PB87-113783
Preliminary Assessment of Air Emissions from
Aerated Waste Treatment Systems at Hazardous
rfaste Treatment Storage and Disposal Facilities
Research Triangle Inst.
Research Triangle Park," NC
Prepared for
Environmental Protection Agency, Cincinnati, OH
Oct 86
-------
EPA/600/2-86/098
October 1986
PRELIMINARY ASSESSMENT OF AIR EMISSIONS FROM AERATED
WASTE TREATMENT SYSTEMS AT HAZARDOUS WASTE TREATMENT
STORAGE AND DISPOSAL FACILITIES
C. C. Allen, D. A. Green
J. B. White and J. B. Coburn
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, NC 27709
Contract No.: 68-02-3992
Project Officer
Ronald J. Turner
Thermal Destruction Branch
Alternative Technologies Division
HAZARDOUS WASTE ENGINEERING RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
TECHNICAL REPORT DATA
(Plmt rttd Itrmtciicxi on tkt mme tr
NO.
EPA/60JO/2T86/Q38
PRELIMINARY ASSESSMENT OF AIR EMISSIONS FROM AERATED
WASTE TREATMENT SYSTEMS AT HAZARDOUS WASTE TREATMENT
STORAGE ANT) DISPOSAL FACILITIES
O»!t
October 1986
u OMbAMlZATlO* CODl
HWERL
1 Pt«'OliMiMC OBCAMiZAiiOM
C. C.'Allen, et.al
OMCAMIZATIOM MAMf
1C
MtNT NO
Research Triangle Institute
Research Triangle Park, North Carolina
D109
27711
63-02-3992/68-03-3149-
12 f»D«»O"l*6 ACINCT MAMt
Hazardous W?ste tngir-een'ng Research Laboratory
Off ice-of Researcli and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
u TYPE or niroKT AMD »«».oc
Final report
* ft»ON»O*l*»C AGf *C* CCOI
EPA/600/12
It
Methods for estimating emissions resulting from mass transfer of volatile
organic compounds from dilute wastewaters into air are reviewed and applied to
full-scale and pilot-scale treatment systems.
«»O*Dt
Jl MO O> »»il
t. Dl*TKI»iSTlO* S1ATCMINT
Release
Unclassif "i
to HCW«i1» Ct-ASt
•' e i
*. 4.77)
-------
NOTICE
The information in this document has been funded wholly or in part by the
United States Environmental Protection Agency under contracts 68-02-3992 and
68-03-3149 to Research Triangle Institute. It has been subject 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.
ii
-------
FOREWORD
Today's rapidly developing and changing technologies and industrial
products and practices frequently carry with them the increased generation
of solid and hazardous wastes. These materials, if improperly dealt with,
can threaten both public health and the environment. Abandoned waste sites
and accidental releases of toxic and hazardous substances to the environment
also have important environmental and public health implications. The Hazardous
Waste Engineering Research Laboratory assists in providing an authoritative and
defensible engineering basis for assessing and solving these problems. Its
products, support the policies, programs and regulations of the Environmental
Protection Agency, the permitting and other responsibilities of State and local
governments and the needs of both large and small businesses in handling their
wastes responsibly and economically.
This report presents and evaluates kinetic and predictive methods for
estimating volatile organic compound emissions from aerated wastewater treat-
ment processes. The available theoretical and semiempirical models were used
to predict emission'- within limits of + 20 percent in the absence of competing
removal mechanisms. Five general treatment plant configurations are described
in this study: trickling filtration, activated sludge (mechanical aeration),
activated sludge (diffused aeration), activated sludge-aerated lagoons (mechanical
aeration), spray ponds (mechanical spray). The results should be of interest to
regulatory agencies, industrial, and other sources involved in VOC emissions
control. / .
Requests for further information regarding this study should be directed
to the Project Officer at the Hazardous Waste Engineering Research Laboratory,
Cincinnati.
Thomas R. Hauser
Director
Hazardous Waste Engineering Research Laboratory
Cincinnati, Ohio
iii
-------
ABSTRACT
Aerated wastewater treatment unit operations are used for the removal of
organic compounds from hazardous waste and other industrial wastewater streams.
In some operations, aeration is required to supply oxygen for aerobic decompo-
sition of organics; in other operations, air/water contact occurs as the wind
blows across the surface of the ponded or flowing wastewater. Methods for esti-
mating emissions resulting from air stripping of volatile organic compounds
that can accompany aerated treatment are reviewed and applied to full seal"
and pilot plant wastewater treatment plants.
Typical treatment plants which may be used for aerated treatment of indus-
trial waste are provided for systems employing trickling filters, activated
sluJge and aerated lagoons. The size and configuration of actual treatment
plants are highly variable, corresponding to the highly variable volume and
strength of industrial wastewater. Typical size and residence time specifica-
tions for unit operations within these plants are included.
The recommended mathematical models are used to generate predictions of
the fate of volatile organic compounds in wastewater treatment systems. Where
full scale plant data are available, predictions generally agree with measure-
ments (within the limits of accuracy which result from variations in sampling
and chemical analysis). For this reason, the mathematical models can be used
to estimate emissions for those sy.-tems whers no field data are available.
This report was submitted in fulfillment of Contract 68-03-3149, Work
Assignment 25-2, and Contract 68-02-3992, Work Assignment 22, by Research
Triangle Institute under sponsorship of the U.S. Environmental Protection
Agency. This report covers the period from December 10, 1984 to May 31, 1985
and work was completed as of October 31, 1985.
iv
-------
. CONTENTS
Number
1 Introduction and Objectives 1
2 Summary and Conclusions 4
3 Description of Aerated Wastewater Treatment
Processes 6
Introduction 6
Description of Industrial Processes 10
Petroleum Refining 11
Organic Chemical Industry 12
Iron and Steel Industry 14
Textile Industry. . 14
Pulp and Paper Industry 16
Model Treatment Plants 19
Model Plant //I: Trickling Filtration 22
Conventional Activated Sludge System 25
Model Plant #2: Conventional Activated Sludge
(Mechanical Aeration) 28
Model Plant #3: Conventional Activated Sludge
(Diffused Air: Coarse and Fine Bubble) 29
Model Plant //4: Aerated Lagoons (Mechanical
Air) 31
Model Plant //5: Spray Evaporation Ponds 31
Physical-Chemical Treatment System 31
Dissolved Air Flotation (DAF) 36
Neutralization (Equalization) Process 36
Description of Removal Mechanisms 38
Background 38
Biological Oxidation 38
Classical Approach.. 39
Rates from the EPA Data Base 40
-------
CONTENTS (continued)
Number Page
. Mass Transfer in Aerated Wast.ewater Treatment Systems..... . 45
Available Equations 45
Liquid Mass Transfer 46
Mass Transfer from Falling Droplets 50
Thin Falling Films 53
Gas Mass Transfer 55
Subsurface Aeration 58
Adsorption 62
Predictive Fate Models 63
5 Mathematical Models For Aerated Waste Treatment 70
Selection of Models 70
Pretreatment 73
Clarif iers 74
Surface Aeration.. 75
Secondary Clarifier 78
Trickling Filter 78
Subsurface Aeration 80
Equalization 81
Storage Tanks 81
Dissolved Air Flotation 82
Spray Ponds 82
Cooling Towers 83
A Comparison of the Predictions of the Mathematical
Models with Reported VOC Losses from Aeration
Processes 83
Parameter Specification 84
Results of Calculations 89
Discussion of the Model Applicability.... 89
6 Nomenclature 101
7 References 104
vi
-------
FIGURES
Number Page
1 Model trickling filter plant 24
2 Conventional activated sludge with mechanical aeration.... 27
3 Conventional activated sludge with diffused aeration
model plant 32
4 Aerated lagoon model plant 34
5 ^ Flow scheme for Union Carbide plant 86
6 WERL pilot plant flow scheme 91
\
-------
TABLES
Number . Page
1 Organic Priority Pollutants Detected in Refinery
Discharges at More Than One Location 13
2 Toxic Organic Concentrations Achievable by Two Stage
Activated Sludge Systems 15
3 Priority Pollutants Detected in Raw Wastewater from
Textile Manufacturing at More Than One Location 17
4 Organic Compounds Present in Wastewater from the
Pulp, Paper, and Paperboard Industry 18
5 Generators and TSD by Industry Type 20
6 Model Plant Operational Parameters - Trickling Filter 26
7 Model Plant #2 Conventional Activated Sludge with
Mechanical Aeration. Operational Parameters 30
8 Model Plant //3 Conventional Activated Sludge with
Diffused Aeration. Operational Parameters 33
9 Model Plant #4 Aerated Lagoon. Operational Parameters.... 35
10 Model Plant #5 Evaporation Pond. Operational Parameters.. 35
11 Operational Parameters - Dissolved Air Flotation 37
12 Biodegradation Data from Fitter 41
13 Bioxidation Rates from EPA Data Base 43
14 Constants for Drag Coefficient Calculation 52
15 Fraction of Benzene Remaining in a Droplet Falling
at Terminal Velocity 54
16 Total Distance Fall Before Evaporation to Dryness 54
-------
TABLES (continued)
Number Page
17 A Comparison of Selected Predictions of Mass
Transfer Coefficients for Benzene Emissions From
a Clarifier (Model Plant 1) 57
18 Concentration of Benzene Absorbed During Bubble
Rise in Water (Mole Fraction) 61
19 Clusters: Examples of Their Compounds <.nd Fate
Description 64
20 Cluster Characteristics Used as .3 Basis in Grouping
Compounds 65
21 Equations for Estimating Fate of 'Organic Priority
Pollutant Compounds 66
22 Predicted Fate of Organic Compounds According to
Cluster Number 68
23 Model Data Requirements 72
24 Model Input Data for Union Carbide Plant 85
25 Physical Property Data Used in Mathematical Model 87
26 A Comparison of Measured and Theoretical Partition
Coefficients 88
27 A Comparison of the Experimental Loss of Volatiles
from the Union Carbide Plant Field Test to the
Predicted Loss from Mathematical Models 90
28 -A Comparison of the Experimental Loss of Volatiles
From a WERL Pilot Study to the Predicted Loss From
Mathematical Models 92
29 Model Input Data for WERL Pilot Plant Study 93
30 A Comparison of the Experimental and Predicted
Loss of Volatiles From a WERL Pilot Facility 94
-------
TABLES (continued)
Number . Page
31 Model Input Data for WERL Pilot Study 95
32 A Comparison of the Reported Fractional Loss of
Components From Primary Wastewater Treatment and the
Predicted Losses, Willow Island Plant 96
33 Model Input Data For Willow Island Wastewater Study 96
34 A Comparison of th
-------
SECTION' 1
IKTKODUCTIOS AND OBJECTIVES
Undev the 1976 Resource Conservation and Recovery Act (RCRA) the EPA
Office ot Air Quality Planning and Standards (OAQPS) is developing regulations
to control air emulsions iron haurdous waste treatment, storage, and disposal
facilities (TSDF). The purpose of the air emissions regulations is to protect
human health and the emirou-vent (roa eeistaoas of volatile organic compounds
(V'OCr.), particulars, and aerosols.
The sources of TSDF emissions include storage tattkti, treatment processes,
surface impoundments, lagoons, lanJiilts, land treatment, and drua storage and
handling facilities. Those processes involving the* utie of air (biological
treatment and cooling) or subject to the introduction of air (stirred equali-
zation aad neutralization} say emit VOC a* a consequence of air stripping.
In order to further understand and better eeti&dtir the source and extent
of VOC emissions froa TSDFft, tooUel Aerated treatment facilities are investi-
gated. Various aacheaatical »xlt:ls are presented tu predict the role and extent of
VOC eraisulons during each of tt«s different process conditions oncounCerod. Tho
models, include a tsethodology for csti»a'ing the* relative ieport.ance of cctapet-
ing removal pathways (i.e., adsorption and biological oxidation). The study
is limited to those biological processes and physical chenical processes with
potential to vert VOC to the air.
The basic problea in evaluating air emissions froa Aerated waste treat-
ment processes is to deteraine the importance of coapetiag mechanises. In an
activated sludge biological treatment process, for exasple, the aajor tischan-
isms for renoval of dissolved contaeajnants frca the aqueous waste are biologi-
cal oxidation, adsorption on bioasass, and Bass transfer into the air. In a
rough analysis of "removal efficiency'', concentrations of the contaainant in
the aqueous influent and effluent are eeasured and the fraction of the contam-
inant which disappears in the process is reported as an efficiency. This
procedure gives no information about the relative itsportance of coapcttng
removal Hsechanises.
Where on* or nore of tiie reraoval etechanissk is destructive of the conta«-
inant, e.g., cheoical or biological oxidation, and other eech^nisas, e.g., air
-------
stripping or aerosol dispersal, are nondestructive, measurements of the relative
importance of the competing mechanisms become complicated. In such a case,
the contaminant removed by the non-destructive mechanism trust be recovered and
the balance assumed to have been destroyed. Errors in chemical analyses, and
particularly sampling errors, will have a great impact on calculations of the
fraction of the contaminant which has been destroyed.
Alternatively, data may be obtained under controlled laboratory conditions
and adapted to provide reasonable estimates of emissions expected under field
conditions. One way to do this is to conduct experiments in covered vessels
to permit accurate sampling of vent gases and accurate metering of .vent gas
flow rates. Another approach involves inhibiting the destructive removal
mechanisms by taking steps to chemically sterilize the system or to substitute
an inert gas such as nitrogen for air. In the latter approach, the mass
transfer to the gas phase will continue to occur, eliminating all of the
destructive mechanisms. The elimination of the destructive mechanisms may
cause a higher concentration and inflate mass transfer. This effect can be
accounted for in the data analysis. Influent and effluent aqueous phase
sampling can be used to determine the extent of removal resulting from stripping.
Vapor phase sampling of such a laboratory system should close the mass balance
for the system. The extent to which under-recovery or over-recovery of contami-
nants is found can be used to evaluate the adequacy of laboratory vapor phase
sampling techniques.
When adsorption is unimportant (typically true with VOCs), then any
difference between influent and effluent contaminant levels may be attributed
to air stripping and biological oxidation. Thus, if a rate expression describ-
ing mass transfer through air stripping is available, it can be used to calcu-
late contaminant levels which are susceptible to destructive removal mechanisms.
Conversely, if rate data for the destructive removal mechanisms (e.g., biolog-
ical oxidation kinetics) are available, these can be used to estimate the
relative importance of air stripping.
One approach to the problem described above is to measure the total
disappearance in the field and then apply mathematical models obtained from
laboratory data to simultaneously estimate the rate of stripping and biological
oxidation. When these models account for significantly more or significantly
less contaminant removal than the observed total removal, several steps can be
-------
taken. The simplest method is to assume that the mass transfer model is more
accurate than the biological oxidation mcHel and account for biological oxida-
tion (and other destructive removal mechanisms) by difference. The mass
transfer models are in fact more accurate predictors than the biological
oxidation models because they are less affected by changes in operating condi-
tions and influent wastewater composition and thus are less site-specific. In
some cases, air sampling programs must be conducted to verify the prediction
of stripping models. The interpretation of data obtained under field conditions
is subject to considerable error.
The objectives of this study are to present and evaluate the various
kinetic and predictive fate models that are applicable to aerated wactewater
treatment systems. Four different waste treatment systems are con^^iered.
Thes3 are (1) trickling filters, (2) activated sludge, (3) aerated lagoons,
and (A) sprsy ponds. When possible, model predictions are compared with
experimental data.
The approach taken is to investigate mathematical models and correlations
available in the literature for mass transfer in systems similar to wastewater
treatment units. Classical engineering approaches to mass transfer in strippers
and absorbers, packed bed design, and stream and lake reaeration are considered
and adapted as necessary to wastewater treatment systems.
-------
SECTION 2
SUMMARY AND CONCLUSIONS
The available theoretical and seroiempiricai methods for predicting the
rates of mass transfer of volatile organic compounds from dilute solution into
w
air can be used to accurately (± 20%) predict emissions frc;n aerated wastewater
treatment processes in the absence of competing removal mechanisms. Where
liquid phase concentration data and system operating conditions are well
characterized, VOC emissions from aerated lagoons, activated sludge processes,
trickling filters, and clarifiers associated with wastewater treatment systems
can be estimated within the accuracy of sampling and chemical analysis results.
The presence of competing removal mechanisms, the most important of which
is bioxidation, reduces the accuracy of emissions predictions based on plaat.
influent concentration data because direct mass balances cannot be obtained.
If bulk wastewater concentrations are known for individual treatment units,
then reasonably accurate (± 20%) predictions of the fate of VOCs can be made.
However, when only the concentration of the influent to the entire plant is
known, accurate rate data for biooxidation becomes more important and the
absence of this data has a greater influence on the accuracy of emissions
estimation (uncertainties of ± 50% may result).
Biooxidation rate data is available in the literature for various organic
compobii'U ',/hich may be present in industrial wastewater. The data are highly
dependent on the experimental conditions under which they are obtained and are
not easily adapted to real systems which may have different initial concentra-
tions, biomass loadings, nutrient and inhibitor concentrations, etc. A table
of bioxidation rate data obtained from the literature has been included in
Section 4 to provide a starting point for calculations.
Industrial waste treatment systems vary widely in design, reflecting the
wide variation in the waste streams to be treated. The recommended predictive
mathematical models are adaptable to widely varying volumetric flow rates.
Residence time, which in actual process system designs accounts for the concen-
tration and increases with the difficulty of removal of organic compounds and
suspended solids, must be specified from actual system designs or estimated
(less accurately) from removal rate data. Model plant designs have been
developed in this study which fit within the wide range of suitable designs.
-------
The best mathematical models, as determined in this study, have been compared
with experimental measurements of full-scale and pilot-scale wastewater treat-
ment systems. Sample calculations are included in Appendix B.
The study concluded that reliable emissions estimates (i.e., estimates
within the accuracy which is expected to result from variations in sampling
and chemical analysis) can be made using selected mathematical models and
correlations. Where plant operating conditions including residence times,
aeration rates, dimensions, etc., are available, limitations in applying the
mathematical models result from lack of accurate bj^oxidation and partition
coefficient data. When descriptions ot specific plant operating parameters
are not available, rough estimates can be made based on model plant desr.gns,
adjusted for wastewater flow rates.
-------
SECTION 3
DESCRIPTION OF AERATED WASTEWATER TREATMENT PROCESSES
INTRODUCTION
Aerated wastewater treatment systems are primarily designed to promote
decomposition of organic contaminants by aerobic organisms. Complete decomposi-
tion of hydrocarbons is accomplished by oxidation to carbon dioxide and water.
Intermediate steps in this oxidation process may result in the conversion of
larger organic molecules to smaller organic molecules. Organic nitrogen and
phosphorus present are ultimately converted to nitrate and phosphate. A
portion of the carbon, nitrogen, and phosphorus removed from the wastewater is
converted into additional organisms.
Only oxygen which is dissolved in the wastewater is available for use in
the reaction. At common treatment temperatures, under air at one atmosphere
pressure the solubility of oxygen in the wastewater is limited to approximately
5 to 10 mg/L. Oxygen must be transferred into the wastewater on a continuous
basis to replace that used in the oxidation reactions. A variety of processes
have been developed by environmental engineers to promote the oxidation reac-
tions and supply oxygen to the wastewater. From an engineering standpoint,
these processes differ with respect to the environment of the oxidizing organisms
and the means of oxygen transfer.
Aerobic biological oxidation processes include variations of the activated
sludge process, trickling filters and aerated lagoons. In addition, because
the dissolved air flotation process involves aeration of wastewater for separa-
tion of solids, biological oxidation may take place simultaneously with solids
separation but this does not influence process design. Wastewater treatment
systems include other process units which are not aerated but permit air/water
contact which can result in transfer of volatile chemicals out of the waste-
water. These process units include both turbulent and quiescent contact over
a range of liquid residence times.
All municipal wastewater treatment plants and some industrial wastewater
treatment pleats have a bar screen or an equivalent device to remove large
solids (>0.5 cm) from the influent wastewater. Typically wastewater flows at
approximately C.5 m/sec through a screen of vertical metal bars spaced approxi-
mately 0.5 cm apart. The. turbulence resulting from this screening promotes
-------
air-water contact and potential emissions of volatile material may result.
Although the amount of air-water contact is limited at 'chis point (due to the
short residence time) the influence of mass transfer if increased because the
bar screen occurs at the beginning of the treatment process. For extremely
volatile contaminants a significant fraction of the total concentration may
volatilize at the first opportunity.
Grit chambers are process units designed for removal of dense solids
(e.g., sand) by sedimentation. Their purpose is to simplify further treatment
processes. A very short residence time is sufficient for grit removal. For
the purposes of this report, bar screens aud\ grit chambers have been lumped
together as "pretreatment" and are considered to operate in plug flow.
Industrial waste treatment plants designed to treat wastewater that is
variable in contaminant concentration may include an equalization tank. The
purpose of using such a process unit is to avoid shocking the biological
treatment systems with intermittent high concentration waste loads. Typically,
the equalization tank is a large ope.n tank with sufficient volume to smooth
out variations in influent concentration. The tank may be agitated n.echanically
to promote mixing. Volatilization of contaminants may occur in such a unit as
a result of either normal evaporation or the agitation.
Wastes that are strongly acidic or basic may be neutralized to promote
t
biodegradation. Neutralization may take p]2~e in an. open agitated mixing tank
with the addition of base or acid. Emissions from this process unit will be
similar to those from an agitated equalization basin.
Process units designed for gravity separation of suspended solids are
ordinarily designed to Minimize agitation to prevent resuspension of soli..'? .
These clarifiers and sedimentation tanks provide a large surface area for
evaporation of volatiles and may be designed for several hours of residence
time. In some plant designs, sedimentation tanks may be placed before or
after biological treatment units: before, for the removal of solids present
in the influent waste stream, and after, for separation of biomass' floes. If
a high degree of volatile compound removal is achieved (either through biooxi-
datioii or volatilization) in the biological oxidation units, emissions from
the secondary.clarifiers (i.e, those designed to separate biomass from treated
effluent) will be relatively unimportant.
-------
Aerated treatment units specifically designed for biological treatment of
contaminants in wastewater have the greatest potential for emission of volatiies
because of the extent of air/water contact and the length of residence times
provided. Such units include roughing filters, trickling filters, activated
sludge units and aerated lagoons. Several different methods of aeration are
available for activated sludge units and aerated lagoons.
The activated sludge process consists of an aeration tank in which waste-
water is contacted with microorganisms. The nature of the aeration system
promotes mixing, and as an approximation, the aeration tank can be considered
a completely mixed reactor. The microorganisms responsible for oxidation of
wastewater components are present in the form of floes of biomass. These
floes are suspended in the aeration tank and are continually agitated by the
aeration system. The oxidation of the wastewater components results in the
conversion of hydrocarbons to carbon dioxide and water. In addition, new
biomass is formed from a portion of the v:astewater components. Biomass floes
leave the aeration tank along with the treated effluent and are typically
separated from the effluent by gravity sedimentation in secondary clarifiers.
Commonly, some fraction of the separated biomass (sludge) is returned to the
aeration tank (return sludge) to maintain or increase the number of microorgan-
isms available for oxidation of the organic material in the wastewater. The
remainder (waste sludge) is removed from the system, dewatered and disposed of
as solid waste. The dewatering of the waste sludge is important for volume
reduction; the separated liquid portion is returned to the treatment plant.
Some minor VOC emissions are possible in this solid separation process (e.g.,
in centrifuging or in filtering the waste).
Oxygen must be continuously supplied to the wastewater in the aeration
tank by mass transfer from air to the liquid phase. This is generally accom-
plished using either submerged aerators or surface aerators. Submerged aera-
tors operate through the discharge of air through diffusers near the bottom of
the aeration tank. The oxygen transfer occurs as the air bubbles rise to the
surface of the tank. Volatile components may diffuse into the air bubbles as.
they rise and be released at the surface. Bubble size may vary with smaller
bubbles resulting in a greater degree of oxyj-en transfer to the wastewater, as
-------
well as a more efficient stripping of organics. Smaller bubbles cost more to
produce because of is the increased compressor horsepower required by the
diffusers.
Surface aerators promote transfer of oxygen into the wastewater through
vigorous agitation of th^ surface of the aeration v.ank. Typically, a rapidly
rotating partially submerged impeller is used to create a large amount of
turbulence in the wastewater at the top of the aeration tank. The wastewater
in the vicinity of the impeller is rapidly saturated with oxygen and a high
degree of mixing circulates this saturated water throughout the volume of the
tank. Activated sludge units typically operate as very nearly completely
mixed systems whether equipped with diffused or surface aerators.
Trickling filters promote biological oxidation of wastewater components
by microorganisms in the form of a slime layer supported on a coarse filter
medium. Wastewater is sprayeil or trickled over a bed of crushed rock (particle
size is approximately 5 cm) up to approximately 3 m deep. The sprayer is
usually a rotating arm with nozzles spaced along its length. The arm slowly
rotates over the (generally circular) filter bed at a height of approximately
20 cm. The media is continuously wetted with wastewater and a layer of biomass
accumulates on the surface of the rocks. In some cases, specially designed
plastic filter media have been used as a support for the biomass. Oxygen is
supplied by transfer frosi the ambient air (the pores between the coarse media
are filled almost entirely with air with the exception of a thin film of
wetted biomass on the surface of the media). Air flows through the filter,
induced by a thermal density gradient, at a rate sufficient to prevent oxygen
depletion. This air velocity can vary widely (as well as change directions),
due to weather conditions but is typically on the order of 15 cm/sec.
Biomass loading on trickling filters eventually reaches a steady state.
Additional biomass produced is roughly equivalent to that which sloughs off
and leaves the filter with the treated effluent. This biomass is removed in a
secondary clarifier prior to discharge. A much lower mass is produced in
trickling filters than in the activated sludge process.
Operation of aerated lagoons is similar to that of activated sludge units
in that both are well mixed (completely mixed as a first approximation) systems
with the microorganisms responsible for biological oxidation suspended in the
-------
wastewater. A consequence of the high degree of mixing is that the bulk
concentration approximates that of the effluent concentration. Typically,
aeration is supplied with either surface aerators or diffused aeration systeiis
similar to those of the activated sludge process. The distinction between
activated sludge processes and aerated lagoon processes is not precise.
A type of aerated lagoon referred to as a spray pond is aerated by pumping
a portion of the wastewater through a spray nozzle into the air. As the
droplets rise and fall back into the spray pond they absorb oxygen which is
distributed throughout the bulk wastewater by diffusion and dispersion. The
degree of mixing provided by systems of this type is somewhat less than that
provided by diffused or surface mechanical aerators, and some regions of the
pond may be relatively stagnant and lower in. dissolved oxygen concentration
than the bulk wastewater. Aerosol formation and wind loss (of liquid droplets
as opposed to evaporation losses) may be important removal mechanisms for
spray pond systems.
