility.
PROJECT OBJECTIVES
The objectives of the Ambersorb adsorbent technology demonstration project included the
following:
• Demonstrate that Ambersorb adsorbents can offer a cost-effective alternative to
GAC treatment, while maintaining effluent water quality that meets maximum
contaminant levels (MCLs), as established in the National Revised Drinking Water
Regulations (40 CFR 141.61).
• Validate design parameters and system performance to be used for scale-up to
full-plant scale, including the evaluation of service cycles and establishing steam
regeneration efficiency, superloading, and ease of phase separation. Superloading
refers to the process whereby the aqueous condensate from the steam regeneration
of an Ambersorb 563 adsorbent service column is treated using a smaller column
containing Ambersorb 563 adsorbent. Following superloading treatment, the
-------
aqueous condensate is discharged as part of the treated water stream, The
superloading process is not typically used for GAC system.
• Evaluate the performance/cost characteristics of the Ambersorb adsorbent
groundwater remediation system.
-------
SECTION 2
CONCLUSIONS
Based on the results of the Ambersorb 563 adsorbent technology demonstration, the
following conclusions were developed:
1 Ambersorb 563 adsorbent is an effective technology for the treatment of
groundwater contaminated with chlorinated organics. The effluent groundwater
from the Ambersorb 563 adsorbent system consistently met drinking water
standards.
2. Direct comparison of the performance of Ambersorb 563 adsorbent with
Filtrasorb* 400 GAC, based on the number of bed volumes treated to the MCL,
indicated that Ambersorb 563 adsorbent was able to treat approximately two to
five times the bed volumes of water as Filtrasorb 400 GAC while operating at five
times the flow rate loading [1/5 the empty bed contact time (EBCT)].
3. On-site steam regeneration was successfully conducted during the demonstration
and yielded an easily separable condensate consisting of a VOC-saturated aqueous
stream (top layer) and a concentrated organic phase (bottom layer). The steam
regenerations recovered approximately 73% to 87% of the total VOC mass
adsorbed on the Ambersorb 563 adsorbent column during the service cycle. The
organic phase contained approximately 88% to 93% of the total VOC mass
recovered. The majority of VOC recovery was shown to occur within 3 bed
volumes of steam as condensate.
4. The principle of superloading was demonstrated as an effective treatment method
for the aqueous condensate layer resulting from the steam regeneration of the
Ambersorb adsorbent. A condensate stream containing 700,000 pg/L VOCswas
treated to below the drinking water standards using a superloading column
containing Ambersorb 563 adsorbent.
5. Preliminary cost estimates of the installed costs for a 100~gpm treatment system
using Ambersorb 563 adsorbent were significantly greater than those using GAC.
However, the annual operating cost of the Ambersorb 563 adsorbent system was
significantly lower than the GAC system. The total present worth cost analysis
showed that after approximately 2 years, the Ambersorb 563 adsorbent system
would be more economical because of its lower operating costs.
5
-------
6. The demonstration study enhanced the existing database for the Ambersorb 563
adsorbent technology and helped validate process design parameters and system
performance for scale-up to full-scale treatment systems. Information pertaining
to key parameters of process configuration. EBCT or flow rate loading, vessel
configuration, and steam regeneration conditions was developed or confirmed as
part of the demonstration project.
7. The removal of particulate matter from the influent groundwater prior to the
adsorbent columns must be considered as part of the treatment system design.
During the demonstration project, orange-brown particulate matter (likely iron
precipitates) was observed to accumulate on the column inlet screens, causing
higher than expected pressure drops. The particulate matter was passing through
the pilot unit prefilters or precipitating out from a dissolved state after the filters.
No negative impact on the performance of the Ambersorb 563 adsorbent or
Filtrasorb 400 GAC was observed due to the particulate matter.
8. Based on a comparison of the measured performance results obtained during the
demonstration project and the performance results predicted by the breakthrough
capacity model developed by Rohm and Haas, the breakthrough capacity model
is a useful tool in predicting the adsorption capacity and service cycle times to
support full-scale system design and cost analysis for the Ambersorb 563
adsorbent technology.
9. The accurate quantification of vinyl chloride in the influent groundwater is critical
in establishing the service cycle time for process operations of the Ambersorb
adsorbent and GAC treatment systems. Based on the Rohm and Haas predictive
model, levels of vinyl chloride in the groundwater result in significant decreases
in adsorbent performance as compared to groundwater containing no vinyl
chloride. As measured in the study and predicted by the model, incremental
increases in vinyl chloride concentration result in decreases in adsorption capacity.
10. A 22% to 40% decrease in the number of bed volumes treated to the MCL was
observed for certain contaminants (VC and 'TCE) following one steam regeneration
of the virgin Ambersorb 563 adsorbent. The reduction in bed volumes treated to
the MCL may be the result of the increase in influent vinyl chloride concentration
during the study. Additional steam regenerations and service cycles with
relatively constant vinyl chloride concentration are needed to estimate the
long-term effect of multiple steam regenerations on Ambersorb 563 adsorbent
performance.
-------
SECTION 3
RECOMMENDATIONS
Ambersorb 563 adsorbent technology should be considered as an alternative treatment
method to GAC for the remediation of groundwater contaminated with chlorinated organics.
Specifically, for 100~gpm pump-and-treat systems that are expected to operate for several years,
the Ambersorb 563 adsorbent technology is expected to perform as well as or better than GAC
and at a lower overall cost. In addition, on-site regeneration of the Ambersorb 563 adsorbent
columns provides the option of recycling or direct disposal of the contaminants recovered in the
condensate organic phase. During feasibility studies for sites that require groundwater
remediation, Ambersorb 563 adsorbent technology should be included among the list of viable
treatment technologies considered for evaluation.
-------
SECTION 4
EXPERIMENTAL DESIGN AND PROCEDURES
EXPERIMENTAL DESIGN
Overview
The Ambersorb adsorbent technology demonstration employed a 1-gpm continuous pilot
system. The pilot unit included prefilters to remove suspended solids, two adsorbent columns
that could be operated in parallel or series, one superloading column, and a steam regeneration
system.
The steam regeneration system enabled the direct, on-line regeneration of the Ambersorb
adsorbent columns on-site and included a steam generator, condenser, collection/separation vessel,
and vapor phase Ambersorb adsorbent trap for the condenser vent discharge. Steam was passed
through the beds in a downflow mode to minimize condensate holdup in the vessels. To conduct
a countercurrent regeneration, both adsorbent columns used an upflow, fixed bed configuration.
The Ambersorb adsorbent technology demonstration consisted of four service cycles, three
steam regenerations, and one superloading test. During the first service cycle, the columns were
operated in parallel for direct comparison of the performance of virgin Ambersorb 563 adsorbent
to virgin Filtrasorb 400 GAC. For the remaining cycles, two Ambersorb 563 adsorbent columns
were operated in series to investigate the effect of multiple service cycles and steam regeneration
on Ambersorb adsorbent performance.
Breakthrough Capacity Model
A breakthrough capacity computer model, developed by Rohm and Haas, was used to
predict the service cycle times for the demonstration study based on the average contaminant
concentrations measured in the Site 32/36 wells during the Stage 5 IRP at Pease AFB.
Liquid-phase static adsorption isotherms are commonly used to estimate adsorption
capacity for organic contaminants from water over a range of concentrations. Although these
isotherms cannot simulate an adsorbent's performance under dynamic conditions for a
multicomponent system, isotherms are valuable tools in helping to predict service cycle time.
The linear Freundlich equation is commonly used to represent adsorption isotherms for
GAC. Rohm and Haas has found, however, that the linear Freundlich isotherm is not appropriate
-------
for the curved isotherms that are typically obtained for Ambersorb adsorbents over the low part
per million (ppm) concentration range.'* A quadratic equation can model this behavior. An
investigation of several isotherm functions showed that the Dubinin-Astakov (DA) equation was
the optimum equation for representing typical VOCs.17
As a tool to assist in predicting estimated service cycle time, Rohm and Haas developed
a computer model based on the DA equation.'* Using the contaminants and. respective
concentrations for a given influent water analysis, the model provides an estimate of the number
of bed volumes that can be treated to a 50% stoichiometric breakthrough point for a given
contaminant (i.e., C^C,, = 0.5, where Ce is the effluent concentration and C0 is the influent
concentration). The model also predicts the first component to breakthrough based on the
contaminant load.
Specifically, for this study, the model predicted that vinyl chloride would be the first
component to break through and that service cycle times would be significantly affected by small
changes in the influent vinyl chloride concentration. During the first 7 days of the Ambersorb
adsorbent demonstration study, however, no detectable levels of vinyl chloride (<5ng/L) were
measured in the influent groundwater. Because of high TCE concentrations in the influent stream,
the influent samples needed to be diluted 10-fold thus increasing the minimum level of detection
for vinyl chloride from 0.5 |ig/L to 5 ng/L. Based on the lower than expected influent VC
concentrations (assumed to be zero), the Rohm and Haas model predicted that service cycle times
would be almost twice the duration previously estimated. Therefore, after 7 days of operation,
the estimated process flow rates for Cycle 1 were doubled. After final evaluation of all influent
and effluent VOC concentrations measured during the demonstration study, it was estimated that
vinyl chloride was present in the influent stream at concentrations ranging from 3 to 11 (Jg/L
based on a volume-weighted average during the entire study.
The influent contaminant levels measured during the first service cycle were then used
as input to set the operating parameters for subsequent cycles. The contaminant concentrations,
specifically vinyl chloride, had a significant impact on breakthrough time and other performance
parameters, including leakage during each cycle.
OPERATING CONDITIONS
Service Cycles
Operating conditions for each of the service cycles are presented in Table 1. During the
first service cycle, the columns were operated in parallel for direct comparison of the
performance of virgin Ambersorb 563 adsorbent (A563) to virgin Filtrasorb 400 GAC (F400).
Initially, the virgin Ambersorb 563 adsorbent column designated by column identification number
A563A and the virgin Filtrasorb 400 adsorbent column, designated by column identification
number F400, were operated at flow rates of approximately 0.44 and 0.29 gpm (EBCT of 2.7 and
15.8 minutes), respectively. During the first 7 days of operation, no detectable levels (<5 jig/L)
-------
TABLE 1. OPERATING CONDITIONS FOR SERVICE CYCLES*
Service Cycle
Adsorbent Column
Bed Geometry
Diameter, inches
Length, inches
Volume, gallons
Orientation
Process Operations Data}
Total Operation Time, days
Total Volume Treated, gallons
Total Volume Treated, BV
Process Flow Rate, gpm
Flow Rate Loading, BV/hr
Hydraulic Loading, gpm/ft*
Empty Bed Contact Time, min.
Cycle
A563A
4.0
22.0
1.2
up-flow
16.8
16,400
13,700
0.44/0.84
22/42
5.1/9.6
2.7/1.4
It
F400
6.0
37.0
4.5
up-flow
30.6
23,000
5.070
0.29/0.59
3.8/7.8
1 .5/3.0
15.8/7.7
Cycle 2
A563B
4.0
22.0
1.2
up-flow
12.8
15,200
12,700
0.83
41
9.5
I.4
Cycle 3
A563A
4.0
22.0
1.2
up-flow
7.8
IO.300
8,600
0.9 I
46
10.5
1.3
Cycle 4
A5638
4.0
22.0
1. 2
up-flow
12.9
15,300
12,800
0.82
41
9.4
1.5
• During Cycle I. columns were operated in parallel. During Cycles 2,3, and 4, two columns were operated in series. Operating conditions for
Cycles 2,3, and 4 represent system loading to the lead column.
t During Cycle I the process flow rate was doubled after 7 days of operation to decrease service cycle times. VC was below detection limits in the
influent stream.
I Time weighted averages and cumulative totals for the total operating period.
-------
of vinyl chloride were measured in the influent groundwater. Therefore, after 7 days of
operation, process flow rates of the Ambersorb adsorbent and Filtrasorb GAC columns were
doubled to approximately 0.84 and 0.59 gpm (EBCT of 1.4 and 7.7 minutes), respectively.
During Cycle 1, the Ambersorb adsorbent and Filtrasorb GAC columns were operated
well beyond vinyl chloride breakthrough to fully define the breakthrough curves for the
remaining VOCs. Breakthrough for a specific VOC is defined as the condition at which the
column effluent VOC concentration equals one half the influent VOC concentration. The
Ambersorb adsorbent column was terminated after 13,700 bed volumes (16,400 gallons) had been
treated (after 17 days of operation), and the Filtrasorb GAC column was terminated after 5,070
bed volumes (23,000 gallons) had been treated (after 31 days of operation).
Bed volumes as opposed to absolute gallons, are the units typically used to compare the
performance of different adsorbents for varying sized systems. Bed volumes represent the
relative volume of groundwater treated normalized to account for the size of the adsorbent
column. The Ambersorb 563 adsorbent column had a bed volume of 1.20 gallons, and the
Filtrasorb 400 GAC column had a bed volume of 4.53 gallons (i.e., approximately four times
larger than the bed volume of the Ambersorb adsorbent column).
For the remaining cycles, two Ambersorb 563 adsorbent columns were operated in series
to investigate the effect of multiple service cycles and steam regenerations on Ambersorb
adsorbent performance.
After the first service cycle, the exhausted Ambersorb 563 adsorbent column (A563A) was
steam regenerated on-site and placed in the lag position for the second service cycle. After steam
regeneration, the column identification number changed from A563A to A563A-1 to designate
that one steam regeneration was conducted on column A563A. The Filtrasorb 400 GAC column
from the first service cycle was replaced by a new virgin Ambersorb 563 adsorbent column
identical in dimension to the A563A column and placed in the lead position for Cycle 2. The
new virgin Ambersorb adsorbent column was designated A563B.
For Cycles 3 and 4, the newly regenerated lead columns from the previous cycles
(A563B-1 and A563A-2) were also placed in lag positions, and the lag columns from the
previous cycles (A563A-1) andA563B-l) were placed in lead positions to simulate the operating
mode in a full-scale system. For Cycles 3 and 4, therefore, the lead Ambersorb adsorbent
columns were preloaded with VOC leakage from the previous service cycles.
During Cycles 2, 3, and 4, the Ambersorb adsorbent columns were operated in series at
flow rates ranging from 0.83 to 0.91 gpm corresponding to 1.3- to 1.5-minuteEBCTs for one
column or 2.6- to 2.9-minute EBCTs for two columns in series. Each cycle was operated well
beyond vinyl chloride breakthrough in the lead column to defme the breakthrough curves for the
remaining VOC's Cycles 2, 3, and 4 were terminated after 13, 8, and 13 days of operation,
respectively. During Cycle 3, the system was shutdown for approximately 2 days, resulting in
8 days of actual operating tune.
