FPA Report Number 68-03-002
-1 May, 1982
Air and Steam Stripping
of Toxic Pollutants
Volume I
by
D. J. Goldstein
Hater Purification Associates
238 Main Street
Cambridge, MA 02142
for
U.S.E.P.A.
Industrial Environmental Research Laboratory
Cincinnati, Ohio 45263
Contract No. 68-03-3002 Task 4
Project Officer Harry Bostian
-------
Table of Contents
Summary 1
List of Figures ; H
List of Tables 12
1. Introduction and Outline 13
2. Conclusions !5
3. Recommendations 16
4. VAPOR LIQUID EQUILIBRIA
4.1 Introduction to the Theory of Vapor-Liquid Equilibria of
Organic Molecules in Dilute Aqueous Solution 18
4.2 Activity Coefficients for Slightly Soluble Compounds 21
4.3 Vapor Pressure as a Function of Temperature 24
4.4 Activity Coefficient as a Function of Temperature 25
4.5 Extrapolation of Activity Coefficients Measured at High
Concentration - Activity Coefficient as a Function of
Concentration 26
4.6 Direct Measurement of Henry's Law Constant 31
5. THE THEORY OF STRIPPING
5.1 Organization of the Section, Batch and Continuous Stripping,
Isothermal and Adiabatic * 33
5.2 Evaporation from a Lake 36
5.3 Isothermal Batch Stripping 36
5.4 Adiabatic Batch Stripping with Air 39
'5.5 Isothermal, Counter Current Stripping and Estimation of
the Height of an Equivalent Theoretical Plate 45
5.6 Number of Equilibrium Stages in Continuous, Isothermal
Stripper 51
5.7 Pollutants Classified by Ease of Stripping 57
-------
5.8 Adiabatic Continuous Stripping with Air 57
5.9 The Use of Air-Steam Mixtures 70
5.10 The Effect of Recirculating the Water 70
6. THE EFFICIENCY OF STRIPPING EQUIPMENT
6.1 Introduction and Summary 77
6.2 Factors Effecting the Efficiency of a Bubble Cap Tray 77
6.3 Factors Effecting the Efficiency of a Packed Bed 79
6.4 Two Film Theory of Mass Transfer 80
6.5 Choice of Equipment 84
6.6 Choice of Packing or Trays 85
7. COMPARISON OF PREDICTIVE CALCULATIONS TO STRIPPING DATA AT LOW
CONCENTRATIONS
7.1 Methodolory of Comparison 88
7.2 Stripping Data at Low Concentration 89
7.3 Data Analysis and Results 91
7.4 Development of a Correlation for the "Effect of Liquid-diffusivity
on Tower Sizing and Organic Removal Efficiency 97
References 102
Appendices
1. Adiabatic Batch Stripping Using Air 106
2. Calculation of Water Temperature in a Continuous Air
Stripper 109
3. Henry' s Law Constants 114
ii
-------
SUMMARY
A compilation has been made showing the Henry's Law constant (B)
as a function of temperature for all the organic pollutants on the toxic
pollutant list. Henry's Law is
(concentration in the vapor)(total pressure) =
(concentration in water solution)(B)
so that B is a measure of the apparent vapor pressure over a solution in water.
For poorly soluble compounds the water molecules try to expel the organic
molecules and B is much higher than the vapor pressure over the pure organic.
The ratio (B/vapor pressure) is called the activity coefficient. The activity
coefficient is inversely proportional to solubility for poorly soluble
compounds and often has a value in the thousands or higher.
Most values of B have been .determined from data on the effect of
temperature on vapor pressure and from a knowledge of the solubility at one
temperature. The solubility has been converted to an activity coefficient
and the activity coefficient has been extrapolated to other temperatures.
This procedure is believed to be the most reliable available. When B has
been measured at one temperature it is still the activity coefficient, and
not B, which has been extrapolated to other temperatures.
For the few toxic pollutants which are miscible with water the
activity coefficient has been estimated from the vapor-liquid equilibrium
data or from the azeotrope data (azeotropes are common) using the technique
of van Laar.
We believe the estimates of B to be reasonably satisfactory for all
except 33 of the 186 listed pollutants.