DESCRIPTION OF INDUSTRIAL PROCESSES
Most municipal wastewater treatment systems in the United States include
aerated biological oxidation processes for removal of biochemical oxygen
demand from wastewater. Many municipal wastewater treatment plants accept
untreated or partially treated industrial wastewater; in some cases, this
wastewater may contribute more than 50% of the total flow through the treat-
ment system. Industrial wastewater treatment plants commonly employ aerated
biological waste treatment processes for removal of biodegradable wastewater
components. For wastes which are amenable to treatment processes of this
type, aerobic biological treatment methods are preferred to physical chemical
treatment methods for reasons of cost. Process effluents which contain valuable
byproducts or unreacted materials would first be subjected to recovery processes;
the dilute effluent from the recovery processes would then be subjected to
biological oxidation.
A study bv SCS Engineers (1979) found that industrial wastes suitable for
biological treatment are primarily produced by five major industries. These
industries are petroleum refining, organic chemicals manufacturing, pulp and
paper, iron and steel (coal conversion processes are included in this category
because of the importance of coke manufacturing as a source of wastewater) and
10
-------
textile manufacturing. Extensive studies of water pollution and water reuse
possibilities in these industries have been sponsored by EPA. Additional
industrial wastewater sources are less important in terms of total volume
produced but may represent important individual sites for emissions of volatile
organic compounds during wastewater treatment processes.
Petroleum Refining
Wastewater production from petroleum refining was studied for the develop-
ment of effluent guidelines (U.S. EPA, 1974a). Raw wastewater volumes and
compositions were determined for individual plant sectors. In general, the
wastewaters are contaminated with oil fractions including the crude oil input
to the refinery and various refined products. Other organic contaminants
present in the combined plant effluent include alcohols, phenols, and aromatics
(benzene, toluene, xylene, etc.). Individual components of the highly complex
crude oil and product fractions may result in varying levels of paraffins,
olefins, naphthenes and polynuclear aromatic hydrocarbons dissolved or suspended
in the wastewater.
Pretreatment of wastewater involves separation of suspended oil droplets
and suspended solids using clarifiers or dissolved air flotation. Aerated
biological treatment can involve aerated lagoons, activated sludge units,
trickling filters, or cooling towers which are operated with biomass ou the
tower packing. Where land is readily available, large shallow oxidation ponds
may be used (with very long residence times) without aeration equipment.
Influent biochemical oxidation demand is typically lowered by 75 to 95% using
aerated lagoons, 80 to 90% using activated sludge and 60 to 85% using trickling
filters (U.S. EPA, 1974a). The effluent guidelines document gives no estimate
of the fraction of removal due to stripping.
The influent wastewater to activated sludge units and aerated lagoons is
reported by EPA to contain up to 40 mg/L of phenols and up to 200 mg/L of BOD
under typical conditions. Refinery process upsets or partial shutdowns can
result in either much higher or much lower influent concentrations. In addition,
failure or upsets in upstream wastewater pretreatment equipment such as clari-
fiers and oil separators can increase the BOD of the influent to the aeration
units.' Waste biomass is produced in the activated sludge and trickling filter
* *r
processes. Water associated with this sludge, is typically separated and
-------
returned to the treatment system. The dewatered sludges may contain organic
compounds of both low volatility and low biodegradability which may eventually
volatilize to some extent if applied to land as a final disposal technique.
Rosenburg, et al. (1976) surveyed hazardous waste production and management
in the refinery industry. Waste sludge from activated sludge, trickling
filters, and aerated lagoons was found to contain between 1.7 and 10.2 mg
phenols per kg dry solids. The benzo-a-pyrene content of this material ranged
between 0.002 and 0.005 mg/kg dry solids. At present the waste sludge is
typically disposed of by landfilling or land spreading.
The 1979 EPA effluent guidelines survey (EPA 440-l-74-Ol4a) for petroleum
refineries covered 309 plants, which included (as of 1976), 63 dissolved air
flotation systems, 50 activated sludge systems, 10 trickling filter systems,
and 73 aerated lagoon systems were in service. A list of organic priority
pollutants from the 1982 Effluent Guideline Survey (EPA-440/1-82-014) and the
maximum levels observed is given in Table 1. -
Organic Chemical Industry
The 1975 EPA effluent guidelines study of the organic chemical industry
covered a total of 98 wastewater treatment and disposal facilities. Among the
aerated waste processes in use, the study identified activated sludge systems,
aerated lagoons, and trickling filters. Of the 98 facilities, 28 discharged
to municipal treatment plants, with or without some degree of pretreatment.
The composition and concentration of wastewater subjected to aerated
treatment processes varies widely depending upon the products manufactured and
production processes in use. Wastewater production ranged from 0.05 m3 waste-
water per ton of chlorobenzene produced by chlorination of benzene to 121 m3
wastewater per ton of chlorotoluene produced by chlorination of toluene for
the processes investigated in the 1975 EPA study. The biochemical oxygen
demands of the wastewater ranged from 3 mg/L for production of propylene
glycol by hydrolysis of propylene oxide to 303,000 mg/L for production of
cresol by methylation of phenol.
Erickson et al. (1980) estimated uncontrolled secondary emissions from
treatment of wastewater from organic chemical manufacturing at 650 g of vola-
tile organic emissions per ton of product. Based on total synthetic organic
chemical production of 108 Mg in 1978, the estimated air .emissions from waste-
-------
TABLE 1. ORGANIC PRIORITY POLLUTANTS DETECTED IN
REFINERY DISCHARGES AT MORE THAN ONE LOCATION'"
Maximum Concentration
Compound M8/L
chlorobenzene 31
chloroform 100
1,1 dichloroethane 54
ethylbenzene 18,000
toluene 43,000
2,4 dimethylphenol 18,300
2,4 dinitrophenol 11,000
phenol 33,500
acenaphthene 665
anthracene 1750
butyl benzyl phthalate 16
4 chlorophenyl phenyl ether 30
chrysene 30
diethylphthalate 38
fluoranthene 812
naphthalene 3750
pyrene 16
fluorene 495
isophorone 3550
phenanthrene 1750
*Pestioides excluded
SOURCE: Effluent Guidelines, EPA-440/1-82-014
-------
water treatment are 6.5 x 104 Mg/year. These estimates were based on a model
plant which consisted of a 3 n deep clarifier and an activated sludge system
with a retention time of three hours.
Iron and Steel Industry
The EPA effluent guidelines study of the iron and steel industry (1982)
covered options for suspended solids removal, oil and grease removal, dissolved
metal removal, and organic chemical re.noval from aqueous waste streams. The
greatest opportunities for emissions of VOCs occur in the organic removal
processes. Biological oxidation systems (typically activated sludge systems)
were in use at eighteen coke-making operations. Well operated activated
sludge systems were found to i.emove more than 90% of toxic organics and nearly
100% of phenolics, naphthalene and xylene. No data were provided on possible
air emissions from those processes. The average concentration in the effluent
from one individual biological treatment system treating byproduct coke uianu-
facturing wastewatcr was 3 ppb phenol, < 0.05 ppb naphthalene, 2 ppb benzo(a)-
pyrene and < 0.05 ppb benzene. These low concentrations may reflect stripping
and adsorption losses as well as biooxidation. EPA estimates of achievable
concentrations in the effluent from two stage activated sludge systems are
given in Table 2.
Suspended solids removal techniques in use in the iron and steel industry
include settling basins of up to 400 acres surface area, but the typical size
range is between 0.01 and 10 acres for the 140 settling basins found at 39
steel plants. Settling basin retention times are on the order of days.
Conventional clarifiers are also used for suspended solids removal with shorter
residence times (hours instead of'days) than settling basins.
Textile Industry
An EPA technology transfer report (EPA-625/3-74-004) on the textile
industry covered cotton, wool, and synthetic fiber processing. Cotton mills
discharge between 170 and 580 gallons of wastewater per kg of finished product.
The wastewater BOD ranges between 175 and 800 mg/L. Large variations in waste
volume and composition occur from day to day due to batch p.rocess operations.
Suspended solids can range uu to 100 mg/L. Cotton processing wastewater
contains fine suspended fibers which are difficult to remove in primary clari-
fiers, and consequently, these fibers are present in the aeration basin.
-------
TAB;*. 2. TOXIC ORCAKIC CONCENTRATIOS'S ACHIEVABLE BY
TWO STAGE ACTIVATED SLUDGE 5VSTCM
Acbi<-vable Concentration (pp>>)
Priority' Pollutant
~~""
"^Acrylonitt i le
*Benzrne
*2,i,6-?i iiiiiurophrmsl
^Ch I ore fora
*2"Chloropnenol
*2 , 4-0 ime thy 1 pheno 1
i2,4-Dinitroteluene .
;V2.6-OinitrotolueBe
EtHylbe-rizenc
*F!ucranthene
Isophorone
Naphthalene
2-Ni tropiienol
*4,6-Uinitro-o-cr«*8ol
^'Phenol
*Phthal3to8, Total
*BenEo(») anthracene
*Ben2o(.i (pyrene
^'Chryse.-if
Acenaptthylev.e
Anthracene
Fl'^rem-
Pyrene
•"^etrachiorethylene
*Toluene
Xylene
Biological Uxidat:ion
" — ~'~«-
100
50
50
200
50
5
50
100
25
5
IOC
5
100
25
25
200
5
5
10
10
1
5
10
100
50
too
Source: Development Document (or Effluent Limitations Guidelines and
Standards for the Iron and Steel Point Source Category,
EPA 440/1-S2/024 (May 1982).
"Indicates Appendix "III coepoucds (40CFR261).
-------
Wool processing mills discharge approximately 170 gallons of wastewater
per kg of finished product if unsecured wool (i.e., wool which contains natural
grease and other impurities) is used as the starting, material. This wastewatec
has between 900 and 3000 mg/L BOD and approximately 100 mg/L suspended solids.
Synthetic fiber processing (i.e., scouring, dyeing and finishing but not
production of the fibers) wastewaters vary greatly in volume and compos .lion
depending on the type of fiber (e.g., nylon, acrylic, polyester, etc.) ..nd the
type of processing. Organic chemicals removed from the fibers during scouring
and dyeing or scouring and bleaching operations include sulfcnated oils from
acetate, esters from nylon, formic acid and aromatic amines from acrilics.
Dyeing operations produce wastewater which can contain chlorobenzenes and
phenolic compounds. Much of the organic content of dyeing wastewater is, by
design, non-biodegradable or very slowly biodegradable. Trickling filters are
reported to remove 40-85% of the five day BOD from cotton and synthetic finishing
wastewater (Porter, 1971). The biodegradable portion of the organic loading may
be only a fraction of the total organic loading. BOD removals for activated
sludge units are estimated at 70-95 percent and for aerated lagoons at 50-95
percent (Porter, 1971). For wool processing plants similar figures are given for
trickling filters (80-85 percent) and activated sludge units (85-90 percent). A
list of organic priority pollutants found in textile wastewater is given in
Table 3.
Pulp and Paper Industry
Wastewater volume and composition from pulp and paper production varies
with the type of process used and the end product. These wastewater streams
have been found to contain organic sulfur compounds, phenolic compounds, and
oils (SCS, 1979). A typical integrated pulp and paper mill using the bleached
kraft process produces a total of 340 L of wastewater per Mg of paper produced
(EPA Effluent Guidelines, 1976). Raw wastewater from the different categories
of production using the bleached kraft process contains 240 mg/L BOD on average.
A list of organic chemicals found in pulp and paper manufacturing wastewater
is given in Table 4.
Pulp and paper manufacturing wastewater is subjected to pretreatment
(i.e., bar screens) prior to primary clarification. Effluent guidelines
assume 75 to 85 percent suspended solids removal with an overflow rate of 24
m-3/m2-day. BOD removal is typically accomplished using aerated stabilization
basins, which involve long residence times (e.g., 14 days) and .>re not well
16
-------
TABLE 3. PRIORITY POLLUTANTS DETECTED IN RAW WASTCWATER
FROM TEXTILE MANUFACTURING AT MORE THAN ONE- LOCATION
Maximum concentration
Compound . ^g/L
acenaphthene 273
acrylonitrile . 1,600
benzene . 200
chlorobenzene 296
1,2,4 trichlcrobenzene 14,000
hexachlorobenzene 2
1,2 dichloroethane . 6
1,1,1 trichloroethane 1,200
1,1 dichloroethane 14
1,1,2,2 tetrachloroethane 21
2,4,6 trichlorophenol 94
para chlorometa cresol 29
chloroform . 642
2 chlorophenol 131
1,2 dichlorobenzene 460
1,3 dichlorobenzene 1,700
1,4 dichlorobenzene 760
1,1 dichloroethylene 84
1,2 trans dichloroethylene 360
2,4 dichlorophenol 41
1,2 dichloropropane . 100
2,4 dimethylphenol 190
ethylbenzene 19,000
methylene chloride 2,600
trirhlorof luoromethaiic 45
naphthalene 2,079
4-nitrophenol 240
N-uitrosodiphenylamine 130
pentachlorophenol 310
phenol 4,930
bis 2 ethyl hexyl phthalate 1,450
butyl benzl phlthalate 160
di-N butylphthalate 67
di-N octylphthalate 10
diethyl phthalate 150
dimethyl phthalate 111
anthracene 12
fluorene 15
phenanthrene 12
tetrachloroechylene 1,130
toluene 3,200
trichloroethylene 5,600
17
-------
TABLE 4. ORGANIC COMPOUNDS PRESENT IN WASTEWATER
FROM THE PULP, PAPER, AND PAPERBOARD INDUSTRY
Priority pollutants
benzene
chlorobenzene
1,2-dichloroethane
1,1,1-trichloroethane
1,1-dichloroethane
1,1,2,2-tetrachloroethane
trichlorophenol*
chloroform
2,4rdichlorophenol
ethylbenzene
fluoranthene
methylene chloride
dichlorobromomethane
trichlorofluoromethane
Nonconventior.al pollutants
oleic acid
linoleic acid
lonolenic acid
pimaric acid
isopimaric acid
dehydroabietic acid
abietic acid
chlorodibromomethane
isophorone
naphthalene
phenol
bis(2-ethylhexyl) phthalate
di-n-butyl phthalate
di-n-octyl phthalate
diethyl phthalate
chrysene
anthracene/phenanthrene
tetrachloroethylene
toluene
trichloroethylene
3,4,5-trichloroguaiacol
tetrachloroguaiacol
monochlorodehydroabietic acid
dichlorodehydroabietic acid
9,10-epoxyscearic acid
9,10-dichlorostearic acid
xylenes
*Includes 2,4,5- and 2,4,6-trichlorophenol
SOURCE: EPA 440/1-82-025 (1982)
18
-------
mixed. The long residence times i..ake these systems relatively insensitive to
shock loading. Design specifications assume approximately 1.25 mg oxygen
supplied/mg BOD removed using mechanical aerators which supply 1.06 kg
oxygen/kW-hr.
Conventional activated sludge systems can also be used for bio-oxidation
with approximately 6-8 hour residence times in well mixed systems. These
systems are more susceptible to upsets due to shocks and require secondary
cla-rifiers for removal and return of sludge. Design specifications used on
the effluent guidelines study included an oxygen requirement of 1.0 mg oxygen/
mg BOD removed supplied by aerators with a transfer efficiency of 0.34 Kg
oxygen/Kw-hr. High rate trickling filters have been used upstream of serated
stabilization systems to reduce high organic loadings (Gehm, et al. 1965).
MODEL TREATMENT PLANTS
Hazardous waste generators and management facilities have been reported
to be concentrated in the manufacturing industries. (Westat, 1984). Eighty
five percent of the 14,098 generators and 72 percent of the 4,818 TSD facili-
ties are estimated to be associated with industrial manufacturing codes 2000
through 3999. (See Table 5).
The wastes from these generators include solvent, corrosives, sludges.
etc. They are subject to various forms of handling prior to and during dis-
posal. Those processes involving the use of air (biological treatment and
cooling), or subject to the introduction of air (stirreii equalization and
neutralization) are subject to emit VOC as a consequence of air stripping.
In general, these processes can be categorized as biological or physical-
chemical, depending on the unit process involved. For the purposes of this
report, this distinction will be maintained. Further, only those biological
processes or physical-chemical processes with potential to vent to the air
will be addressed. Anaerobic processes, including anaerobic sludge digestion
have been excluded because the enclosed nature of these systems substantially
reduces air emissions.
Aerobic biological treatment is only economically viable for the destruc-
tion of organics when the waste stream is large, continuous, fairly consistent
in composition and of low toxicity. Consequently, biological systems are
viable for a limited range of treatment, primarily for dilute streams with low
19
-------
.,TABLE 5. GENERATORS AND TSD BY INDUSTRY TYPE
(Westat, 1984)
Description
Fabricated metal products
Chemicals
Electrical equipment
Other metal related products
Nonmanufacturing
Generators
2,636
2,443
1,515
2,222
2,074
TSD*
1,249
547
540
804
800
-'Treatment, storage, and disposal facilities.
20
-------
metal concentrations in order to reduce the potential of biostatic and/or
biotoxic side effects. As such, biological treatment would not be a principal
factor in the treatment of hazardous wastes at generator or TSDF sites.
father they would serve to clean up other aqueous streams which have been
contaminated or which have been combined with low levels of toxicants.
Normally, physical, chemical, or some combination of physical-chemical
treatment process will be used for actual hazardous waste treatment at genera-
tor or TSD sites. These processes may include air stripping, equalization,
cooling, neutralization, evaporation, all of which have potential for air
emissions.
Biological treatment is generally a sub^ot of an overall treatment system
that i.uy also include physical and chemical processes. A typical arrangement
of units will include preliminary treatment processes, primary treatment,
secondary treatment with sedimentation, optional tertiary treatment, sludge
handling, and ultimate disposal.
The design and function of these units have been thoroughly discussed in
the literature. In general, they perform the following functions. Preliminary
treatment involves the screening and degritting of raw wastewater. In this
step coarse materials are collected on bar screens (1.3 to 5.1 cm spacing) for
burial, and heavier inorganic particulates are removed through selective
deposition in velocity control chambers.
Primary treatment involves detention of wastewater ia either rectangular
or circular basins for 1-2 hours in order to allow readily settleable organic
solids to be removed via gravity induced sedimentation. Primary treatment can
also involve the introduction of microscopic air bubbles into the wastewater
as occurs with a dissolved air flotation unit (DAF) in order to float oils and
grease to the surface where they are removed by skimmers for disposal.
Secondary treatment can include some physical and chemical processes, but
typically this is the domain of biological treatment processes. These processes
include trickling filters and modifications of the activated sludge process.
These processes will be farther discussed under each model plant heading.
Tertiary treatment involves the removal of those pollutants not removed
by the conventional biological processes contained in secondary treatment.
These pollutants include suspended solids, BOD (typically less than 10-15
mg/1), refractory organics, nutrients, and inorganic salts. The tertiary
treatment option selection largely depends on the water quality requirement.
21
-------
Sludge handling involves the stabilization of the solid-water mixtures
derived from the primary and secondary clarifier as w<:ll as the excess biomass
from the activated sludge process and chemical reactions:. These mixtures
undergo thickening, anaerobic or aerobic digestion and dewatering prior to
ultimate disposal. Anaerobic digestion is designed for minimal air/sludge
contact. Emissions from the other processes are likely to be insignificant
because the upstream processing units will have provided extensive opportuni-
ties for volatilization prior to the sludge handling operations.
Because the focus of this work involves the identification and modeling
of the competing mechanisms of pollutant removal, five general model systems
have been described that center on the use of air and aeration process. These
models are:
1. Trickling filtration
2. Activated sludge (.mechanical aeration)
3. Activated sludge (diffused aeration)
4. Activated sludge - aerated lagoons (mechanical aeration)
5. Spray Ponds - (mechanical spray)
Although presented as discrete model systems, it should be noted that
portions of these model plants are interchanged with other model plants, and
that all model plants can be composed of unit processes that fit into the
broad categories of treatment previously described. A good example of the
former situation is the use of a trickling filtration unit in the treatment
train ahead of an activated sludge unit. In this capacity, the trickling
filter operates as a roughing filter to pretreat industrial wastes prior to
secondary treatment and not as a secondary treatment process.
It should be emphasized that treatment systems vary widely depending on
the nature of the wastewater, the availability of land, prior regulatory
pressure, the composition and flow rate that may have existed at the time the
system was designed, and many other factors. The model systems that follow do
not represent an "average" system. These systems do fall within the range of
sizes and configurations that may be encountered at TSDFs.
Model Plant ill. Trickling filtration
The typical trickling filter plant consists of the following units: pre-
treatment, primary clarifier, trickling filter, secondary clarifier and post-
22
-------
treatm nt unit (see Figure 1). The heart of the system is the trickling
filter itself, which consists of a circular basin 1-2.4 m deep p.-.cked with a
bed of either rock or plastic media, over which wastewater is sprayed. A
zoogleal slime which attaches to the media assimilates and oxidizes the organics
in the wastewater. Oxygen and organic matter diffuse into the zoogleal mass
and end products of oxidation counter-diffuse back into the flowing liquid or
to the void spaces. The treated water and any particulates from the filter
bed are collected in an underdrain system and sent to secondary clarifiers for
sedimentation.
The packing media is typically dosed with a rotary distributor which
sprays the waste over the media. The media nu'.y be either plastic or rock.
The rock medium represents a traditional approach; the plastic however, offers
advantages such as lower specific weights and higher void spaces and is amenable
to above ground installation.
The performance of the unit is affected by many factors such as hydraulic
and organic loadings, depth and physical characteristic of the media, the
method of wastewater distribution, ventilation, and characteristics of the
applied wastewater (Ecksnfelderj 1970). Municipal wastewater and a wide
variety of industrial wastewaters are amenable to treatment in trickling
filters.
The principal components of the trickling filter process are:
1. The distribution system
2. The filter media
3. The underdrain
4. Final sedimentation
The rotary distributor consists of two or more horizontal arms mounted on a
turntable assembly anchored to a center column. The wastewater is uniformly
distributed over the media through orifices located in the arms. The principal
drive mechanism for these arms is the reaction force from the spray on the
radial arms. The arms are sized to limit velocities to 1.2 m/sec at maximum
flow. The rotation speed of the arms varies with flow rate in the range of
0.1-2 rpm.
Ventilation is extremely important in achieving efficient filter operation.
Usually, if the underdrain is properly sized, the differences in air and water
temperature will provide a natural driving force for ventilation. An air flow
23
-------
Sludge to
digester
Pretreatment
Primary
Clarifier
Recirculation
Trickling
filter
Secondary
Clarifier
Effluent
Sludge to
digester
Figure 1. Model trickling filter plant.
-------
rate of approximately 0.03 rn^/m2 .filter area per minute is required to sustain
aerobic conditions within the bed. When forced ventilation systems are re-
quired, they are typically designed to provide an air flow of 0.3 m3/m2 of
filter arej per minute.
Organic and hydraulic loading determines the classification of the filters:
low-rate, high-rate, or roughing-rate. Low-rate filters are generally not
equipped with recirculation and are rarely used. High-rate filters use recir-
culation to dilute the influent organic strength and to flush the media voids.
This permits higher BOD loadings per volume of media and promotes the return
of activated organisms as a seed. The high-cate filters are generally designed
to accept a continuous flow of wastewater and may be either single stage or
two staged. High-rate filters also have a number of modifications of the
basic recirculation scheme.
Roughing rate trickling filters provide an intermediate stage of treatment
and are used frequently to precede activated sludge units or second stage
filters. The purpose of this operation is to reduce high organic loadings
prior to further treatment. This intermediate stage is typical for industrial
systems.
The trickling filter model is based on a design (EPA-430/9-77-006) currently
in operation in a U.S. municipality (See Figure 1). It represents a single
train of a multi-train high rate process. The operating conditions and speci-
fications fall within the range expected for industrial waste treatment.
The design parameters are given in Table 6.
Conventional Activated Sludge System
The typical configuration of an activated sludge system is pretreatment,
optional primary sedimentation followed by the aeration process including
secondary clarification, and post treatment (e.g., Figure 2). The principal
treatment process is the aeration tank. This is a continuous flow, biological
treatment process characterized by the turbulent suspension of microscopic
aerobes. The turbulence promotes mixing and induces a relatively homogenous
state in which the microbes are able to absorb and oxidize soluble and colloidal
organics. The process involves an aeration step followed by a solid-liquid
separation step in which part of the separated sludge is recycled back to the
system for mixing with the raw influent.
25
-------
TABLE 6. MODEL PLANT OPERATIONAL PARAMETERS - TRICKLING FILTER
Plant Flow: 6.4 MGD (0.28 m3/sec) Plant Performance: 85% BOD Removal.
75% Suspended
Solids Removal
Influent BOD-183 rag/1
Influent Suspended Solids - 188 mg/1
Trickling Filter: Diameter 190 ft (58 m)
Depth 5 ft (1.5 m)
Area 28353 (2640 m2)
Volume 141764 ft3 (3960 m3)
Hy'raulic loading 29 MGD/acre (1.1 m3/m2-hr)
Recii';culation - 190%
Clarifiers: Diameter 100 ft (30 m)
Depth 9.2 ft (2.8 m)
Area 7854 ft2 (730 m2)
Weir height 1 ft (30 cm)
Surface loading - 1350 gal/ft2/day (0.47 m3/m2-day)
Detention time - i.2 hours -.
-------
Equalization
DAF Unit
Aeration Tank
Clarifier
Influent •
Float to
Separator
Bottoms
to
thickener
Sludge
Recycle
To fire pond
' and discharge
Waste
Sludge to
Digester
Figure 2. Conventional activated sludge system with mechanical aeration.
-------
There are many variations of the activated sludge process; however, they
generally can be reduced to looking at either the loading rates in terms of
BOD or the physical arrangement of the process train. The loading is typically
one of three basic types. High rate takes advantage of the settleability of
sludge when the system is loaded at a rate of 0.80-1.15 g of BOD/g of mixed
liquor suspended solids per day. Conventional rate is of the range 0.2 to 0.5
g BOD/g mixed liquor volatile suspended solids per day. This rate is typical
for most larger municipal treatment plants. Extended aeration rate is the
lowest range of process loading ai;d is used in those plants which are small in
size and do not receive 24 hour supervision. As such they are generally
conservatively designed and operate in the range of 0.05-0.15 g of BOD applied/g
of MLVSS/day. Industrial wastewaters vary widely in terms of influent concen-
tration and biodegradability. Thus, a wide range of loading rates are used
depending on individual circumstances.