11
-------
During each service cycle, influent and effluent samples of each column were collected
daily and analyzed for VOCs. In addition, selected influent and effluent column samples were
measured for pH, conductivity, and alkalinity. Process parameters, including groundwater influent
flow rate, temperature, and pressure, were also monitored at periodic intervals throughout the
service cycles.
Steam Regenerations
Steam regenerations were conducted on the Ambersorb adsorbent column at the end of
Cycle 1 and on the lead Ambersorb adsorbent column at the end of Cycles 2 and 3 to evaluate
the effect steam regeneration has on Ambersorb adsorbent performance. The steam regenerations
were also conducted at various temperatures (307 °F, 293 °F, 280 °F) to evaluate the effect of
regeneration temperature on contaminant recovery. Operating conditions for each of the steam
regenerations are presented in Table 2.
Each steam regeneration was conducted directly on the lead Ambersorb adsorbent column
on-site using a portable steam generator. Prior to the introduction of steam to the top of the
column, the Ambersorb adsorbent columns were wrapped in electrical heating tape, insulated, and
preheated to the target regeneration temperature. Steam was then applied to the column in a
downflow direction. Desorbed contaminant vapors and water vapor were then condensed in a
water-cooled condenser and collected in 1-liter graduated glass burettes. The volumetric rate of
condensate produced (regeneration rate) was increased incrementally over a 17- to 19-hour period
from approximately 0.23 BV/hr to 0.82 BV/hr. Depending on the regeneration rate, the 1-liter
burettes were filled and recovered for sample collection every 15 to 60 minutes.
The condensate produced during each regeneration consisted of a visible and separable
concentrated organic phase (bottom layer) and a VOC-saturated aqueous phase (top layer). A
photograph of the condensate phase separation following steam regeneration is presented in
Figure 1. The volumes of both the organic and aqueous layers were measured directly in the
burettes. The organic layer was then drained from the burette into a volumetric flask and diluted
with methanol to a known volume (typically 250 mL) for analytical purposes. Samples of the
diluted organic phase were collected from the volumetric flask and analyzed for VOCs Samples
of the aqueous phase were collected directly from the graduated burette and analyzed for VOCs,
pH, and conductivity.
During the field trial, the regeneration was extended significantly beyond what would be
practiced during full-scale commercial operation, and considerably more samples were taken in
order to fully define the mass recovery curve and to complete an accurate mass balance.
To ensure that there was no VOC vapor discharged during each steam regeneration, a trap
containing Ambersorb 563 adsorbent was used on the vapor discharge from the condenser. At
the end of each regeneration, the vapor trap adsorbent was recovered and extracted twice using
a 2:1 ratio of methanol to adsorbent volume. The two extracts were combined and the final
volume of methanol measured. Duplicate samples of the combined extracts were collected and
analyzed.
12
-------
TABLE 2. OPERATING CONDITIONS FOR STEAM REGENERATIONS
Steam Regeneration
Adsorbent Column
Bed Geometry
Diameter, inches
Length, inches
Volume, gallons
Orientation
Process Operations Data*
Total Operation Time, hours
Total Volume Condensate Generated, gallons
Total Volume Condensate Generated, bed volumes
Column Temperature, °F
Column Temperature, °C
Regeneration 1
A563A
4.0
22.0
1.20
down-flow
17.4
9.1
7.6
307
153
Regeneration 2
A563B
4.0
22.0
1.20
down-flow
17.1
8.4
7.0
293
I45
Regeneration 3
A563A
4.0
22.0
.20
down-flow
18.5
10.7
8.9
280
138
" Time weighted averages and cumulative totals for the total operating period
-------
Figure 1, Photo of Condensate Separation Following Steam Regeneration
14
-------
After steam regeneration was completed, the adsorbent column was allowed to cool to
approximately 194 °F. Then the adsorbent column was flushed with tap water in a up-flow
direction to rehydrate the adsorbent. After reaching ambient temperature, the adsorbent column
was then placed in the lag position and the subsequent service cycle was initiated.
Superloading
A test to demonstrate the use of an Ambersorb 563 adsorbent superloading column to treat
the aqueous condensate from a typical steam regeneration process was also conducted during the
field trial. This test was performed to demonstrate the concept of a closed loop Ambersorb
adsorbent treatment system in which the only discharge is the separable organic layer resulting
from steam regeneration. Ambersorb adsorbent was chosen for the superloading column because
of its high adsorption capacity and superior kinetics while operating at a high flow rate loading.
Operating conditions for the superloading test are presented in Table 3.
Superloading was conducted by passing the saturated aqueous phase from the third steam
regeneration (approximately 4 gallons) through an Ambersorb 563 adsorbent superloading column
(A563S) with a diameter of 2 inches and a bed height of 21 inches. Superloading was conducted
at an approximate rate of 8 BV/hr for approximately 1.8 hours and treated approximately 14 BVs
of saturated condensate. Influent and effluent samples from the superloading column were
collected for VOC analysis initially and every hour during the test.
SAMPLE COLLECTION AND ANALYSIS
Service Cycles
The VOC sampling and analysis program for the service cycles is summarized in Table 4.
Samples were collected from ports located before and after each column directly into sample
containers. Initially during Cycle 1, samples of the influent to the pilot unit and the effluent
streams from each of the two columns were collected three times per day at approximately
7 a.m., 11 a.m., and 3 p.m.
Once the VOC breakthrough curves were defined in Cycle 1, sampling frequency was
reduced for the remaining cycles. During Cycles 2, 3, and 4, column influent and effluent
samples were collected twice per day at approximately 7 am. and 3 p.m. All samples for VOC
analysis were collected in duplicate.
During Cycle 1, two of the three daily sample sets (typically the 7 a.m. and 3 p.m. sample
sets) were analyzed for VOCs. During the remaining cycles, one of the two daily sample sets
(usually the 3 p.m. sample set) was analyzed for VOCs.
During each cycle, column influent and effluent samples were collected once a day
(typically during the 7 a.m. sample set) for pH and conductivity measurements. In addition.
during Cycle 1, selected influent and effluent samples of the Filtrasorb 400 GAC column were
15
-------
TABLE 3. OPERATING CONDITIONS FOR SUPERLOADMG
Adsorbent Column A563S
Bed Geometry
Diameter, inches 2.0
Length, inches 21.0
Volume, gallons 0.29
Orientation up-flow
Process Operations Data*
Total Operation Time, hours 1.8
Total Volume Treated, gallons 4.0
Total Volume Treated, bed volumes 14.0
Process Flow Rate, gpm 0.038
Flow Rate Loading, bed volumes/hr 8.0
Hydraulic Loading, gpm/ft1 1.7
Empty Bed Contact Time, minutes 7.5
.Time weighted averages and cumulative totals for the total operating period.
16
-------
TABLE 4. VOC SAMPLING AND ANALYSIS PROGRAM FOR SERVICE CYCLES
Service
Cycle
1
31t
13
8
13
34
Number of Samples Analyzed per Service Cycle*
Base Samples
Column Column Eflluent
Influent Lead \_aa
62 35 62
13 13 13
888
13 13 13
96 69 96
QA Samples
_Dupli- Matrix Confirm- Blanks
cates Spikes atory Field Trip
16 II II II II
44444
33333
44444
27 22 22 22 22
Total
219
59
39
59
376
. For each service cycle, conductivity and pH were measured on influent and cfllucnt samples collected each day during the 7:00 am sampling event (not
shown in table). For the Filtrasorb 400 GAC column, alkalinity was measured on one initial and one final influent sample and each day on one
effluent sample for the tint 17 days. After 17 days, alkalinity was measured once every other day on the Filtrasorb 400 GAC column.
t During Cycle I, the Ambersorb 563 adsorbent column operated 17 days and Filtrasorb 400 GAC column operated 3 I days.
-------
analyzed for alkalinity as identified in Table 4. Alkalinity was measured on the Filtersorb 400
GAC influent and effluent streams to determine if pH control would be required for the full-scale
design. GAC typically imparts some alkalinity into the effluent stream when treating
groundwater. If the effluent pH increases above the discharge criterion, then pH control will be
required.
Quality assurance samples were collected and analyzed during each service cycle to
determine accuracy, precision, and other data quality parameters. All VOC samples were
collected in duplicate, from which three samples were randomly selected for analysis every 6
days during Cycle 1, and two were randomly selected for analysis every 6 days during the
remaining cycles. Two trip blanks, field blanks, matrix spike/matrix spike duplicates, and
confirmatory samples were collected and analyzed every 6 days during each cycle.
Steam Regenerations
The VOC sampling and analysis program for each steam regeneration is summarized in
Table 5. The condensate produced during each regeneration consisted of a visible and separable
concentrated organic phase (bottom layer) and a VOC-saturated aqueous phase (top layer). The
organic layer was drained from the burette into a volumetric flask and diluted with methanol to
a known volume (typically 250 mL). Samples of the diluted organic phase were collected from
the volumetric flask and analyzed for VOCs. Samples of the aqueous phase were collected
directly from the graduated burette and analyzed for VOCs, pH, and conductivity. In addition,
VOC analyses were performed on the methanol extract of the vapor traps for each regeneration.
Quality assurance samples were collected and analyzed during each regeneration to
determine accuracy, precision, and other data quality parameters. All VOC samples were
collected in duplicate, from which one sample was randomly selected for analysis for each
regeneration. Trip blanks, field blanks, matrix spike/matrix spike duplicates, and confirmatory
samples were collected and analyzed at a frequency of one per regeneration. In addition, a
duplicate VOC sample of the vapor trap extract was analyzed for the first regeneration.
Superloading
The VOC sampling and analysis program for the superloading test is summarized in
Table 6. Steam regeneration was not conducted on the superloading column. Influent and
effluent VOC samples were collected from the superloading column every 30 minutes during the
test.
Quality assurance samples were collected and analyzed for the superloading test to
determine accuracy, precision, and other data quality parameters. All VOC samples were
collected in duplicate, from which one sample was randomly selected for analysis. One trip blank,
field blank, and confirmatory sample were also collected and analyzed for the superloading test.
18
-------
TABLE 5. VOC SAMPLING AND ANALYSIS PROGRAM FOR STEAM REGENERATIONS
Regeneration
1
2
3
Total
Column I.D.
A563A
A563B
A563A-I
--
Number of Samples Analyzed per Regeneration*
Base Samples
Condensate Phases Vapor Trap
Aqueous Organic Extract
23 19 1
17 13 1
26 14 1
66 46 3
Quality Assurance Samples
Dupli- Confimn-
cates atory Field Trip
2t 1 I 1
1 1 1 1
1111
2333
Total
46
35
45
126
. During each regeneration, conductivity and pH were measured on the aqueous condensate collected during each sample event (not included in table).
t During the first regeneration, one VOC duplicate from the vapor trap extract was analyzed.
-------
TABLE 6. VOC SAMPLING AND ANALYSIS PROGRAM FOR SUPERLOADING
Number of Samples Analyzed*
Base Samples
Influent Effluent
4 4
Quality Assurance Samples
Duplicates Confirmatory Field Blanks Trip Blanks
1111
Total
12
Conductivity and pH were measured on all influent and effluent samples collected (not included in table).
to
O
-------
Analytical Procedures
Table 7 identifies the laboratory, method, and holding time for the parameters tested
during the field trial program. Analyses for temperature, pH. and conductivity were performed
on-site immediately upon collection. Samples collected during the service cycles, steam
regenerations, and superloading were analyzed for target VOCs (i.e., VC, 1,1-DCE, cis- 1.2-DCE.
trans- 1,2DCE, and TCE) at AEL, located on the Pease AFB property. These analyses were
performed on a quick turnaround basis (i.e., 24 to 48 hours). Selected samples were analyzed
for VOCs (i.e., full list) at AEL's off-site laboratory for confirmation purposes.
The testing procedures used for sample analysis were based upon EPA-approved methods.
The deliverables consisted of commercial data packages.
DATA COLLECTION AND ANALYSIS
Data Collection
Process operating data were collected during each sampling event during the
demonstration study and recorded directly onto data spreadsheets or into bound logbooks. The
data parameters measured and recorded on the data spreadsheets were influent stream flow rate
and temperature, column dimensions, pressure drops, and totalizer readings. Observations, notes,
process upsets, key incidents, and influent and effluent stream pH and conductivity measurements
were recorded in the field logbook. At the end of each day of operation, copies of operations data
and VOC analytical results were faxed to WESTON's project engineer for review, data
validation, and key entry into the computer spreadsheets. The original operations data sheets
were kept secured in the field laboratory during each service cycle and, after each service cycle,
were transferred by the project engineer to the project file at WESTON.
In addition, photographs of the site, field laboratory, pilot plant system, and condensate
samples from the steam regenerations were taken during the demonstration project and
maintained in the project files.
Data Summary
Operations data and VOC analytical results were summarized in graphical and tabular
forms using computer-based spreadsheets (Microsoft Excel).
All quantitative data, such as operations data and VOC analytical results, entered into
computer spreadsheets were checked against the original data records to ensure that the correct
values had been transferred. Following this, the data were reviewed and inconsistencies were
resolved by seeking clarification from the study personnel responsible for collecting the data.
21
-------
TABLE 7. ANALYTICAL LABORATORIES AND METHODS
Parameter
Temperature
PH
Conductivity
Alkalinity
VOCs (target list)
VOCs (target list)
VOCs (full list)
Matrix
Groundwater
Groundwater
Groundwater
Groundwater
Groundwater
Aqueous Phase,
organic Phase,
Vapor Trap
Groundwater
Laboratory
On-site
On-site
On-site
On-Site
AEL Pease
AEL Pease
AEL Off-Site
Method
SM 212
EPA 150.1
EPA 120.1
SM2320
SW846 8010
SW846 8010
SW846
Holding Time
Immediately
24 hours
28 days
7 days
14 days
7 days
14 days
-------
Qualitative data, such as field notes recorded in the logbook, were checked by the project
engineer by direct interview with the study personnel recording the notes. Random checks of
sampling and testing conditions were made by the project engineer to confirm the recorded
observations. Peer review also was incorporated into the data summary process, particularly for
qualitative data, to maximize consistency between study personnel.
VOC analytical results for QA/QC samples collected were also checked to assess data
precision, accuracy, and completeness.
Data Analysis
Based on the summarized test results, the data were analyzed to develop specific
information concerning the following key design parameters investigated during the study:
• Confirm the ability of Ambersorb adsorbent to meet drinking water standards
while maximizing flow rate loading.
• Establish working capacity over several cycles.
• Compare performance of Ambersorb adsorbent and GAC in terms of treatment
effectiveness.
• Identify regeneration conditions, including:
Steam temperature
Steam flow rate
Total steam consumption.
The information developed during the demonstration study was used to expand the
existing technical database on the Ambersorb adsorbent technology and to enhance information
on scale-up and estimates of treatment costs.