A brief development of the theory of stripping is given with the intent
of classifying pollutants by the ease with which they can be stripped. A
classification proved possible and will be found on Table 5-2 which is
reproduced at the end of this summary. Compounds which are "very easily
stripped" can be reduced to about 1/1,000th of the feed concentration by
low vapor rates in columns which are quite short (5 to 10 ft), rather like
small cooling towers. Sixty-eight of the 186 listed pollutants are very
easily stripped. "Intermediate" and "difficult to strip" compounds may
require' columns up to twenty times as high and will also need high steam
-------
or air rates. "Very easily stripped" compounds have a half life in a still
body of water on the order of 12 hours.
We compared air and steam stripping at vapor rates which gave approxi-
mately the same cost for the vapor. Air rates were taken at about eight tines
the steam rate. On this basis most compounds can be stripped as easily with
air as with steam because the Henry's Law constants tend to be multiplied by
about eight when the temperature is raised from 20°C to 100°C. In general
air stripping will be used when the contaminant can be released to the
atmosphere and steam stripping will be used when the contaminant must be
recovered.
When air is the stripping vapor, water will also evaporate and the water
stream will cool down which decreases E. Procedures for calculation are
given. It is suggested that a good control is to add steam to the air to
saturate it, particularly on cold, dry winter days.
A brief discussion on available equipment and on the comparison of
predictive calculations to stripping data at low concentration. As expected
and supported by predictive calculations, most stripping data indicated
significant removal of compounds with high Henry's Law Constant (H @ 100°C >
20 atm). Deviations from this behavior were observed for certain compounds. .
In case of packed towers stripping a waste stream consisting of
compounds with Henry's Law Constant higher than 100 atm, the height of
packing or the removal efficiency was found to be nearly independent of the
magnitude of Henry's Law Constants. Liquid-phase resistance (diffusivity)
was found to be affecting the performance. Developing better contacting
devices in terms of reducing liquid-phase resistance through more efficient
packing and better liquid distribution in plate columns seems to be the
probable solution to improving the performance of strippers.
Available data on stripping and stripping equipment is so limited that
meaningful comparisons between predictive calculations and operating data
are very difficult to make, particularly when there is so much uncertainty
in making on-site analyses of organic compounds at low concentrations.
-------
Table 5-2 TOXIC POLLUTANTS CLASSIFIED BY EASE OF STRIPPING
Very easily stripped
H(100°C)>100 atm
H(20°C) 9 13 atm
5-4"1" 2-Chloroethyl vinyl ether
7-13 Triethyl amine
9-1 Benzene
9-2 Chlorobenzene
9-3 1,2-DiChlorobenzene
9-4 1,3-Dichlorobenzenj _
9-5 1,4-Dichlorobenze
9-6 1,2,4-Trichlorobenzene
9-7 Hexachlorobenzene
9-8 Ethyl benzene
9-10 Toluene
9-15 Benzyl chloride
9-16 Styrene
9-18 Xylenes
10-1 2-Chloronaphthalene
** 10-3 Benzo(b)fluoranthene
** 10-4 Benzo(k)fluoranthene
** 10-5 Benzo(a)pyrene
10-9 Acenaphthene
** 10-12 Chrysene
** 10-13 Fluoranthene
10-15 Naphthalene
10-17 Byrene
4- the left hand column are the code numbers used in the Treatability
Manual and in Appendix 3.