Physical arrangements are of three types; the complete nix activated
sludge, plug flow activated sludge, and activated sludge with reaeration. In
the complete mix arrangement, the return sludge and the wastewater are uni-
formly introduced into the aeration basin through several points in order to
obtain a homogeneous mixture. In a plug flow arrangement both the untreated
wastewater and the return sludge are introduced at the head of the plant and
flow through the plant in a modified plug flow. Such plants are often compart-
mentalized to maintain the plug flow regime. Activated sludge with sludge
reaeration constitutes a rearrangement of process streams. In this instance
the sludge is compartmentalized and aerated prior to its contact with the
untreated waste.
Model Plant #2 Conventional Activated Sludge (Mechanical Deration)
The principal component of the mechanical aeration system is the aerator.
There are two types in general use today, surface aerators and turbine aerators.
The surface aerator is highly developed and widely used, particularly in the
treatment of industrial waste. The surface aerators may either float or be
mounted on supports in the aeration basin. They enhance the entrainment of
atmospheric air in the aeration basin by producing a region of high turbulence
? ^ '•-
around the periphery of"the aerator. Oxygen transfer efficiency of these -
aerators increases with the depth of submersion, as does power cost; conse-
quently, there is a trade off between efficiency and cost.
-------
Sin e 1950, the gubaerged turbine has been widely used in the chemical
industry. It offers an attractive eeans of upgrading existing facilities to
handle increased loads. These aerators are u-sed because of improved oxygen
transfer efficiency and lover hotsepover requirements. Oxygen transfer effi-
ciency for aerators, as rated in ter»s of chatueal aerator approach is tound in large open basins particularly
in tiiose plants operating tn » rooplcte eia, conventional activated sludge
mode. The turbulence intro^t .ed by the rotary action of the aerator blades
proaotfn a hossogfnroue aixibg jnd enhances the overall complete mix node ot
operation.
Hod;IPS ant 63: Conventional Activated Sludge (Diffused Air:_ Coarse and
Ftftc Bubbjrj
A second approach to aeration is the use of diffuser systems which are
generally used in plug flew r.yateas and sludge reaeration systems, the roost
coaizan types of aeration rysteses used in activated sludge plant. The distri-
bution Byatea consists of au array of diffusers situated near the bottom of
the btsia. These diffusers are designed to produce either coarse or fine
bubbles and are supplied with air by compressors. In the period from 1950-1978,
the fine bubble systems were in wide use. At that time, it was felt that the
increase ia oxygeo transfer efficiency of the smaller bubble diameter (8% vs.
S» for the coarse bubble) vas iatportant. Later, however, inefficiencies such
as clogging decreased the overall.attractiveness of these systems.
The ao»t coa»ou type of fine bubble diffusers are nylon or dacron socks
and e«ran wrapped tubes. Other systems include porous ceramic plates that
generate Mull diaaelcr bubbles. Coarse bubble diffusers can be tubes covered
with synthetic fabric or wound with filaments, and sprayers with multiple
openings created by drilling holes in pipes or loosely attaching plates or
discs to a supporting piece of pipe. Although the oxygen transfer efficiency
i* lover, coarse bubble diffusers do not suffer from clogging and have lower
29
-------
TABLE 7. MODEL PLANT //2 CONVENTIONAL ACTIVATED SLUDGE WITH MECHANICAL
AERATION. OPERATIONAL PARAMETERS
Plant Flow: 1.5 MGD (0.066 m3/sec) Plant Performance 85% BOD Removal
Plant Loading: Influent BOD - 80 rag/1
Aeration Basin (Mechanical Aeration)
(56 Kw each)
(43 m)
(3.7 m)
(21 m)
(910 m2)
Detention time - 14 hours at average flow
aerators
length
depth
width
area
- 2, 75 Hp each
- 141 ft
- 12 ft
- 69 ft
- 9715 ft2
ClariJier
Length
Depth
Area
Weir height
Surface loading
Detention time
75 ft
9.8 ft
740 ft2
1 ft
2030 gal/ft2/day
0.5 hours
(23 m)
(3.0 m)
(40 m2)
(30 cm)
(140 m3/m2-day)
30
-------
initial cost aud maintenance. Many treatment plants are reported to have
switched to the coarse buLMe systems in order to take advantage of these
features.
The model plant for this system based on typical organic chemical manufac-
turing wastewater (EPA-450/3-80-'>25) is described in Figure 3 and Table 8.
Model Plant #4: Aerated Lagoons n-iechanical Air)
Aerated lagoon systems are medium depth basins designed for biological
treatment on a continuous flow basis. They are equipped with surface aerators
and are primarily used to treat wastes of low-medium strength in areas where
land is inexpensive. They are not as widely used as stabilization poads, but
their feasibility has been fully demonstrated and they may represent an upgrading
of an oxidation pond.
Aerated lagoons have detention times on the order of 3-10 days. Aerated
lagoons are staged in series and are designed to achieve partial mixing.
Consequently, aerobic and anaerobic stratification can occur. A large fraction
of the incoming solids may in fact settle out near the head of the plant.
The model proposed for this type of treatment system (See Figure 4) is
based on a system at an existing manufacturing facility (Alsop, et al. 1984)
(see Table 9).
Model Plant #5: Spray Evaporation Ponds
Spray evaporation ponds are used primarily to reduce the amount of water
contained in a waste. Biodfgradation in these ponds may be relatively unimpor-
tant. These are basically ponds equipped with submersible pumps attached to
vertical pipes ending in standard irrigation spray headers. Water is pumped
through this system and dispersed in the air. As the droplets fall back to
the pond they are enriched with oxygen and subjected to evaporative processes.
These ponds occur primarily in waste treatment systeirs involving large quanti-
ties of recycled water. Operational parameters are given in Table 10. Evapora-
tion ponds without spray headers are termed solar evaporation ponds and are
not intentionally aerated.
Physical-Chemical Treatment System
Physical-chemical processes are generally defined as those operations
which effect the removal and/or destruction of undesirable constituents in
wastewater by means other than biological degradation or biological conversion.
31
-------
Pretrcatment
Primary Clarifier
Aeration
JUXU
Secondary Clarifier
. Effluent
Recycle
OJ
Figure 3. Conventional activated sludge with diffused aeration model plant
-------
TABLE 8. MODEL PLANT //3 CONVENTIONAL ACTIVATED SLUDGE WITH DIFFUSED
AERATION. OPERATIONAL PARAMTERS
Plant flow 4.5 MGD (0.2 m3/sec) Plant Performance 85% BOD Removal
Plant loading: :
Influent BOD 2000 mg/L
Aeration Basin (Sparger)
Av scfm 21,600 (612 m3/nin)
length 260 ft (79 m)
depth 20 ft- (6 m)
area 68,000 ft2 (6,300 m2)
Detention time 2 days @ average flow in aeration basin
Clarifier
Diameter 49 ft (15 m)
Depth 9.8 ft (3 m)
Area 1,900 ft2 (180 m2)
Weir height 1 ft (30 cm)
Detention time 3.0 hours
33
-------
Equalization
Effluent
Clarification
Aeration
o
o
1
1
o
o
—
000
o o o
1
0 (
0 C
1
DO
) O
u>
Figure 4. Aerated lagoon model plant.
-------
TABLE 9. MODEL PLANT #4 AERATED LAGOON. OPERATIONAL PARAMETERS
Plant Flow 5.3 MGD (0.23 mVsec) Plant performance: 80% BOD Removal
Plant Loading: Influent BOD = 2000 mg/L
Aeration Basins
2 basins
aerators
10 (MOO hp each,
5 @ 75 hp each
length 560 ft (170 m)
depth 12 ft (3.6 m)
width 280 ft (85 m)
Area 157,000 ft2
Detention time - 250 hours
Clarifiers (2)
Depth 10 ft (3 m)
Area 5675 ft2 (527 m2)
Diameter 85 ft (26 m)
Weir height 1 ft (30 cm)
Surface loading - 148 gal/ft2/day
(6.1 m3/m2-day)
Detention time - 13 hours
Preliminary Hold'ing
Dejjth
Length
Width
Area
Detention time -
(Equalization)
14 ft (4.3 m)
310 ft (94 m)
180 ft (55 m)
56,000 ft2
51 hours
TABLE 10. MODEL PLANT #5 SPRAY EVAPORATION POND. OPERATIONAL PARAMETERS
Pond capacity:
1 MGD (0.043 mVsec)
Pond loading:
Influent BOD 100 mg/1
Influent suspended solids
N/A
Pond
Depth
Length
Width
Area
10 ft
(3.0 m)
134 ft (41 m)
100 ft (30 m)
13369 ft2 (1240 m2)
Nozzle capacity 40 gal/sin
(2.5 L/sec)
6 feet above pond surface
(1.8 m above pond surface)
35
-------
Physical-chemical processes include a wide array of traditional and innovative
processes. For the purpose of this report, however, the focus is placed on
those processes which are intentionally aerated. These processes include
dissolved air flotation, and mechanically agitated equalization/neutralization
basins. These processes can be used as adjuncts to the model plant flow
charts presented in this section."
Dissolved Air Flotation (DAF)
Dissolved air flotation is widely used in industry to remove suspended
solids by flotation. The flotation of the particulate is induced by microscopic
air bubbles attaching to the particulate or agglomerate and giving it buoyancy.
Particles are floated to the surface where they are removed by skimmers for
further treatment.
The DAF system generates a supersaturated solution of wastewater and air
by pressurizing either the influent wastewater (or a side stream of the influent
wastewater) and introducing compressed air. The pressure is then released in
the detention tank generating the numerous microscopic bubbles which adhere to
particulates or are trapped by any floe which may be present. The design
criteria for a model DAF unit are contained in Table 11.
Neutralization (Equalization) Process
Although neutralization and equalization units perform different functions,
i.e., pH neutralization vs. flow equalization, these operations can be considered
together as they permit similar modes of air/water contact. Primarily, these
units are open basins or tanks with varying size depending upon the desired
retention time. Mechanical agitation by stirrers is used to assure a homoge-
neous mixture. The design criteria for these processes are dependent on the
variation in influent composition. For example, when the objective is equali-
zation, more erratic fluctuations in the influent composition necessitates
longer residence times.
36
-------
TABLE 11. OPERATIONAL PARAMETERS - DISSOLVED AIR FLOTATION
Flow
Pressure
Detention time
Surface loading . -
Recycle
Solids loading
Air to solids ratio -
Depth
5 MGD
25-70 psig
0.25 - 1 hr
500-8000 gpd/ft2
5-120%
.5-5 Ib/ft2/hr
.01 - .1 Ib/lb
4-9 feet
(0.22 m3/sec)
(1.7 - 4.8 atm gauge)
(0.17 - 2.8 m3/m2-day)
(2.4 - 24 kg/m3 hr)
(.01 - 0.10 kg/kg)
(1-3 m)
37
-------
SECTION 4
DESCRIPTION OF REMOVAL MECHANISMS
BACKGROUND
VOCs are removed from wastewater through destructive and nondestructive
mechanises. The major destructive mechanism is biological oxidation. Of less
importance are chemical reactions, which also destroy specific VOCs through
conversion to other materials. The major nondestructive removal mechanism is
mass transfer from the liquid phase to the atmosphere. This takes place in a
variety of processes that offer the opportunity for air/water contact. A less
important nondestructive removal mechanism results from preferential adsorption
of VOCs on surfaces (chiefly biomass solids). The effect of adsorption on
solids can be determined through accurate measurements and analyses of net
solids removal from process units.
A generalized mass balance applicable to either individual processes or
complete treatment systems is given in Equation 1.
(Mass In) - (Mass Out) = (Mass Destroyed) + (Mass transferred to air)
+ (Mass transferred to solids) (1)
The following sections deal with biological oxidation, mass transfer, and
adsorption.
BIOLOGICAL OXIDATION
Biological wastewater treatment processes involve contacting contaminated
wastewater with microorganisms capable of using the organic contaminants
present in the wastewater as food; Aerobic processes have been used most
often for wastewater treatment. In these processes, which include activated
sludge, trickling filters, aerated lagoons and spray ponds, microorganisms
present in the wastewater (or supported on media in the case of trickling
filters) use oxygen dissolved in the wastewater to convert the contaminants to
more oxidized forms ac.d to produce more organisms. The ultimate oxidation
products of organics are carbon dioxide and Water and when reactions go to
completion, they proceed in accordance with the chemical equations for complete
combustion of hydrocarbons. Intermediate products include biomass and partiall;
oxidized hydrocarbons.
38
-------
The rates of reaction for biological oxidation are dependent upon many
factors including the presence or absence of nutrients, the presence or absence
of inhibitors, the other oxidizable compounds present in the wastewater, the
temperature, and the concentration of biomass in the system. Various investi-
gators have attempted to determine the biodegradability of specific organic
components under aerobic conditions. No standard procedure has been adopted
for experiments of this kind and extrapolations of laboratory and pilot scale
data to real systems, either for design calculations or emissions estimation
must be made with caution.
Classical Approach
The conceptual approach to biooxidation which is used most commonly by
sanitary engineers follows the model of Lawrence and McCarty (1970). The
production of biomass is related to the utilization (and thus disappearance)
of the organic material as
dX dC
B = Y —£ - k X (21
dt B dt
-------
. k, ,,. = pollutant concentration at which the rate of pollutant consumption
per mass of biomass is one-half of the maximum rate, mg/L.
Equation (2) describes a kinetic irate expression in which the order of
rate dependency is variable depending on the concentration of the pollutant.
While this is inconvenient, it has been found to accurately explain the biode-
gradation of organic material in a wide variety of systems and is generally
accepted. The four constants describing the system; Y_, k,, k , and K. ., ,.
O Q HldX 113 j. I
are available for a very limited number of pure compounds and mixed industrial
wastes. A laboratory procedure for obtaining these constants is given by
Metcalf and Eddy (1972). Kincannon and Stover (1983) have reported experimen-
tal values for certain chemical and plastic manufacturing wastewaters.
Rates from the EPA Data Base
While the literature data on biodegradability were not obtained under
standardized procedures, some qualitative judgments regarding individual
compounds can be made. Where several components were tested using the same
procedure, the observed rates of biodegradation can be ranked in relative
order. The most extensive compilation available is that of Fitter (1976)
which includes biodsgradation data for 123 organic compounds (see Table 12).
Fitter's work involved pure solutions of chemicals formulated to represent 200
mg/L of chemical oxygen demand. Carefully acclimated inocula were used.
Blackburn et al. (1982) conducted an extensive literature review of biodegrad-
ability and biodegradation rate studies. A list of biodegradation rates
converted to a basis roughly comparable to that of Fitter is given in Table 13.
Specific compound removal rates have also been converted to theoretical oxygen
demand (ThOD) removal rates in this table. In practice, these rates are
applicable to suspended growth reactors such as activated sludge and aerated
lagoon processes. The relative rates may be applicable to trickling filter
processes.
One of the difficulties in interpreting biodegradation data obtained by
different investigators is the limitation on rates imposed by the influent
substrate concentration. Where the compound in question is completely degraded
over the course of the test, a minimum rate is obtained because had more of
the compound been present, a greater amount might have been oxidized in the
same period of time. Similarly, although biooxidation of organic chemicals is
-------
TABLE 12. BIODEGRADATIOK DATA FROM FITTER (1976)
Compound '
Ammonium oxalate
n-butanol
Sec. butanol
Tert. butanol
1 ,4-Butanediol
Diethylene glycol
Diethanolamine
Etbylene diamine
Ethylene glycol
Glycerol
Glucose
n-Propanol
Iso-Propanol
Triethylei.e glycol
Borneol
Caprolactam
Cyclobcxanol
Cyclopeutanol
Cyclohexanone
Cyclopentauone
Cyclohfxaoolone
1 ,2-Cyclobexanediol
Dimetbylcyclobexanol
4-Metbylcyclohexanol
4-.1ethylcyclohexanone
Menthol
Tetrabydrofurfuryl alcohol
Tetrahydrophthalioide
Tetrabydrophthalic acid
Aniline
Amioopbenolsulpbonic acid
Acetanilide
p-Astinoacetanil ide
o-Aninotol uene
n-Aninotoluene
p-Aoinotoluene
o-Aminobenzoic acid
n-Aainobeazoic acid
p-Aminobrnzoir acid
o-Aainophenol
B-Aainophenol
p'Aaiaopbenol
Benzenesulpbonic acid
n-Benzenedisulpbonic acid
fienzaldebyde
Benzoic acid
o~Cresol
B-Cresol
p-Cresol ,
D-Ctlorampbenicol
o-ChJoropbenol •
p-Chloropbenol
o-Cblorojniline
n-Chloroaoi 1 inr
p-Chlorcanilfne
2-Chloro-4-nitrophenol
2t4-Dichloropbenol
1 ,3-Dinitrobenzene
1 t4-Dinitrobenz»ne
Rate of
biodegradation
(ng COD g~' h"1)
9.3
84.0
55.0
30.0
40.0
13.7
19.5
9.8
41.7
85.0
180.0
71.0
52.0
27.5
8.9
16.0
28.0
55.0
30.0
S7.0
51.5
66.0
21.6
40.0
62.5
17.7
40.0
0
0
19.0
7.1
14.7
11.3
15.1
30. P
20.0
27.1
7.0
12.5
21.1
10.6
IS. 7
10.6
3.4
119.0
88.5
54.0
55.0
55.0
3.3
25.0
11.0
16.7
6.2
5.7
5.3
10.5
-
•
Conpo^d
2 , 3-Dimethylphenol
2 ,4-Dimetcylphenol
3,4-Dinetbylphenol
3 ,5-Uinethylphen"!
' 2,5-Dimethylphenol
2 ,6-Diawthylphenol
3 ,4-Dimethylaniline
2 , 3-Dinethylaniline
2 ,5-Dimethylani 1 ine
2,4-Diaminophencl
2,5-Dinitropheaol
2t6-Dinitropbenol
2,4-Dinitrophenol
. 3,5-Dinitrobenzoic acid
3,5-Dinitrosalicylic acid
Furfuryl alcohol
Furfurylaldehyde
Gallic acid
Gentisic acid
p-Hydroxybenzoic acid
Hydroquinone
laophtbalic acid
Hitol
Napbtoic acid
1-Naphtbol
V-Napbthylanine
-1-Napbthaleneaulfonii acid
l-Kapbtbol-2-sulphonic acid
l*Napbthylaaine-6-fiulpbonic acid
2-Napbthol
p-Nitroacetophenone
Kit-robenaene
o-Nit ropbenol
n-Nitrophcnol
p-Nitropbenol
o-Nitroluene
•-Nitrotoluene
p-Nitrotoluene
o-Nitrobenza Idehyde
m-Nitrobenzaldehyde
i-Nitrobenzaldebyde
-Nitrobenzoic acid
-Nitrobenzoic acid
I'Nitroben^oic acid
-Nitroaniline
-Nitroaniline
>-Kitroaniiine
Pbtbalimide
Fhtbalic acid
Phenol
Pbloroglucinol
Pyrocatechol
Resorcinol
Salicyclic acid
Sulphosalicyclic acid
Sulphanilic acid
Thymol
p-Toliienesulpbonic acid
2t4(6-Trinitropbenol
Rate of
biouegrajation
(og COD g"1 h"1)
35.0
28.2
13.4
11.1
10.6
9.0
30.0
12.7
3.6
12.0
0
SO
6.0
-
-
41.0
37.0
20.0
80.0
1CO.O
54.2
76.0
0.8
15.5
38.4
0
18.0
18.0
0
39.2
S.2
14.0
14.0
17.5
17.5
32.5
21.0
32.5
13.8
10.0
13. S
20.0
7.0
19.7
0
0
0
20.8
78.4
80.0
22.1
55.5
37.5
95.0
11.3
4.0
15.6
8.4
0
-------
typically regarded as a reaction which proceeds with a zero order dependence
on che concentration of the organic material, at very low organic concentra-
tions this may not be true.
The biocxidation rates listed in. Table 13 should be considered as minimums
for well designed well operated systems. Better acclimated full-scale systems
might perform more efficiently under some circumstances. More importantly,
poorly designed, poorly operated systems can be expected to have much lower,
or in some cases zero, rates of biooxidation. Ths presence of toxic or inhibit-
ing materials in the wastewater or lack of nutrients can adversely affect
removal rates. Shock loadings of material which may be rapidly biodegradable
under normal conditions can cause process upsets which greatly reduce treatment
plant efficiency. In addition, low temperatures will depress the biooxidation
rate and cause seasonal fluctuations in plant performauce.
The best estimates of treatment efficiency to be expected under real
conditions will be based on mass balance data from existing plants treating
similar wastes. Air emissions data are rarely available, however, and the
actual destruction efficiency of the treatment plant generally must be esti-
mated by calculating expected air emissions based on effluent concentration
data. Where specific organic compounds are present in the effluent in concen-
trations below the detection limit, this calculation is not possible.
The application .of laboratory and field data to emissions estimation is
imprecise because the laboratory data are obtained under conditions which may
not accurately reflect the varying and uncontrolled nature of full-scale
systems. Data obtained from larger scale systems may be inaccurate due to
difficulty in sampling. Deviations from ideal flow (e.g., completely mixed
stirred tank) may result in errors due to nonrepresentative sampling. Real
systems may have influent flow rates which vary; calculation of residence times
associated with particular effluent samples may be subject to uncertainty.
Notwithstanding the uncertainties inherent in the bioxidation data abstrac-
ted from the literature, these data are the best alternative for use in predic-
tive emissions models if no full-scale system using similar waste in a similar
flow scheme is available. Where no data are available for a particular compound,
the approach of Strier (1982, 1985) can be used to obtain an estimate of the
relative importance of biological oxidation, air stripping", and adsorp'tion on
sludge in a wastewater treatment plant. A description of .this approach-ic
given in the section titled Predictive Fate Models.
-------
TABU! 1J. BIQOXFDATIOS JUTES J'K«W STA OAU EASE
t'captf*wi*«t
( t h y 1 •< r t 4 1 *
I , ?-IM«'h t«f ti»f fc>n4>
Hk*o*?l
l,J-l»ltUI..|.*. ...•«»•
C«oj»n«
frrtolft
Mtrof.t.ttn*
(JlrMor »?j>»i*f*»l
At ry lunllrl It*
H«(hyl*iir rhlurldff
i'llmnul
lotlliJ
'*'.
.'***
l«
I.WO
.»
110
ion
»;
I8J
litn
JOO
1.000
loo
8 1 nio.1 A a
n 4 oJ*
I.UW*
. .«:«*
j'.m
i!»wy*
I.PWI*
% r >w%
1 . ?'.«
I.WHJ*
LOW"
i.wo"
l.ixw"
I.(HK)
1 . /*<)
I.WJO
*3Bt^'«-v
: i ' j*1
),' Mb
*.» ll>
10 l«.h
& t i J
V V .J *
J* «*•
d.» •.»"
19 "it*
2) '>>"
J? '>Sb
1.0 1.!.
1? Ifl
i.» 12
ii :i
BT MU,W
tl t'» »l And (' | W AIMXW ,
it c**'!'! OS^4 it(*'»|'e*rr «»»4 I'lnr^np^o,
BatlaV t J *h«itt <-t «| . * IVi 7
t-ii-vrr *nd *lfit«nnu;i.
tit *»vr I »t»
-------
TABLE 13 (continued)
Compound
Formaldehyde
Acetic ncid
Oxalic acid
Trinitrotoluene
n-Propanol
1-Propanol
n-Butanol
i-Butanol
2-B»tonol
t-Butanol
Amlnobenzolc acid
Clilorolicnzolc acid
Formic acid
Valeric acid
Caprnlc acid
lleptanolc acid
Benzole acid
Methyl cellulose
Pentnchlornphenol
Aniline
nimetliylanUlne
Chloronnl 1 ine
N,N-dlmethyl-l-
naphty lamlne .
Initial
concentration
mg/L
720
600
200
1 , 500
2,000
25
500
500
500
500
500
500
20
20
200
1,500
200
200
200
200
70
30
20
20
20
20
B 1 omn s s
concentration
">g/1.
2,600
2,500
350
1,440
3,500
,180
,000
,000
, 000
,000
,000
170
30
30
350
1,440
350
350
350
350
5,000
2 , 000
10,000
10,000
10,000
10,000
Biodepradation
mg compound/g *ir mg
0
4.0
>1.5
88
36
28
12
11
10
10
12
2.6
220
82
>1.5
26
>1.5
>1.5
>1.5
>1.5
0.03
(1.85
0.18
0.10
0.07
0.02
Rate
ThOU/'g hr
0
43
>l.6
94
4.6
3.8
29
26
26
26
31
6.7
475
121
>0.5
9.0
>3.0
>3.3
>3.5
>3.0
0.04
n.w
0.37
0.32
0.15
0 . 06
Reference
Placek el nl., 1947
Pl.icek et al. , 1947
Englebrecht et a] .,
1957
Sato et nl., 1972
Nelson et al. , 1954
Nay et al., 1974
McKlnney et al . , 1955
McKinney et al. , 1955
McKlnney et al., 1955
McKinney et al. , 1955
McKtnney et al., 1955
McKlnney et al . , 1955
Gerlke et al., 1980
Gerlke et al., 1980
F.nglebreclit ct n].,.
1957
Sato ct nl. , 1972
Englebrecht et al.,
1957
Englebrecht et al . ,
1957
Englebrji-ht et al. ,
1957
F.ngK'-brechl et al.,* .
1957 -.' '
Blnnchnrd ct al. , 1976
Klrsli et nl., 1973
B.iird et nl . , 1977
Halrd et al., 1977
llnlrd et a] . , 1977
Baird et al. , 1977
Assumed vnluo.
Actual measured COD value.
-------
MASS TRANSFER IN AERATED WASTEWATER TREATMENT SYSTEMS •
This section describes various equations used to estimate water-to-air
mass transfer coefficients. Alternative equations are available in several
situations (K. for lagoons and settling basins; K for windswept sources).
When alternative equations are available, the advantages and limitations are
discussed. The various equations are summarized in Appendix A.
Available Equations
Hwang (1982) published an article describing reliable quantitative methods
of estimating toxic emissions from land disposal facilities. This article
provides many useful equations to estimate mass transfer coefficients. The
following paragraphs describe some of the recommendations of this paper.
The overall rate of mass transfer, Km is predicted by combining the
UJj
resistance of the liquid and the resistance of the gas.
KOL = (1/kL * 1/CKkg))'1 (4)
where
kT = the mass transfer coefficient for the liquid, g. mol/cm2-sec
LJ
k = the mass transfer coefficient for the gas, g. mol/cm2-sec
O
K = the partition coefficient, mole fraction gas/mole fraction
liquid
K_T = the overall mass transfer coefficient based on liquid concen-
trations, g. mol/cm2-sec
This equation is valid for each of the aerated waste treatment processes
previously described, with the exception of multiphase resistances (oil slick,
surfactant layer). The resistance of an oil layer is relatively easy to
predict, but in *.he case of volatile organic compound (VOC) release, the
effect of surfactants is less well known. Effects such as electrostatic
repulsion cannot be predicted without experimental data.