23
-------
SECTION 5
EQUIPMENT AND MATERIALS
PILOT UNIT
The key equipment item required to conduct the technology demonstration was the pilot
unit. This unit, designed and owned by Rohm and Haas, consists of a 1-gpm, transportable
assembly that is designed for 24-hour continuous operation with two adsorbent columns that can
operate either in parallel or in series. The pilot unit also includes a self-contained steam
generator for direct on-line, on-site steam regeneration. A schematic of the pilot system is
presented in Figure 2, and a photograph is presented in Figure 3.
The portable 1-gpm rig is housed in a 4-ft-wide by 7-ft-high enclosure that can be easily
moved using a forklift truck. The enclosure has a vent fan, a rubber-lined roof for protection
against rain, and a front door that can be locked for security. The steam generator is enclosed
in a separate 4-ft-wide by 4-ft-high container.
Key equipment for the pilot unit includes the following:
1. Two lo-micron cartridge filters to remove particulate matter.
2. Two glass adsorption columns (4- or 6-inch diameter) that can operate either in
series or in parallel. Each is equipped with a flow meter, influent and effluent
pressure gauges, and sampling ports.
3. Self-contained portable steam generator.
4. Condenser.
5. Condensate collection burette (phase separation vessel).
6. Vapor trap containing Ambersorb adsorbent to capture any gaseous emissions.
SITE REQUIREMENTS AND UTILITIES
The pilot unit was set up within the fenced area surrounding the existing Site 32/36
treatment plant at Pease AFB The pilot unit was located on the northern side of and
24
-------
_o
(X
I
o
C/5
•e
o
CO
-------
PORTABLE AMIEftSOM AOSOMCMT
DCMONSTKA^ON STSTIN
Figure 3. Photo of Ambersorb® Adsorbent Pilot Unit
26
-------
adjacent to the air stripping towers at the existing treatment plant. The existing treatment plant
at Site 32/36 was in normal operation during the demonstration period.
Untreated groundwater was delivered to the pilot unit from an existing 10,000-gallon
holding tank at the site, which was used for flow equalization and storage of contaminated
groundwater recovered by several remediation wells installed in the Site 32/36 area.
The treated effluent from the pilot unit was passed through a GAC polishing filter prior
to discharge to the site sewer to ensure that there was no VOC discharge from the pilot unit.
City water was available for flushing and rehydrating the columns prior to starting each
cycle and to provide water for the portable steam generator and condenser (minimum flow rate
of 5 gpm at a pressure of 50 to 60 psi).
The electrical service required for the pilot unit and portable steam generator consisted
Of:
• Two 208-40~amp hookups (three-phase).
• Six ground fault interrupted circuits (GF'IC) with 20-amp breakers.
A portion of the on-site building, currently used by WESTON on-site operations staff at
Pease A F B was used as an office for on-site personnel from the WESTON/Rohm and Haas SITE
project team. This building is located within 0.5 mile of the Site 32/36 treatment plant. The
office area was equipped with a desk, table, chairs, telephone, fax machine, copy machine, and
personal computer.
ON-SITE FIELD LABORATORY
The control room for the Site 32/36 treatment plant was used as an on-site laboratory for
sample storage, equipment calibration and storage, conducting pH and conductivity
measurements, and logbook and data sheet entry and filing.
Daily influent and effluent samples, collected for VOC analysis, were stored in two 4-
cubic-foot refrigerators. Influent and effluent samples collected during service cycles were
separated from the highly contaminated samples, such as steam regeneration samples and vapor
trap extracts. VOC samples were stored until analysis by AEL or held for a maximum of 14 days
or until data validation was completed for that day's VOC analytical results. VOC samples
selected for analysis were transferred directly from the refrigerators to small coolers with blue
ice for transport to AEL.
27
-------
SECTION 6
RESULTS AND DISCUSSION
SERVICE CYCLE RESULTS
Cycle 1
Cycle 1 was a direct comparison of the performance of the Ambersorb 563 adsorbent and
Filtrasorb 400 GAC. Cycle 1 process operations data are presented in Table 8 and include the
influent average VOC concentrations measured over the total operating period for each column.
Because of analytical limitations such as elevated TCE concentrations (as discussed in Section
4, page 9) influent average VC and 1 ,1-DCE concentrations were estimated, based on the mass
of VC and 1,1-DCE that was subsequently recovered during the first steam regeneration.
Process operations data presented for each service cycle are reported as time-weighted averages
and cumulative totals for the total operating period. Time-weighted averages were calculated by
integration of the cumulative operating time and process operating parameter (such as flow rate)
measured during each service cycle.
In addition, during Cycle 1 and throughout the entire study, orange-brown particulate
matter (likely iron precipitates) was observed to build up on the column inlet screens, causing
higher than expected pressure drops. The particulate matter was either passing through the pilot
unit pre-filters or precipitating out from a dissolved state after the pre-filters. The particulate
matter was periodically cleaned from the column inlet screens during the study. In spite of the
presence of particulate matter, there was no negative impact on the performance of the
Ambersorb 563 adsorbent or Filtrasorb 400 GAC.
During Cycle 1, the virgin Ambersorb 563 adsorbent column (A563A) was operated for
17 days at an average flow rate of 0.68 gpm (1.8-minuteEBCT) treating a total of 13,700 bed
volumes (16,400 gallons) of groundwater. The virgin Filtrasorb 400 GAC column (F400A) was
operated for 31 days at an average flow rate of 0.52 gpm (8.7-minute EBCT) treating a total of
5,070 bed volumes (23,000 gallons) of groundwater.
Cycle 1 process operations data show that the average VOC concentrations in the influent
stream exceeded the MCL, except for 1,1-DCE. In addition, the pH of the influent groundwater
28
-------
TABLE 8. CYCLE PROCESS OPERATIONS DATA+
Ambersorb 563 Adsorbent
Filtrasorb 400 GAC
Column I.D.
Bed Geometry
Diameter, inches
Length, inches
Volume, gallons
Orientation
Process Operations Data
Total Operation Time, hours
Total Volume Treated, gallons
Total Volume Treated, bed volumes
Process Flow Rate, gpm
Flow Rate Loading, bed volume s/hr
Hydraulic Loading, gpm/ft2
Empty Bed Contact Time, minutes
Column Skin Temperature, °F
Pressure Drop Across Bed, psi
Influent Characteristics
pH, standard units
Specific Conductance, (imhos/cm
Alkalinity, mg/L as CaC03
VOC Concentrations, jig/L
Vinyl Chloride
I, I -Dichloroethene
cis- 1,2-Dichloroethene
trans- 1,2-Dichloroethene
Trichloroethene
Effluent Characteristics
pH, standard units
Specific Conductance, ^mhos/cm
Alkalinity, mg/L as CaC03
A563A
4.0
22.0
1.20
up-flow
403
16,400
13,700
0.68
34
7.8
1.8
62
8.4
7.3
575
200
3.4f
0.3 It
312
102
4,330
7.3
574
NAJ
F400
6.0
37.0
4.53
up-flow
735
23,000
5,070
0.52
6.9
2.7
8.7
64
9.3
7.2
606
200
3.9t
0.3 If
329
101
4,120
7.2
608
203
. Time weighted averages and cumulative totals for the total operating period.
f VC and 1.1 -DCE concentrations estimated based on the mass recovery results for the first steam regeneration of column A663A.
I NA = not analyzed.
29
-------
during Cycle I ranged from 5.9 to 8.1. The pH of the effluent stream from the Ambersorb 563
adsorbent column ranged from 6.3 to 7.8 and the pH of the effluent stream from the Filtrasorb
400 GAC column ranged from 5.8 to 8.1 during Cycle 1. The average conductivity of the
influent groundwater and effluent streams ranged from 574 to 608 micromhos per centimeter
(umhos/cm). The average alkalinity of the influent groundwater and effluent stream from the
Filtrasorb 400 GAC column was 200 and 203 mg/L as CaCO3, respectively. No significant
difference was observed between the influent and effluent pH. conductivity, and alkalinity of each
column during Cycle 1. Cycle 1 performance results, based on treatment to the MCL, are
presented in Table 9. The number of bed volumes treated to the MCL were determined by
analysis of the VOC breakthrough and leakage curves for each column. Cycle 1 VOC
breakthrough curves for the Ambersorb 563 absorbent and Filtrasorb 400 GAC columns are
presented in Figures 4 and 5, respectively. Cycle 1 VOC leakage curves, presented in Figures
6 and 7, expand the values of the ordinate (concentration levels) to a maximum concentration of
20 ng/L, which shows the effluent quality of each column more clearly.
Cycle 1 performance results show that both Ambersorb 563 adsorbent and Filtrasorb 400
GAC adsorbents achieved effluent water quality below the MCL for each VOC. Specifically, the
Ambersorb 563 adsorbent column treated approximately 8,120 bed volumes before the first VOC
(VC) broke through at a concentration above the MCL. For the Filtrasorb 400 GAC column,
approximately 1,730 bed volumes were treated before the first VOC (VC) broke through at a
concentration above the MCL. During Cycle 1, concentrations of 1 ,1-DCE and trans- 1 ,2-DCE
in the effluent of the Ambersorb 563 adsorbent column and trans-l,2-DCE in the effluent of the
Filtrasorb 400 GAC column never exceeded the MCL.
A comparison of bed volumes treated to the MCL for each VOC shows that, while
operating at approximately five times the flow rate (1/5 the EBCT), Ambersorh 563 adsorbent
treated approximately two to five times the bed volumes of groundwater as Filtrasorb 400 GAC.
Cvcle 2
Cycle 2 was conducted using two Ambersorb adsorbent columns in series. A virgin
Ambersorb 563 adsorbent column (A563B) was placed in the lead position, and the steam
regenerated Ambersorb 563 adsorbent column (A563A-1) from Cycle 1 was placed in the lag
position.
Cycle 2 process operations data are presented in Table 10 and include the influent average
VOC concentrations measured over the total operating period. Because of the analytical
limitations discussed in Section 4, page 9, the influent average 1,1-DCE concentrations were
estimated based on the mass of that was subsequently recovered during the second
steam regeneration.
30
-------
TABLE 9. CYCLE I PERFORMANCE RESULTS
Volatile Organic Compound
Vinyl Chloride
,1 -Dichloroethene
cis- 1,2-Dichloroethene
trans- 1,2-Dichloroethene
Trichloroethene
MCL+
Hg/L
2
7
70
100
5
Bed Volumes
Ambersorb 563 Adsorbent
8,120
>13,700
9,690
>13,700
8,190
Treated to MCL
Filtrasorb 400 GAC
1,730
>5,070
3,710
5,040
4,850
Difference
Factorf
4.7
-2.7
2.6
>2.7
7
* Maximum Contaminant Levels from National Revised Primary Drinking Water Regulations, 40 CFR 14 1.6 I.
t Difference Factor = (BV Treated by Ambersorb 563 Adsorbent)/(BV Treated by Filtrasorb 400 GAC).
-------
o
o
o
o
o"
o
o
o
1 i
iv a
5
3
«
1
= 1
0 1
1 1 j
S = *
u -a „
>* 9
.£ -
> - i
1
1 1 1
0 C
o c
r- v
u
I |
\ o
I «
i 9
J
3 is
1
1
§ i
3 v
J
^
J
>
•>
U
c
D
:
1
3
^
1
C
T
(O
U
1
H
H
U
03
U
§ i
r f
» c
s e
1 r
1
3 C
3 C
v|
1
i
1
3 <
. o
oo
u
0 3
- 8 "5
? >
T3
0 «
§
0
TT~
*- 1
O_
m"
O
z. o
°.
o
_.
3
U
00
8
I
o
(O
•o
SO
?
o
S2
00
32
-------
U)
500
450
400
300
e
'&
sj
e
o
o
150
100
50
• Vinyl Chloride
•6 1,1 -Dichloroethene
* cis- 1,2-Dichloroethene
•trans- 1,2-Dichloroethenc
* Trichloroethene
EBCT = 8.7 minutes
89 B^SAtt-fi-SfegMB
500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500 5,000 5,500
Bed Volumes Note Scale
Figure 5. Cycle Filtrasorb 400 GAG VOC Breakthrough Curves
-------
c
TJ
TJ
V
•a
n
0.
X
U
U
*" »
= ->
1 o
> >
•8 |
CD •£
o
>n
•e
-------
20
18 -
16
14
o
g 10 -----
u
u
§
U
2 -I
!
— — O— — trans- 1,2-D
3BCT = 8.
7 minutes
r^
>ride
roethene
hloroethene
ichloroethene
thene
. ... _.
^ m^^M
Wt—^A
Mm— \Emt~wm-
J
1,000 1,500 2,000 2,500 3,000 3.500 4,000 4,500
Bed Volumes
5,000 5.500
Note Scale
Figure 7. Cycle Filtrasorb 400 GAG VOC Leakage Curves (Expanded Ordinate)
-------
TABLE IO. CYCLE 2 PROCESS OPERATIONS DATA'
Column I.D.
Bed Geometry
Diameter, inches
Length, inches
Volume, gallons
Orientation
Process Operations Data
Total Operation Time, hours
Total Volume Treated, gallons
Total Volume Treated, bed volumes
Process Flow Rate, gpm
Flow Rate Loading, bed volumes/hr
Hydraulic Loading, gpm/ft*
Empty Bed Contact Time, minutes
Column Skin Temperature, °F
Pressure Drop Across Bed, psi
Influent Characteristics
pH, standard units
Specific Conductance, fimhos/cm
VOC Concentrations, ng/L
Vinyl Chloride
1.1 -Dichloroethene
cis- 1,2 -Dichloroethene
trans- 1,2 Dichloroethene
Trichloroethene
Emuent Characteristics
pH, standard units
Specific Conductance, ^mhos/cm
Lead
A563B
4.0
22.0
1.20
up-flow
307
15,200
12,700
0.83
41
9.5
1.4
70
15.0
6.7
654
4.9
6.33t
353
122
4,510
6.7
654
Lag
A563A-
4.0
22.0
1.20
up-flow
307
15,200
12,700
0.83
41
9.5
1.4
70
8.0
6.7
654
3.1
0.1 Of
29
1
18
6.7
653
Series
A563B & A563A-
4.0
44.0
2.39
up-flow
307
15,200
6,370
0.83
20.8
9.5
2.9
70
23.0
6.7
654
4.9
6.33t
353
122
4,510
6.7
653
. Time weighted averages and cumulative totals for the total operating period.
t I.I-DCE concentrations estimated based on the mass r~~vcry results for the first steam regeneration of column A663B.
36
-------
During Cycle 2. the system was operated for 13 days at an average flow rate of 0.83 gpm
and treated a total of 15,200 gallons of groundwater. For the individual lead or lag columns, this
corresponds to operating at a 1.4-minute EBCT for a total 12,700 bed volumes. For the total
system in series, this corresponds to operating at a 2.9-minute EBCT for a total 6,370 bed
volumes.