** probably, but data is poor
-------
(very easily stripped - continued)
11-1 Aroclor 1016
11-2 Aroclor 1221
11-3 Aroclor 1232
11-4-- Aroclor 1242
11-5 Aroclor 1248
11-6 Aroclor 1254
11-7 Aroclor 1260
12-1 Methyl chloride
12-2 Methylene chloride
12-3 Chloroform
12-4 Carbon tetrachloride
12-5 CM oroe thane
12-6 1,1-Dichloroethane
12-7 1,2-Dichloroethane
12-8 1,1,1-Trichloroethane
12-S 1,1,2-Trichloroethane
12-10 1,1,2,2-Tetrachloroethane
12-11 Hexachloroethane
12-12 Vinyl chloride
12-13 1,2-Dichloropropane
12-14 1,3-Dichloropropene
12-15 Hexachlorobutadiene
12-16 Hexachlorocyclopentadiene
12-17 Methyl bromide
12-18 Dichlorobromomethane
12-19 Chlorodibromomethane
12-20 Bromoform
12-21 Dichlorodifluoromethane
12-22 Trichlorofluoromethane
12-23 Trichloroethylene
-------
(very easily stripped - continued)
12-24 1,1-Dichloroethylene
12-25 1,2-Trans-dichloroethylene
12-26 Tetrachloroethylene
12-27 Ally! chloride
12-30 Ethylene dibromide
13-20 Heptachlor
13-25 Toxaphene
13-37 Isoprene
13-46 Carbon disulfide
14-4 Amyl acetate
14-5 n-Butyl acid
14-13 Vinyl acetate
15-1 Methyl mercaptan
15-3 Cyclohexane
Easily stripped
H(100°C) 2Q to 100 atm.
H(20°C) 2 to 13 atm.
5-3 Bis(2-chloroisopropyl) ether
7-7 Acrylonitrile
9-19 Nitrotoluene
10-10 Acenaphthylene
10-14 Fluorene
10-16 Phenanthrene
13-8 Aldrin
13-9 Dieldrin
13-24 Chlordane
13-26 Cap tan
14-1 Acetaldehyde
14-16 Acrolein
14-18 Propylene oxide
** probably, but data is poor
-------
Intermediate
H(100°C) 8 to 20 atm.
H(20°C) 1 to 2 atm.
5-2. B1s(2-chloroethyl) ether
8-5 Pentachlorophenol
8-5 2-Ni trophenol
9-9 - Nitrobenzene
9-11 2,4-Dinitrotoluene
9-12 2",6-Dinitrotoluene
10-11 Anthracene
12-31 Epichlorohydrin
13-12 4.4'-DDD
13-21 Heptachlor epoxide
Difficult to strip
H(100°C) 4 to 8 atm..
H(20°C) 0.5 to 1 atm.
8-2 2-chlorophenol
8-3 2,4-Dichlorophenol
8-4 2,4,6-Trichlorophenol
14-15 Crotonaldehyde
15-4 Isophorone
* this is for air; a higher category for steam
** probably, but data 1s poor
-------
Very difficult to strip
H(100°C) 2 to 4 atm.
6.2 Diethyl phthai ate
8-1 Phenol
8-10 2,4-Dimethylphenol
8-13 4,6-Dinitro-o-cresol
8-14 Cresol
9-17 Quinoline
14-3 Ally! alcohol
Cannot be stripped
5-7 Bis(2-chloroethoxy) methane
6-1 Dimethyl ph thai ate
7-10' Ethylenedlamine
8-7 4-N1trophenol
8-8 2,4-Dinitrophenol
8-9 Resorcinol
9-13 Aniline
9-14 Benzole acid
10-8 Benzo(ghi) perylene
13-2 Endosulfan sulfate
13-13 Endrin
** 13-18 Diurone
13-22 Carbofuran
** 13-28 Coumaphos
13-29 Oiazinon
13-30 Dicamba
13-31 Dichlobenil
13-32 Malathion
13-33 Methyl parathion
13-34 Parathion
* this is for air; a higher category for steam
** probably, but data is poor
-------
(cannot be stripped - continued)
13-35 Guthion
13-38 Chlorpyrifos
13-39 Dichlorvos
13-41 Disulfoton
13-43 Mexacarbate
13-44 Trichlorfon
14-2 Acetic acid
14-6 Butyric acid
14-7 Formaldehyde
14-8 Formic acid
14-12 Propionic acid
14-14 Adi pic acid
15-2 Dodecyl benzenesulfonic acid
15-5 Strychnine
15-7 Zinc phenol sulfonate
Poor data, but probably difficult to strip
5-5 4-Chlorophenyl phenyl ether
6-3 Di-n-butyl phthalate
6-5 Bis(2-ethylhexyl) phthalate
7-1 N-nitrosodimethyl amine
7-2 N-nitrosodiphenylamine
7-5 3,3'-Diphenylhydrazine
7-6 1,2-Diphenylhydrazine
13-11 4,4'-DDT
13-17 Kepone
13-40 Diquat
13-42 Mevinphos
14-9 Fumaric acid
14-10 Maleic acid
-------
Poor data; better data worth obtaining as may be stn'ppable
5-6 4-Bromophenyl phenyl ether
6-4 Di-n-octyl phthalate