A modified equation for the overall mass transfer coefficient can be
written as a modified form of Equation 3.
)] + [l/(Kkg)]3 (5)
45
-------
where
6 = thickness of film, cm
D_ = diffusion coefficient of VOC in oil, cm2/sec
p = density of oil, g/cni3
MW = average molecular weight of oil, g/gmol
Kf. = film-liquid partition coefficient.
Equation 3 can also be modified to account for area modifying controls such as
floating balls, etc. In any case, the equation for the controlled emissions
should be a modified version of the uncontrolled emissions equation so that a
common basis can be used for evaluating the control effectiveness.
Liquid Mass Transfer
The removal rate of VOCs in a flowing body of water after Cadena, et al.,
(1984) follows the equation:
Kref n
C _ v , OL _ UVOC . .
ln r -Kzt K~ ~K2t on (6)
o OL
where
K2 = the oxygen reaeration rate constant in base e, sec l
C - the concentration of interest, mg/L
C = the initial concentration (at t = 0), mg/L
t = time, sec
DV__ = the VOC diffusion coefficient, cm2/sec
ref
K..T = the overall mass transfer coefficient for a tracer,
OL , / o
g-mol/cm^-sec.
According to information presented by Metcalf & Eddy, (1972), rapids and
waterfalls would have a K2 value of >1.I/day. This would be similar to rates
of mass transfer expected with some biological processes. Cadena et al.
(1984) suggest a range of K2 of 3.36 to 47.0 day * for rapid shallow streams
(one foot deep). The value of K2 is relat.-'d to the depth of the stream by the
46
-------
power of 1.67 (Churchill, 1962) or 1.85 (Owens, 1964). Taking an intermediate
value reaeration rate constants can be adjusted for different depths as follows:
v ref h 1.7
K|_=(-pl) (7)
where .
h = stream depth, cm.
h _ = depth of reference stream, cm
The 30 cm (1 foot) depth of the shallow stream would suggest (by this relation-
ship) that 300 cm (10 feet) deep streams would have lower values of Kg by a
factor of 50. This suggests that the values of Catiena are not inconsistent
with those of Metcalf & Eddy (1972). For rapid agitation in a biotreatment
system a value of K- of 17 day 1 is not unreasonable.
The valut. of the reaeration rate constant can also be estimated by a
correlation of Owens et al. (1964) for stream flow velocities of 3-600 cm/sec.
K2 = 50.5v°-6V-85 (8)
where
K2 = base e reaeration constant, hr
v = velocity, cm/sec.
If a holding basin has a depth of 3 meters (10 ft) and a length of 10
meters (30 ft), with a residence time of 1.0 hour, the velocity would equal
0.28 cm/sec. This is less than the range of the correlation, and therefore,
the estimate of K2 is subject to increased uncertainty. Nonetheless, substi-
tuting into the above equation, the value of K2 would equal 3.6(10 2) day 1
Using the equation of O'Connor (1958) for a stream
K2 = 12.9 v°-5/h1>5 (9)
where
v = velocity, ft/sec.
h = depth, ft.
The predicted K2 for the holding basin would equal 0.038 day 1.
These equations for flowing streams are not directly applicable to holding
.basins because the average velocity is less than the applicable range of the
velocity in the equations, and the turbulence (due to the mixing of the influ-
ent) is not accounted for. The reaeration rate constant for holding basins
would likely fall in the range between that of a sluggish stream (0.23/day)
47
-------
and that of a swift stream (1.15/day). A value appropriate for holding basins
could be on the order of 0.5/day, which has been assumed for comparison pur-
poses.
A typical mass transfer coefficient for benzene is approximately 7 x 10
gmol/cm2-sec (see Table 17). Based upon a depth of 300 cm and a water molar
density of 1/18 gmol/cm3, the time constant of VOC decay frcni a surface impound-
ment would be 3 days. This value is not inconsistent with the stream reaera-
tion rate constants (0.36/day vs. 0.5/day).
Another method of estimating liquid mass transfer coefficients is based
upon wind friction. Mackay and Yeun (1983) suggest that the liquid mass
transfer coefficient can be written as a function of the liquid velocity and
the Schmidt number, where
where
k' = 1.0(10"6)'+ O.Ol44(U*)2-2(ScT)~0'5 (10)
LI LI
I", = liquid mass transfer coefficient, m/sec
U-'; = (U10)(6.1 + 0.63Ul0)°-5/100
UK) = air velocity 10 m above the surface, m/sec
Sc, = liquid Schmidt number, dimensionless.
Equation (9) is valid where U» < 0.3. Where U» > 0.3, McKay and Yeun (1983)
suggest,
k' = 1.0(10"6) + 0.00341U*(ScT)~°'5 (11)
LI LI
Assuming a wind velocity of 447 cm/sec (equivalent to 10 miles/hr),
ScT = 1,000, U* = 0.133 and IL = 6.4(10" ) m/s-?c. The time constant for loss
from the liquid would be estimated as:
0.0000064 m/sec 3,600 sec _ n -1
O * . .\ , — u.uu//nr .'
meters depth) hr
This calculated value of the time constant is equivalent to a mass transfer
coefficient of 3.5 x 10 5 g-mol/cra2-sec and is somewhat greater than the
coefficients predicted by the correlations of Owens &nd Edwards (Table 17,
1.5 x 10~ g-mol/cm2-sec).
The estimated rate of volatiles loss from the relationship of Mackay
(1983) is lower than that predicted from Chuirhill (1962) based on liquid flow
only. It should be noted that the functionality (and therefore the predictions)
of the equations of Churchill and Mackay are different. The absence of a
functionality of depth of liquid may limit the applicability of the equation
of Mackay for relatively deep bodies of water.
48
-------
In some wastewater treatment facilities impellers or other devices are
used to agitate the liquid surface. The entire tank may be agitated, or
agitation may be carried out on only one portion of a body of Wat-i:. A correla-
tion for the mass transfer coefficient under agitation was presented by Thibodeaux
(1973) whare the mass transfer coefficient was proportional to the power of
agitation and the square root of the ratio of the diffusion coefficient of the
specieb to that of oxygen (Equation B, Appendix A).
In the case of liquid mass transfer without agitation from lagoons and
other related facilities, the equation of Owens et al., (1964). and Gibbs
(1964) can be useful (Equation C, Appendix A). The important variables in the
Owens correlation are temperature, depth of the liquid, surface velocity, and
the diffusion coefficient of the component of interest. The surface velocity
is assumed to be the product of 0.035 and the wind velocity for a nonaerated
surface or 0.03 m/s outside the effective aeration region for an aerated
process. Values of liquid mass transfer coefficients calculated from the
correlation of Owens, et al., (1964) are higher than those calculated from the
correlation of MacKay and Yeun (1983).
Cohen, Cocchlo, and MacKay (.1978) investigated the volatilization of
benzene from water in the presence of wind waves. The mass transfer coeffi-
cient is a function of the roughness Reynolds number, which in turn is a
function of the wind velocity and other parameters (Equation A, Appendix A).
The correlation is based on benzene in an unagitated system. The ratio
of the diffusion coefficients between a species of interest and benzene is an
important parameter which is used to extend their results to otner components.
Based on the work of Roberts.am! Dandiiker (1983), the volatilization rate is
proportional to thr. ratio of the diffusion coefficients to the 0.66 power.
.Although Cohen, Cocchlo, and MacK.ny (1978) discuss the theoretical basis for
their correlation, it should be regarded as primarily empirical with some
reservations as to its applicability to larger systems, such as lagoons.
The velocity of the surface away from the aeration region can be estimated
by the correlation of Kullenberg (1976).
v = 0.3048 (3.2809 U/100)(l.506 + 1.5748 II) (12)
where
U = the velocity of the wind, m/s
v = the bulk velocity of the liquid, m/s
-------
Mass Transfer from Falling Droplets
The air emissions from a falling droplet are of potential importance from
a variety of wastewater treatment sources. Some important aspects of droplet
emissions are (1) the mass transfer coefficients are substantially greater in
falling droplets than from surfaces and (2) some of the droplets may never
reach the earth. As the droplet loses water by evaporation, its velocity
decreases. Eventually some of the droplets may evaporate to dryness, forcing
all of the pollutants (VOCs , seraivolatiles, non-volatiles , inorganics) to be
emitted into the atmosphere, either as vapors or aerosols. There will be very
little biological decay during the short time of droplet fall. The sources
which produce falling droplets include roughing filters, spray ponds, and
cooling towers.
Lenard (1921) reported that a droplet could not fall faster than 8 m per
second and this 'maximum droplet velocity is known as the terminal velocity.
The droplet forms an inverted cup and breaks up at the higher velocities. A
free falling droplet (without wind resistance) would require approximately 0.8
seconds to achieve this velocity. The travel distance would be 3-i oeters , a
distance comparable to that present in many wastewater treatment systems
employing spray ponds.
The rate of droplet fall can be estimated by means of a general equation
for the terminal velocity, v of spheres (McCabe and Smith, 1967).
4 A B1+" (p -p) 1/(2-°'
where ?
A^ = (Re J2~n - l- - — (14)
e P 4 p(pp-p)
The values of bj and n depend on the values of the particle Reynolds
number, Re , which is calculated as:
P Dvp
50
-------
where
D = the particle diameter, cm
p = density of air, g/cm3
p = density of droplet, g/cm3
(J = viscosity of the air, g/cm-sec.
The necessary constants are given in Table 14 for three specific ranges
of Reynolds numbers. The Reynolds number is calculated iteratively on the
basis of an assumed velocity. The velocity is then calculated and a corrected
Reynolds number is obtained.
The overall mass transfer coefficient for a falling drop can be calculated
(Treybal, 1968) as,
KOL = ShavCDAB/Dp
where
D.R = diffusion coefficient of the compound in water, cm2/ sec
C = concentration, gmol/cni3
and
where
Sc = Schmidt number = p/(pD _)
and
In cases where,
ScgD3p2/!J2 > 108
Equation (18) should be used instead of Equation (17):
where
Gr = Grashof number = (980 cm/sec2)D3p2/M2.
= ShQ + 0.347[Re (Sc)0'5]0'62 (17)
Sh = 2.0 + 0.569(GrSc)°-25° (18)
o
Sh = 2.0 + 0.0254(Gr.)0'333(Sc)0-577 (19)
o
51
-------
TABLE 14. CONSTANTS FOR DRAG COEFFICIENT CALCULATION
Re range bj n
Re <2 2.4 1
P
2
-------
This equation is similar to the relationship proposed for mass transfer from a
sphere in a packed bed (Wakao and Fanayku, 1978). The velocity due to porosity
of the bed would account for some of the differences in the two equations.
Sh = 2 + 1.1 (Re Sc'5)°'6 (20)
a V p
When the velocity of a falling drop can be estimated, the mass transfer
coefficient can be estimated. Both mass transfer for water and VOCs should be
calculated, since the droplet can evaporate to dryness before reaching the
earth. Treybal (1968) states that spray ponds are subject to high windage
losses of water. Cooling towers have mist eliminators which can reduce some
of the windage.
Table 15 presents the results of a numerical simulation of the VOC loss
from a falling droplet using the equation suggested by Treybal (1968). Small
droplets readily lose their VOG components into the atmosphere and larger
droplets lose significant amounts of VOCs over a relatively short distance of
travel. In addition, the mass transfer of water away from the droplet surface
is relatively rapid, with small droplets evaporating to dryness after several
meters fall (see Table 16).
Thin Falling Films
Evaporation from cooling towers occurs from a thin falling film of liquid.
The film Reynolds number is defined as:
Ref = 4F/p
where .
F = the mass flow rate of liquid per unit film width, g/cm-sec.
The film thickness, 6 (cm), can be determined as follows:
where ,
g = gravitational constant =980 cm/sec2.
The overall mass transfer coefficient (for Re,. > 100) is
577 = Shav = <|* I ReSc * (22>
AB
53
-------
TABLE 15. FRACTION OF BENZENE REMAINING IN A DROPLET
FALLING AT TERMINAL VELOCITY
Falling distance (m)
0
0.21
0.68
1.26
1.60
2.03
2.67
TABLE 16. TOTAL DISTANCE
Initial diameter (cm)
Initial droplet
0 . 1 cm 0 . 2 cm
0.04
0.47
0.21
0.18
0.10
FALL BEFORE EVAPORATION
Distance (cm)
diameter
0.5 cm
0.80 -
0.71
0.68
0.60
0,53
TO DRYNESS
0.05 350
0.10 1500
54
-------
where
L = length of column (or section thereof), cm.
For Rer < 100, Sh Zl.Ul.
f av
Mass transfer in falling films, such as those present when wastewater
flows over a clarifier weir, can be considered by the method of Owens (1964),
modified for 90 percent less turbulence in free fall as opposed to continuous
shear against a vertical surface. An example is presented in Appendix B, part
Illc. The gas phase mass transfer is estimated by the correlation of MacKay
and Yeun (1983). The falling film is considered as plug flow (well-mixed
perpendicular to the direction of fall). This approach can be considered a
semi-empirical one, since the value of the multiplier applied no the correla-
tion of Owens, (0.1) can be adjusted as additional experimental data are
obtained.
Gas Mass Transfer
The classic approach to predictions of mass transfer coefficients, (Sher-
wood & Pigford, 1952), is presented in Equation H for comparison purposes.
The exponent of the diffusion coefficient in Equation H (Appendix A) agrees
roughly with the recent work of Roberts and Dancliker (1983); this suggests
the possiblity that the mass transfer mechanisms are similar, whether they are
for the turbulent gas-phase or the turbulent liquid-ph-ise transfer processes.
It is also interesting to note that the classical methods predict mass transfer
coefficients that are similar to those observed in recent research. It should
be noted however, that the exponents in some of the recent correlations differ
from those used in the Colburn equation (Sherwood and Pigford, 1952).
Equation I (Appendix A), from Thibodeaux and Parker (1974), predicts the
gas-phase mass transfer coefficients from surface impoundments where the gas
mass transfer coefficient is •» function of the wind velocity, the diffusion
coefficient, and the length of the impoundment. For the case where the liquid
surface is agitated, the gas-phase mass transfer coefficient can be estimated
from the correlation of Rinehart (1977). The gas mass transfer coefficients
are a function of the system geometry and various dimensionless numbers describ-
ing it.
55
-------
In general the more recent correlations are similar to the classic equa-
tion (Sherwood and Pigford, 1952). Considering the thickness of the boundary
layer in turbulent flow in circular conduits, the classical equations for
turbulent flow should be applicable to planar systems; moreover, some of the
more recent correlations lack the extensive experimental support that the
classical equations have. It is recommended that the Colburn type equation
(Equation H, Appendix A) be used for the gas phase transfer from surfaces.
Mackay and Leinonen (1975) concluded that the mass transfer resistance of
the gas phase is negligible for evaporation of volatiles. For semivolatiles
(H < 0.0001 atro M3/mol) the gas phase mass transfer resistance can be significant.
Their conclusions were based upon reported values of the gas phase mass transfer
coefficient of 3000 cm/hr (O.G5 g mol/cm2-sec) . This is substantially greater
than the liquid phase mass transfer coefficient in unagitated systems, and for
situations where K > 1, the liquid film resistance would likely control the
rate of mass transfer.
Mackay and Yeun (1983) recommend the following gas phase mass transfer
coefficient.
k* (m/s) = 0.001 + 0.000462 U* Sc ~°-67 (23)
& &
where
U* = (6.1 + 0.63 U10)°-5 U10/100
= wind velocity at 10 meters, m/sec
Sc = gas Schmidt number [viscosity/ (density x diffusion coefficient)]
o
This correlation predicts lower mass transfer rates than the Reinhart
equation (Equation J, Appendix A). The Reinhart equation should predict rates
higher than those from the Macl:ay equation since there should be greater mass
transfer associated with splashing and mechanical agitation than from wind
agitation alone. Thus, the equation of MacKay and Yeun (1983) is recommended
for wind agitated systems while the Reinhart equation is preferred for mechani-
cally agitated systems.
Table 17 presents a summary of the results from using various equations
to predict the mass transfer coefficients from a clarifier. In general, the
more recent correlations show close agreement for both gas and liquid phase
resistances.
56
-------
TABLE 17, A COMPARISON OF SELECTF-0 PREDICTION'S OF MASS
TRANSFER COSFFICIFJ.TS FOH BCSZEXE EMISSIONS FROM A CLARIF1ER
(Model Plant 1. Wind Velocity 447 cs/t>*c)
Gas phase coefficients |g eol/(cs2 sec)}
"J factor" type (Sherwood aed Pigtord. 1952)
Hackay and Matsuga (5973)
Thibodeauit ,in4 Parbrr t!<*74)
Mackay and Yeun (1983)1
Reinhart (1977)2
fg
Ovens, Ktiuattic, and Cibbs (i'564)
Cohen, Corchlo, an«l flackay n• 10
3.J * IO'S
1.7 * lo"5
Z.2 ^ 10*^
18.0 x |Q
*5
fc.J
10
7.2 '" 10
4.9
3.S
5.0 ^
*
*
-------
Subsurface Aeration
The mass transfer into a bubble is often viewed from the perspective of
the liquid: the change in concentration of the liquid is modeled. From the
perspective of air emissions, the concentration in the off-gas is of importance.
Since a close approach to equilibrium is expected, the overall mass transfer
coefficient for the liquid cannot be used to calculate the bubble concentration
.under a variety of conditions. • .
As bubbles rise through wastewater containing VOCs, the rate controlling
mechanism for the flux of VOCs into the bubble is controlled by the liquid
mass transfer resistance. The liquid mass transfer coefficient, K, is pre-
dicted by the relationship of Calderbank (1967).
K.(Sc)0'5 = 0.42 v°.33 (24)
Lt all
Simulation of bubble rise using the above relationship indicates that the mass
transfer is controlled by the liquid phase resistance and that equilibrium is
achieved rather quickly.
This theoretical prediction was confirmed experimentally. Smith (1980)
has shown that the rate of loss of VOC is controlled by the liquid phase
resistance, except for low volatility compounds. Truong and Blackburn (1984)
demonstrated that the rate of stripping in a dispersed bubble system could be
predicted closely by equilibrium considerations: that is, the bubble system
was approximately equilibrium controlled, not mass transfer controlled. Their
equation was of the following form:
where
KOLa V/Q = °-00371 H (25)
Knla = stripping rate constant (hr-1)
UL - -
V = volume of liquid (L)
Q = air flow rate (L/hr)
H = Henry's law constant (torr L/g-mole)
This relationship is based on toluene, MEK, 1,4-dichlorobenzene, and
phenol in water, so the data encompasses semi-volatiles as well as volatiles.
In addition, the predicted kinetics agree well with various reported literature
data (Truong and Blackburn, 1984) further substantiating the concept that
dispersed bubble strippers are an equilibrium process.
58
-------
The diameter of a bubble produced by a subsurface aerator is a function
of the power input to the aerator. Using Calderbank'.s correlation (Freeman,
1981):
Dfe = [4.15(ST)°'6 (PI/V)'0'4 (PL)"0'2 ghfl + 0.09 (26)
where
D, = bubble diameter, cm
ST = surface tension, g/cm
PI/V = power input per basic volume, g-cm2/(cm3-sec)
g, .. = gas holdup fraction = v /v,
v = superficial gas velocity, cm/sec
6 . .
v, = actual bubble velocity, cm/sec.
The ratio of the stripping rate constant of volatile hydrocarbons and
chlorinated hydrocarbons to the oxygen reaeration rate constant in a bubble
system was found to be proportional to the square root of the ratio of the
critical volume of oxygen to the critical volume of the volatile compound.
For more spherical molecules, the ratio of stripping rates was proportional to
the cube root (Matter- Muller, 1980). The system used in this study may have
been small enough for the process to be kinetically controlled.
The rate of stripping is generally equilibrium controlled except for very
highly volatile compounds; the gas in the rising bubbles would be saturated
with the VOCs in the water. The rate of loss of VOCs into the atmosphere can
be predicted by a relationship such as that recommended by Blackburn and
Truong (1984). The major data required would be the VOC concentration in the
aerated water, the gas flow rate, and the equilibrium partition constant
(Henry's constant).
Although most compounds are expected to reach equilibrium within the
bubbles before the bubbles reach the surface, the concentrations of highly
volatile materials in the bubbles such as vinyl chloride and oxygen are expected
to be kinetically controlled. The mole fraction of these compounds in the gas
phase is large compared to the mole fraction in the liquid phase. Therefore,
the liquid phase resistance and bubble residence time in the liquid determine
the concentration. For these volatile compounds, the concentration can be
59
-------
estimated by specifying the bubble size, calculating the rate cf bubble rise,
estimating the liquid mass transfer coefficient with an appropriate correlation
(Calderbank, 1967), and numerically integrating to obtain the concentration
within the bubble as a function of time. Providing smaller bubbles and a
longer rise distance promotes the approach to equilibrium.
Mass transfer in a bubble column was investigated by Munz (1982) and the
mass transfer was related to the Henry's constant for a series of VOCs. The
overall mass transfer coefficient for stripping (of five chlorinated hydrocar-
bons) was found to be linearly related to the overall mass transfer coefficient
for oxygen absorption over a wide range of operating conditions.
Roberts and Dankliker (1983) caution against using the assumption of
liquid phase resistance control in laboratory studies. They suggest that the
gas phase resistance may be important based on laboratory studies with chloro-
form, and that the gas phase resistance may be important in mechanically
agitated liquids.
The data of Truong and Blackburn (1984) agree well with the concept of
equilibrium being established rapidly in bubble columns. In a benzene strip-
ping experiment they observed a stripping rate constant of 1.13 hr 1 with 2
L/min air flow through a liquid volume of 25.7 liters. The corresponding
value of H is calculated as follows:
(25.7 L liquid)'(1000 g/L)(24.0 L/mol)(1.13/hr)
(2 L gas/min)(60 min/hr)(18 g H20/mol)
323
This compares to a reported value of Henry's constant for benzene, of 295
(Freeman, 1982). It is interesting to note that surfactants and biomass tend
to depress the Henry's constant for toluene, but for 1,4-dichlorobenzene, the
addition of biomass and surfactants did not depress the Henry's constant.
Table 18 illustrates the rapid approach to equilibrium of air bubbles
rising in water. Bubbles produced by fine diffusers are on the order of 0.1
cm in diameter. Except perhaps for large bubbles in laboratory equipment,
equilibrium can be assumed for VOC absorption from water. The same assumption
for oxygen uptake by the water is not necessarily valid. Small bubbles sparged
into the water would theoretically tend to maximize oxygen absorption while
controlling VOC loss.
60
-------
TABLE 18. CONCENTRATION OF BENZENE ABSORBED DURING BUBBLE RISE IN
WATER (mole fraction)
Bubble size (cm)
0.1
Concentration, Equilibrium 0.03
Concentration (0 . 1
Rise distance (cm)
Concentration (0.2
Rise distance (cm)
Concentration (0.4
Rise distance (cm)
Concentration (1.0
Rise distance (cm)
sec) 0.0236
at 0.1 sec 1. 1
sec) 0.0286
at 0.2 sec 2.22
see)
at 0.4 sec
sec)
at 1.0 sec
0.3 0.6
0.03 0.03
0.012 0.0068
3.0 4.2
0.0193 0.012
6.0 8.4
0.0261 0.0193
12 16.86
0.0277
42.1
61
-------
ADSORPTION
Organic compounds can be removed from wastewater by adsorption on surfaces.
Potentially adsorptive surfaces include suspended solids present in the waste-
water, soil in unlined surface impoundments, polymeric liners in lined surface
.impoundments, tank, walls of steel or concrete, and bicmass generated from
decomposition of organic compounds.
With the exception of the suspended solids and suspended biomass, adsorp-
tive surfaces will rapidly reach equilibrium with wastewater of constant
composition. Once the surfaces are "loaded," no further net adsorption will
take place. The biomass or secondary sludge stream removed from the treatment
system may be enriched in VOCs with respect to the effluent wastewater. In an
activated sludge system, a waste sludge stream that is about 1 percent dry
solids is removed from the process. Upon dewatering, the sludge may become
more concentrated in certain organics than the water which was removed from
it. '
In absolute terms, the importance of adsorption on sludge as a pathway
for removal of VOCs from the system is decreased by the relatively low percent-
age of total solids in the sludge. About 99 percent of the mass of the sludge
is water; the VOC content of this water is approximately that of the bulk
liquid (in a completely mixed system this concentration is approximatley that
of the effluent). Thus, if the solids in the sludge were enriched 100-fold in
VOCs with respect to the wastewater, the resulting sludge stream would still
have a VOC content of only about twice that of the solids-free wastewater.
Estimation of the partition coefficient of VOCs between sludge solids and
water can be based on the octanol-water partition coefficients for specific
compounds as reported by Hansch and Leo (1979). Matter-Muller et al. (1980)
determined that a linear relationship exists between the log of the octanol-
water partition coefficient and the log of the distribution coefficient between
activated sludge solids and the aqueous phase. This relationship was observed
for chlorinated aromatic compounds, but on a theoretical basis it implies that
the sludge solids act similarly to an immiscible organic phase in equilibrium
with water. In the absence cf data from direct measurements for the specific
compounds of interest, it may be assumed that the following relationship,
described by Matter-Muller et al. (1980), is valid:
62
-------
log K = 0.67 log P - 2.61 (27)
s,y ow
where
K = concentration of VOCs in sludge solids/concentration of VOCs in
s'y wastewater [ (mg/kg)/(mg/:u3)]
P = concentration of VOCs in octanol/concentration of VOCs in water.
ow
PREDICTIVE FATE MODELS
The removal mechanisms can be divided into three pathways: (1) adsorption,
(2) evaporation or air-stripping, and (3) chemical and biological oxidation.
Chemical and biological oxidation is treated as one pathway due to the difficul-
ties in experimentally differentiating between the two mechanisms.
Structure-activity correlations have been developed by Strier (1983,
1985) in an effort to predict the relative importance of each removal pathway
in activated sludge systems. By grouping compounds according to their ability
to adsorb on activated carbon, and considering their solubilities, molecular
weights, chemical structure and other similarities, Strier formed six "clusters"
of compounds. Table 19 lists examples of chemicals found in each cluster.
Table 20 gives a general description of the chemical and physical properties
that were used as a basis for the gro iping of each cluster.
By using data obtained from field tests at publicly owned treatment works
and pilot studies conducted at EPA, Strier developed empirical equations to
predict the percent removal of an organic compound by adsorption and by air-
stripping in a well-aerated activated sludge treatment system. These equations
are presented in Table 21. The appropriate equation to use for each cluster
is also outlined in Table 21, though there is some question as to which air-.
stripping equation to use for Cluster V compounds. For those compounds in
Cluster V which are not particularly amenable to biodegradatlon or have a
high Henry's law constant (i.e., greater or equal to the Henry's law coeffi-
cient, II, of benzene), Equation 3 should be used. For those Cluster V com-
pounds which are biodegradable and/or have a moderate-to-low Henry's law
constant (i.e., H < H of benzene), Equation 4 should be used. Obviously, the
most important parameters are the log octanol-water partition coefficient (log
P ) and the Henry's law constant. Even without % metabolized data, Equation
A could still be used to estimate the % removal by air-stripping since the
dependence of Equation 4 on the % metabolized term is small.