Cycle 2 process operations data show that the average VOC concentrations in the influent
stream exceeded the MCL, except 1,1-DCE. In addition, the pH of the influent and effluent
streams for each column ranged from 6.2 to 7.6, and the average conductivity of the influent and
effluent streams was approximately 654 umhos/cm. No significant difference was observed
between the influent and effluent pH and conductivity of each column during Cycle 2.
Cycle 2 performance results, based on treatment to the MCL, are presented in Table 11.
The number of bed volumes treated to the MCL were determined by analysis of the VOC
breakthrough and leakage curves for the lead column. VOC breakthrough and leakage curves for
the lead column, representing a 1.4-minute EBCT, arc presented in Figures 8 and 9, respectively.
The VOC leakage curves for the lag Ambersorb adsorbent column, representing a 2.9-minute
EBCT, are presented in Figure 10.
Cycle 2 performance results show that both the lead and regenerated lag Ambersorb
adsorbent columns achieved effluent water quality below the MCL for each VOC. Specifically,
the lead Ambersorb adsorbent column treated approximately 8,320 bed volumes before the first
VOC (VC) broke through at a concentration above the MCL. During Cycle 2, concentrations of
1.1-DCE and trans-1 ,2-DCE in the effluent of the lead Ambersorb adsorbent column never
exceeded the MCL. Because virgin Ambersorb 563 adsorbent was loaded in the lead column and
the influent VOC concentrations were similar to those measured in Cycle 1, the bed volumes
treated to the MCL during Cycle 2 are similar to the Cycle 1 results.
Evcle
Cycle 3 was also conducted using two Ambersorb 563 adsorbent columns in series. The
lag Ambersorb adsorbent column (A563A-1) from Cycle 2 was placed in the lead position for
Cycle 3. The steam-regenerated lead Amber-sorb adsorbent column (A563B-1) from Cycle 2 was
placed in the lag position.
Cycle 3 process operations data, which are presented in Table 12, include the influent
average VOC concentrations measured over the total operating period. Because of the analytical
limitations discussed in Section 4, the influent average 1 ,1 -DCE concentrations were estimated
based on the mass of 1.1-DCE that was subsequently recovered during the third steam
regeneration.
37
-------
TABLE I 1. CYCLE 2 PERFORMANCE RESULTS
Volatile Organic Compound
Vinyl Chloride
1, 1 -Dichloroethene
cis- 1 ,2-Dichloroethene
trans- 1 ,2-Dichloroethene
Trichloroethene
MCL'
Hg/L
2
7
70
100
5
Bed Volumes Treated
to MCL
8,320
>12,700
10,600
>12,700
9,400
Maximum Contaminant Levels from National Revised Primary Drinking Water Regulations, 40 CFR 141.61
38
-------
o
o
o
o
§
o
o
o
o
S
I
o
[
«
c
v 4>
fti *3
U J= «
1 I i
•c § 3 -S
o s .a Q
J= S Q cs
- 5 ^ T"
.£ 3 A S
^ *•* o •£
j
•
*
u
>
g
D
0
j
C
•i^
i
»
1
i
i
if?
3
'i
m«
II
' H "
U
00
UJ
t
i
1
!
i
!
!
'
I
I
!
1
I
|
I
A
|
b
I
I
'
,
^
U
o
8
m
u
o
o
U
g
•o
£
o
I
E
(N
U
O
O
O
O
S
(N*
O
O
O
OO
S
o
OO
O
-------
20
18
16
14
12
§
i 10
c
u
u
d 8
"• Vinyl Chloride
-fr—- 1,1 -Dichloroethene
* cis-1,2-Dichloroethene
• trans-1,2-Dichloroethene
"* Trichloroethene
EBCT * 1.4 minutes
0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000 11,000 12,000 13,000
Bed Volumes
Figure 9. Cycle 2 Ambersorb 563 Adsorbent Lead Column VOC Leakage Curves
-------
«•££•
•QUO
•™ Q ^i
U
u
Q
o
--
A
ii
O;
-------
TABLE 12. CYCLE 3 PROCESS OPERATIONS DATA'
Column I.D.
Bed Geometry
Diameter, inches
Length, inches
Volume, gallons
Orientation
Process Operations Data
Cycle Operation Time, hours
Cycle Volume Treated, gallons
Cycle Volume Treated, bed volumes
P .:!oad Volume Treated, bed volumest
T. -al Volume Treated, bed volumes
Pr .ess Flow Rate, gpm
Flow Rate Loading, bed volumes/hr
Hydraulic Loading, gpm/ft1
Empty Bed Contact Time, minutes
Column Skin Temperature, °F
Pressure Drop Across Bed, psi
Influent Characteristics
pH, standard units
Specific Conductance, jimhos/cm
VOC Concentrations, ng/L
Vinyl Chloride
1,1 -Dichloroethene
cis-,l,2-Dichloroethene
trans- 1 ,2 Dichloroethene
Trichloroethene
Effluent Characteristics
pH, standard units
Specific Conductance, ^mhos/cm
Lead
A563A-I
4.0
22.0
1.20
up-flow
188
10,300
8.600
4,000
12,600
0.91
46
10.5
1.3
68
16.0
7.0
628
5.7
373
116
3.600
6.8
631
Lag
A563B-I
4.0
22.0
1.20
up-flow
188
10,300
8,600
_"
8,600
0.91
46
10.5
1.3
68
6.1
6.8
631
5.8
0.1 5t
70
5
157
6.9
624
Series
A563A- I & A563B- I
4. 0
44.0
2.39
up-flow
188
10,300
4,300
4,300
0.91
22.9
10.5
2.6
68
22.1
7.0
628
5.7
6.10-t
373
116
3,600
6.9
624
. Time weighted averages and cumulative totals for the total operating period.
t Preloaded volume treated based on Cycle 2 lead column VC leakage profile.
J I.I-DCE concentrations estimated based on the mass recovery results for the first steam regeneration of column A563A-I,
-------
During Cycle 3, the system was operated for approximately 8 days at an average flow rate
of 0.91 gpm and treated a total of 10,300 gallons of groundwater. For the individual lead or lag
columns, this corresponds to operating at a 1.3-minute EBCT for a total 8,600 bed volumes. For
the total system in series, this corresponds to operating at a 2.6-minute EBCT for a total 4.300
bed volumes.
As shown in Table 12, a preload volume of 4.000 bed volumes was added to the cycle
volume treated for the lead Ambersorb adsorbent column (A563A-1). The preload volume
accounts for the bed volumes of water treated during the previous cycle (Cycle 2) when A563A- 1
was in the lag position and was loaded with VOC leakage from the Cycle 2 lead column.
Cycle 3 process operations data show that the average VOC concentrations in the influent
stream exceeded the MCL, except for 1,1 DCE In addition, the pH of the influent and effluent
streams for each column ranged from 6.5 to 7.5, and the average conductivity of the influent and
effluent stteams ranged from 624 to 631 umhos/cm No significant difference was observed
between the influent and effluent pH and conductivity of each column during Cycle 3.
Cycle 3 performance results based on treatment to the MCL are presented in Table 13.
The number of bed volumes treated to the MCL was determined by analysis of the VOC
breakthrough and leakage curves for the lead column, which include the estimated preload
volume. VOC breakthrough and leakage curves for the lead column, representing a 1.3-minute
EBCT, are presented in Figures 11 and 12, respectively. The VOC leakage curves for the lag
Ambersorb adsorbent column, representing a 2.6-minute EBCT, are presented in Figure 13.
Cycle 3 performance results show that both the regenerated lead and regenerated lag
Ambersorb adsorbent columns achieved effluent water quality below the MCL for each VOC.
Specifically, the lead Ambersorb adsorbent column treated approximately 5,130 bed volumes
before the first VOC (VC) broke through at a concentration above the MCL. During Cycle 3,
concentrations of 1,1-DCE and trans-l,2-DCE in the effluent of the lead Ambersorb adsorbent
column never exceeded the MCL. The estimated average vinyl chloride concentration in the
influent increased from 4.9 ng/L during Cycle 2 to 5.7 ng/L during Cycle 3, which may have
decreased the number of bed volumes treated to the MCL during Cycle 3.
Cycle 4
Cycle 4 also was conducted using two Ambersorb adsorbent columns in series. The lag
Ambersorb adsorbent column (A563B-1) from Cycle 3 was placed in the lead position for
Cycle 4. The steam-regenerated lead Ambersorb adsorbent column (A563A-2) from Cycle 3 was
placed in the lag position.
43
-------
TABLE 13. CYCLE 3 PERFORMANCE RESULTS
MCL+ Bed Volumes Treated
Volatile Organic Compound ng/L toMCLt
Vinyl Chloride 2
1, I-Dichloroethene =*12.600
cis-1,2-Dichloroethene 70
trans-1 ,2-Dichloroethene 100 >12,600
Trichloroethene 5
• Maximum Contaminant Levels from National Revised Primary Drinking Water Regulations, 40 CFR 141.61
t Includes bed volumes preloaded during Cycle 2.
44
-------
R
450
400 -
350
300
* 250
g
200 -
150 -
100
50
0 -
•• Vinyl Chloride
•& 1, l-Dichloroethene
* cis- 1,2-Dichloroethene
• trans- i ,2-DichIoroethene
Trichloroethene
Preloaded Bed Volumes
7
0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000 1 1,000 12.000 13.000
Bed Volumes
Figure 1 I. Cycle 3 Ambersorb 563 Adsorbent Lead Column VOC Breakthrough Curves
-------
o
o
o
I 8
O r-
O
§
O
g
ts"
U
o
00
.
ra
u
c
J3
O
O
"8
U
I
vo
u*>
•e
o
E
U
-------
o
o
o
o
o
o
o
o
o
so*
3
u
*»
60
u
p
c
E
"5
U
*y is
E -3
1 1
T3
§ a
o
o
o
o
o
8
rl*
o
o
o
o
>>
O
Si
o
cs
1/Srf 'UOJ184U33U03
47
-------
Cycle 4 process operations data are presented in Table 14 and include the influent average
VOC concentrations measured over the total operating period. Because of the analytical
limitations discussed in Section 4. page 9, the influent average VC and 1,1-DCE concentrations
were estimated based on reanalysis of selected influent samples at lower dilutions.
During Cycle 4, the system was operated for approximately 13 days at an average flow
rate of 0.82 gpm and treated 15,300 gallons of groundwater. For the individual lead or lag
columns, this corresponds to operating at a 1.5-minute EBCT for a total 12,800 bed volumes. For
the total system in series, this corresponds to operating at a 2.9-minute EBCT for a total 6,390
bed volumes.
As shown in Table 14, a preload volume of 4,000 bed volumes was added to the cycle
volume treated for the lead Ambersorb adsorbent column (A563B-1). The preload volume
accounts for the bed volumes of water treated during the previous cycle (Cycle 3) when A563B- 1
was in the lag position and was loaded with VOC leakage from the Cycle 3 lead column.
Cycle 4 process operations data show that the average VOC concentrations in the influent
stream exceeded the MCL, except for 1 ,I-DCE and trans-1,2-DCE. In addition, the pH of the
influent and effluent streams for each column ranged from 6.9 to 8.0 and the average
conductivity of the influent and effluent streams was approximately 666 ^mhos/cm. No
significant difference was observed between the influent and effluent pH and conductivity of each
column during Cycle 4.
Cycle 4 performance results, based on treatment to the MCL, are presented in Table 15.
The number of bed volumes treated to the MCL was determined by analysis of the VOC
breakthrough and leakage curves for the lead column, which include the estimated preload
volume. VOC breakthrough and leakage curves for the lead column, representing a l.Sminute
EBCT, are presented in Figures 14 and 15, respectively. The VOC leakage curves for the lag
Ambersorb adsorbent column, representing a 2.9-minute EBCT, are presented in Figure 16.
Cycle 4 performance results show that both the regenerated lead and the twice-regenerated
lag Ambersorb adsorbent columns achieved effluent water quality below the MCL for each VOC.
Specifically, the lead Ambersorb adsorbent column treated approximately 5,010 bed volumes
before the first VOC (VC) broke through at a concentration above the MCL. Concentrations of
trans-l,2-DCE in the effluent of the lead Ambersorb adsorbent column never exceeded the MCL
during Cycle 4. Furthermore, the influent average VC concentration increased from 5.7 |jg/L
during Cycle 3 to 10 jag/L during Cycle 4, which may have decreased the number of bed
volumes treated to the MCL during Cycle 4. The leakage curve for the lag column in Cycle 4,
shown in Figure 16, indicates some leakage of VC above the MCL after 7,500 bed volumes.
This may be because the previous steam regeneration was performed at the lowest temperature.
48
-------
TABLE 14. CYCLE 4 PROCESS OPERATIONS DATA*
Column ID.
Bed Geometry
Diameter, inches
Length, inches
Volume, gallons
Orientation
Process Operations Data
Cycle Operation Time, hours
Cycle Volume Treated, gallons
Cycle Volume Treated, bed volumes
Preload Volume Treated, bed volumes!
Total Volume Treated, bed volumes
Process Flow Rate, gpm
Flow Rate Loading, bed volumes/hr
Hydraulic Loading, gpm/ft1
Empty Bed Contact Time, minutes
Column Skin Temperature, °F
Pressure Drop Across Bed, psi
Influent Characteristics
pH, standard units
Specific Conductance, nmhos/cm
VOC Concentrations, ng/L
Vinyl Chloride
1, 1-Dichloroethene
cis- 1 &Dichloroethene
trans- 1,2 Dichloroethene
Trichloroethene
Effluent Characteristics
pH, standard units
Specific Conductance, (imhos/cm
Lead
A563B- I
4.0
32.0
1.20
up-flow
311
15.300
12.800
4.000
16,800
0.82
41
9.4
1.5
68
14.0
7.7
666
10.1
0.13-t
350
85
3,920
7.7
666
Lag
A563A-2
4.0
22.0
1.20
up-flow
311
15,300
12,800
--
12,800
0.82
41
9.4
1.5
68
5.6
7.7
666
8.8
120
7
268
7.7
667
Series
A563B- I & A563A-2
4.0
44.0
2.39
up-flow
311
15,300
6,390
—
6,390
0.82
20.6
9.4
2.9
68
19.7
7.7
666
10.1
0.1 3f
350
85
3,920
7.7
667
• Time weighted averages and cumulative totals for the total operating period
t Preloaded volume treated based on Cycle 3 lead column VC leakage profile
JVC and 1,1-DCE concentrations estimated based on re-analysis of selected influent
at lower dilution.