6-6 Butyl Benzyl phthalate
7-3 N-m'trosod1-n-propylam1ne
7-4 Benzldlne
7-9 Di ethyl ami ne
8-12 p-Chloro-m-cresol
10-2 Benz(a)anthracene
10-6 Indeno(1,2,3-cd)pyrene
10-7 D1benzo(ah)anthracene
13-10 4,4'-DDE
13-27 Carbaryl
14-17 Furfural
15-6 2,3,7,8-Tetrachlorodibenzo-p-dioxin
j?ata inadequate for comment
7-8 Butyl amine
7-11 Monoethylamine
7-12 Monomethylanrine
7-14 Trimethylamine
12-28 2,2-Dichloropropionlc acid
13-1 s-Endosulfan
13-3 S-Endosulfan
13-4 o-BHC
13-5 8-BHC
13-6 6-BHC
13-7 T-BHC
13-14 Kelthane
13-15 Naled
13-16 Dichlone
13-19 Endrin aldehyde
• 9
-------
(data inadequate for comment - continued)
13-23 Mercaptodimethur
13-36 Ethion
13-45 Propargite
14-11 Methyl methacrylate
Decompose in water
5-1 Bis(chloromethyl) ether
12-29 Phosgene
10
-------
List of Figures
Figure 5-1.
Figure 5-2.
Figure 5-3.
Figure S-4.
Figure 5-5.
Pictorial of packed and plate towers 35
Half life in a lake as a function of Henry's
Law constant 37
Affect of air quantity on residual concentration
when stripping o-nitrotoluene 41
Water temperature and water vaporized in batch
stripping
42
Effect of air quantity on residual concentration
when stripping nitrobenzene 43
Nomenclature for a packed tower 47
Simple stripping tower showing equilibrium stages 52
Relationship between stripping vapor rate and
number of plates 55
Water temperature when stripping water water with
air ".
Water temperature when stripping cold water with
air
67
68
Figure 5-6.
Figure 5-7.
Figure 5-8.
Figure 5-9.
Figure 5-10.
Figure 5-11.
Figure -5-12.
Figure 5-13.
Figure 6—1.
Figure 6-2.
Figure 7-1.
Figure 7-2.
101
Figure A2-1. A water stripping column or cooling tower 110
Stripping in a cooling tower with circulated water 71
The effect of recycle when the organic is stripped 73
The effect of recycle when the organic is con-
centrated 76
Two film theory of mass transfer 81
Bubble-cap tray 86
Dependency of NTD on Stripping Factor and Removal Efficiency 98
The Effect of Liquid Diffusivity on the Height of a Packed
Bed
11
-------
List of Tables
Table 4-1. Henry' s Law Constants for Gaseous Contaminants 32
Table 5-1. Isothermal Batch Stripping 40
Table 5-2. Toxic Pollutants Classified by East of Stripping 58
Table 7-1. Summary of Available Stripping Data at Low Concentration ... 90
Table 7-2. Organic Removal Efficiency Calculations for Plant
No. 1290-010 of the EPA Organic Data Base 93
Table 7-3. Organic Removal Efficiency Calculations for Plant
No. 2930-035 of the EPA Organic Data Base 95
Table 7-4. Organic Removal Efficiency Results of the Water Factory Data 96
Table A2-1. Nomenclature for Water Stripping Column Ill
12
-------
1. INTRODUCTION AND OUTLINE
The transference of toxic pollutants from solution in water to the vapor
phase or to the atmosphere is of interest for several reasons.
(1) It will occur naturally. Many pollutants have a half life in
solution in a slowly flowing river, pond or lake which is controlled by the
rate at which they vaporize into the atmosphere. Bubbling air through the
water or a waterfall will greatly enhance the rate of vaporization.
(2) Stripping can be used as a deliberate method of cleaning the
water. If the contaminant is considered to be harmless in the atmosphere,
air stripping can be used. Air stripping is particularly useful for removing
small amounts of chlorinated hydrocarbons from drinking water. Air stripping
is used at Lake Tahoe to remove dilute ammonia from treated sewage effluent.