63
-------
TABLE 19. CLUSTERS:
EXAMPLES OF THEIR COMPOUNDS AND FATE DESCRIPTION
(Strier, 1982)
Cluster
Example compounds
Mol. wt.
Fate description
I Hexachlorobenzene
2,3,7,8-TCDD
PCBs
284.8
332
275-420
These are among the most incompatible and least soluble
organic pollutants. They are remc/ed to at least 70% in
primary and secondary sludges, with little air emissions. .
II . N-Nitrosodimethylamine 74.1
1,2-Benzanthracene 228.3
Pyrene 202
III 1,2,4-Trichlorobenzene 181.5
Hexachlorocyclopentadiene 273
Anthracene 178.3
Fluorene 152.2
IV Benzidine 184.2
1,4-Dichlorobenzene 147
Hexachlorobutadiene 260.8
4,6-Dinitro-O-cresoi 198.1
V Benzene 78.1
Carbon tetrachloride 153.4
Chlorobenzene 112.6
1,1,2,2-Tetrachloroethane 167.9
1,2- & 1,3-Dichlorobenzene 147
Nitrobenzene . 123.1
Phenol 94.1
Tetrachloroethylene 165.8
VI Acrolein 56.1
1,2- & 1,1-Dichloroethane 99.0
1,1,1-Trichloroethane 133.4
1,1,2-Trichloroethane 119.2
Chloroform 119.2
Methylene chloride 84.9
Trichloroethylene 131.4
Vinyl chloride 62.5
These compounds are essentially nonbiodegradable with little
air emissions. They are adsorbed by primary and secondary
sludges to the extent of 70%.
These compounds are biodegradable up to 50%. They are
adsorbed up to 50-60% by combined priirary and secondary
sludges. There is little air stripping of these compounds,
except for hexa~hlorocyclopentadiene, which is potentially
highly air-slrippable.
These compounds are mostly biodegraded with the exception
of the highly chlorinated compounds. Removal by adsorption
on combined primary and secondary sludges is no better than
30%. The butadiene is capable of being highly air-stripped
and is also highly chemically reactive.
These compounds are adsorbed on sludges by less than 25%.
The highly chlorinated compounds in this cluster will be
air-stripped. The phenols will be highly biodegraded,
while the benzenes will be preferentially air-stripped.
Compounds in this cluster are essentially not adsorbed on
primary or secondary sludges. These compounds are essen-
tially not biodegraded, except for acrolein. Compounds in
this cluster are removed primarily by air-stripping.
-------
TABLE 20. CLUSTER CHARACTERISTICS USED AS A BASIS IN GROUPING COMPOUNDS
(Strier, 1982)
Cluster Cluster, characteristics
I ETEL 1.0 MS/1; water insoluble; log P range of 5.6-7.7; high
molecular weights (250-350); nonbiodegradable; nonvolatile;
melting point range of 440-570 °K.
II ETEL =1-10 (Jg/1; low water solubility; generally lower molecular
weights and melting points than cluster I; nonbiodegradable;
nonvolatile.
Ill ETEL = 10 |Jg/l; increasing solubility and biological activity
over cluster II; log P range of 3.0-5.3; mid-range molecular
weights (150-275); mostly low volatilities.
IV ETEL = 25 |Jg/l; increased water solubility: log P range of
1.5-3.5; typically biodegradable; mid-range molecular weights
(150-260); high boiling points/low volatility.
V ETEL = 50 |Jg/l; log P^ range of 1.5-3.0; mostly liquids at
ambient temperatures;"molecular weight range fo 80-1801 typically
biodegradable; lower boiler points and increased volatility than
cluster IV.
VI ETEL = 100 - 1,000 |jg/l; low log PQ (0-2.5); low molecular
weights (50-170); only slightly biodegradable; high Henry's law
constant/volatile/low boiling point.
a
ETEL is the estimated effluent concentration, assuming an influent concentra-
tion of 1 mg/1, from an activated carbon column. This is the primary basis
for developing the clusters.
65
-------
TABLE 21. EQUATIONS FOR ESTIMATING FATE OF ORGANIC PRIORITY POLLUTANT COMPOUNDS
(Strier, 1985)
Function For cluster(s) Equation
1. % Adr. vs. log P I-IV if Eqn. 1 % Ads. = 14.1 log P
°W % Ads. >70%
. 2. % Ads. vs. MW + MP + log P I-VI if Eqn. 1 % Ads. = 28.2 - 0.01 MP + 0.05 MW
°W % Ads. <70% + 14.1 log P
ow
3. % Air-strip, vs. H + log P V-VI % Air-strip. = 8.8 + 10.8 H + 9.8 log P
4. % Air-strip, vs. H + log P III-V % Air-strip. = -6.1 + 1.5 H + 3.5 log P
+ % metabolized °W 0.07 % metabolized OW
A
o\ Abbreviations: MW = Molecular weight (g/mol)
<* MP = melting point (°K)
H = Henry's law constant (Atm-m3xi,000/mol)
P = Octanol-1/water partition coefficient
. ow i r
Refers to experimentally determined fraction of compound biodegraded.
-------
The predicted fate of compounds according to their clusters is given in
Table 22. Note that, in general, each cluster has its own characteristic
removal pathways. This allows useful generalizations to be drawn about a
compound's fate knowing only the cluster number to which it belongs. Further-
more, the ability to make more accurate predictions about the fate of an
organic pollutant using a limited number of known parameters and the mathemat-
ically simplistic equations of Table 21 is a great advantage of the structure-
activity correlations.
Although Strier does not present any equations which allow the prediction
of percent removal by biochemical oxidation, he does outline some useful
generalizations. Compounds which have intermediate solubilities (log P =
ow
1.5-3.5) are excellent candidates for biochemical oxidation. Compounds which
have a log P value greater than 3.5 tend to be insoluble and are readily
adsorbed on sludge. Compounds which have a log P value of less than 1.5 are
often volatile. Thus, compounds with a low log P and a high Henry's law
constant are typically air-stripped.
Even though a compound is soluble in water and nonvolatile, there is no
guarantee that it can or will be biodegraded. That is, there are chemical
structures and specifically chemical substituents that gre=>t-iy inhibit bio-
chemical oxidation. These substituents are electron-attracting in nature and
are referred to as electrophilic. Examples of these are the halogens (Cl, Br)
and nitro (N02) and nitrosamine (>N-N=0) groups. Their presence on an organic
compound reduces that compound's ability to be biodegraded. Further substitu-
tion of these electrophylic groups on an organic compound further reduces its
ability to biodegrade.
Conversely, presence of a nucleophylic or electron-donating substituent
will enhance biochemical oxidation. Examples of these functionsl groups are
hydroxyl (-OH), amino (-HN2), methyl (-CH3), and ethyl (-C2H5) groups. Unlike
the electrophylic substituents, further addition of a nucleophylic substituent
has little additional effect on biochemical oxidation rates. However, multiple
nucleophylic substituents on an organic compound typically enhance that com-
pound's reactivity. In the presence of both nucleophylic and electrophylic
functional groups, the compound is usually biodegradable or at least chemically
reactive.
67
-------
TABLE 22. PREDICTED FATE OF ORGANIC
COMPOUNDS ACCORDING TO CLUSTER NUMBER
(Strier, 1982)
Cluster
I
II
III
IV
V
VI
% sludge
<80
>70
50
30
0-25
0-5
% biodeg.
0
0
50
<70
0-96
>0
% air
0
0
0
0
0-100
100
68
-------
Even though the structure-activity correlations are simple and often
accurate, they are not without their limitations. As stated previously, no
equation was presented by Strier to determine the percent-removal by biochemical
oxidation. It was assumed that the rate of either adsorption or air-stripping
was much faster than that of biochemical oxidation. Although this is typically
the case, the structure-activity correlations may not be predictive when the
rate of biochemical oxidation competes with the rate of adsorption or air-strip-
ping. Furthermore, the equations presented were developed from data of well-
aerated activated sludge treatment systems. Consequently, due to the empirical
nature (or lack of theoretical backing) of the equations, they may be of
questionable validity for predicting the fate of organic pollutants in different
aerated treatment facilities (i.e., trickling filters, aerated lagoons, evapora-
tion ponds, or even activated sludge systems operating under different process
conditions than those at which the data had been collected and th. equations
fitted). Nonetheless, the structure-activity correlations presented give
remarkably accurate fate predictions for a number of organic compounds. Also,
Equation 4 can be used for Clusters III-VI to predict the decreased removal by
air-stripping under conditions of less aeration.
The equations presented by Strier (1982, 1985) can give good estimates on
the importance of air and adsorption removal mechanisms for well-aerated
activated sludge treatment facilities. This approach may also be employed to
predict pollutant fates for different aerated biological treatment systems,
however, this may necessitate the development of different empirical equation
coefficients.
69
-------
SECTION 5
MATHEMATICAL MODELS FOR AERATED WASTE TREATMENT
SELECTION OF MODELS
This section describes mathematical models which are recommended for
estimating instantaneous air emissions from aerated waste treatment processes.
Aerated waste treatment includes the following processes:
Lagoons
Clarifiers
Dissolved air floatation (DAF)
Surface aerators
Diffused air
Trickling filters
Agitated holding tanks
Spray ponds
Cooling towers
These aerated waste treatment processes are expected to encompass most of the
air emission sources associated with wastewater treatment technologies employed
at hazardous waste management facilities.
The models of the wastewater treatment processes can be formulated by
combining equations for each of the physical phenomena which control the
release of pollutants into the atmosphere. The equations for the significant
phenomena which govern the various mass transfer mechanisms include the follow-
ing:
A. Liquid phase resistance in a body of water
B. Liquid phase resistance in an agitated system
C. Liquid phase resistance in a flowing film
D. Liquid phase resistance in a droplet
E. Liquid phase resistance around a rising bubble
F. Gas phase resistance over a surface
G. Gas phase resistance over an agitated surface
H. Gas phase resistance over a falling droplet
I. Gas phase resistance in a rising bubble
J. Gas phase resistance over a flowing film
Parameters that are required for calculation of the mass transfer rate
include the following:
1. Surface area
2. Flow rate of liquid
3. Liquid volume
70
-------
4. Flov rate of gas
5. Equilibrium p^rtit iooi-»« into gas
6. Biological decay rates
7. Adsorption partitisming ioto
8. Residence tic*
9. Fractios of the cut face agitated
10. Practise of drops lest to drift.
The retired eta>d«l» for the aerated waste treatment processes iael«»<.!ogical l«> fur predicting tfi*1 flust
of volatilce troai all of the above AA«» tr<>a*fer i&?chanis«» ecrrpt Hhieh charactcrC'ae wa»t«? lre<*?.ttent
ies caa b<> assuscMi. estieaterf, or eeasured at specific facilities. Th«?
Ibriua pdftil toning into the vapor phase ran be estiadted within t 30
percent. The biological <5ecav fates *rp even tsaore variable and uucrrVjin.
Whea a proceed Ua* zeaea of distinct.ly dtf(er«&t POIRSIOQ ra'.e*. tl l»
possible to eatiaato the tftdividujil r
-------
TABLE 23. MODEL DATA REQUIREMEXTS
Complexity
level
Of B»d»i
J
2
3
4
S
Waste
Input
et
E2
«»
K2
K1
Weather
Et
£1
E2
E2
32
Source
Cbararterlstice
El
E3
Ml
M2
M3
Operation
El
Ei
£3
Ml
»2
El
E2 s
.11 = evaoureiJ on ette
?t2 » spssareil ov«-r a p*»rtc.
72
-------
using only limited data; Models with a conplexity level of 5 predict detailed
transient emissions froa various areas of specific facilities. Operating
conditions at one specific site may vary considerably froo day-to-day and from
Bonth-to-aonth. Data representing one specific time at a specific site may be
useful in estimating the validity of the model.
The level of model complexity selected depends on several factors:
I. Regulatory needs
2. Resources available to collect data
3. Equal ions/taethodoiogy available to analyze the data.
la general, the level of oodel selected should not be limited by the computa-
tional cooplexity of the calculations and data Analysis since this can be done
by computer. However, theoretical setbads are not available to perfora the
calculations necessary fo/ the most coapiex eodels (e.g., level 5 Models,
Table 23). In the diocuanion that follovu the influence of the second And
third t actors on data collection and tcodel choice IB considered. Regulatory
nerds will, not be discussed ia this report but may influence the ultimate
selection of a oodel.
Pretre-itswct, including bar screens and grit ch^abers, resoves sand,
grit, and other it etas froa the waste stream. Since there is little a.ixiag,
the pretreatsscnt process is modeled as a "plug flow" syotco: the wa&tevater
flows across the vessel, losing coae of the volatiles as it proceeds. The
loss of VOCa froa the unit ia the integral of the loss froo an "average"
elettent of the waste a* it flows through the pretreausent process with tiote.
The "average" tine an eleaent is retained in the pretreatsent process is:
I - V/Q (28)
where
V - voluae of pretreat»eut unit, a3
Q = volumetric flow rate, nn/sec.
The rate of decrease in mole fraction, X, of a given compound is:
73
-------
where
Knr = overall mass transfer coefficient, g mole/cm2-sec
UL
a = interfacial surface area, cro2/cm3
p' = molar density, g mole/cm3. .
CB
Substituting (A/V) for a , where A and V are the surface area (cm2) and volume
(cm3) of the treatment unit and integrating yields:
ln(Xout/Xin) = -KOLA/(1°6 <»:> ' . (30)
The fraction of VOC lost in the process, F is given by
F = 1 - exp(-KOLA/(106 Qp^)] (31)
The values of A and Q can be estimated for the actual plant or obtained from a
erode! plant. Either the individual values or the ratio A/Q is sufficient.
The molar density of water p' is assumed to be (1/18 g-mol/cm3). The value of
m
KQ, is waste specific and depends on the value of the gas mass transfer coeffi-
cient, the liquid mass transfer coefficient, and the partition coefficient
(see Equation 4). The diffusion coefficiertts are VOC compound specific, and
the partition coefficients are both VOC compound specific and waste specific.
The gas phase coefficient is estimated by the correlation with wind speed
of MdCkay and Yeun (1983, Equation H, Appendix A). The liquid phase resistance
nay be estimated by several techniques.
1. assume that the liquid phase resistance is zero
2. use a flowing stream model
3. u»e an assumed mixing rate
Because there are few experimental data to support approach 3 and approach 1
ig inaccurate for sany compounds, approach 2 will be used based upon the
equation of Owens et al. The partition coefficient is unknown but will be
ectiaated using Henry's law for single components in water.
Clarifiern
• Clarifirrs are designed to remove suspended solids and thus the solids
concentration increase: with depth. Concentrations of dissolved VOCs are not
expected to vary significantly with depth. The "plug flow" node I will be
used, as given in Equation (29). Equation 31 is applicable to the fraction of
-------
each VOC lost in the clarifier. The primary clarifier is assumed to have no
microorganisms present, and consequently, no biodegradation is occurring. The
gas mass transfer coefficient is estimated by the correlation of Mackay and
Yeun (1983) given in Equation 23. The liquid mass transfer coefficient is
based upon the correlation of Owens et al. (1964). The values of mass transfer
coefficients from the correlation of Owens et al. are substantially greater
(400%) than other predictions, (MacKay, 1983; Owens, et al. 1964). However,
since the induced turbulence at the entrance to the clarifier can lead to
higher rates of mass transfer than would be present in a stagnant system, such
an inflated predictor may be justified for clarifiers. As observed from the
apparent agreement of the model (see detailed calculation in Appendix B) with
the reported losses of VOCs from clarifiers which follow, there is little to
suggest that the predictions can be improved with a different trquaticn.
The flow of water out of the clarifier occurs around the perimeter of the
vessel creating a thin film of water. A modified form of the Owens' equation
was used to model the VOC lost from this waterfall. Since the flow is freefall
in air and not against a surface, the turbulence is expected to be reduced.
Therefore, an arbitrary reduction of 90% was imposed on the liquid mass transfer
coefficient. Even with the 90% reduction, the loss of VOCs from the weir was
predicted to account for 20 to 40 percent of the clarifier losses. The value
of the partition coefficient can be estimated from Henry's constant.
Surface Aeration
The model for VOC release resulting from surface aeration should account
for the following factors:
• Different rates of VOC release in the agitated zone and the rela-
tively still liquid zones.
Biological decay
Predictive functionalities (power input, number of agitators, geometry)
which influence emissions and are potentially subject to regulatory
control.
Adsorption rates of VOCs on primary and secondary sludges.
The surface aeration vessel is assumed to be well mixed and at a uniform
concentration. Under these conditions, the biological decay rate is expected
to be a constant as is the rate of VOC loss into the atmosphere for o given
75
-------
influent concentration. This surface aeration model applies to continuous
flow systems and does not apply to batch processing systems. In batch systems,
some of the VOCs are lost in the first few hours and the rate of emissions
decreases until the next batch charge. For batch operations where the biomass
concentration in the aeration vessel is approximately constant, the rate of
biooxidation may still be approximately constant in many cases. The losses of
VOCs in such systems can be calculated by numerical integration over time,
with the VOC concentration, and thus the driving force for volatilization
decreasing between charges.
The biological rate is assumed to be a constant for the continuous flow
system, but the rate is a function of the operational variables (parameters)
B = B(X.,Xn ,M,T) (32)
1 1 v/2
where
B. = rate of removal of component i (sec ')
X, = mole fraction dissolved oxygen
X. = mole fraction component i
M = biomass concentration (g/L)
T = temperature.
In addition, the biological decomposition process c:m generate VOCs by forming
intermediate decomposition products, with rate constant G. (sec ).
G. = G(X.,Xn , M,T) . (33)
* I. ^2
In general, these products are oxygenates and may be photochemically active
and, therefore, cf potential regulatory concern.
The material balance for component i can be written as the stun of the
amount flowing out of the system, the amount adsorbed on the sludge, the
amount lost to the atmosphere, and the amount destroyed biologically; these
are equal to the amount of a VOC entering the aeration vessel.
QX. . = QX. •+ K..X. .A/UOO p1) * B.VX. + Sk X. . (34)
i,in i.out OL i,out rn i i.out s i.out
76
-------
where
S = sludge removal rate, m3/sec
k = sludge partition coefficient, (gVOC/cc wet sludge) (gVOC/cc effluent).
S
Therefore.
Xi,out = ^i.in'W * KOLA/(1°° Pm> + BiV * Sks]
The sludge partition coefficient can be experimentally determined, or estimated
using Equation (27), The fraction of the VQf" entering the aeration vessel
which is lost to the air, F, is given by
I- = K_.X. _A/(100 p'QX. . ) (36)
OL i.out rra i,in
For the generation of VOCs from biotreatment (intermediate cecomposition
products), the total VOC lost car be estimated from the sum of individual VOCs
lost. The concentration of specific VOCs in the system can be obtained by a
material balance.
It is assumed that the concentration of the bioproduct entering the
system is negligible or that the rate of generation of bioproduct exceeds the
amount entering the system by a substantial amount. The component balance for
bioproducts can then b«.> written as:
where
VG. = QX. + 0.01 K_tX. A/p' + B.VX. + Sk X. (37)
i i.out OL i.out *in i i.out s i.out
C. = rate of generation of intermediate, gaole/m3-sec.
The effluent concentration of the intermediate is then given by
Xi,out = VV((> * °-01 KOL,iA/P; * BiV + Sks) (38)
The ratio of the amount emitted to the atmosphere to the amount discharged
from the process in the liquid effluent can be obtained from the ratio of the
rate of air emissions to the rate of discharge.
E-tio = iSM! {39)
0)
Note that an evaluation of K_. .A and the effluent concentration would permit
an estimation of air emissions, independent of bio-oxidation rates and byprod-
uct generation rates.
Freeman t'1982a) predicted that a minimum rate of emissions would occur
whenever just enough1 aeration' was present to support the biological activity
77
-------
of the waste treatment process. Materials such as vinyl chloride that are
resistant to biological decay are more rapidly lost at higher agitation rates.
It should be noted that these results are dependent on the rate constants of
the respective processes, as well as the models used. The Gerber kinetic
model used in Freeman (1982a) assumes first order kinetic behavior with respect
to the oxygen content and may not apply for some systems.
The overall mass transfer coefficient can be estimated by the equations
of Thibodeaux (1978) (Appendix A, Equation B) and Reinhart (1977) (Appendix A,
Equation J); for the regions outside the agitated zone, the correlations of
Owens, et al. (1964) could be used. These equations are discussed in Section
4 and summarized in Appendix A.
Secondary Clarifier
The rate of emissions from a secondary clarifier is estimated in a similar
manner to that of a primary clarifier. There will be some biological activity
in the secondary clarifier but since the aeration is limited and the biomass
settles, the biodegradation rate is much lower than that in the aeration basin
and can be neglected. The model is, therefore, identical to that of the
primary clarifier.
Trickling Filter
In a trickling filter, water is sprayed over a bed of solids. There are
potential emissions as the water dvoplets fall to the bed. Based upon the
model of Caiderbank (see equation 24) approximately 12% of the benzene in a
dilute solution would be lost as a G.S cm droplet of water fell 53 cm. As an
example, for a typical trickling filter fall distance of 23 cm (9 inches), the
loss of benzene would be approximately 6%.
Most of the VOC emission; would be expected to come from mass transfer on
the packing of the trickling filter. With an average packing diameter of 7 cm
(2 3/4 inch) the surface area of a randomly packed (porosity 50%) bed would
equal 1.7 cm2/cm3. A typical VGC liquid phase mass transfer coefficient,
K' a, for a wetted packing is 0.5 to 1.6 min"1 (Stallings, et al. 1985). The
K'.a for a specific system is measured for actual operating conditions, since
K',a is a function of the liquid and gas flow rates, as well as the system
geooietry.
The mole fraction of a VOC remaining in the treated wastewater, X. ,
is estimated by the following equation:
78
-------
p'QTX. . = p'QTX. .. + K' aV(X.-X) p1 (40)
*in'L i,in rm L i.out OL i i *in
*
where X. is the equilibrium concentration of VOC. iu the liquid phase at the
air water interface, and QT denotes the wastewater flow rate.
*
Assuming X. = 0 (liquid transfer rate controlled) and the model plsnt parame-
ters of
V = 4 (103) m3
K'a = 1.0 min"1
UL»
QT = 17 m3/min,
ii
the fraction of VOCs emitted by mass transfer from the packing of the trickling
filter is:
- (1 min"1) (4)(103) m3
F = 1-exp = 1 . (41)
17 m3/min
This indicates that at high air flow rates the filter would strip most of the
VOCs out of the water if the biological decay rate is sufficiently slow.
No verified models are available for combining the rate of stripping with
biological decay; however, the following equations illustrate a potential
approach. The rate of loss of VOCs in a trickling filter may be described
with the same equations as used for stripping columns (Truong and Blackburn,
1984),
f =K'La(X-X*) (42)
where
K' a = the liquid based rate constant, min
- UJ, ' . . - -
X - the mole fraction of .VOC in wastewater
* = the hypothetical mole fraction at equilibrium with the gas.
The solution assuming X* = 0 (liquid transfer rate controlled) is:
ln*-=-KOLat
o
If biological decay is significant, Equation 36 can be modified to estimate
the emissions loss
79
-------
The solution is:
dX = K^aXdt + XBdt . (44)
X = XQ exp [-(K£ta + B)t] (45)
Where the biological decay rate constant is independent of the VOC concen-
tration: . •
IT + K' a X
In - ^ - = K'at (46)
B' + K' a X °L
OL o
where B' is the biological rate constant based on mole-fraction.
The fraction lost to the atmosphere assuming a constant biological decay
rate constant would be estimated by the following relationship:
Ka
{l-exp[-(K'a + B')t]} (47)
(K^a + IT) °L
The other limiting case occurs where the gas is approximately at equilib-
rium with the wastewater. From a mass balance,
(Xin ' VVm<10^= Vng^in (48)
where Q is the gas flow rate, m3 ;ec, and p' is the gas density, g/cm3 and K
is the gas/water partition coefficient based on mole fraction. Solving the
preceeding equation for the fraction lost to the atmosphere,
= Qgp;gK/(106QLPm> "(49)
Subsurface Aeration
In the process of subsurface aeration, air is bubbled under the surface
of an activated sludge basin and the bubbles rise to the surface. It is
assumed that the bubbles have attained an equilibrium concentration based upon
the thermodynamic partitioning of the VOCs into the gas phase. VOCs can have
several potential fates:
• Pass through the system (water or sludge)
Volatilize into the bubbles
Volatilize at the surface
Biodegrade
80
-------
The mass transfer of compound i to the atmosphere, flux, includes both a
surface diffusion term and a bubble equilibrium term:
fluXi (gmol/sec-cm2) = K^X./ + p^ IQ^Q /A (50)
The fraction lost to. the atmosphere may be calculated as the sum of the fraction
lost to bubbles and the fraction lost from surface volatilization.
(0.01 K_TA + p1 KQ 106)
*
F=
(QLp; +0.01 KQLA + 106 p^ KQg + B.V^)
The surface mass transfer coefficients, k, and k in subsurface aeration
systems are estimated by methods similar to that of surface aeration systems.
Equalization
Equalization takes place in large mixing tanks and insures that concen-
trated surges of material do not shock the biooxidation systems. It is assumed
that the equalization tank is well mixed. The model of MacKay (1983) is used
to predict the liquid and gas phase mass transfer coefficients. The rate of
biological decay and adsorption in the equalization basin is assumed to be
negligible. .
Storage Tanks
The atmospheric emissions from storage tanks are estimated by mass transfer
coefficients, and the average emissions were estimated using an exponential
decay model given in Appendix A (equations K and L) . To convert the mass
transfer coefficients to emission models for the cases of mixing and no mixing,
assumptions were developed for the size and residence time.
The emissions from storage tanks (Equation D of Appendix A) are assumed
to be working losses, with turnover factor of 1.0 (dimensionless). The working
losses predicted by this equation are substantially greater than the emissions
from a storage tank with a floating roof (which can be considered an emission
control device) or from a storage tank with only hreathing losses. In .— «.
these situations, Equation D does not properly model the emissions
in cases where wind Is permitted to blow through the storage .ie
tanks because these systems can emit VOCs even when the tank is not
81
-------
being filled. la these situations, Equation D could seriously underestimate
the emissions. Also, for the special situation in which a waste liquid was
handled more often than assumed, more VOCs could be emitted than predicted by
the model. Equation D should be used to estimate the emissions from each
tank.