49
-------
TABLE IS. CYCLE 4 PERFORMANCE RESULTS
MCL* Bed Volumes Treated
Volatile Organic Compound Hg/L toMCLt
Vinyl Chloride 2
,I-Dichloroethene 7
cis-1,2-Dichloroethene 70 11,140
trans-1,2-Dichloroethene 100 >16,800
Trichloroethene 5
• Maximum Contaminant Levels From National Revised Primary Drinking Water Regulations, 40 CFR 14 1.61
t Includes bed volumes preloaded during Cycle 3.
50
-------
U
«B
£
ffi
I
u
TJ
«a
c
U
X)
5
V)
•u
|
I
j>
u
U
*»'
o
o
o
o
-------
g
20
18 --
16
14 -
12
10
g
S «
6 -
Vinyl Chloride
1,1-Dichloroethene
cis-l,2-Dichloroethene
trans-1,2-DichIoroethene
Trichloroethene
EBCT = 1.5 minutes
Preloaded Bed Volumes
2,000
4,000 6,000
8,000 10,000 12,000 14,000
Bed Volumes
16,000 18,000
Figure 15. Cycle 4 Ambersorb 563 Adsorbent Lead Column VOC Leakage Curves
-------
g
i
Z.U
IX
1A
M_i
n
in
-
ft 4
/
— /
• Vinyl C
•— — O trans- 1 ,
EBCT -
2.9 minut
/"
hloride
hloroethci
Dichloroc
2-Dichlor
•oethene
es
ie
thene
oethene
^
**
/
-*
- -
r
\
^^^ r\
s'
^-^
w
s
•
\
/
A^.
0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000 11,000 12.000 13,000
Bed Volumes
Figure 16. Cycle 4 Ambersorb 563 Adsorbent Lag Column VOC Leakage Curves
-------
STEAM REGENERATION RESULTS
Steam regeneration was conducted on the Ambersorb 563 adsorbent column at the end
of Cycle 1 and on the lead Ambersorb adsorbent columns at the end of Cycles 2 and 3 to
evaluate the steam regeneration efficiency and the effect on subsequent Ambersorb adsorbent
performance. The steam regenerations were also conducted at various temperatures to evaluate
the effect of regeneration temperature on contaminant recovery. Process operations data for the
steam regenerations are presented in Table 16.
The condensate produced during each regeneration consisted of a visible and separable
concentrated organic phase and a WC-saturated aqueous phase. To ensure that there was no
VOC vapor discharge during each steam regeneration, a trap containing Ambersorb 563 adsorbent
was used on the vapor discharge from the condenser. The VOC mass recovery results reflect the
VOC levels measured for each phase (aqueous, organic, vapor).
The VOC mass recovery results, presented in the following subsections, were based on
the VOC mass adsorbed onto the lead Ambersorb adsorbent column during each service cycle
and the VOC mass recovered from each subsequent steam regeneration. The VOC mass
adsorbed onto the lead Ambersorb adsorbent column was calculated by integration of the
cumulative volume and VOC concentrations measured in the influent stream during each service
cycle. The VOC mass recovered during each subsequent steam regeneration was calculated by
integration of the cumulative volumes and VOC concentrations measured in each phase (aqueous,
organic, and vapor). Integration was conducted using the trapezoid rule. VOC concentrations
reported as less than the detection limit were assigned a zero value for purposes of integration.
Steam Regeneration 1
Steam Regeneration 1 was conducted on column A563A at an average temperature of
307 °F over a 17-hour period and generated approximately 9.1 gallons (7.6 bed volumes) of
condensate. Steam flow rates (as condensate) were increased incrementally over the operating
period from 0.23 BV/hr to 0.82 BV/hr, as shown in Table 16.
VOC mass recovery results for Steam Regeneration I (see Table 17) show individual
VOC mass recoveries for the 3 bed volumes of condensate and for the total bed volumes
of condensate produced. Table 17 also shows the VOC mass recoveries for each condensate
phase. Total VOC mass recovery profiles for Steam Regeneration 1 are presented in Figure 17.
The VC and 1.1-DCE mass recoveries were assumed to be 100% as the basis for estimating
Cycle 1 influent VC and 1,1-DCE concentrations.
Steam Regeneration 1 recovery results show that 73% of the total VOC mass was
recovered in the first 3 bed volumes and that 78% was recovered overall. Approximately 85%
of the total VOC mass recovered was collected in a separable organic phase.
54
-------
TABLE 16. STEAM REGENERATIONS PROCESS OPERATIONS DATA*
Steam Regeneration
Column I.D.
Bed Geometry
Diameter, inches
Length, inches
Volume, gallons
Orientation
Process Operations Data
Total Operation Time, hours
Total Volume Condensate Generated, gallons
Total Volume Condensate Generated, bed volumes
Column Temperature, °F
Steam Generator Pressure, psi
Column Inlet Pressure, psi
Condensate pH, standard units
Condensate Conductivity, umhos/cm
Steam Regeneration Flow Rate If
Steam Flow Rate as Condensate, BV/hr
Time at Reported Flow Rate, hours
Steam Regeneration Flow Rate 2f
Steam Flow Rate as Condensate, BV/hr
Time at Reported Flow Rate, hours
Steam Regeneration Flow Rate 3f
Steam Flow Rate as Condensate, BV/hr
Time at Reported Flow Rate, hours
Regeneration 1
A563A-0
4.0
22.0
1.20
down -flow
17.4
9.1
7.6
307
58
52
4.5
489
0.23
7.1
0.4 I
6.3
0.82
4.0
Regeneration 2
A563B-0
4.0
22.0
1. 20
down -flow
17.1
8.4
7.0
293
54
46
4.1
344
0.25
5.9
0.35
5.9
0.80
4.5
Regeneration 3
A563A-1
4.0
22.0
1. 20
down -flow
18.5
10.7
8.9
280
53
41
5.5
280
0.28
6.5
0.43
6.3
0.82
5.7
• Time weighted averages and cumulative totals for the total operating period.
t Average value for specified time interval.
-------
TABLE 17. STEAM REGENERATION 1 VOC MASS RECOVERY RESULTS
Volatile Organic Compound
Vinyl Chloride
I, l-Dichloroethene
cis-l,2-Dichloroethene
trans- 1,2-Dichloroethene
Trichloroethene
Total VOCs
After
3 Bed Volumes
Mass Recovery, %
Aqueous
Phase
27.9
0.0
12.3
8.1
7.1
7.4
Organic
Phase
0.0
0.0
54.5
68.2
66.2
65.6
Vapor
Phase
72.1
100.0*
1.3
2.3
0.2
0.3
Total
Phases
100.0*
100.0*
68.2
78.6
73.4
73.2
Fraction
in Organic
Phase, %
0.0
0.0
80.0
86.8
90. 1
89.5
After 7.6
Bed Volumes
(Total)
Mass Recovery, %
Aqueous
Phase
27.9
0.0
12.4
8.1
11.4
11.3
Organic
Phase
0.0
0.0
54.5
68.2
67.0
66.3
Vapor
Phase
72.1
100.0'
1.3
2.3
0.2
0.3
Total
Phases
100.0*
100.0*
68.2
78.6
78.5
78.0
Fraction
in Organic
Phase, %
0.0
0.0
80.0
86.8
85.3
85. I
*VC and I.I-DCE total recovery assumed to be 100% as basis for estimating Cycle I
VC and I ,1-DCE concentrations.
-------
e
v
TJ
e
o
u
ts
w
K>
R}
a.
o
a.
S3
O
V
3
O"
o
tx
£•
u
O
s
o
W5
1
*o
*o
05
eu
.H
*c
«
oo
6
««
1
c
o
1
1
so
{2
E
SB
55
r-C
«
W
§5
p
o
O
•x
O
o
•o
o
-------
Steam Regeneration 2
Steam Regeneration 2 was conducted on column A563B at an average temperature of
293 °F over a 17-hour period and generated approximately 8.4 gallons (7.0 bed volumes) of
condensate. Steam flow rates (as condensate) were increased incrementally over the operating
period from 0.25 BV/hr to 0.80 BV/hr, as shown in Table 16.
VOC mass recovery results for Steam Regeneration 2 (see Table 18) show individual
VOC mass recoveries for the first 3 bed volumes of condensate and for the total bed volumes
of condensate produced. Table 18 also shows the VOC mass recoveries for each condensate
phase. Total VOC mass recovery profiles for Steam Regeneration 2 are presented in Figure 18.
The 1.1 -DCE mass recovery was assumed to be 100% as the basis for estimating Cycle 2 influent
1.1 -DCE concentration.
Steam Regeneration 2 recovery results show that 7 1% of the total VOC mass was
recovered in the first 3 bed volumes and that 73% was recovered overall. Approximately 90%
of the total VOC mass recovered was collected in a separable organic phase.
Steam Regeneration 3
Steam Regeneration 3 was conducted on column A563A-1 at an average temperature of
280° F over a 19-hour period and generated approximately 10.7 gallons (8.9 bed volumes) of
condensate. Steam flow rates (as condensate) were increased incrementally over the operating
period from 0.28 BV/hr to 0.82 BV/hr, as shown in Table 16.
VOC mass recovery results for Steam Regeneration 3 (see Table 19) show individual
VOC mass recoveries for the first 3 bed volumes of condensate and for the total bed volumes
of condensate produced. Table 19 also shows the VOC mass recoveries for each condensate
phase. Total VOC mass recovery profiles for Steam Regeneration 3 are presented in Figure 19.
The 1.1-DCE mass recovery was assumed to be 100% as the basis for estimating the Cycle 3
influent 1 ,1-DCE concentration.
Steam Regeneration 3 recovery results show that 79% of the total VOC mass was
recovered in the first 3 bed volumes and that 87% was recovered overall. Approximately 80%
of the total VOC mass recovered was collected in a separable organic phase.
Summary of Steam Regeneration Results
Total VOC mass recovery results for the steam regenerations, as summarized in Table 20,
include the average pH measured for the condensate aqueous phase. Total VOC mass recovery
profiles for each steam regeneration, are presented in Figure 20. Condensate pH profiles for the
operation period of each steam regeneration are presented in Figure 21.
58
-------
TABLE 18. STEAM REGENERATION 2 VOC MASS RECOVERY RESULTS
Volatile Organic Compound
Vinyl Chloride
1,1-Dichloroethene
cis- 1,2-Dichloroethene
trans- 1 ,2-Dichloroethene
Trichloroethene
Total VOCs
After
3 Bed Volumes
Mass Recovery, %
Aqueous
Phase
15.6
0.0
15.3
7.0
4.3
5.1
Organic
Phase
0.0
99.0
81.8
64.4
64.2
65.4
Vapor
Phase
25.5
1.0
1.0
1.4
0.1
0.2
Total
Phases
41.1
100.0*
98. 1
72.8
68.7
70.7
Fraction
in Organic
Phase, %
0.0
99.0
83.4
88.5
93.6
92.5
After 7.6
Bed Volumes
(Total)
Mass Recovery, %
Aqueous
Phase
15.6
0.0
15.3
7.1
6.9
7.4
Organic
Phase
0.0
99.0
81.8
64.4
64.6
65.7
Vapor
Phase
25.5
I.O
1.0
1.4
0.2
8.2
Total
Phases
41.1
100.0*
98.2
72.9
71.6
73.4
Fraction
in Organic
Phase, %
8.0
99.0
83.3
88.4
90.2
89.5
I, I-DCE total recovery assumed to be 100% as basis for estimating Cycle 2 influent VC and I, concentrations
-------
c
M
"O
C
o
E
as
4>
E
V
CO
ra
0,
u.
O
CL
to
3
O
3
CT
«S
0.
C
ft
on
o
SO
0.
O
o
sx
O
u
u
oi
to
O
eo
w
Of
E
«
«
33
60
tZ
o
o
o
so
o
SO
o
60
-------
TABLE 19. STEAM REGENERATION 3 VOC MASS RECOVERY RESULTS
Volatile
Vinyl
Organic Compound
Chloride
1, I -Dichloroethene
cis- 1
trans-
2-Dichloroethene
I , 2-Dichloroethene
Trichloroethene
Total
VOCs
.Afier
3 Bed Volumes
Mass Recovery, %
Aqueous
Phase
14.1
29.2
23.1
u4.4
6.4
8.1
Organic
Phase
0.0
66.6
48.2
41.1
73.1
69.9
Vapor
Phase
28.0
4.2
2.4
3.6
0.8
1.1
Total
Phases
42. I
100.0*
73.7
59. I
80.3
79.1
Fraction
in Organic
Phase, %
0.0
66.6
65.4
59.6
91.0
88.4
After 7.6
Bed Volumes
(Total)
Mass Recovery, %
Aqueous
Phase
14.1
29.2
23.3
14.8
15.4
16.1
Organic
Phase
0.0
66.6
48.2
41.1
73.2
70.0
Vapor
Phase
28.0
4.2
2.4
3.6
0.9
1.2
Total
Phases
42.1
100.0'
73.9
59.5
89.5
87.2
Fraction
in Organic
Phase, %
0.0
66.6
65.2
69.
81.8
80.2
I ,l -DCE total recovery assumed 10 be 100% as basis for estimating Cycle 3 influent VC and 1.1-DCE concentrations
-------
c
V
o
en
£
m
o
T3
U
ca
o
D.
3
O
3
CT
a,
_o
'E
Si
ra
£
_
o
o
b~
a,
£•
*>
o
u
1)
o£
S
U
O
o
c
.2
E
v
c
U
M
v
o;
E
V
)H
3
o
3QA
62
-------
Avenge pH 5.5
I
Average pH 4 5
3456
Bed Volumes of Steam (as condensate)
=—regeneration 1 @ 307°F
Regeneration 2 @ 293° F
Regeneration 3 @ 280°F
Figure 20. Steam Regenerations Total VOC Mass Recovery Profiles
-------
Q.
7.5 n
7.0
6.5 *\—
6.0
5.5
5.0
4.5
4.0
3.5
3.0 -L
3 4 5
Bed Volumes of Steam (as condensate)
Regeneration I@307°F
Regeneration 2 @ 293°F
Regeneration 3 @280°F
Figure 2 1. Steam Regenerations Condensate Aqueous Phase pH Profiles
-------
The overall recovery results show that maximum recovery was achieved during Steam
Regeneration 3, which was operated at the lowest temperature (280 °F) and generated condensate
with the highest range of pH values (4.3 to 7.2) of all three steam regenerations. VOC mass
recovery decreased with decreasing condensate pH, based on the results in Table 20. An
explanation for the differences observed for the three different steam regenerations may be related
to the mechanism of dehydrohalogenation. Chlorinated organics under elevated temperatures may
dehydrohalogenate and thereby produce an acidic stream containing hydrochloric acid (HC1).
Therefore, the lower recoveries observed during the higher temperature regenerations may be due
to dehydrohalogenation resulting in a reduction of chlorinated organic concentration in the
condensate.