When recovery of the contaminant is required, such as when stripping high
concentrations of ammonia or hydrogen sulfide from industrial wastewaters,
steam stripping is usually used.
The ease with which a particular compound is stripped or naturally
volatilizes depends on the volatility, which in turn depends on two
properties of the contaminant - its vapor pressure and its solubility in
water. That compounds with high vapor pressures are easily stripped is
expected. The effect of solubility can also be explained quite simply.
A compound which is not much soluble in water is a compound whose molecules
are not compatible with water molecules. In dilute solution each molecule
of organic is surrounded by water molecules which want to push the organic
molecule away. The apparent vapor pressure of a poorly soluble organic can
be thousands of times higher over an aqueous solution than over the pure
organic.
The measure of the apparent vapor pressure over solution is called
the Henry's Law constant. Explanations and techniques for determining the
Henry's Law constant are given in Section 4. A compilation of Henry's Law
constants for all lisred organic toxic pollutants is given in Appendix 3.
Appendix 3 is the single most important contribution of this report.
In Section 5 the theory of batch and continuous stripping is given and
the ease with pollutants can be stripped is determined. Pollutants will be
found ranked by ease of stripping on Table 5-2. Of 185 listed toxic
pollutants, 68 are very easily strippable; that is, not much vapor is needed
13
-------
and the period of contact or height of contacting tower between vapor and
liquid is snail. For these 68 pollutants, stripping is a useful treatment.
Forty-eight compounds cannot be stripped and the rest can either be stripped
with varying degrees of difficulty or information is lacking to make a
judgment.
When air is used as the stripping vapor, water will also evaporate.
This cools the water and decreases the volatility of the organic relative
to water. Approachs for allowing for this effect are also given in
Section 5.
This report is not a design manual. A brief introduction to the choice
of equipment is given in Section 6 and the reader is referred to manufacturers'
manuals for details of design.
Comparison of predictive calculations to stripping data at low
concentration is discussed in Section 7. Also presented in this section
is the development of a correlation for the effect of liquid diffusivity
on tower sizing and organic removal efficiency.
14
-------
2. CONCLUSIONS
The Henry's Law constant is known or can be reasonably estimated
for all except 33 of the 185 toxic organic pollutants listed. About 68
pollutants can be very easily stripped; they have a half life in a lake of
the order of 12 hours and can be reduced to l/100th or 1/1,000th of their
original concentratin by stripping with air in a single stage device such
as a small cooling tower.
An additional 36 pollutants can be stripped with more difficulty.
Design of air stripping equipment for these compounds is difficult because
the water temperature falls on dry, winter days when water is evaporated
by the stripper. The addition of enough steam to saturate the air is a
possible method of control. Stripping with pure steam may also be used.
Although steam is much more expensive than air, the increased temperature
so increases volatility that less steam than air is required for the same
job. Steam stripping is cheapest when the water stream is already hot.
About 48 compounds cannot be stripped or probably cannot.
15
-------
3. RECOMMENDATIONS
The following compounds, for which volatility data is inadequate, are
probably sufficiently volatile that measurement of Henry's Law constant is
of interest.
5-6 4-Bromophenyl phenyl ether
6-4 Di-n-octyl phthalate
6-6 Butyl benyl phthalate
7-3 N-m'trosodi-n-propylamine
7-4 Benzidine
7-9 Diethyl amine
8-12 p-Chloro-m-cresol
10-2 Benz(a)anthracene
10-6 Indeno(1,2,3-cd)pyrene
10-7 Dibenzo(ah)anthracene
13-10 4,4'-DDE
13-27 Carbaryl
14-17 Furfural
15-6 2,3,7,8-Tetrachlorodibenzo-p-dioxin
The following compounds, for which volatility data is inadequate, are
probably not sufficiently volatile that measurement of Henry's Law constant
is of interest.