For open tanks with-mixing (Equation K, Appendix A), the emissions can be
estimated from an equation that accounts for the depletion of volatiles with
time as the open tank is agitated. The overall mass transfer coefficient is
calculated much like the zone of agitation in agitated impoundments. In a
similar fashion, the emissions from open tanks with no mixing (Equation L,
Appendix A) are estimated by using an overall mass transfer coefficient obtained
from the conditions of che surface impoundment.
Dissolved Air Flotation
In a dissolved air flotation unit, air bubbles pass through the wastewater
and volatile components are lost to the atmosphere through these bubbles. The
partition coefficients are used to estimate the concentration of the bubbles.
The mass transfer coefficients are obtained from McKay (1983) (Appendix A,
Equation N) for the gas phase and Churchill (1962) for the liquid. The air
emissions from the bubbles are assumed to be added to the emissions from the
surface.
Spray Ponds
Spray ponds are simply shallow ponds, typically three feet deep, with a
system of pumps and nozzles to spray wastewater into the air. Most of the VOC
emissions will result from mass transfer from the sprayed wastewater droplets
to the air. The actual emissions will depend upon the flow rate of the pumps
and the type of spray nozzle employed. These operating parameters will deter-
mine the droplet velocity, size distribution, and projection (i.e., air resi-
dence time).
If the droplet size and velocity are known, then the equation of Treybal
(1968) (see Equation 16) can be used to estimate the overall mass transfer
coeffcient. The rate of droplet fall can be estimated by equations from
McCabe and Smith (1967). When the droplet size and falling distance is such
that the droplet evaporates to dryness, it can be assumed that all of its
82
-------
contents, volatiles, and nonvclatiles, may be released into the atmosphere as
either vapors or aerosols. Consequently, spray ponds can be a major source of
VOC emissions.
The pond may contain biologically active material and could be modeled in
a manner similar to the surface aeration process previously discussed. The
agitated zone would be the surface area in which the wastewater droplets
return to the pond. The power input term in Equation B, Appendix A (Thibodeaux,
1978) would be estimated from the kinetic energy of the falling liquid droplets.
Such a correlation between droplet kinetic energy and agitator power require-
ments is not readily available, but it may be developed to apply Thibodeaux's
equation for such an aerated system. Nonetheless, the mass transfer coefficients
of the droplets are much greater than those for either the agitated or the
nonagitated surfaces. Consequently, droplet evaporation accounts for almost
all of the VOC emissions that occur from a spray pond.
Cooling Towers
A cooling tower is a large column in which water is evaporatively cooled.
The water forms a thin film on the walls and packing of the tower which can
greatly enhance mass transfer due to the high surface area. Typically, baffles
are employed to enhance liquid mixing and heat transfer. If the cooling water
contains VOCs (e.g. recycled wastewater or conttninated groundwater), emissions
can be significant for the following two reasons: (1) a large mass transfer
coefficient exists due to the short diffusional pathlength, and (2) the recir-
culation of the cooling water from the tower, to the process, and back to the
tower increases the effectiveness of the air-stripping process.
Mass transfer coefficients for falling films can be estimated by equations
given by Treybal (1968) (refer to Equation 16). The model employed would be
adapted to account for the recirculation rate and the number of baffles for
the specific cooling tower geometry. In general, however, cooling water does
not contain large amounts of VOCs. Thus, air stripping of VOCs froai cooling
towers, though effective, is not typically significant.
A COMPARISON OF THE PREDICTIONS OF THE MATHEMATICAL MODELS
WITH REPORTED VOC LOSSES FROM AERATION PROCESSES
This section provides examples of the use of the mathematical models
described in this document to predict the VOC losses from aeration systems for
83
-------
which VOC emission data are available. A description is provided of (1) how
the input parameters were specified, (2) the results of the calculations, and
(3) the success of the models at predicting VOC losses.
Parameter Specification ' .
The model provides a specification of the fractional loss of VOCs to the
air, to biological decay, and to the outflow of the process unit. This reli-
ance on unit processes provides flexibility; different process configurations
can be simulated with the standard default settings, and the input concentra-
tions are not needed to calculate fractional losses to the atmosphere. The
unit model approach does have disadvantages; the user roust select which unit
operations are of concern and perform some simple calculations to obtain air
emission rates. For example, if 20 percent of the VOC of interest is lost in
the primary clarifier and 30 percent is lost in the following equalization
basin, the overall loss would be the sum of 20 percent and (1-.2) (.3)xlOO or
44 percent. If the flow of benzene into this system is 30 g/sec, then the air-
emissions would be calculated as 30 x 0.44 or 13.2 g/sec. In a detailed
estimation of national emissions, these site specific calcuations could be
performed using the model, with a few changes in the program.
An example of .tiodel plant parameters which must be specified are identi-
fied in Table 24. The parameters shown are obtained from actual plant data or
from model plants. The component properties are selected from an internal
data file. The user can scan these component data files and select the components
of interest (e.g., benzene, trichloroethane, tetralin). The model was applied
to an industrial waste system (Union Carbide) described by Alsop, et al.
(1984). A simplified schematic is given in Figure 5. No independent biokinetic
data were available for this study and default values of negligibly slow
biorates were assumed. The process parameters for this plant were used as
default values for the other plants described in this report.
Table 25 presents the physical properties input to the mathematical
models. The partition coefficient was estimated from the ratio of the vapor
pressure to the solubility. These data are compared to experimentally obtained
partition coefficients in Table 26. In some cases, the theoretical values
vary significantly from experimental data. The specification of partition
coefficient is, however, an input to the model which may be specified based on
the best available information. Theoretical data were used in this report.
' . '•. " - .."•'••'.• • '84' • • • ; .' ' v .•. .
-------
TABLE 24. MODEL INPUT DATA FOR UNION CARBIDE PLANT
(Alsop, et al. 1984)
Temperature (deg C) 25
Air Velocity (crn/s) 200
Concentration of Benzene (mole tract) .001
Concentration of Naphthalene (mole fract) .005
Concentration ot kthylbenzene (mole fract) .001
Concentration of Methyl cellosolve (mole fract) .0002
Concentration of Tetralin (mole fract) .0001
Biorate of Benzene (hours) 400
Biorate of Naphthalene (hours) 400
Biorate of Ethylbenzene (hours) 400
Biorate of Methyl cellosolve (hours) 800
Biorate of Tetralin (hours) 200
Waste flow rate (ra3/sec) .07
Area of pretreatraent basin (m2) 50
Depth of pretreatment basin (m) 3
Diameter of clarifier (m) 19.4
Depth of clarifier (m) 2.4
Number of clarifiers 1
Length of aeration basin (m) 170
Width of aeration basin (m) 170
Depth of aeration basin (m) 3.6
Area of agitators, each (m2) 96
Number of agitators 30
Power of agitation, each (hp) 75
Impeller diameter (cm) 30
Impeller rotation (rpm) 2000
Number of secondary clarifiers 1
Diameter of secondary clarifier (m) 37
Depth of secondary clarifier (m) 3
Area of equalization basin (m2) 5185
Depth of equalization basin (m2) 3
Air flow in DAF (pretreat) (sm3/s) 0
Flow of submerged air (m2/sec) .4
Power of.sub-agitation, each (hp) • 50
Length of sub-aeration basin (m) 33.5
Width of sub-aeration basin, (r.) 8.4
Depth of sub-aeration basin (m) 8.5
Area of sub-agitators (each) (m2) 70.1
Number of sub-units in series 3
Impeller rotation (.rpm) 70
Impeller diameter (cm) 243
Weir height (cm) 30
Weir flow thickness (cm) 1
Enter {1} for covered sub-agitators 1
-------
Influent
Primary
Clarifier
Equalization
B-.sin
Aeration
Effluent
00
ON
. Secondary
Clarifier
Source: Alsop et at., 1984
UNOX Train
Figure 5. Flow scheme for Union Carbide plant.
-------
TftBLE 25. PHYSICAL PROPERTY DATA USED IN MATHEMATICAL MODEL
Water diffusivity
Compound cm2/sec
BEKZENE
TOLUENE
XYLENE
PHENOL
METHANOL
111 TCE
ACETONE
BUTANOL
ETHANOL
ACRYLONITRILE
METHYL CHLORIDE
METHYLENE CHLORIDE
VINYL CHLORIDE
CARBON TETRACHLORIDE
OXYGEN
ETHYL BENZENE
2 ETHYL HEXANOL
NAPHTHALENE
TETRALIN
D1CHLOROETHANE
METHYL CELLOSOLVE
CELLOSOLVE
METHYL CELLOSOLVE ACETATE
CELLOSOLVE ACETATE
i-OCTF.NE
CHLOROBENZENE
CHLOROFORM
112 TRICHLOROETHANE
TETRACHLOROETHANE
TETRACHLOROETHYLENE
TRICHLOROETHYLENE
DICHLOROETHYLENE
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.000010
.000009
.000008
.000009
.000016
.000003
.000011
.000010
.000013
.000012
.000014
.000010
.000012
.000010
.000025
.000008
.000007
.000008
.000008
.000010
.000010
.000010
.000008
.000008
.000008
.000009
.000007.
.000008
.000008
.000008
.000009
.000010
Air diffusivity Molecular
cm2/sec weight
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.088
.087
. 063
.082
. 149
.064
. 103
.079
.112
. 102
. 114
. 100
.093
.063
. 178
.070
.090
.059
.070
.080
.080
.080
.070
.070
.070
.090
. o'ao
.070
.060
.060
.060
.075
78
92
106
94
32
133
58
74
46
41
50
99
63
154
32
106
130
128
132
99
76
90
118
132
112
113
120
133
168
166
131
97
Partition
coefficient.
305
366
300
0
0
273
0
1
0
4
2044
177
30000
1679
4000
271
1
65
600
61
0
0
0
0
53
224
186
47
24
1547
650
856
-------
TABLE 26. A COMPARISON OF MEASURED AMD THEORETICAL PARTITION
COEFFICIENTS (Petrasek et al., 1983)
Component
Dichloroethene
Chloroform
Carbon tetrachloride
1,1,2 Trichloroethane
Benzene
1,1,1 Trichloroethane
Chloi-oben^ane
Tetrachloroethane
Toluene
Ethyl Benzene
Partition Coefficient (Y/X)
Experimental
150
164
1138
16.7
795
2409
81
54
792
859
Modelb
856
186
1679
47
SCO
273
224
24
329
271
*1
Obtained from the ratio cf the concentration in the off-gas to the ftxit
liquid concentration.
Obtained from the ratio of the vapor pressure to the solubility in water.
-------
Results of Calculations
Table 27 presents the comparison of the calculations of the model with
the reported VOC losses for benzene, ethylbenzene, toluene, dichloroethane,
naphthalene, and tetralin. The losses fron the clarific.-r were reported as the
sum of the loss from the surface and the loss from the overflow. The UNOX
losses are based upon a closed system.
An investigation of VOC losses in the Water Environmental Research
Laboratory (WERL) pilot facility (Petrasek, 1981) is compared to the results
of the model. A schematic of the WERL pilot plant is shown in Figure 6. The
results of this comparison are presented in Table 28. The parameters specified
in the model are presented in Table 29. Many of the critical parameters were
available from the report.
Additional WERL data were published by Petrasek (1983). Table 30 presents
a comparison of the loss of VOCs from the WERL ;:ilot facility based upon the
measured VOC emission rates and the model rates. Two numbers are presented in
the experimental column, representing the results from the two parallel treat-
ment units. Analyses of the off-gas from the aeration process indicated that
there was an average VOC loss of 43 percent due to volatilization in the
aeration basin. Table 31 presents the input parameters for the models.
Tables 32 and 33 present the comparisons of the predicted and reported
VOC lesser, from the American Cyanamid Willow Island Plant and the parameter
assumptions, respectively. In a similar fashion Tables 34 and 35 describe a
comparison for the: Sisterville Plant, and Tables 36 and 37 describes a compari-
son for the du Pont Belle Plant (Lassiter, P.E., Memo 1985).
Discussion of the Model Applicability
The rate of loss of volatiles from the Union Carbide Plant (Alsr.p, et al.
1984) as predicted by the mathematical models agreed well with measured results,
with the exception ot the UNOX biological treatment unit. The rate of removal
in the UNOX system w^o overpredicted by the model, particularly in the case of
tetraiin. It is unknown whether analytical procedures or nonlinear K values
for tetralin were responsible for the variation. Though the loss in the
aeration basin was overpredicted by the model for tetralin and naphthalene
(89% was the observed loss in both cases), the absolute errors were still only
89
-------
TABLE 27. A COMPARISON OF THE EXPERIMENTAL LOSS OF VGLATILES FROM
THE UNION CARBIDE PLANT FIELD TEST TO THE PREDICTED
LOSS FROM rlATHEMATICAL MODELS (Alsop, et al., 1984)
Unit/volatile
Clarifier surface
Benzene
Ethyl benzene
Toluene
Dichloroethane
Naphthalene
Tetraiin
Clarifier overflow
Benzene
Ethyl benzene
Toluene
Dichloroethane
Naphthalene
Tetraiin
Total Clarifier
Benzene
Ethyl benzene
Toluene
Lichloroethane
Naphthalene
Tetraiin
Equalization basin
Benzene
Ethyl benzene
Toluene
Dichloroethane
Naphthalene
Tetraiin
Aeration basin
Benzene
Ethyl benzene
Toluene
Dichloroethane
Naphthalene
Tetraiin
UNOX
Benzene
Ethyl benzene
Toluene
Dichloroethane
Naphthalene
Tetraiin
Fractional loss
(experimental)
.058
.103
.13
.143
.045
.225
.391
.35
.281 /
.389
.269
.115
.9984
.993
.993
.987
.892
.905
.26
.44
.19
-.08
.31
.05
Fractional less to air
(.predicted)
.16
.15
.15
.13
.12
.15
.01
.01
.01
.01
.006
.01
.17
.1C
.16
.14
.116
.16
.31
.29
.29
.31
.27
.29
.998*
.997*
.993*
.994*
.994*
.998*
.58*
.55*
.62*
.23*
.26*
.73*
*Ip.wludes biological loss (approximately O.Oi).
-------
Influent •
Primary
Clarifisr
Aeration Train
Based on Petrasek, 1981
Effluent ^
Secondary
Clarifier
Figure 6. WERL pilot plant flow scheme.
-------
TABLE 28. A COMPARISON OF THE EXPERIMENTAL LOSS OF VOLATILES
FROM A MERL PILOT STUDY TO THE PREDICTED LOSS FROM
MATHEMATICAL MODELS (Petrasek, 1981).
Fractional loss Fractional loss
Unit/Volatile (experimental) (predicted)
Clarifier surface
Meth/lene chloride .065
Benzene .063
Trichloroethane .063
Dichloroethene .067
Carbon tettachloride . .066
Clarifier overflow
Methylene chloride .005
Benzene .005
Trichloroethane .005
Dichloroethene .005
Carbon tetrachloride .005
Total Clarifier
Methylene chlorj.de .26 .070
Benzene .32 .068
Trichloroethane .85 .068
Dichloroethene .51 .072
Carbon tetrachloride .18 .071
Aeration basin . .
Methylene chloride .84 .77
Benzene .88 .85
Trichloroethane -.12 .84
Dichioroethene .95 .94
Carbon tetrachloride .91 .97
92
-------
TABLE 29. MODEL INPUT DATA FOR WERL PILOT PLANT STUDY
(Petrasek, 1981)
Temperature (deg C) 25
Air Velocity (cm/s) 300
Concentration of Benzene (mole fract) .001
Concentration of Methylene chloride (mole fract) . .005
Concentration of 1,1,1-TCE (mole fract) .W.
Concentration of Methyl cellosolve (mole fract) .0002
Concentration of Carbon tetrachloride (mole fract) .0001
Biorate of Benzene (hours) 200
Biorate of Methylene chloride (hours) 200
Bicrate of 1,1,1-TCE (hours) 200
Biorate of Methyl cellosolve (hours) 200
Biorate of Carbon tetrachloride (hours) 200
Waste flow rate (m3/sec) .0022
Diameter of clarifier (m) 3
Depth of clarifier (m) 3.6
Number of clarifiers 1
Number of secondary clarifiers 1
Diameter of secondary claritier (m) . 3.6
Depth of secondary clarifier (m) 3.6
Flow of submerged air (m3/sec) .057
Power of sub-agitation, each (hp) 10
Length of sub-aeration basin (m) 3
Width of sub-aeration basin (m) 1.35
Depth of sub-aeration basin (m) 3.6
Area of sub-agitators (each) (m2) 4
Number of. sub-units in series 4
Impeller rotation (rpm) 100
Impeller diameter (cm) 200
Weir height (cm) . 9
Thickness of weir flow (cm) 1
93
-------
TABLE 30. A COMPARISON OF THE EXPERIMENTAL AND PREDICTED LOSS OF
VOLATILES FROM A WERL PILOT FACILITY (Petrasek et al., 1983)
Loss
Clarifier
Experimental
Methyene chloride
Dichloroe thane
Chloroform
Carbon teLrachloride
Trichloroethylene
Benzene
Trichloroethane
Toluene
Ethyl Benzene
Tetrachloroethylene
27 , 25
65,57
33,--
47,47
31,36
--,16
48,27
29,22
70,16
88
Predicted
7.0
7.2 .
5.6
7.1
6.7
6.8
6.8
6.3
6.3
6.3
of VOC (%)
Subaeration
Experimental
88.7,99.5
94.5,99.4
94.4,97.6
94.0,99.3
93,97.8
88,99.6
97,99.7
93.5,99.6
97,99.7
93
Total a
predicted
77.4
94.3
78.3
97.0
92.6
85.5
84.1
87.6
84.0
96.7
air emissions (lost to bubbles only) plus biological decomposition.
94
-------
TABLE 31. MODEL INPUT DATA FOR WERL PILOT STUDY
(Petrasek. 1983)
Temperature (cleg C) 25
Air Velocity (cm/s) 300
Concentration of Benzene (mole fract) .001
Concentration of Methylene chloride (mole fract) .005
Concentration of 1,1,1-TCE (mole fract) .001
Concentration of Methyl cellosolve (mole Tract) .0002
Concentration of Carbon tetrachloride (mols fract) .0001
Biorate of Benzene (hours) 200
Biorate of Methylene chloride (hours) 200
Biorate of 1,1,1-TCE (hours) 200
Biorate of Methyl cellosolve (hours) 200
Biorate of Carbon tetrachloride (hours) 200
Waste flow rate (m3/sec) .0022
Diameter of clarifier (m) 3
Depth of clarifier (m) 3.6
Number of clarifiers 1
Number of secondary clarifiers 1
Diameter of secondary clarifier (m) 3.6
Depth of secondary clarifier (m) . 3.6
Flow of submerged air (m3/sec) .057
Power of sub-agitation, each (hp) 10
Length of sub-aeration baein (m) 1.35
Width of sub-aeration basin (in) 3
Depth of sub-aeration basin (m) 3.6
Area of sub-agitators (each) (m2) . 100
Number of sub-units in series 4
Impeller rotation (rpm) 100
Impeller diameter (cm) 200
Weir height (cm) 9
Thickness of weir flow (cm) 1
95
-------
TABLE 32. A COMPARISON OF THE REPORTED FRACTIONAL LOSS OF COMPONENTS
FROM PRIMARY WASTEWATER TREATMENT AND THE
PREDICTED LOSSES, WILLOW ISLAND PLANT
(Lassiter, P.E., Memo, 1985), American Cyanamid
Fractional loss of VOC
Component Reported Predicted
Ethyl Cellusolve
Methanol
1-octene
0.071
0.41
0.13
0.0023
0.0152
0.259
TABLE 33. MODEL INPUT DATA FOR WILLOW ISLAND WASTEWATER STUDY
(Lassiter, P.E., Memo, 1985)
Temperature (deg C) 25
Air Velocity (cm/s) 200
Concentration of Methanol (mole fract) .001
Concentration of Ethyl cellosolve (mole fract) . .005
Concentration of 1-octene (mole fract) .001
Biorate of Methanol (hours) 500
Biorate of Ethyl cellosolve (hours) 500
Biorate of 1-octene (hours) 500
Waste flow rate (m3/sec) .127
Area of pretreatment basin (m2) 3
Depth of pretreatment basin (m) ' 3
Diameter of clarifier (m) 19.4
Depth of clarifier (m) 2.4
Number of clarifiers 1
Weir distance (cm) 30
Thickness of weir flow (cm) 1
96
-------
TABLE 34. A COMPARISON OF THE ESTIMATED FRACTIONAL VOC LOSS AND THE
PREDICTED VOC LOSS FOR THE SISTERVILLE PLANT
(Lassiter, P.E., Memo, 1985)
Compound
Ethanol
Methanol
Toluene
Fractional
Estimated+
.005 '
.005
.128
Loss
Predicted-M-
0.0008
0.0016
0.157
+ Plant emissions
++ Primary treatment, no equalization
TABLE 35- MODEL INPUT DATA FOR SISTERVILLE WASTEWATER STUDY
(Lassiter, P.E., Memo, 1985)
Temperature (deg C) 25
Air Velocity (cm/s) 200
Concentration of Ethano] (mole fract) .001
Coucentration of Methanol (mole fract) .005
Concentration of Toluene (mole fiact) .001
Biorate of Ethanol (hours) 400
Biorate of Methanol (hours) 400
Biorate of Toluene (hours) 400
Waste flow rate (m3/sec) .08
Diameter of clarifier (m) 19.4
Depth of clarifier (m) 2.4
Number of clarifiers 1
Weir height (cm) 30
Thickness of weir flow (cm) 1
-------
TABLE 36. A COMPARISON OF THE ESTIMATED AND PREDICTED FRACTIONAL LOSS OF
VOCs TO THE AIR FROM PRIMARY AND SF.CO,VDARY TREATMENT
AT THE DU PONT BELLE PLANT (Lassiter, P.E., Memo 1985)
Component
Methanol
Acetone
Chloroform
Butanoi
Primary
Reported Predicted
0.015
0.008
0.62
0.0007
0.024
0.035
0.375
C.120
Secondary
Reported Predicted
0.000001
0.0015
0.0062
NR
0.0005
0.0009
' 0.393
0.0045
+ Pretreatment, clarifier, equilization
NR = not reported
•H- aeration, submerged and covered
-------
TABLE 37. MODEL INPUT DATA FOR DUPONT BELLE WASTEWATER STUDY
(Lassiter. P.E., Memo, 1985)
Temperature (deg C) 25
Air Velocity (cm/s) 200 .
Concentration of Methanol (mole fract) .001
Concentration of Butanol (mole fract) .005
Concentration of Acetone (mole fract) .001
Concentration of Chloroform (mole fract) .0002
Biorate of Methanol (hours) 400
Biorate of Butanol (hours) 400
Biorate of Acetone (hours) 400
Biorate of Chloroform (hours) 400
Waste flow rate (m3/sec) ..08
Area of pretreatsnent basin (m2) 50
Depth of pretreatment basin (m) 3
Diameter of clarifier (m) 19.4
Depth of clarifier (m) 2.4
Number of clarifiers 1
Number of secondary clarifiers 1
Diameter of secondary clarifier (m) 37
Depth of secondary clarifier (m) 3
Area of equalization basin (in2) 5185
Depth of equalization basin (m2) 3
Air flow in DAI' (pretreat) (sm3/s) 0
Flow of submerged air (m3/sec) .4
Power of sub-agitation, each (hp) 50
Length of sub-aeration basin (m) 33.5
Width of sub-aeration basin (m) 8.399999
Depth of sub-aeration basin (m) 8.5
Area of sub-agitators (each) (m2) 70
Number of sub-units in series 2
Weir height (cm) 30
Thickness of weir flow (cm) 1
-------
5-10%. Therefore, it may be concluded that the models were successful in
simulating the field conditions for the plant, and that most of the VOCs were
lost to the atmosphere.
In an EPA pilot investigation (Petrasek, 1981) a number of VOCs were
processed in a ciarifier followed by an aerated biological oxidation unit.
Enough data ware provided to specify the input parameters of the model.
Although the treatment plant was not directly exposed to the atmosphere, air
was circulated and the mass transfer is considered to be liquid phase con-
trolled. For both the ciarifier and the aerator, the model predicted VOC
losses in close agreement with the pilot data. The model predicted that most
of the VOCs were lost to the circulating air. While there was good agreement
between the predicted and experimental emissions rates for subsurface aeration
in more recent work at che WERL Pilot Facility (Petrasek, 1983), the experimental
results for the ciarifier VOC losses appear anomalously high. However, Petrasek did
iot discuss this possible discrepency and, therefore, is it not possible to determine
whether the modelled or experimental results are in error.
The fractional loss of VOCs to the air was estimated for the du Pont
Belle plant (Table 36). The model predictions for methanol, acetone, and
chloroform were within a factor of 5 for the primary process, but did not agree
with the estimated losses of butanol. The secondary treatment estimations did not
agree with model predictions. Not enough information was provided to determine
the source of the difference. At the Union Carbide Sisterville plant, the
predicted fractional loss was within a factor of 5 of the estimated values.
Only primary treatment was assumed; no information was provided dbout the
source characteristics except for the flow rate. The reported values of the
VOCs lost from the American Cyanamid Willow Island plant did not agree with
the predicted values. No information about the source characteristics (except
flow rate) was available.
Iii the two wastewater treatment units for which source descriptions as
well as dependable field measurements were available, the agreement of the
mcdel with the reported VOC losses was acceptable (considering sampling and
analysis errors) for estimation of emissions. In situations where the source
characteristics were not known the prediction of the model was less accurate.
The accuracy of the reported values has not been established; no changes in
the model are recommended based on these limited data.
100
-------
SECTION 6
NOMENCLATURE
a interfacial surface area/bed volume, cm2/cm3
A surface area, cm2
A effective surface area of a particle, cm2
C concentration, gmol/cm3
C concentration at time = 0, gmol/cm3
D impeller diameter, cm
D effective diameter of nonagicated surface, m
D.g diffusion coefficient of component A in B, cm2/sec .