The incomplete mass recovery of VOCs may be due to the following:
• Volatilization of VOCs during sampling of the condensate aqueous and organic
phases.
• Inaccuracies during analysis of the steam regeneration samples.
• VOCs retained in the highest energy micropores of the Ambersorb adsorbent were
not removed during steam regeneration.
• Dehydrohalogenation of the chlorinated organics.
SUPERLOADING RESULTS
The superloading process operations data (see Table 21) include the influent average VOC
concentrations and effluent maximum VOC concentrations measured over the total operating
period for the superloading column.
During superloading, the virgin Ambersorb 563 adsorbent superloading column (A563S)
was operated for 1.8 hours at an average flow rate of 0.038 gpm (7.5~minute EBCT) and treated
4 gallons (14 bed volumes) of VOC-saturated condensate generated during Steam Regeneration 3.
The superloading process operations data show that the influent stream consisted of
73,000 pg/L cis-l,2-DCE, 7,500 pg/L trans-1,2-DCE, and 621,000 jig/L TCE. The pH of the
influent and effluent streams were 5.9 and 4.3, respectively. The average conductivity of the
influent and effluent streams were 286 and 485 pmhos/cm, respectively.
Superloading performance results based on treatment to the MCL are presented in
Table 22. Superloading VOC leakage curves are presented in Figure 22. The performance
results show that the Ambersorb 563 adsorbent superloading column treated 14 bed volumes of
VOC-saturated condensate (700,000 pg/L total VOC's to an effluent water quality below the
MCL for each VOC TCE was the only VOC detected in the effluent stream and was first
detected at a concentration of 2.5 pg/L after 14 bed volumes had been treated.
65
-------
TABLE 20. SUMMARY OF STEAM REGENERATIONS TOTAL VOC MASS RECOVERY RESULTS
Steam Regeneration
Column Temperature, °F
Total Bed Volumes Generated
Total VOC Mass Recovery @ 3 BV, %
Total VOC Mass Recovery @ End, %
Total VOC Fraction In Organic Phase @ End, %
Condensate Aqueous Phase pH
Regeneration
307
7.6
73.2
78.0
89.5
4.5
Regeneration 2
293
7.0
70.7
73.4
92.5
4.1
Regeneration 3
280
8.9
79. I
87.2
88.4
5.5
Os
-------
TABLE 21. SUPERLOADING PROCESS OPERATIONS DATA*
Ambersorb 563 Adsorbent
Column ID.
Bed Geometry
Diameter, inches
Length, inches
Volume, gallons
Orientation
Process Operations Data
Total Operation Time, hours
Total Volume Treated, gallons
Total Volume Treated, bed volumes
Process Flow Rate, gpm
Flow Rate Loading, bed volumes/hr
Hydraulic Loading, gpm/ftl
Empty Bed Contact Time, minutes
Pressure Drop Across Bed, psi
Influent Characteristics
pH, standard units
Specific Conductance, fimhos/cm
VOC Concentrations, ng/L
Vinyl Chloride
1,1 -Dichloroethene
cis- 1,2-Dichloroethene
trans- l,2_Dichloroethene
Trichloroethene
Effluent Characteristics
pH, standard units
Specific Conductance, nmhos/cm
A563S
2.0
21.0
0.29
up-flow
1.8
4.0
14.0
0.038
8.0
1.7
7.5
5.9
286
0
0
72,888
7,469
620,510
4.3
485
Time weighted averages and cumulative totals for the total operating period
67
-------
TABLE 22. SUPERLOADMG PERFORMANCE RESULTS
Volatile Organic Compound
Vinyl Chloride
-Dichloroethene
cis- 1,2-Dichloroethene
trans- 1,2-Dichloroethene
Trichloroethene
MCL*
HE/L
2
7
70
100
5
Bed Volumes Treated
to MCL
>14.0
> 14.0
> 14.0
> 14.0
> 14.0
: Maximum Contaminant Levels from National Revised Primary Drinking Water Regulations, 40 CFR 141.61
-------
o
u
o
U
LU
1R
1 C
M_
O
0
8
6
u li
— 0— trs
EBC
Infli
T = 7.5m
lent Aver
nyl Chloride
1-Dichloroethene
s- 1,2-Dichloroethene
ins- 1,2-Dichloroethene .
richlorocthene
inutes
age Total
VOC Cot
,_
icentratic
>n = 700.C
__
)00 ug/L
— -
^
... . ...
-"
-
-
10
11
34 6789
Bed Volumes
Figure 22. Ambersorb 563 Adsorbent Superloading Column VOC Leakage Curves
12' 13
14
-------
DATA QUALITY REVIEW
Overview
The overall QA objective for the project was to produce data of sufficient and known
quality to evaluate the effectiveness of Ambersorb 563 adsorbent technology for the treatment
of VOCs in groundwater. Specifically, the primary objective of the project was to demonstrate
that the effluent from the Ambersorb adsorbent columns contained VOCs at concentrations less
than the MCL.
A list of the target VOCs, along with their MCLs and the detection limits achieved for
the column influent and effluent samples, is provided in Table 23. The detection limits for the
target VOCs achieved in the effluent samples were below the MCL values and, as such, the data
were useable to evaluate Ambersorb adsorbent performance. The higher detection limits for the
target VOCs in the influent samples resulted from the sample dilutions required because of the
contaminant levels, primarily TCE, present in the groundwater.
Table 24 provides a summary of the sample analysis program for the service cycle,
regeneration, and superloading phases of the demonstration. A total of 404 field samples were
analyzed for target VOCs by Method SW846 8010. This equates to 2,020 data points that were
available to evaluate treatment performance during the project
Analytical data packages for the field samples as well as the QA samples are available
for review at EPA
Accuracy
The accuracy of the analytical data was monitored by the use of the matrix spike and
matrix spike duplicate samples. The percent recovery acceptance limits for each of the target
VOCs was established in the QAPP. Forty-four matrix spike/matrix spike duplicate samples,
representing 220 data points, were analyzed during the project. Table 25 provides a summary
of the accuracy data for the matrix spike/matrix duplicate samples.
Overall, 95% of the data points associated with spike samples were within the established
recovery range. Eleven data points fell outside the established recovery acceptance limits. TCE
was the compound that typically was outside the recovery acceptance limit. For these data
points, recoveries (i.e., 62% to 76%) were consistently below the lower end of the recovery range
of 77%.
70
-------
TABLE 23. TYPICAL DETECTION LIMITS FOR TARGET VOCs
Compound MCL Detection Limit (pg/L)
(Mg/L)
Influent Samples Effluent Samples
Vinyl Chloride 2 5 0.5
1,1-Dichloroethene 7 10 1
cis-1.2-Dichloroethene 70 5 0.5
trans- 1,2-Dichloroethene 100 5 0.5
Trichloroethene 5 5 0.5
71
-------
TABLE 24. SUMMARY OF SAMPLE ANALYSIS PROGRAM
Test Phase
cycle 1
Cycle 2
Cycle 3
Cycle 4
Regemration 1
Regeneration 2
Regeneration 3
Superloading
Total
Field Confirmatory
samples samples
159
42
33
45
44
21
42
8
404
11
4
3
5
1
0
1
I
26
Field
Duplicates
16
3
3
4
2
2
1
2
33
No.
Laboratory
Duplicates
1
0
2
0
0
0
0
0
3
of Samples
Matrix Matrix Spike
Spikes Duplicates
11
3
3
5
0
0
0
0
22
11
3
3
5
0
0
0
0
22
Field
Blanks
11
4
3
5
1
1
1
1
27
Trip
Blanks
11
5
3
5
1
1
J.
28
lustrument
Blanks
35
23
10
15
12
6
7
3
111
-------
TABLE 25. SUMMARY OF ACCURACY DATA FOR TARGET VOCs
Target VOCs
Test
Phase
Cycle 1
Cycle 2
Cycle 3
Cycle 4
Total
Project
Measurement
QAPP Acceptance Criteria (% Recovery)
Range (% Recovery)
Average (% Recovery)
Number of Samples
Number of Samples Meeting Criteria
Range (% Recovery)
Average (% Recovery)
Number of Samples
Number of Samples Meeting Criteria
Range (% Recovery)
Average (% Recovery)
Number of Samples
Number of Samples Meeting Criteria
Range (% Recovery)
Average (% Recovery)
Number of Samples
Number of Samples Meeting Criteria
Range (% Recovery)
Average (% Recovery)
Number of Samples
Number of Samples Meeting Criteria
Vinyl chloride
MS* MSB*'
59-141
66-99
81
11
11
75-89
81
3
3
65-81
72
3
3
62-97
75
5
5
62-99
78
22
22
60-1 17
80
11
11
79-92
86
3
3
71-84
78
3
3
66-93
76
5
5
60-1 17
80
22
22
1, 1-Dichloroethene cis-l,2-Dichloroethene
MS MSD MS MSD
63-137
79-109
94
11
11
101-122
111
3
3
99-1 16
107
3
3
73-121
92
5
5
73-122
98
22
22
64-139
77-125 77-1 17 69-116
95 97
11 11
11 11
98
11
11
106-1 18 100-1 12 107-1 10
110 106
3 3
3 3
108-1 15 87-12
111 100
3 3
3 3
108
3
3
87-123
103
3
3
71-1 10 73-123 67-122
89 100
5 5
5 5
96
5
5
71-125 73-123 67-123
98 99
22 22
22 22
100
22
22
trans- 1 ,2-Dichloroethene
MS MSD
64-139
69-121
90
11
11
85-1 14
101
3
3
97-1 19
106
3
3
78-1 10
92
5
5
69-121
94
22
22
68-1 14
89
11
11
90-111
100
3
3
100-1 19
112
3
3
80- 108
89
5
5
68-1 19
93
22
22
Trichloroethene
MS MSD
77-123
62-112
88
11
8
81-101
94
3
3
7/0-101
90
3
3
71-132
98
5
3
62-132
91
22
17
66-1 15
88
II
8
80-111
92
3
3
75-1 13
88
3
1
77-126
98
5
4
66-126
91
22
16
* MS - matrix spike.
**MSD matrix spike duplicate.
-------
Precision
The precision of the analytical data was assessed by the use of field duplicate and
laboratory duplicate samples. The acceptance criteria for precision data established in the QAPP
was a relative percent difference (RPD) value of s 50%. Thirty-three duplicate samples,
representing 165 data points, were analyzed during the project. Table 26 provides a summary
of the precision data for the duplicate samples. Overall. 95% of the data points associated with
duplicate samples met the established precision criteria.
Other Data Quality Measures
Surrogate Recovery--
Surrogate compounds were added to each sample to assess the efficiency of the analysis.
The percent acceptance limits for each of the surrogate compounds were established in the QAPP.
During the project, 673 samples were analyzed, which represented 2,019 surrogate recovery data
points. Overall, greater than 99% of the surrogate recovery data points were within the
established recovery range.
Confirmatory Samples--
Selected influent and effluent samples were analyzed for VQCs (i.e., full list) by Method
SW846 8260 for confirmation purposes. Twenty-six confirmatory samples were analyzed during
the project. A comparison of the confirmatory sample results with the associated field sample
results showed agreement in both the compounds detected and the measured concentrations for
detected compounds.
Blank Samples--
Field, trip, and instrument blank samples were incorporated in the project to provide field
and laboratory checks of data quality.
During the project, 27 field blanks, representing 135 data points, were analyzed. Overall,
95% of the field blank data points consisted of nondetectable values for the target VOCs. TCE
was the compound most frequently detected in the field blanks, typically at low ppb values.
Twenty-eight trip blanks, representing 140 data points, also were analyzed. Overall, 97%
of the trip blank data points consisted of nondetectable values for the target VOCs. TCE was
the compound detected in the trip blanks, typically at low ppb levels. TCE contamination in trip
blanks may be due to cross-contamination from the influent or steam regeneration samples
containing elevated levels of TCE. Trip blank contamination could have occurred during storage,
transport, or analysis.
-------
TABLE 26. SUMMARY OF PRECISION DATA FOR TARGET VOCs
Test
Phase
Cycle 1
Cycle 2
Cycle 3
Cycle 4
Measurement
QAPP Acceptance Criteria (% Recovery)
Average (% Recovery)
Number of Samples
Number of Samples Meeting Criteria
Range (% Recovery)
Average (% Recovery)
Number of Samples
Number of Samples Meeting Criteria
Range (% Recovery)
Average (% Recovery)
Number of Samples
Number of Samples Meeting Criteria
Range (% Recovery)
Average (% Recovery)
Number of Samples
Number of Samples Meeting Criteria
Vinyl chloride
£50
0-105
10
16
15
0-120
52
3
2
0-15
9
3
3
0-29
7
4
4
1 . I-Dichloroethene
£50
0
0
16
16
0
0
3
3
0
0
3
3
0-100
25
4
3
Target VOCs
cis-I ,2-Dichloroelhene
£50
0-42
5
16
16
0-15
6
3
3
0-5
3
3
3
0-5
3
4
4
trans- 1 ,2-Dichloroethene
£50
0-49
14
16
16
0-9
3
3
3
0-67
32
3
2
0-67
22
4
3
Trichloroelhene
£50
0-67
13
16
14
0-76
26
3
2
0-18
9
3
3
0-11
6
4
4
-------
TABLE 26. SUMMARY OF PRECISION DATA FOR TARGET VOCs
(Continued)
ON
Test
Phase
Regeneration
1
Regeneration
2
Regeneration
3
SUPER-
LOAD
Total
Project
Measurement
QAPP Acceptance Criteria (% Recovery)
Range (% Recovery)
Average (% Recovery)
Number of Samples
Number of Samples Meeting Criteria
Range (% Recovery)
Average (% Recovery)
Number of Samples
Number of Samples Meeting Criteria
Range (% Recovery)
Average (% Recovery)
Number of Samples
Number of Samples Meeting Criteria
Range (% Recovery)
Average (% Recovery)
Number of Samples
Number of Samples Meeting Criteria
Range (% Recovery)
Average (% Recovery)
Number of Samples
Number of Samples Meeting Criteria
Vinyl Chloride
£50
0-24
12
2
2
0
0
2
2
0
0
1
1
0
0
2
2
0-120
12
33
31
1,1 -Dichloroethene
£50
0-29
14
2
2
0
0
2
2
5
5
1
1
0
0
2
2
0-100
4
33
32
Target VOCs
cis- 1 .2-Dichloroethene
£50
5-20
12
2
2
7-23
15
2
2
0
0
1
1
0-8
4
2
2
0-42
6
33
33
trans- 1, 2-Dichloroethene
<50
17-24
20
2
2
15-22
19
2
2
2
2
1
0-15
8
2
2
0-67
16
33
Trichloroethene
£50
4-14
9
2
2
10-23
16
2
2
3
3
1
1
4-40
22
2
2
0-76
13
33
30
-------
During the project, 111 instrument blanks, representing 555 data points, were analyzed.