5-5 4-Chlorophenyl phenyl ether
6-3 Di-n-butyl phthalate
6-5 Bis(2-ethylhexyl) phthalate
7-1 N-nitrosodimethyl amine
7-2 N-nitrosodiphenylamine
7-5 3,3'-Diphenylhydrazine
7-6 1,2-Diphenylhydrazine
13-11 4.4--DDT
13-17 Kepone
13-40 Di qua t
13-42 Mevinphos
14-9 rumaric acid
14-10 Maleic acid
16
-------
For the following compounds there is insufficient data to judge whether
or not they are likely to be volatile.
7-8 Butyl amine
7-11 Monoethylamine
7-12 Monomethylanrine
7-14 __ Trimethylanrine
«
12-28 2,2-Dichloropropionic acid
13-1 o-Endosulfan
13-3 6-Endosulfan
13-4 a-BHC
13-5 B-BHC
13-6 6-BHC
13-7 Y-BHC
13-14 Kelthane
13-15 Naled
13-16 Dichlone
13-19 Endrin aldehyde
17
-------
4. VAPOR LIQUID EQUILIBRIA
4.1 Introduction to the Theory of Vapor-Liquid Equilibria of Organic
Molecules in Dilute Aqueous Solution.
In order to design a stripper one must have information on the volatility
of the material to be stripped. In this section we discuss whether the
volatility (more correctly the vapor-liquid equilibrium compositions) can be
estimated from properties of the pure components. In fact, estimates cannot
be made for most organic molecules in water and some experimentally determined
information is required. Very occasionally one has direct measurements at the
pressure, temperature and concentration of interest. Sometimes there is
distillation data at concentrations much higher than those of interest and
one needs to extrapolate to low concentrations. Usually the only data
available is the vapor pressure of the pure organic and its solubility in
water. This data can also be used to approximate the vapor-liquid equilibrium
relationship. .
The theory and estimation of vapor-liquid equilibria is given in many
texts; for example, by Lewis and Randall , Gilliland and Reid, Pransnitz
and Sherwood . The discussion given here is limited to a brief introduction
to nomenclature and estimation of the equilibria for dilute solutions of
organics in water at close to atmospheric pressure and in the temperature
range of liquid water.
The property which describes the "escaping tendency" of a compound
was called "fngacity" by G. N. Lewis1. At equilibrium, and by definition
of fugacity, the fngacity of every component, i, in the liquid equals the
fugacity in the vapor:
4 ' 'I
Fugacity is logically defined so that at the given temperature and at such
a low total pressure that the vapor is an ideal gas, the fugacity equals
the vapor pressure. Thus, fugacity has the units of pressure and may be
regarded as an "ideal" or "corrected" vapor pressure.
Gas Phase
The "mole fraction" of a component, i, in the gas phase, written y^
equals the vapor pressure divided by the total pressure and is, therefore,
a convenient expression of concentration. Since the pressure of interest
18
-------
to us is low, the fugacity equals the vapor pressure and we have
V
f . = Vapor Pressure = y.P
where ? is the total pressure.
This equation is called Dalton's law and it can be used for all cases
of interest to us here.
Liquid Phase
The fugacity of a compound in the liquid phase must be related to the
vapor pressure of the pure liquid, written p*. Also, the fugacity is
generally a function of pressure, temperature and concentration. However,
the pressure of interest to us is low and it is customary, for low pressures,
to define an "activity coefficient", Y., and to write
In writing this equation several decisions have been made. The
correction necessary for high pressure has been omitted. This correction
is called the "Poynting effect" in some texts. The activity coefficient,
which is a function of composition and temperature, has been defined. The
expression of concentration, x., has been chosen to be the mole fraction.
This is so that the ratio y./x. (concentration in the vapor divided by
concentration in the liquid) shall be dimensionless. In dilute solutions
x = moles/1 x 0.018
and x = mg/1 x 18 x 10~ /MW
where MW •= molecular weight of the organic.
A solution is called "ideal* if Y » 1. In this case
f* - x. p*
i i *i
which equation is called Raoult's law. In solutions of interest to us
Gilliland ). We cannot use Raoult's equation and 7. must be determined
Raoult's law holds for water but not for the organic (see, for example,
Gilliland2). W«
experimentally.
19
-------
The Form of the Vapor-Liquid Equilibrium Relationship.
The equations introduced so far give
y. = VL x. p*/P
In this equation P is total pressure and p? is the vapor pressure of
the pure organic (a function of temperature) . The activity coefficient,
which is dimensionless, is a function of concentration and temperature.