D. diffusion coefficient in film, cm^/sec
DV diffusion coefficient of VOC, cm2/sec
e emission rate, gmol/sec
F. fraction of waste in agitation impoundment from source
A
s
(
F fraction o£ VOC lost to the atmosphere
Fr Froude number
g,,. gas holdup fraction, v /Vv
Gi rate of VOC (component i), generation, gmol/cm3-sec
G gravitational constant
h liquid height or depth, cm
H Henry's Law Constant, cm3-atm/g-mol
J oxygen transfer rating
K,T film-liquid partition coefficient
XL*
K partition coefficient, Y/X
k mass transfer coefficient of gas, g-mol/cm2-sec
O •
k1 mass transfer coefficient of gas, cm/sec
O
101
-------
k' liquid mass transfer coefficient k' = kT(MW)/p', era/sec
L» Lf it m
k_ mass transfer coefficient of liquid, gmol/cm2-sec
k' overall mass transfer coefficient, cm/sec
\jLi
K.T overall mass transfer coefficient of a tracer, g-mol/cm2-sec
UJ_»
KnT overall mass transfer coefficient based on liquid concentrations,
OL
g-mql/cm2-sec
-1
K2 oxygen reaeration rate constant (base e), s
L characteristic length, cm
MW molecular weight, g/gmol
MW Average molecular weight of liquid, g/gmol
p1 molar density, g-mol/cm3
p particle density, g/cm3
p density, g/cm3
p. vapor pressure of component i, atm
PI power input to the agitator or aerator, hp or g-cm2/sec3
POWR power of agitators
Po Power number
P n-octanol/water partition coefficient
q. atmospheric emission rate of component i, g/sec
qT total atmospheric emissions, g/sec
Q volumetric flow rate, cm3/sec
R gas constant = 82.05 cm3 atm/g-mol
Re Reynold's number
JU
Re roughness Reynold's number
Re particle Reynold's, number
ST surface tension, g/sec2
102
-------
T residence time, sec
t time, sec
T temperature, °C or °K
|j viscosity, g/cm-sec
U* correlated wind velocity = (6.1 + 0.63 U10) ' U10/100, in/sec
U wind velocity, cm/sec
UIQ wind velocity at 10 m above surface, m/sec
v velocity, cm/sec
v terminal velocity, cm/sec
V . volume, cm^
w rotation speed, rpm
W mass flow rate, g/sec
X mole fraction in liquid phase
Y mole fraction in gas phase
F liquid mass flow rate per unit film width, g/cm-sec
6 film thickness, cm
Subscripts
g denotes gas phase
L denotes liquid phase
b bubble
p particle
i • denotes component i
o denotes initial condition at t = 0.
103
-------
SECTION 7
REFERENCES
Alsop, G.M., R.L. Berglund, T.W. Siegrist, G.M. Whipple, and B.E. Wilkes.
"Fate of Specific Organics in an Industrial Biological Wastewater
Treatment Plant," Draft Report. Prepared for U.S. Environmental
Protection Agency, IERL. June 29, 1984.
Baird, Rodger, Louis Carmona, Richard L. Jenkins, "Behavior of Benzidine and
Other Aromatic Amines in Aerobic Wastewater Treatment," Journal WPCF,
Vol. 49, pp. 1609-1615 (July 1977).
Blackburn, J.W., J.L. Fleming, W.L. Troxler, G.S. Sayler, and M.W. Cantrell,
Final Report: Treatabilitv of Organic Chemicals, Draft, Submitted
to U.S. EPA under Contract 68-03-3027 (1982).
Blanchard, F.A., I.T. Takahashi, H.C. Alexander, "Biodegradability of (14C)-
Methylcellulose by Activated Sludge," Applied and Exp. Microbiology
32(4), pp. 557-560 (October 1976).
Cadena, F., G.A. Eveman, and V.J. Vandiver. "Removal of Volatile Organic
Pollutants From Rapid Streams." J. W.P.C.F., 56(5):460. 1984.
Calderbank, P.H. "Mass Transfer." Chapter 6 of Mixing. Uhl ind Gray, eds.,
Academic Press, New York. 1967.
Capestany, George J., John McDaniels, Jarl L. Opgrande, "The Influence of
Sulfate on Biological Treatment of Phenolbenzaldehyde Wastes," Journal
WPCF, pp. 256-261 (February 1977).
Churchill, M.A., H.L. Elmore, and P.A. Buckingham. "Prediction of Stream
Reaeratioa Rates." Proc. ASCE. J. San. Eng. Div., SA4, p.l, July 1962.
Cohen, Y., W. Cocchlo, and D..MacKay. "Laboratory Study of Liquid-Phase
Controlled Volatilization Rates in Presence of Wind Waves." Envir. Sci.
and Technol. _12(5):553. 1978.
Durham, J.F., OAQPS Memo, July 19, 1985.
Engelbrecht, Richard S., Ross E. McKinney, "Activated Sludge Cultures Devel-
oped on Pure Organic Compounds," Sewage and Industrial Wastes 29(12),
pp. 1350-1362 (December 1957).
Erikson, D.G., J.J. Cudahy, V. Kalsevic, and R.L. Standifer. "Organic
Chemical Manufacturing." Vol. 3: Storage Fugitive and Secondary Sources.
EPA-450/3-80-025. December 1980.
Freeman, R.A. "Comparison of Secondary Emissions From Aerated Treatment
Systems." Presented at AIChE Meeting. Paper 5c. Orlando, FL, February
1982.
.104
-------
Freeman, R.A. "Secondary Emissions From Subsurface Aerat?d Treatment Systems."
Env. Prog. 1,2:117. 1932.
GCA Corporation, Hazardous Waste TSDF Waste Process Sampling. EMB Report
85-HWS-3. October, 1985
Gehm", H.W. et al. J. Water Pollution Control Fed. 37(10): 1393. 1965.
Gerike, P; , W.K. Fischer, W. Holtmann, "Biodegradability Determinations in
Trickling Filter Ui its Compared with the OECD Confirmatory Test," Water
Research, Vol. 14, pp. 753-758 (March 1979).
Hansch, E., and A.J. Leo, Substituent Constants for Correlation Analysis in
Chemistry and Biology. Wiley, New York (1979).
Hwang, S.T. "Toxic Emissions From Land Disposal Facilities." Env. Prog.
1,1:46. 1982.
Kincannon, D.F., and E.L. Stover, Determination of Activated Sludge Biokinetics
Constants for Chemical and Plastic Industrial Wastewater, EPA-600/2-83-073,
NTIS No. PB83-245233.
Lassiter, P.E. "West Virginia Wastewater Plant Data.11 U.S. EPA, OAQPS.
Memorandum, April 8, 1985.
Lawrence, A.W. and P.L. McCarty, A Unified Basis for Biological Treatment
Design and Operation, J.S.D., ASCE, % SA3 (1970).
Lenard, *. Ann. Physik, 6j>:629. 1921.
Kullenberg, G.B. "On Vertical Mixing and the Energy Transfer From the Wind to
the Water." Tellus Institute for Physical Oceanography, University of
Copenhagen, Denmark, Vol. XXIII, 1976.
MacKay, D. and P.J. Leinonen. "Fate of Evaporation of Low - Solubility
Contaminants From Water Bodies to Atmoaphere." Environ. Sci. Technol.
9(11):78. 1975.
MacKay, D. and R. S. Matsugu. "Evaporation Rates of Lrqiiid Hydrocarbon Spills
on Land and Water." Can. J. Chem. Eng. 5J:434. 1973.
MacKay, D. and A.T.K. Yeun. "Mass Transfer Coefficient Correlations for
Volatilization of Organic Solutes From Water." Environ. Sci. Technol.
1983.
McKinney, Ross E., John S. Jeris, "Metabolism of Low Molecular Weight Alcohols
by Activated Sludge," Sewage and Industrial Wastes 27(6), pp. 728-735
(June 1955).
Matter-Muller, W. Gujer, W. Geiger, and W. Stumm. "Non-biological Elimination
Mechanisms in a Biological Sewage Treatment Plant." Prog. Wat. Tech.
12:299-314. 1980.
105
-------
Metcalf and Eddy, Inc. Wastewater Engineering. McGraw-Hill. 1972.
Munz, C. and P.V. Roberts. "Mass Transfer and Phase Equilibrium in a Bubble
Column." Proc. of AWWA Conf., Miami Beach, FL. May 1982. pp. 633-637.
Nay, Marshall W., Jr., Clifford W. Randall, Paul H. King, "Biological Treata-
bility of Trinitrotoluene Manufacturing Wastewater," Journal WPCF 46(3),
pp. 485-497 (March 1974).
Nelson, Donald J., Leo M. Reading, C.W. Christenson, "Destruction of Oxalic
Acid by Activated Sludge," Sewage and Industrial Waste 26(9), pp. 1126-
1129 (September 1954).
Ottinger, R.S. et al. "Recommended Methods of Reduction, Neutralization.
Recovery, or Disposal of Hazardous Waste." Vol. IV. Disposal Process
Descriptions, Biological and Miscellaneous Waste Treatment Processes,
PB-224583, August 1973.
Owens, M., R.W. Edwards, and J.W. Gibbs. "Some Reaeration Studies in Streams."
Inter. J. Air & Water Poll., 8:496. 1964.
.Petrasek, A.C., Jr. "Removal and Partitioning of the Volatile Priority
Pollutants in Convention.-il Wastewater Treatment Plants, " Capsule
Report. Prepared for U.S. Environmental Protection Agency, MERL.
October 1981.
Petrasek, A.C., Jr., B.M. Austern, and T.W. Neiheisel. "Removal and Partitioning
of Volatile Organic Priority Pollutants in Wastewater Treatment."
Presented at Ninth U.S.-Japan Conference on Sewage Treatment Technology.
Contribution No. 46. Tokyo, Japan, September 1983.
Pfeil, 3ernhard H. G. Fred Lee, "Biodegradation of Nitrolotiiacetic Acid in
Aerobic Systems," Environmental Science and Technology 2(7), pp. 543-546
(July 1968).
Fitter, P., Determination of Biological Degradability of Organic Substances,
Water Research, 10, 231-235 (1976).
Placak, O.R., C.C. Ruchhoft, "Studies of Sewage Purification, XVII.-The Utili-
zation of Organic Substrates by Activated Sludge," Sewage Works Journal
19(3), pp. 423-440 (May 1947).
Porter, J. "State of the Art of Textile Waste Treatment, 1971." cited in
U.S. EPA Office of Technology Transfer, EPA-625/3-74-004. 1974.
Radhakrishnan, I., A. K. Sinha Ray, "Activated Sludge Studies with Phenol
Bacteria," Journal WPCF 46(10), pp. 2393-2418 (October 1974).
Reinhart, J.R. "Gas-Side Mass Transfer Coefficient and Interfacial Phenomena
of Flat Bladed Surface Agitators." University of Arkansas. Ph.D. Thesis.
1977.
106
-------
Roberts, P.V. and P.G. Dandliker. "Mass Transfer of Volatile Organic Contaminants
from Aqueous Soluton to the Atmosphere During Surface Aeration." Environ.
Sci. Technol. _17(8):484-489. 1983.
Rosenberg, D.G., P.J. Lofy, H. Cruse, E. V/eisberg, and B. Beutler. "Assessment
of Hazardous Waste Practices in the Petroleum Refining Industry."
EPA-SW-129c. NTIS // PB 259097, June 1976.
Sato, Takahiko, Takashi Akiyama, "Biochemical Oxidation of Volatile Organic
Acids by Activated Sludge," Water Research, Vol. 6, pp. 1059-1072 (1972).
SCS Engineers. Selected Biodcgradation Techniques for Treatment and/or
Ultimate Disposal of Organic Materials. EPA-600/2-79-006. NTIS # PB295394.
March 1979.
Sherwood, T.K. and R.L. Pigford. "Adsorption and Extraction." 2nd ed.
McGraw Hill, New York. 1952.
Smith, J.H., B.C. Bomberger, and D.L. Haynes, Production of the VolatiJ.ization
Rates of High-Volatility Chemicals from Natural Water Bodies, Environ.
Sri, and Tech., 14, 1232 (1980).
Stallings, R.L. T.N. Rogers, and M.E. Mullins. "Air Stripping of Volatile
Organics." Proceedings of the Institute of Environmental Sciences. Las
Vegas, NV. April 30-Hay.2, 1985. pp. 170-174.
Stamoudis, V.C., R.G. Luthy, W. Harrison, "Removal of Organic Constituents in
a Coal Gasification Process Wastewater by Activated Sludge Treatment,"
ANL/WR-79-!, Argonne National Laboratory (June 1979).
Stover, Enos L., Don F. Kincannon, "Biological Treatability of Specific Organic
Compounds Found in Chemical Industry Wastewaters," Presented at the 36th
Purdue Industrial Waste Conference, Purdue University. West Lafayette,
IN, May 12-U, 1981.
Strier, M.P., "Use of Structure-Activity Correlations to Establish the Fate of
Organic Priority Pollutants in Activated Sludge Treatment Systems,"
Memorandum to James D. Gallup. Effluent Guidelines Division, U.S. Environ-
mental Protection Agency, Washington, DC 20460, July 8, 1982.
Strier, M.P., "Treatability of Organic Priority Pollutants," Part F--Supplenent
I, "The Removal and Fate of Organic Priority Pollutants by Activated
Sludge Treatment—Estimated Percentage Removal Pathways," U.S. Environ-
mental Protection Agency, Effluent Guidelines Division, Washington, D.C.
20460, May 1985.
Swisher, R.D., M.M. Crutchfield, P.W. Caldwell, "Biodegradation of Nitrilotri-
acetate in Activated Sludge," Environmental Science and Technology
1(10), pp. 820-827 (October 1967).
Thibodeaux, I..J. "Air Stripping of Organics From Wastewater: A Compendium."
Proceedings of the Second National Conference on Complete Water Use.
Chicago, IL. May 4-8, 1978.
107
-------
Thibodeaux, L.J. and Parker. ''Desorption Limits of Selected Industrial Gases
and Liquids From Aerated Basins." Paper No. 30d. AIChE Conference,
Tulsa, OK. March 1974.
Treybai, R.E. "Mass Transfer Operations." 2nd ed. McGraw Hill, New York.
1968.
Truong, K.N. and J.W. Blackburn. "The Stripping of Organic Chemicals in
Biological Treatment Processes." Env. Prog. 3(3):153-152. 1984.
U.S. Environmental Protection Agency. Office of Technology Transfer. Wastewauer
Treatment Systems: Upgrading Textile Operations to Reduce Pollution.
EPA-625/3-74-004. October 1974.
U.S. Environmental Protection Agency. Effluent Guidelines Division. Synthetic
Organic Products Segment of Organic Chemical Manufacturing Industry.
EPA-440/1-75-045. 1975.
U.S. Environmental Protection Agency. Municipal Operations Branch. Process
Control Manual for Aer.cbic Biological Wastewater Treatment Facilities.
EPA-430/9-77-006. March 1977.
U.S. Environmental Protection Agency. Effluent Guidelines Division. Development
Document for Proposed Effluent Limitations Guidelines, New Source Performance
Standards and Pretreatment Standards for the Petroleum Refining Point
Source Category. EPA-440/1-82/014. October 1982.
U.S. Environmental Protection Agency. Development Document for Effluent Limi-
tations Guidelines and Standards for the Pulp, Paper, and Paperboard and
the Builders' Paper and Board Mills Point Source Category. EPA 440/1-82-025.
U.S. Environmental Protection Agency. Development Document for Proposed Effluent
Limitations Guidelines and New Source Performance Standards for the Organic
Chemicals and Plastics Industry. EPA-440/1-83/009 B. February 1983.
U.S. Environmental Protection Agency. Organic Chemical Manufacturing Volume 3:
Storage, Fugitive, and Secondary Sources. EPA-450/3-80-025.
U.S. Environmental Protection Agency. Effluent Guidelines Division. Development
Document for Effluent Limitations Guidelines, New Source Performance
Standards and Pretreatment Standards for the Iron and Steel Manufacturing
Point Source Category. EPA-440/1-82-024. May 1982.
Wakao, N. and T. Funazkri, Chem. Eng. Sci. 33:1375. 1978.
Westat, Inc. National Survey of Hazardous Waste Generators and Treatment
Storase and Disposal Facilities Regulated Under RCRA on 1981.
EPA-530/SW-84-005. April 1984.
108
-------
Appendix A
Emission Equations
-------
APPENDIX A
EMISSION EQUATIONS
A. LIQUID MASS TRANSFER WITH WIND
KL = .J.3 (Re°-195 - 0.57)(DIwater/DbenzWater>0-67
B. LIQUID MASS TRANSFER WITH AGITATION
KLaV - [J(POWR) a 106 (1.024)T-20/165.04J (Diwater/Doxgnwater)°•5
C. LIQUID MASS TRANSFER WITHOUT AGITATION
KL = 3.12 (1.024)T-20 VQO.G7 (Diwater/Doxgnwater)°-66/(h/3)0.85
T = the temperature, 23 °C
V0 = the surface velocity, 0.035 (14.7 ft/sec) natural surface. 0.1 ft/sec
out.si.1e region of effect of agitation in biological treatment
h = the depth of the liquid, ft
"oxgnwater = *ne diffusion coefficient of oxygen in water
D. STORAGE TANKS, FIXED ROOF
Lw " 0.01089(MWlp(TP)
TP = tank throughput, gal/day
p = pressure (psia) ' '
Assumes all emissions are from tank working losses.
References
Based on Cohen, Coechlo,
and MacKay (1978)
Thibodeaux (1978)
Based on Owens, Edwards,
and Gibbs (1964)
AP42, modified (1981)
\
Continued
-------
APPENDIX A (Continued)
Refsrences
E. SURFACE IMPOUNDMENTS
qT (g/sec) = A{cm2)KoL(MW)X
X = mole fraction of the component in the liquid
H = Vapor pressure (torr)/7CO * (MW)/solubii;ty (mg/L)
K0r = (1/kr + 1/Kkg)"1
V IJ Irf £^
k£, = from Owens et al. (1964) K = 10s /MW
Kg = from Thibodeaux and Parker (1974;
»
i '
° F. AERATION
«i ° KoLcL . "
CL = the concentration in the liquid (g/cm^j
KoL (agitation) = (l/kL » l/lipK)"1
where k^ from Thibodeaux (1978)
kg from Reinhart (.1977)
KoL (nonagitation) = (l/kL + 1/kgK)"1
where k[, from Owens et aj . (1P64)
kg frptn MacKay and Muts;iga (1973)
Continued
-------
APPENDIX A (Continued)
References
A] = area agitated
A2 = area away from agitation
G. GAS MASS TP.ANSFER COEFFICIENT
kg - 0.0908 U°-7tf palr/t28.97(Sc)°-67(De)°'11]
where U = wind velocity, m/hr
De = effective diameter of non-agitated surface, m
P = density, lb/ft3
H. GAS MASS TRANSFER COEFFICIENT
k = 4.03 • 10~6 u-804 L~-196 D;66? (CoJburn type equation)
i -3 1 r
I. GAS MASS TRANSFER COEFFICIENT .
kg = 9.9 U
MacKay and Matsuga (1973)
Sherwood and Pigford
(1952)
Thibodcaux & Parker
i1974)
J. GAS MASS TRANSFER COEFFICIENT
k - 0.00039 D- (Re)1 -42(PN)°-4 (Sc)°-5/[D°- l (Fr)°-21 ]
g . pg
where k Js given in lb/(ft.2 hr)
Reinhardt (1977)
Continued
-------
APPENDIX A (Continued)
References
K. OPEN TANKS WITH MIXING
QT (S/day) - w (g/day)[l-exp(-KoL x T/h)]
q-j .= total emissions, g/day
W •= mass flowi-ate. g/day
T = residence time (seconds)
h = depth, cm
KQL " the overall mass transfer coefficient obtained from agitated
impoundments, zone of agitation (cm/sec)
L. OPEN TANKS WITH NO MIXING
qt '(g/day) * W (g/day111 - exp(-KoL x T/h)J
q-[ = total emissions, g/day
W = mass flowrate, g/day
T = residence time (seconds)
h <= depth, cm
KQ*L = the overall ir.ass transfer coefficient obtained from agitated
impoundments, zone of agitation (cm/sec)
Continued
-------
APPENDIX A (Continued)
References
M. COOLING TOWERS •
For volatile components (vapor pressure > 3 torr), the rate of emissions is the . •
rate of generation times 0.98. For the other components, use the ratio.
emission/generation = vapor pressure/3 torr x 0.98.
Assume the drift loss is .2 percent of recirculation rate.
Assume a recycle of 10 times. The drift loss would then be .02 times the rate of
waste generation.
Assume that cooling towers would not be used for organic steams (fraction
nonwater > .1) or for sludges (fraction solids > .1).
o> -..•'.
N. GASS MASS TRANSFER COEFFICIENT .
kg (m/sec) = 0.001 + 0.462 U*/(Sc)°-67 MacKay and Yeun (1903)
where U* = (6.1 H- 0.63 U10)°-5 U10/100
-------
Appendix B
Sample Calculations
-------
APPENDIX B: SAMPLE CALCULATIONS
The sample calculations are divided into three sections. The first
illustrates the use of each of the gas mass transfer correlations; the second
illustrates the use of each of the liquid mass transfer correlations. The
third .section demonstrates the overall mass transfer coefficient and
fractional removal calculations for each of the different process units
modeled.
All of the sample, calculations are performed for the cr,innone.nt benzene in
the Union Carbide (Alsop, et al., 1984) plane. The input parameters for the
plant are presented in Table 24. The physical constants for benzene are
presented in Table 25. Other constants needed for the sample calculations are
given below:
Density = p = 1.185 x 10~3 ±/CTS?
axr
-4 2
Viscosity = u = 1.78 x 10 g/cm-sec
Molecular wt. = MW =28.8 g/g mol
3
Density - pR Q = 1.0 g/cm .
Viscosity = Vu n = 1.0 x 10 .g/ciL-sec
H-U
Molecular wt. = MW =18 g/g mol
-52
Diffusivity •= Dn u n ~ 2.5 x 10 cm /sec.
w
1. GAS PHASE MASS TRANSFER COEFFICIENTS
A. MacKay and Yeun (1983) (Eq. N, Appendix A)
kg' (m/sec) = 10~3 + 0.0462 U* Sc_~°'6?
O
V*- = (6.1 + 0.63 U10)°'5U10/100 .
Uj_ = air velocity = 200 cm/sec = 2.0 m/sec .
U* •= (6.1 + 0.63(2))°'5(2/100) =0.0542.
B-2
-------
u -4
_ , ., ., „ air 1.78x10 . 7f.-
Schmidt No. = Sc = - - - - 5 - = 1.707
8 Pair i-air (1.185x10 ) (0.088)
so that Kg' = 10~3 + 0.0462 (0.0542) (1.707)~°'67 = 2.75 x 10~3 m/sec
or Kg = kg'(P /MW ) = (2.75 x 10~3 x — cm)(1.185 x 10"3/28.8)
g & m
-5.2
Kg = 1.13 x 10 g-mol/cm -sec.
B. Thibodeaux and Parkp.r (1974) (Eq. I, Appendix A)
Kg(g-mol/cm2-sec) = 9.9 x 10~7 U r>
Again, U = 200 cm/sec;
L « length of square impoundment or diameter of round
impoundment — e.g., for primary clarifier = 1,940 cm.
Kg = 9.9 x 10~7 (200)(0.088)/(1,940)0>1
—A 9
Kg = 8.2 x 10 g-mol/cm -sec.
C. MacKay and Matsuga (1973) (Eq. G, Appendix A)
Kg(lb-mol/ft2-hr) = 0.0958 U°'78(m/hr) (Sc)~°'67
x Ddn)'0*11? (ib/ft3)/MW
g 8
U = 200 cm/ sec (3,600 sec/hr) . J, m • - 7,200 m/hr
100 cm
Sc = 1.707 (previously calculated in I. A.)
&
D = 19.4 m for primary clarifier
p = 1.185 x 10~3 g/c-n3 x 62.43 Ib-cm3/g-ft3 = 0.074 lb/ft3
6
MW = 28.8 g/g-mol = 28.8 Ib/lb-mo1.
&
Kg = 0. 0958(7, 200)°'78(1.707)~°t67(19.4)~°'11(0.074)/(28. 8)
= 0.1267 lb-mol/ft.2-hr
or Kg (g-mol/cm2-sec) = 0.1267 (1.356 x 10~4
Ib-mol-cm '-sec
-5 2
Kg = 1.72 x 10 g-mol/cm -sec.
-------
I). Sherwood and Pigford (1952) (Eq. 4, Appendix A)
„/•-,/ 2 . . ,,_ lf.-6 ..0.804 -0.196 0.667
Kg(g-mol/cm -sec) = 4.03 x 10 U L D.
i ™ 3 1 r
Again, modeling the primary clarifier (constants given in I.B. cgi units)
Kg - 4.03 x 10-6 (200)°-604 (1940)-°-i%(0.088)0-667
Kg = 1.28 x 10~5 g-mol/cm2-sec.
Reinhart (1977) (Eq. J, Appendix A)
' , p (lb/Ct3)D (ft3/hr)
Kg(lb/ft2-hr) = 3.9 x 10'4 -B - . 1-air -
dimp(ft)
D 1.42r -0.21-0.4,. 0.5
x Re Fr Po 3c
p = 1.185 x 10~3 g/cm3 = 0.074 lb/ft3
O
D, . = 0.088 cm2/sec = 0.341 ft2/hr
Vi = 1.78 x 10~ g/cm-sec = 1.2 x 10~ Ib m/ft-sec
o
For the aeration basin: Power Input (PI) = 75 HP = 41,250 ft-lb/sec
Impeller diameter = d^ = 30 cm = 0.987 ft
r imp
Angular velocity = u = 2,000 rpm = 2,000(2ir/60) = 209.4 rad/sec
Reynolds No. = Re - ^8 = (0.987)2(209.4) (0.074) _ K2H ^ 1Q6
UC (1.2 :•: 10"5)
Froude No. = d, u2/g = (0.987) (209. 4)2/32. 17 = 1,345
imp . c
Power No. = Po = (PI)g ,
c L imp
PL = 1 g/cra3 = 62.4 lb/ft3
Po = 41, ?50(32.. 17)/(62.4)(0.937)5(209.4)3
Po = 0.00247
Sc = 1.707 (as before in I. A.)
B-4
-------
Kg(lb/ft2-hr) - 3.9 x UT* (°-°™><-3A1) d.258 x I06) 1 ' A2(1.345)-°-21
x (C.00248)0'4(l./07)°'5
«=• 119 Ib/ft2-hr
or in terms of mole (28.8 Ib air = 1.0 Ib mole air),
Kg = 4.13 Ib mol/ft2-hr.
—4 2
Converting to cgs units Kg = 4.13 x (1.356 x 10 ) g mol/ca -sec
/ O
or Kg •= 5.6 x 10 g mol/cm -sec.
II. LIQUID PHASE MASS TRANSFER COEFFICIENTS
A. Owens et al. (1964) using reaeration constant (Eq. 7)
K2(base e hr'1) = 50.5 v0'69^1'85
for the clarifier weir
h = 1 cm
v = Q/x-sectional area of flow = Q/irD x h
= 0.07 m3/sec x 106 cm3/ni3/ir(l ,940) (1) = 11.5 cm/sec
K2 = 50.5 (li.5)°'69/(l)1'85 = 259 1/hr
The reaeration constant, K., is related to the liquid phase mass transfer
coefficient as fellows:
(D^,, 0/DQ _H 0>°' 7
= (1) (259/3, 600) (1/18) (1 x 10~5/2.5 x 10~5)°'7
1^ ~ 2.1.1 x 10~3 g-mol/cm2-sec.