All of the instrument blank data points consisted of nondetectable values for the target VOCs.
Summary
The data quality review of the VOC analytical data for the Ambersorb adsorbent
demonstration project indicates that the acceptance criteria established in the QAPP were met on
a consistent basis. The QA objectives identified in the QAPP are presented in Table 27. As a
result, the overall completeness goal of 95% was achieved. The detection limits achieved and
the accuracy, precision, and internal quality control checks indicate that the field sampling and
laboratory analysis methods used throughout the course of the study generated data that was
representative, comparable, and of sufficient quality for its intended use in the Ambersorb
adsorbent technology demonstration project.
COMPARISON OF AMBERSORB ADSORBENT AND FILTRASORB GAC
PERFORMANCE
The performance results for Cycle 1, presented in Table 9, are a direct comparison of
Ambersorb 563 adsorbent and Filtrasorb 400 GAC performance. A direct comparison of
Ambersorb 563 adsorbent and Filtrasorb 400 GAC VC and TCE leakage curves is also presented
in Figure 23. The results for Cycle 1, as illustrated in Figure 23, show that, while operating at
approximately five times the flow rate (1/5 the EBCF), Ambersorb 563 adsorbent treated
approximately two to five times the bed volumes of groundwater as Filtrasorb 400 GAC.
Specifically, the bed volumes treated to the TCE MCL (5 jig/L) using Ambersorb 563 adsorbent
were 8,190, whereas the bed volumes treated using Filtrasorb 400 GAC were 4,850
(approximately two times the water). For the VC MCL (2 ng/L), Ambersorb 563 adsorbent
treated 8,120 BVs, whereas Filtrasorb 400 GAC treated 1,730 BVs (approximately five times the
water).
SUMMARY OF AMBERSORB ADSORBENT PERFORMANCE
Performance results for the four service cycles are summarized in Table 28. Virgin
Ambersorb adsorbent was used in Cycle 1, and the adsorbent used in Cycle 1 was regenerated
and used in Cycle 3 (Cycle 2 also used virgin Ambersorb adsorbent, and Cycle 4 used the
regenerated adsorbent). Therefore, Cycles 1 and 3 and Cycles 2 and 4 are grouped together in
the table to facilitate evaluating the effect of one steam regeneration on Ambersorb 563 adsorbent
performance. In addition, Table 28 presents the average influent VC concentrations for each
service cycle.
Performance results show there was a 37% to 40% decrease in bed volumes treated to the
VC MCL and a 22% to 37% decrease in bed volumes treated to the TCE MCL after the first
steam regeneration of Ambersorb 563 adsorbent. For the remaining VOC's however, there was
no consistent decrease in the capacity of Ambersorb 563 adsorbent after steam regeneration,
based on bed volumes treated to the MCL.
77
-------
TABLE 27. QA OBJECTIVES FOR PRECISION, ACCURACY, AND MDL FOR TARGET VOCS
Matrix
Groundwater (influent
and effluent)
Aqueous Phase,
Organic Phase
Method
SW846 8010
SW846 8010
Reporting
Units MI
Precision
)L (RPD)
Mg/L 1* *50
H^k 100 i50
Accuracy
(96 Recovery)
50-150t
NAS
Completeness
95
95
*Minimum MDL listed canbe attained for undiluted samples only. MDL will increase when dilutions are necessary
to meet instrument response linearity (i.e., likely for influent groundwater samples).
criteria for the critical VOCs in matrix spike samples are:
Svike Comvound
Vinyl Chloride
l,l-Dichloroethene
cis-l,2-Dichloroethene
Trans-l,2-Dichloroethene
Trichloroethene
Percent Recovery Accevtance Limits
59-141
63-137
64-139
64-139
77-123
|NA - Not applicable, matrix spike and matrix spike duplicates analyses were not performed for the aqueous and
organic phase samples from steam regeneration.
-------
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o"
o
o
o
o
o
o
o
o
o
o
o
u
<
o
o
o
"*f
JO
.k.
o
T3
e
CB
C
J3
w
O
«
"a
Xl
w
o
«
w
V
x>
E
o
M
o.
o
u
3
80
'UOIJBJJU33U03
79
-------
TABLE 28. SUMMARY OF AMBERSORB 563 ADSORBENT PERFORMANCE RESULTS
oo
O
MCL*
Mg/L
Column I.D.
Volatile Organic Compound
Vinyl Chloride 2
1 , 1 -Dichloroethene 7
cis- ,2-Dichloroethene 70
trans- 1 ,ZDichloroethene 100
Trichloroethene 5
Influent VC Concentration, ug/L
Bed Volumes Treated to MCL Changet Bed Volumes Treated to MCL Changet
Cycle 1 Cycle 3 J % Cycle 2 Cycle 4J %
A563A A563A-I A563B A563B-I
8,120 5.130 -37 8,320 5,010 -40
>13,700 > 12,600 —8 > 12,700 16,600 <31
9,690 8,810 -9 IO.600 11,140 5
>13,700 > 12,600 ...8 > 12.700 > 16,800 -32
8,190 5,160 -37 9,400 7,350 -22
3.4§ 5.7 4.9 10.l#
'National Revised Primary Drinking Water Standards Maximum Contaminant Levels (MCL). 40 CFR 141.61
t Change = (Performance Before Steam Regeneration - Performance AAer Steam Regenerationj/Performance Before Steam Regeneration • 100
} Includes bed volumes preloaded during previous cycle.
§ VC concentration estimated based on the mass recovery results for the first steam regeneration of column A663A.
K VC concentration estimated based on reanalysis of selected influent samples lower dilution.
-------
The reduction in bed volumes treated to the VC and TCE MCLs is partially attributed to
the increase in influent VC concentration during the study. Influent VC concentrations almost
doubled between each steam regeneration cycle. As predicted by the Rohm and Haas computer
model, small increases in influent VC concentration result in significant decreases in adsorption
capacity.
After the first regeneration, the adsorption capacity for most adsorbents, including GAC.
will be reduced. Additional steam regenerations and service cycles with constant influent VC
concentrations are needed to determine the long-term effect of multiple steam regenerations on
Ambersorb 563 adsorbent performance. Studies conducted by Rohm and Haas show that after
multiple service cycles of VOC loading and steam regeneration, Ambersorb 563 adsorbent does
not lose its adsorptive capacity. (Memorandum from D.N. Smith to S.G. Maroldo, "EDC Multi-
Cycling Summary Report," SR-94- 137. 29 April 1994.)
COMPARISON OF PREDICTED AND MEASURED PERFORMANCE RESULTS
Predicted and measured performance results for Filtrasorb 400 GAC and Ambersorb 563
adsorbent, based on bed volumes treated to VC MCL, are presented in Table 29. Table 29 also
presents the average influent VOC concentrations measured during each service cycle. The
results show that the breakthrough capacity model underestimated the number of bed volumes
actually treated to the VC MCL during Cycles 1 and 2 by 32% to 45% and overestimated the
number of bed volumes treated to the VC MCL during Cycles 3 and 4 by 8% to 12%. The bed
volumes treated by Filtrasorb 400 GAC during Cycle 1 and Ambersorb 563 adsorbent during
Cycles 1 and 2 may have been underestimated due to the use of estimated VC and 1,1-DCE
concentrations from the steam regeneration recoveries.
The model appears to be a useful tool in predicting service cycle time when actual
contaminant levels can be used as input. This emphasizes the importance of obtaining accurate
analyses for VC, 1 ,1-DCE, and other less strongly adsorbed contaminants, especially in the
presence of other high concentration VOCs. Based on these results, the Rohm and Haas
breakthrough capacity model is a useful tool in predicting adsorbent capacities and service cycle
times for bench- and pilot-scale column studies and for full-scale system design and cost analysis.
SCALEUP PARAMETERS
The information developed during the demonstration study enhanced the existing database
for the Ambersorb 563 adsorbent technology and helped validate process design parameters and
system performance for scale-up to full-scale treatment systems. The key process operating
parameters for the preliminary engineering design of an Ambersorb 563 adsorbent system are:
• Process configuration.
• EBCT or flow rate loading.
• Vessel configuration.
81
-------
TABLE 29. COMPARISON OF PREDICTED AND MEASURED PERFORMANCE RESULTS
oo
Service Cycle
Column ID.
Influent VOC Concentration, |ig/L
Vinyl Chloride
1, 1 -Dichloroethene
cis- 1 ,ZDichloroethene
trans- 1 ,2-Dichloroethene
Trichloroethene
Bed Volumes Treated to VC MCL
Predicted
Measured
Difference, %§
Cycle 1
F400
3.9*
0.31*
329
101
4,120
1,200
,730
44
Cycle 1
A563A
3.4*
0.31*
312
102
4,330
6,160
8,120
32
Cycle 3
A563A- 1
5.7
6.10*
373
H6
3,600
5,860
5,130 J
-12
Cycle 2
A563B
4.9
6.33'
353
122
4,510
5,740
8,320
45
Cycle 4
A563B-1
10.lt
0.1 3t
350
85
3,920
5,440
5,010 J
-8
• VC and/or 1,1-DCE concentrations estimated based on the mass recovery results for each subsequent steam regeneration.
f VC and 1,1-DCE concentrations estimated based on reanalysis of selected influent samples at lower dilution.
J Includes bed volumes preloaded during previous cycle.
§ Difference = (Measured BV - Predicted BV)/Predicted BV • 100
-------
• Steam regeneration conditions.
A full characterization of the contaminants in the influent, as well as the effluent
discharge limitations, are required to predict service cycle time for the Ambersorb adsorbent
system operations. The values provided below are typical and are for preliminary purposes only.
Design parameters for a full-scale system need to be derived specifically for that treatment
application.
Process Configuration
The decision to use a single column or two columns operating in series depends on a
number of factors, including the need for continuous operation, space constraints for downstream
regeneration equipment, effluent criteria, and service cycle time constraints or operation logistics.
Typically, the recommended process configuration consists of two columns operating in
series. Such a design offers the following advantages:
• The system can remain in operation at full flow while the lead column is being
regenerated.
• A lag column provides extra insurance that the effluent water quality will meet
extremely stringent effluent criteria.
• A lag column also allows higher utilization of adsorption capacity for the lead
column. In a single column mode, the vessel would have to be regenerated once
the effluent quality exceeded the effluent criteria. With two columns operating in
series, the lead column can operate to 50% stoichiometric breakpoint for the first
contaminant measured in the effluent (i.e., CJC0 = 0.5, where Ce is the effluent
concentration and C0 is the influent concentration). Depending on the contaminant
load, this could have a significant impact on operation by greatly increasing
service cycle time prior to regeneration.
• The downstream regeneration equipment is smaller for the regeneration of one
column at a time.
The field trial demonstrated the viability of operating in a lead/lag mode.
The impact of operating to a 50% stoichiometric breakthrough point, rather than when the
effluent quality exceeded the MCL, is clearly shown in the breakthrough profiles for Ambersorb
563 adsorbent for every service cycle. For instance, in Cycle 1, the actual bed volumes treated
to the vinyl chloride MCL break point was 8,120 bed volumes, whereas the unit was taken off
line for regeneration after 13,700 bed volumes when the vinyl chloride level was approximately
9 \ig/L. This operations approach allowed an additional 5,600 bed volumes to be treated prior
to regeneration.
83
-------
EBCT or Flow Rate Loading
Empty bed contact time or flow rate loading is used to estimate the volume of adsorbent
required. The recommended flow rate loading depends on the effluent criteria, service cycle time
constraints, pressure drop, or other site constraints.
Typically, an EBCT of 3.0 minutes is recommended for preliminary process designs. This
EBCT translates to a flow rate loading of 2.5 gpm/ft1 of adsorbent required.
The field trial results showed that even while operating at a short EBCT of 1.3 minutes,
the Ambersorb 563 adsorbent column could produce effluent water quality that was below the
MCL for each contaminant, The total system EBCT for the field trial ranged from 2.6 to 3.0
minutes. Multicycling performance, taking into account the changing concentrations of vinyl
chloride and 1,1-dichloroethene, showed no loss in ability to consistently produce effluent water
quality below the MCL for each contaminant. The demonstration study results reinforce the
recommendation of an EBCT of 3.0 minutes (flow rate loading of 2.5 gpm/ft3) as a conservative
starting point for estimating adsorbent requirements.
Vessel Configuration
The height to diameter ratio of the adsorber vessels is a function of flow distribution
requirements, pressure drop, or space constraints.
A minimum bed height of 2 to 3 feet is typically recommended for each adsorber vessel.
For treatment applications where the influent contains contaminants that are less strongly
adsorbed (such as vinyl chloride), bed depths of 4 to 6 feet may be advantageous. A deeper bed
provides a margin of safety by providing a larger treatment zone for the less strongly adsorbed
compounds. The deeper bed also enhances flow distribution and water contact within the
adsorption vessel.
Typically, the maximum hydraulic loading (i.e., linear flow rate) recommended for process
operation is 30 gpm/ft*. The hydraulic loading for the field trial ranged from 7.8 gpm/ft1 to 10.5
gpm/ft2.
The estimated pressure drop for a hydraulic loading of 10 gpm/ft2, based on the
Ambersorb adsorbent technical literature, is approximately 1.5 psi per foot of bed depth. The
field trial used a bed depth of 22.0 inches (1.83 feet), equating to an estimated pressure drop
across the bed of 2.7 psi. The actual pressure drop across the beds measured during the study
ranged from 8.4 to 16 psi. This higher pressure drop is attributed to accumulation of orange-
brown particulate matter (likely iron precipitates) at the influent screen of the column. The
presence of particulate matter did not, however, result in a negative impact on effluent water
quality or service cycle time.
84
-------
Steam Regeneration Conditions
The temperature/pressure, flow rate, and total volume of steam required for regeneration
of the Amber-sorb adsorbent is dictated by the contaminants present, effluent criteria, time
constraints, and space or manpower issues for the regeneration equipment.
Depending on the chlorinated organic contaminants present, a starting regeneration
temperature of approximately 300 °F (150 °C) is typically recommended. The field trial results
for the three steam regenerations were conducted at three different temperatures (307 °F, 293 °F,
and 280 °F). The results showed that the percent mass recovery and subsequent cycle
performance were not adversely affected at the lower steam regeneration temperatures.
The field trial results also clearly showed that the bulk of the mass desorbed during
regeneration occurred after the first 3 bed volumes of steam (as condensate). The steam flow
as condensate used during the regenerations for the demonstration project was incrementally
increased over a period of approximately 20 hours from approximately 0.25 BV/hr to 0.80 BV/hr.
85
-------
SECTION 7
CONCEPTUAL DESIGN AND PRELIMINARY COST ESTIMATE
The results of the Ambersorb adsorbent demonstration study were used to develop
conceptual designs and preliminary cost estimates for full-scale groundwater treatment systems
(average design flow of 100 gpm) using Ambersorb 563 adsorbent and GAC. The discharge
criteria for the effluent from the treatment systems were assumed to be drinking water standards
(i.e, MCL).
CONCEPTUAL DESIGN
Design parameters for the Ambersorb 563 adsorbent and GAC treatment systems are
presented in Table 30. Key design assumptions derived directly from the pilot study included:
• Average influent groundwater quality as measured during Cycle 1 (see Table 8).
• EBCT consistent with Cycle 1 (1.5 minutes for each Ambersorb 563 adsorbent
unit and 9.6 minutes for each GAC unit).
• Adsorbent performance as measured during Cycle 1 (see Table 9).
The designs for each treatment system also provide filters for the removal of particulate
matter upstream of the adsorbent columns.
A process flow diagram and major equipment list for the Ambersorb 563 adsorbent
treatment system are provided in Figure 24 and Table 31, respectively. The Ambersorb 563
adsorbent system is designed as an up-flow, fixed bed system, with two 660-lb adsorbent beds
in series, each having a 1.5-minute EBCT at 100 gpm. In addition, the Ambersorb 563 adsorbent
system includes on-line steam-regeneration and a condensate treatment superloading system. The
lead Ambersorb adsorbent bed is regenerated approximately every 8 days or 8,000 bed volumes.
A process flow diagram and major equipment list for the GAC treatment system are
provided in Figure 25 and Table 32, respectively. The GAC adsorbent system is designed with
four 1,800-lb adsorbent beds (two parallel systems of two GAC beds in series). Each GAC bed
has a 9.6-minute EBCT at 50 gpm. In addition, the GAC system uses commercially available
transportable GAC units, as manufactured by Carbtrol Corporation, that are replaced
approximately every 11 days or 1,600 bed volumes.
86
-------
TABLE 30. DESIGN PARAMETERS FOR IOO-GPM TREATMENT SYSTEMS
Ambersorb 563 Adsorbent
Design Parameter Treatment System GAC Treatment System
Number of Adsorbent Vessels 2 6 (2 spare)
Vessel Construction Stainless S teel Commercially Available Units
(Carbtrol Corporation)
Arrangement Series Two parallel systems of
two vessels in series
Orientation Upflow As designed
Bed Geometry Each Vessel)
Diameter, ft 2.0 3.8
Depth, ft 6.5 5.8
Area, ft2 3.1 11
Volume, ft* 20 64
Adsorbent Weight, Ib 660 1,800
Process Operations (Each Vessel)
Process Flow Rate, gpm 100 50
EBCT, minutes 1.5 9.6
Hydraulic Loading, gpm/ft2 32.3 4.5
Flow Rate Loading, gpm/ft3 5.0 0.78
BV Flow Rate, BV/hr 40 6.3
Volume Treated to Break- 8,000 1,600
through, BV
Time to Regeneration/ 8.3 10.6
Replacement, days
87
-------
Boiler
Feedwatcr
Groundwater
from Wells
00-gpm Average
System
Feed
Pumps
Equalization
Tank
Ambersorb
Adsorbent
Vessels*
Bag
Filters
oo
00
V
Cooling
Water
Treated Water
Discharge
Condensate
Transfer Tank
'Vessels configured in series. Lead/lag mode reversed
after every steam regeneration cycle.
100-gpm
Average
Ambcrsorb
Adsorbent
Supcrloudcr
I'hasc
Separation
Tank
Recovered
Organic
Phase HOT
Disposal
Figure 24. Process Flow Diagram for 100-gpm Ambersorb 563 Adsorbent Treatment System
-------
TABLE 31. MAJOR EQUIPMENT LIST FOR 100-GPM AMBERSORB 563
ADSORBENT TREATMENT SYSTEM
Item
Number
Function
Groundwater Equalization Tank
Feed Pump
Bag Filter
Adsorber*
Superloader Adsorber
Condenser
Condensate Transfer Tank
Condensate Transfer Pump
Phase Separation Tank
Aqueous Phase Transfer Pump
Steam Generator
Water Softener
2
2
1
Provide up to 30 minutes detention of influent
groundwater at plant throughput
Provide sufficient head to pump groundwater
through filters and adsorbs
Remove influent suspended solids
Remove VOCs from groundwater
Remove VOCs from aqueous phase of condensate
from steam regeneration
Condense steam regenerant containing desorbed
VOCS
Collect condensate for transfer to phase separator
rank or superloader
Provide sufficient head to pump condensate to
phase separation tank or aqueous phase to
superloader
Provide for phase separation of condensate from
steam regeneration
Provide sufficient head to pump aqueous phase to
condensate transfer tank
Provide sufficient steam at 300 °F for regeneration
of Ambersorb adsorbent
Treat boiler feedwater
*See Table 30 for details.
89
-------
U Si
o£
u a
< c.
c
rf
k
C
rf
k,
^
o ~
ni . i rf
%v
o ~
>
J
1
>
1
c
rC
i
C
-C
k
p
y «
oc!j
DJ
VI
w
w
v^
u
>lu
c
u
a
o
c.
so
S
S?
5
o
E
I
-------
TABLE 32. MAJOR EQUIPMENT LIST FOR 100-GPM GAC TREATMENT SYSTEM
Item Number Function
Groundwater Equalization Tank Provide up to 30 minutes detention of influent
groundwater at plant throughput
Feed Pump 2 Provide sufficient head to pump groundwater
through filters and adsorbers
Bag Filter 2 Remove influent suspended solids
Adsorber* 6 Remove VOGs from groundwater
*See Table 30 for details,
91
-------
PRELIMINARY COST ESTIMATE
Preliminary estimates (+30% to- 15%) for the total installed costs of the Ambersorb 563
adsorbent and GAC treatment systems are provided in Tables 33 and 34, respectively. The cost
estimates assume the construction location to be Pease AFB. New Hampshire.
The installed costs of the Ambersorb 563 adsorbent treatment system ($526100) are
significantly greater than the installed costs of the GAC treatment system ($336,800). The
installed costs of the Ambersorb 563 adsorbent system are greater for the following reasons:
• Additional costs for engineering and design of the adsorber vessels and steam
regeneration and superloading systems.
• Higher costs of materials compatible with the steam regeneration process (i.e.
stainless steel).
• The higher cost of Ambersorb 563 adsorbent media.
The lower installed costs for the GAC treatment system result primarily from the use of
commercially available predesigned units ($86,860 for six vessels) with off-site regeneration.
Such a system requires minimal engineering design and less costly construction materials.
Present worth cost estimates for the Ambersorb 563 adsorbent and GAC treatment systems
are provided in Tables 35 and 36, respectively. These estimates are based on a discount rate of
7.0%. Present worth costs are provided for 5, 10, 15, and 20 years of operation and include the
installed, operating, maintenance, and replacement costs, as well as salvage value for the
treatment system. The operating costs for the GAC system includes the routine replacement and
off-site regeneration of the GAC adsorber units.
The total present worth costs of the Ambersorb 563 adsorbent and GAC treatment systems
are plotted in Figure 26. This analysis indicates that, after approximately 2 years, the total
present worth cost of the Ambersorb 563 adsorbent treatment system is less than the GAC
treatment system. The reduced costs over time result from the significantly lower operating costs
for the Ambersorb 563 adsorbent system as compared to the GAC system.
92
-------
TABLE 33. INSTALLED COSTS FOR 100-GPM AMBERSORB 563 ADSORBENT
TREATMENT SYSTEM
Item
Nunter
Total Installed Cost
Major Process Equipment
Groundwater Equalization Tank
Feed Pump
Bag Filter
Adsorber*
Superloader Adsorber*
Condenser
Condensate Transfer Tank
Condensate Transfer Pump
Phase Separation Tank
Aqueous Phase Transfer Pump
Steam Generator
Water Softener
SUBTOTAL
Other Equipment
Control Building (Pre-Engineered Structure)
Effluent Flowmeter and Totalizer
Process Piping
Electrical
SUBTOTAL
TOTAL EQUIPMENT (ROUNDED)
Other Project Direct and Indirect Costs
Engineering and Design Fee
Project Construction and Facilities
Mobilization and Demobilization
Construe tion Equipment
Small Tools and Consumable Items
Permits and Fees
SUBTOTAL
10,030
8,620
28,780
85,600
12,670
6,200
5,230
1,980
6,540
2,060
22,400
3.420
193,530
28,000
4,500
37,700
14,710
84.910
278.400
60,000
22,300
3,100
8,400
3,400
101,400
*Includes cost of adsorbent media
(Continued)
93
-------
TABLE 33.
(Continued)
Item Number Total Installed Cost ($1
Project/Construction Contract Costs
General and Administrative Overhead Costs 36.100
Contractor Markup and Profit 41.600
SUBTOTAL 77,700
CONTINGENCY (15%) 68.600
TOTAL PROJECT COST 526,100
94
-------
TABLE 34. INSTALLED COSTS FOR IOO-GPM GAC TREATMENT SYSTEM
Item Number Total Installed Cost (S)
Major Process Equipment
Groundwater Equalization Tank 1 10,030
Feed Pump 2 8,620
Bag Filter 2 28,780
Adsorber* 4 61,600
Spare Adsorber* 2 25.260
SUBTOTAL 134,290
Other Equipment
Control Building @e-Engineered Structure) 1 28,ooo
Effluent Flowmeter and Totalizer 1
Process Piping
Electrical 14.710
SUBTOTAL 77,850
TOTAL EQUIPMENT (ROUNDED) 212,100
Other Project Direct and Indirect Costs
Project Construction and Facilities 17,000
Mobilization and Demobilization
Construction Equipment
Small Tools and Consumable Items
Permits and Fees 3,200
SUBTOTAL 31,100
Project/Construction Contract Costs
General and Administrative Overhead Costs 23,100
Contractor Markup and Profit 26.600
SUBTOTAL 49,700
CONTINGENCY (15%) 43.900
PROJECT TOTAL COST 336,800
"Includes cost of adsorbent media.
95
-------
TABLE 35. PRESENT WORTH COSTS FOR 100-GPM AMBERSORB 563 ADSORBENT TREATMENT SYSTEM*
Cost Category
Installed Costs
Operating Costs
Maintenance Costs
Replacement Costs
Salvage Value
Total Present Worth
5-Year
$526,100
$104,391
$75,854
$0
($17,825)
$688,520
1 0-Year
$526,100
$ 178,820
S 129,936
$7,005
($10,167)
$83 1,695
16-Year
$526, 100
$23 1,887
$ 168,496
$12,000
($5,437)
$933,047
20-Year
$526.100
$269,724
$ 195,989
$15,561
($2,584)
$1,004,789 1
• Estimate based on discount rate of 7.0%.
-------
TABLE 36. PRESENT WORTH COSTS FOR 100-GPM GAC TREATMENT SYSTEM+
•4,1
Cost Category
Installed Costs
Operating Costs
Maintenance Costs
Replacement Costs
Salvage Value
Total Present Worth
5-Year
$336,800
$556,684
$75,854
$0
($3,565)
$965,773
lll-Y ear
$336,800
$953,592
$129,936
$7,005
($2,033)
S 1 ,425,300
15-Year
$336,800
$ 1,236,58 I
$168,496
$12,000
($1,087)
$ ,752,790
20-Year
$336,800
S 1 ,438,349
$ 195,989
$15,561
($5 17)
$ 1 ,986, 1 82
. Estimate based on discount rate of
-------
£ 8
(S
V
c
u
U
<
O
§
!i
o
o
I i
i 8
x i
O
I
98
-------
REFERENCES
WESTON (Roy F. Weston, Inc.). 20 January 1994. Work Plan for Demonstration of
Ambersorb' 563 Adsorbent Technology. West Chester, PA.
2. WESTON (Roy F. Weston, Inc.). 30 June 1994. Quality Assurance Project Plan for
Demonstration of Ambersorb" 563 Adsorbent Technology. West Chester, PA.
3. Clark, R.M., C.A Fronk, and B. W. Lykins, Jr. 1988. Environ. Sci. Technol. 22(10): 1126-
1130.
4. "Groundwater Treatment Know-How Comes of Age." Chemical Week 20 May 1987. pp.
50-51.
5. Neely, J.W. and E.G. Lsacoff. 1982. Carbonaceous Adsorbents for the Treatment of Ground
and Surface Waters. Marcel Dekker, Inc., New York, NY.
6. Kollman, C.J. 1977. U.S. Patent No. 4,064,043. "Liquid Phase Adsorption Using Partially
Pyrolyzed Polymer Particles."
7. U.S. Patent No. 4,839,331. Dated 13 June 1989.
8. U.S. Patent No. 4,957,897. Dated 18 September 1990.
9. U.S. Patent No. 5104,530. Dated 14 April 1992.
10. U.S. Patent No. 5,094,754. Dated 10 March 1992.
11. Rohm and Haas Company. 1992. Ambersorb* Carbonaceous Adsorbents Technical Notes,
30G03, August, 1992.
12. Vandiver, M. and E. G. Isacoff. 1994. "THM Reductions with Ambersorb 563 Adsorbent."
Paper presented at 41st Annual Conference, Society of Soft Drink Technologists,
Albuquerque, New Mexico. 19 April 1994.
13. Parker, Jr., G.R. 1992. "Comparison of Amber-sorb 563 Carbonaceous Adsorbent and
Granular Activated Carbon for the Removal of TCE from Water at Short Empty Bed Contact
Times. " Paper presented at American Institute of Chemical Engineers Annual Conference,
Miami Beach, Florida. 4 November 1992.
99
-------
14. Isacoff, E.G., S. M. Bortko and G. R. Parker, Jr. "The Removal of Regulated Compounds
from Groundwater and Wastewater Using Ambersorb 563 Carbonaceous Adsorbent. Paper
presented at American Institute of Chemical Engineers Conference. Miami Beach, Florida.
3 November 1992.
15. Parker, Jr., GR. and S. M. Bortko. 1991. "Groundwater Remediation Using Ambersorb
Adsorbents." Paper presented at the Florida Environmental Chemistry Conference. 30
October - 1 November 1991.
16. Speth, T. F. and R. J. Miltner. 1990. "Technical Note: Adsorption Capacity of GAC for
Synthetic Organics," Journal of 'American Water Works Association, 82: 72-75.
17. Dubinin, M. M. and V. A. Astakov, 1971. Izv. Akad. Nauk SSSR, SER KfflM. 71( 197 1).
page 5.
18. G.R. Parker. 1995. "Optimum Isotherm Equation and Thermodynamic Interpretation for
Aqueous 1,1,2-Trichloroethene Adsorption Isotherms on Three Adsorbents." Adsorption, 1
(2): in print.
100
------- |