Thus, the expression for k. directly is
\ •
B-5
-------
B. Modified Owens et al., (Eq. C, Appendix A)
k, (Ib-mol/ft2-hr) = 3.12(1.024)T-2°.v °'
^ °
T = 25°C
Di-H.,0/D02-H20 = 1/2.5 = 0.40
h = 240 cm = 7.87 ft (for primary clarifier)
For windswept surface,
v = U(ft/sec) x 0.035
o
= 200 cm/sec (1/2.54 x 12)) x 0.035
= 0.23 ft/sec;
then
^ = 3.12(1.024)5(0.23)°'67(2.623)~°*85(0.40)°p66
1^ = 0.315 Ib-mol/ft2-hr
-5 2
or K. = 4.28 x 10 g mol/cm -sec.
For nonagitated regions,
v =0.1 ft/sec
o
so that
.1^ = 3.12(1.024)5(0.10)°'67(2.263)~0>85(0.40)0<66
= 0.1'JO Ib mol/fc2-hr
-5 2
or 1C = 2.45x10 g mol/cm -sec.
C. Mackay and Yeun (1983) (Fq. 9, p. 22)
k^ (m/SRc). = 10~6 + 0.0144 U*2'2 ScL~°'5 for U* 5 0.3 m/sec
= 10~6 + 0.00341 U* Sc~°'5 for U* > 0.3 m/sec
- * L*
where U* = (6.1 + 0.63(U10))°'5 U10/100
B-6
-------
with UIQ = 200 cm/sec = 2 m/sac
U* = (6.1 + 0.63(2))°'5 2/100 = 0.0542 m/sec.
Since U* < 0.3 m/sec, the first equation is ussd.
SCL = V/PVDl-H20=10"2/1(1°"5) =
kJ = 10~6 + 0. 0144(0. 0542)2<2(1,000)~°'5 = 1.75 x 10~6m/sec
or \ ° •ki, X 10°KPL/MWL) = 1-7-'i x 10~A(1/18)
f\ *y
le = 9.7 x 10 g mol/cm -sec .
For comparison purposes, k_ calculated by the second equation is as
follows:
Id = 10~6 + 0. 00341(0. 0542)(l,000)~°'5
= 6.84A6 x 10~6 m/sec
1 6.8^4 x 10~4/18 = 3.8 x 10~5 g mol/cm2-sec .
D. Thibodeaux (1978) (Eq. B, Appendix A)
where (POWR) = net agitator horsepower.
For the aeration basin,
POWR = (PI) x eff. = 7-5 -x-O-jfr.-'* 60 Hp
T = 25°C
a = 0.85, a constant
J = 3, a constant
Di-H,0/D0,-H 0 =0.4
2 2
a V ' = area of agitation = 96 m =96 >: 10.76 = 1,033 ft .
. 3(60)(1.024)5(0.85 x 1Q6) ._ ..0.5 ,_0 ., , ,,,2 .
\ '" - (165.04X1.033) - (°'4) " 639 Ib/mol/ft -hr
l = 639(1.356 x 10~4) = 0.0867 g-mol/cm2-sec .
B-7
-------
III. PROCESS SPECIFIC CALCULATIONS
In this section, the overall mass transfer coefficients and fractional
removal calculations for each process are outlined. However, it is first
instructive .to derive the fractional removal equations that were presented in
the text and subsequently used here. The fractional removal equations
employed Depends on the model used to describe tho process. There are two
different models that are employed for these calculations, the plug flow
model, and the perfectly mixed reactor model.
In plug flow, there is no back or radial mixing in the liquid steam.
Thus, ideally any differential segment can be thought of as a batch reactor
with a reaction time equal to the residence time of the fluid in that reactor
segment. Mathematically, this implies:
/c
out dc
4~ r
III-l
where T = . residence time, sec = V/Q;
r =. rate of appearance, g mol/cm -sec;
c " concentration of desired species, g mol/cm ;
3
V = reactor volume, m ; and
Q " volumetric flow rate, m /sec.
The reaction rate, in the cases to be considered, is simply the
evaporation or mass transfer rate as given below:
r = -K . aX III-2
cL
2 3
where a = the surface area per volume, cm /cm ;
*>
K = the overall mass transfer coefficient, g mnl/cm^-sec; t,nd
OL
X = the mole fraction of the desired species.
Mote that we can express C in terms of X as follows:
B-8
-------
c = (PL/MWL)X = PL x
where MW. -• molecular weight of the waste stream g/g-raol total,
L«
p = average density of the waste stream g/cm , and
L
p1 = molar density = p /MW , g mol total/cm .
Li Li LI
Then dC = ?'.
-------
X — X
or R . - -i2_— = K.A/Qp' III-6
air X oL x L
where R = ratio of loss (air emissions) to discharge.
sir
X, X X,
i_ in - out in ,
Note that Rair = — - - = — - 1
out . out.
Xout/Xin '-
So that the fraction of influent lost the air is:
Fair ' 1 - Xo.,t/Xln ' l " 1/(Rair + » ' (1 + Rair)/(1 + Fair>
- 1/(1 + R.ir>
F , = R . /(I + R . ). TTT-?
air air air --- '
The denominator in Equation III-7 can be thought of as the sum of the
ratios (the ratio of discharge to discharge must always be one). Other
reaction terms can be included (e.g., biooxidation) . Thus, in general,
F . = R . /ZR . III-8
air air
Other mechanisms are included in the reaction rate expression. For example,
if biooxidation is included in the perfectly mixed reactor mass balance,
Flow in - Flow. out. •= loss to air + loss by biooxidation
Qp,' (XJ - X ) = K t A X + B . Vp,' X
L in out oL out i L out
tx _ A I\_A
• <- * -i""* - -r v
x/Ut Li »*
where B = Biological decay constant, sec .
K A
Note that the ratio of air loss to mrss discharged is still P. . «• n i
air ^^t
The ratio of biooxidation loss to the mass discharged is R_. .. = ^JT«
Thus,
F . •= R . /ZR = R./(1 + XD. ,+R.).
air air air Biol ait
5-10
-------
A. Pretreatment
For estimating Kg, the correlation of MacKay and Yeun is used. Thus, Kg
-52
= 1.13 x 10 g mol/cm -sec (refer to I. A.).
The correlation of Owens ct al. (TI-A) is used to estimate 1C. .
For benzene in the pretreatment unit:
Di-H/ V-H9o = °'4;
z / z 2
h = 3 in = 300 cm, A = surface area = 50 m ;
v = Q/X-sec Area = 0.07 x 102/(50)°*5(3) = 0.33 cm/dec;
^ = 7.33 x 10~* (0.33)C'67(300)"°-85(O.A)0-7
1C, = 1.44 x 10 g mol/cm -sec.
Using the partition coefficient for benzene, K = 305, the overall mass
transfer coefficient is:
"5"1
KOL = (1/XL + ^(KW' = d/1.44 * 10~ + 1/(305)(1.13 x ID))
— 1\ 9
K «». 1.A4 x 10 g mol/cm -sec ,
Ulj
Assuming plug flow through the pretreatment unit, Equation III-5 applies.
1 •? exp[-1.44 x 10~°(50)/(0.07 x 100/18)]
1-0.998
F " 0.0002.
air
B. Clarifier ... . ... .-
The entire volume of the clarifier will not be available for the flow of
wastewater. Therefore, the stream depth is some fraction, f, of the clarifier
depth. Letting h, = clsrifier depth (cm) and h = stream depth (cm), then h,
is related to h as follows:
s
h = fh, .
s d
Using the stream depth in the reaetation equation of Owens et al. (1964)
(Eq. 8) yields:
v fv -1\ en c °-67//i 1.85.1.85.
K-(Kr ) - 50.5 v /(h, r ).
B-ll .
-------
Converting to the liquid mass transfer coefficient as in sample calculations
II.A:
K. = (50.5/3,600)(1/18) h. v°-67/(h.1-85f1'85) .
L da
Simplifying k_ = 7.79 x 10~4 (v°-67/h ,°'85)(D. u n/D. „ -,)°'7 f"1" 5 .
T. d i-H2o o2-H2n
The inlet to the clarifier is near t!ie center and below the surface by 30 cm
or 1 foot. As the waste water travels to the periphery of the clarifier, the
velocity of the stream decreases. The mass transfer coefficient is
accordingly reduced. The velocity varies in proportion to the inverse of the
radius, R .
c
V = Q(M3/sec)/2ir R f h ) .
The time weighted average mass transfer coefficient can be obtained by
integrating the function predicting the liquid phase mass transfer coefficient
between the limits of the radius.
Thus, the average liquid phase mass transfer coefficient can be written
as follows:
/r=R IL 2rdr
.-0C "V '
Substituting for K, » the following equation can be obtained:
Lt
\ - y*7.79 x 10-4(Dl_H 0/DC _H c)°'7 h/-85 f-1'85 v-67 2r R^2 dr .
Again substituting for the velocity as a function of the radius,
7.79 x 10-4 (DI_H 0/D0 _ri 0)°'7 hd-°-85 r1'85
0 ^\ ' & '7
x 2r R •=— r . dr .
c 2ir rf h,
a
Separating constants from variables before integrating, the following
expression can be obtained:
-------
*L ' 7'79 * 10"4(Di-H20/D02-«2°)0"7 hd~1'52 f"2'52
D
x R "" Q'67(2)(.29189) f C r°'33 dr .
r°':
*'o
After integration,
10-4(D /Dn ,0.7 -1.52 f-2.52
i-H20 02-H20 d
R , 33
/R i
. /b
Finally, ^ = 3.A2 X lO^CD^^/D^.^)0'7 h/1'52 f-2-52(Q/Rc)0-67 .
For benzene in the primary clarifier:
Di-H20/D02-H20 * C'4'
h, = 240 cm
a
Q = 0.07 x 106 cm3/sec .
R = 1/2 (19. 4) (100) = 970 cm.
c
Assuming f = 0 . 1 .
then, i^ = 3.4.2 x 10~4(0.4)°'7(240)~1 >52(0. 1)~2>52(70,000/970)°>67
-4 2
«. 2.53 x 10 g-mol/cm -sec.
The gas phase mass transfer coefficient is estimated from the correlation
of Mackay and Yeun (1983).
K = 1.13 x 10~ g/mol/cm -sec. (Sec. 1, Part A) .
O
Thus KQL = (l/I^ + 1/(K K ))~1 = (1/2.53 x 10~A + 1/(305)(1.13
= (4.243)"1 =2.36 x 10~A .
Assuming plug flow, Equation 30 applies so that
X /X. = exp(-K ,A/100 Op1)
out In r oL • m
= exp -2i36 x 10~4(ir(19.4)2/4)/[(0.07)(100)/18]
= exp(-0.179) = 0.8358 .
B-13
-------
The fraction lost to the air is
F , = 1 - X XX, = 1 - 0.8358
. air out in
F , = 0.1642 .
air
C. Clarifier Waterfall
Estimate 1C using the correlation of Owens et al. (1964), but assuming 90
percent reduction since the turbulence in the freefalling waste stream is
expected to be reduced. That is,
VL = 0.10 1C ' = 0.10(2.11 x 10~3) = 2.11 x 10~4 g-mol/c.fl2-sec.
Again, the gas phase mass transfer coefficient is estimated by the correlation
of MacKay and Yeun (1983).
-5 2
K = 1.13 x 10 g-mol/cm -sec (Refer to Section I, Part A).
O
K T = (1/2.11 x 10~4 + 1/(305)(1.13 x lO'5))"1
OL
-4 2
= 2.0 x 10 g-mol/cm -sec.
Modeling the stream as plug flow, Equation 26 applies. Hence,
Xout/Xin ' e*P<-KoLA/(K)-
A = irDL = 1,940 ir(30) = 182,840 cm2.
Q = 70,000 cm3/sec.
" X XX - exD -2 x 10"A(182,C40) _
Xout/Xin exp 70,000/18 " °'99-
F . •= 1 - X XX, = 0.01.
air out in
D. Equalization Basin
The model of MacKay and Yeun (1983) is used to estimate K . Thus, K =
-52 88
1.13 x 10 g-mol/cm -sec (Refer to I.A.). The modified Owens' equation is
used to estimate K^ . From Section II.B:
1C = 0.717(1.356 x 10~4)(h/3)"°*85
(h in feet)
-------
For the equalization basin, h = 3 m = 9.84 ft.
1C = 3.54 x 10~5 g-mol/cm2-sec.
Then,
Km = (1/3.54 x 10"5 + 1/(305)(1.13 x 10'5))"1
UL
-5 2
= 3.51 x 10 g-mol/cm sec.
Assuming the equalization basin to be well nixed, Equation III-6
applies.
R = 3.51 x 1G~5 (5,185)7(0.07 x 100/18)
3 iJT
R . = 0.468.
air
Assuming no biological activity, the fraction of influent emitted to the
air is
F . = R . /(I -i R J ) = 0.468/1.468 = 0.3188.
air air axr
E. Aeration Basin
1. Agitated Zones: Estimate 1C using the correlation of Thibodeaux
(1978); estimate K using the correlation of Reinhart (1977).
&
7
fL = 0.0867 g-mol/cm -sec. (From Section II, Part D) .
-4 2
K = 5.6 x 10 g-mol/cm -sec. (From Section I, Part E) .
gag
K-. = (J./0.0867 + 1/(305)(5.6 x ID"4))"1 = 0.0575 g-mol/cm2 -sec.
2. Nonagitated Zones; Estimate K, using the modified correlation of
Owens et al. (1964); estimate K using the correlation oj. MacKay and Yeun
(1983).
v = 0.1 ft/sec.
h = 360 cm = 360/(12 x 2.54) = 11.81 ft.
From Section II, Part B;
K/ = 3.12(1. 356 x 10-4)(1.024)T-2°v°-67/(3.937)0-85
Lnon— ag
B-15
-------
- A.23 x ID'* (1.024)5(0.1)°-67(3.937r0-85(0.4)°-66
1C = 1.74 :: 10~5 g-mol/cnr/sec
t Lnon-ag °
From Section I, Part A
K = 1.13 x 10"5 g-aaol/cm -sec
gnon-ag
then K . = (1/)C + 1(KK ))-!
oLnon-ag Lnon-ag gnon-ag
= (1/1.74 x 10~5 + 1/(305)(1.13 x 1C"5))"1
-5 2
K , = 1.73 x 10 g-mol/cm -sec.
oLnon-ag
The average overall mass transfer coefficient is calculated by weighting
the agitated and nonagitated overall mass transfer coefficients by their
respective surface areas.
The tota] surface area of the aeration basin is
Atot = (170)(170) = 28,900 m2.
2
A =96 (No. of r.gitators) = 96(30) = 2,880 m . Therefore, the
ag
The surface area being agitated is:
A =96 (No. of
ag
nonagitated surface area is:
A = -V _ - A = 28,900 - 2,880 = 26,020 m2.
non-ag tot ag
Thus, K T = (K . A + K . A )/A_ _
oL oLag ag oLnon-ag non-ag tot
= [0.0575(2,880) - 1.73 x 10"5(26,020)]/28,900
-3 2
= 5.75 x 10 g-mol/cm -sec.
The aeration basin is assumed to be well (perfectly) mixed. Consequently, the
ratio of air emissions to discharge is given by Eq. 36.
R . - K , A /(Qp1 x 100)
air oL tot m
= (5.75 x 10~3)(28,900)/(7/18) = 427.
The ratio of biologicfl degradation to discharge is simply the biological
decay rate constant times the residence time.
V± = 1/400 hr = 1/(400 x 3,600) = 6.944 x 10~7 sec"1
T «• V/Q <= (170)(170)(3.6)/(0.07) = 1.486 x 106 sec.
B-16
-------
So that
R. . = B.T = 1.032.
bio i
Since the ratio of discharge to discharge must be 1.0, the sum of the
ratios is:
R t = R,. + R,. + R . = 1 + 1.032 + 427
tot dis oio air
R " A29.032.
tot
Thea, the fraction removed by the air Is:
Fair ' Rair/Rtot = «7/«9.032 = 0.9953,
and the fraction removed by biological docay is:
1.032/429.032 - 0.0024.
I.., V,
bio Taio tot
F. Subsurface Aeration
It is assumed that ti-.e bubbles attain an equilibrium concentration based
upon the thermodynamic partitioning of the VOCs into the gas phase. Thus the
mass transfer of VOCs to the atmosphere induces both a surface diffusion and
a bubble equilibrium term as given in Eq. 50.
Flux to air = ^^ + p1 Q K X^A .
The overall mass transfer coefficient, K , is estimated using the same
\)Lj
correlations and averaging techniques as the surface aerated system wher. the
subaeration basin is open to the atmosphere. If the subaeration basin is
covered, then only mass transfer to the bubbles needs to be considered (i.e.,
the gas in the headspace is comprised solely of bubbles already assumed to be
saturated with VOCs; hence no further emissions of VOCs to the gas in the
headspace is possible) .
For the closed UNOX system, K = 0, so that only mass transfer to the
OL
bubbles needs to be considered. However, in the UNOX system, the waste stream
is divided into four reactor trains consisting of three reactors per train.
Fresh air is fed into the first reactor of each train only, and this gas is
B-17
-------
concurrently recirculated to the next two. Offgases of the third reactor are
vented to the atmosphere.
It is assumed that the total liquid and gas flow is divided equally
between the four trains. Consequently, each train has an identical fractional
loss which is equivalent to the overall fractional loss cf the system. This
assumption permits modeling of the UNOX system of four trains as a single
train that has reactors four times as large as the UNOX reactors.
Considering the system as one train:
Where Y, is the mole fraction of (for this example) benzene in the gas
4
stream. We assume each reactor to be perfectly mixed so that the mole
fraction of benzene in Reactor 1 is X_, in Reactor 2 it is X_, and in Keactor
3, it is X.. We further assumed that the outlet gas of each is in equilibrium
with the liquid in that reactor. Therefore, YZ = KX , Y = KX , and Y4 = KX^,.
We assume Y
1
0. Also we assume that each reactor contains the same volume
of liquid, and since Q ramains constant through che train, each reactor has
i-»
the same residence time, T = V/Q . (Note: This volume, V, is the actual
Lt
volume of the reactors times the number of trains in the system.)
Making a mole balance on benzene around each reactor, and accounting for
the biological oxidation rate, B , yields:
Flow in: liquid + gas = Flow out: liquid + gas + amt. biooxidation
Reactor 1 Q^ - Q^z + QgpgV+ ViX2
Noting that t = V/Q , if = 0, Y = KX , and rearranging yields:
Lt ' 1 Z- Z
gg 0) + B± TX2 . ' 111 F-l
Since, for any one component in a train, (Q p'/Q,pJ)K and B T remain constant,
let .
(Q p'/Q p.')K and D
g g L, >-
B.T.
Equation III F-l becomes
Reactor 1
Xl - X2
Ill F-2
B-18
-------
Similarly, for Reactor 2 and 3,
Reactor 2 X2 - X = C(X - X ) + DX ; III F-3
Reactor 3 X - X, = C(X - X ) + DX . Ill F-4
Adding Equations III F-2, III F-3, and III F-4 yields
X - X, = CX, -f D(X + X + X,) . Ill F-5
Solving Equations III F-3 for X« and III F-4 for X_ yields
X2 = X3d + C + D)/(l + C) ; III F-6
X3 = X4(? + C + D)/(l + C) ; III F-7
which implies X2 - X^d + C + D)2/(l + C)2 . Ill F-8
Substituting Equations III F-7 and III F-8 into Equation III F-5 gives:
Xj - X4 = CX4 + DX^d + d + C + D)/(l + C) + d + C + D)2/(l + C)2) .
Total loss = loss to air + loss by biooxidation.
The ratio of air emissions to discharge is
air 44 g g L L
Similarly, the ratio of biooxidative loss to discharge Is:
RBlol = DX^(1 + (1 + C + D)/(l + C) + (1 + C + D)2/(l + C)2)/X4
= D(l + (1 + C + D)/(l + C) + d + C + D)2/(l + C)2). Ill F-10
The series form of Equation III F-10 suggests that, in general, the ratio
of biooxidatior. to discharge is:
= D [(1 + C + D)/(l + C)]"'1 III F-ll
where n = the number of reactors in series (i.e., the number r.f reactors in
one train); ' .
D = B±T
and C = Q (
B-19
-------
For the UNOX system, the overall liquid and gas flow rates are:
Q. = 0.07 m3/sec p.' = 1/13 g-mol/cm3
Lt LJ
Q = 0.40 m3/sec p' = 1.185 x !0~3/23.8 g-mol/cm3.
8 g
Thus, for benzene, the ratio of air to discharge is:
= C = Q p'K/(Q.pT') = 0.4(1.185 x 10~3) (305) (18)/(28.8 x 0.07)
g g L L
!< , <= C = 1.291.
air
The biooxidation rate for benzene is BI = (1/400)(1/3,600) = 6.944 x 10"7
sec" . The residence time, T, is T = V/Q = L x W x D x No. of trains/Q .
For the UNOX reactors L = 8.37 m, W = 8.4 m (Note: In Table 21, the
length has already been corrected for the number of trains. L = 8.37 x 4 =
33.48 m). Thus, T = 8.4 x 33.5 x 8.5/0.07 or T = 34,170 sec. Then,
D = B T = 0.02373
1 + C = 2.291
1 + C + D = 2.3147 .
The ratio of biooxidation to discharge is:
RD. , = 0.0237(1 + 2.3147/2.291 + (2.3147/2.291)2)
DlOl
= 0.0719.
R^ = 1 + R . + R_. , = 1 + 1.291 + 0.0719 = 2.3629.
tot air Biol
Therefore, the fractional losses are:
F . = R , /k_ » 1.291/2.3629 = 0.5464 -
air air tot
FT>, , " RT,., i/R = 0.0719/2.3629 = 0.0304.
DlOl . DlOl ~*"
B-2H
-------
Appendix C
Estimation of Volatilization Rates Based on
Changes in Concentration of a Reference Compound
-------
APPENDIX C
ESTIMATION OF VOLATILIZATION RATES BASED ON CHANGES IN
CONCENTRATION OF A REFERENCE COMPOUND
An approach. suggested by CAQPS (Durham, 1985) involves prediction of
volatilization rates of VOCs in the absence of other pathways based on the
concentration profile of one reference compound as it moves through a series
of unit operations. This approach could be used to indicate the presence of
biooxidation or decomposition if observed concentration ratios deviated from
predictions.
At steady state, assuming a perfectly mixed tank, the component balance
for a wastewater containing two VOCs is:
QC£1 = QC1 + Ej +>! = QCj + KouClA + DI (1)
= QC2 + KQL2C2A + D,, (2)
where
A = surface area of treatment tanks, m2
C = concentration in tanks, g/m3
C, = inlet concentration, g/m3
D = decomposition rate, g/s
E = emission rate, g/s
Q = flow rate, ma/s
KAT = overall mass transfer coefficient, m/s.
ui»
Subscripts refer tc compounds 1 and 2. .•
Rearranging and dividing Equation (1) by (2) gives
K/* ' rt AO fcTi ' • - -
nT-l^l X "»*1 "l
OL1 1 „ 1 1 /-ox
Kr~ ~ n Ar -n l '
r>OL2L2 y aL2 U2
whera AC. = C,.. - C} and AC. = C,_ - C--
If no decomposition is taking place, then D.. = D_ = 0, and
Cj _ K_, AC,
c Y
C2 KOL1
C-2,
-------
If the gas phase resistance is small relative to liquid phase resistance, then
d 0.67
OL2 2
C. d_ 0.67 AC.
cT(df> AC; («
i A.
where d. and d_ are the dif fusivities cf compounds 1 and 2 in water, cm2/sec.
Individual pair of compounds can be analyzed according to Equation 50.
If one of the compounds is not likely to decompose (e.g.,. D.. = 0 in Equation
(48), then D- can be calculated.
Though Equation (51) is theoretically sound in some cases, there were
several assumptions made in its derivation that may limit its utility. First,
process variations often cause significant changes in the concentrations of
components in the influent stream, and consequently these variations invalidate
the steady state component balances of Equations (46) and (47). Second, not
all waste treatment units are perfectly mixed tanks so that the constant
emission rate substitution (i.e., Ej = KjCxA) used in Equations (46) and (47)
are merely approximations. Third, the dependence of the overall mass transfer
coefficients on the component dif fusivities, as expressed in Equation (51), is
not always correct. For surface agitated systems, the liquid phase mass
transfer coefficient varies with the square root of the component diffusivity
as seen by the correlation of Thibideaux (1978) (Equation B in Appendix A).
For covered submerged aeration where the gas is. in equilibrium with the liquid
(as in the UNOX system), the overall mass transfer coefficient is independent
of the component diffusivity (i.e., Km./KnTO = K^ where Kx and K2 are the
ULii \JLi£
partition coefficients). This, for steady-state, well-mixed process units,
Equation (51) may still be inappropriate. In such as case, however, Equation
(4) could be used and the overall mass transfer coefficient ratio could be
calculated by applying the correct model for the specific process unit in
question.
Besides the theoretical limitations of Equation (6), the analysis has
several practical limitations. The foremost of these limitations are the
inaccuracies of the chemical anlysis. If there is either a very large or a
very small reduction in a component's concentration across a process unit,
C-3
-------
then C. or AC., respectively, will approach zero. In such an instance, inaccur-
acies in the chemical analysis can have a large effect on the component ratios,
and Equation (5) would incorrectly suggest the presence of alternative path-
ways. Inaccuracies also exist in the diffusivity data and mass transfer
calculations.
Daily and average daily composite treatment process flows at the Union
Carbide Plant were analyzed according to Equation (6) as follows:
C AC ' d 0.67 = K
2 1 l OL1
The terra R as determined by the ratio of dif fusivities or by the ratio of
calculated overall mass transfer coefficients was in good agreement. These
predicted values of K ranged from 0.8 to 1.1.
The term R as calculated by daily process flows varied widely (-1.4 tu
+110) due to process variations which caused large fluctuations in the component
concentrations. The term R as calculated by the average flow data was much
improved over the R calculated from daily flows (i.e., many R's =1). However,
for the aerated process units, large fluctuations in R persisted. These
aerated process units exhibited large component reductions so that the chemical
analysis error, not biooxidation, is thought to be responsible for the observed
fluctuations.
Due to the theoretical limitations and the inherent inaccuracy of the
chemical analysis and diffusivity data, Equation (6) has very limited applica-
bility to industrial processes. However, Equation (6) may have some utility
on a laboratory scale since the inherent assumptions can be more easily rea-
lized, and the component removal controlled so as to prevent C. or AC. from
approaching zero.
C-4
-------
Page Intentionally Blank
-------
Page Intentionally Blank
-------
Page Intentionally Blank
------- |