An alternative nomenclature is to write
This equation is called Henry's Law and E is called Henry's Law constant.
H is a function of temperature and concentration except that H is independent
of concentration at low concentrations. H has the units of pressure and the
two usual units for reporting H are mm Hg and atmospheres. In reading the
literature one will find reports in which the concentration unit (for which
we use the mole fraction, x, which is dimensionless) is not dinensionless. In
this case H has the units of pressure divided by concentration. Furthermore,
some authors report the reciprocal of our H. The vapor-liquid equilibrium
relationship is
y± = . x. H./P
and
Yet another nomenclature is to write
y. « x. x.
•*
SO
K. is dimensionless.
i
To use any of the forms of the vapor-liquid relationship we must know 7
and p*, or H, or K.
20
-------
4.2 Activity Coefficients for Slightly Soluble Compounds.
Most of the organic compounds of interest to us are only slightly
soluble in water and this is much the most important method of determining
activity coefficients. This is the principal method used by all authors
4
that we have found (see, for examples, Hwang and Fahrenthold , Kavanaugh
and Trussel5, Dilling et al6).
When a sparingly soluble compound is xiissolved in water to saturation,
(x. »x. (saturated)) there will be two liquid phases in equilibrium (or one
liquid and one solid phase) - the water phase and the organic phase. If,
in addition, no vapors other than water and the organic are present
(particularly if air is not present), there will be three phases in equili-
brium. With two components present, there is one degree of freedom and if
the temperature is chosen, the total pressure is specified. When complete
information on two components with three phases in equilibrium is available,
Henry's law constant can be determined accurately. It can be shown that if
Henry's law holds for the organic in the water phase, then Raoult's law
must hold for the water in the water phase (see, for example, Smith and Van
Ness, Ref. 44, page 347). In this case, for equilibrium between the vapor
and water phases.
y P = Y x p* =• x H
•* o o o o o o
y P = x p*
w \rw
where the suffix w means "water" and the suffix o means "organic". Since
y +y = x + x = 1, adding these two equations gives
P "To XoPo * (1 'V Pw
* xo HO * (1 - v *:
or, Ho - P - (1 - *0) P*
xo {1)
xo
Complete equilibrium data is very seldom available. Solubility is
usually measured at pressures below the equilibrium pressure and, in
addition, air is present in the vapor phase. (Expressed another way, we
21
-------
may say that solubilities are usually measured at one atmosphere and at
temperatures below the equilibrium temperature, with air present in the
vapor phase). The approximation usually made is
In this case
H « p*/x (satd) (2)
o o o
Y « 1/x (satd)
o o
Equation (2) is the approximation used throughout this report and by all
the authors quoted who have made similar estimates. In using Equation (2)
we assume that the fugacity of the organic phase is the vapor pressure of
the pure organic at the prevailing temperature. We thus neglect the
solubility of water in the organic phase and assume that the fugacity of
the liquid organic is independent of pressure.
Often the solubility of water in the organic phase in equilibrium with
the water phase is given. In this case another approximation,* which we have
never seen used, is possible. Assume that Raoult's law holds for the
organic in the organic phase. Then
X H » x p*
o,w o 0,0 *o
where x is the mole fraction of organic in the water phase and XO,Q is
the moli'rraction of organic in the organic phase.
H = x p*/x (3)
o 0,0 *o o,w
Smith and Van Ness (Ref. 44 page 361) give complete data for ethyl-ether
and water. Calculations are presented on the table below. Compared to
Equation (1), which is exact, the approximate Equations (2) yields a maximum
error of 4.4%. Equation (3), which requires more data than Equation (2),
yields the same error and does not seem to be preferable.
22
-------
Ethvl-ether and Water System
P
(atm)
Satd.
Hater
Phase
x
34 1.000 0.0123
50 1.744 0.0103
70 3.195 0.0075
90 5.514 0.0058
Satd.
Ether
Phase
x
(atm)
H (atm)
o
(atm) Eqn. (1) Eon. (2) Eon. (3)
0.9456
0.9348
0.9212
0.9107
0.053
0.121
0.306
0.691
0.983
1.679
3.018
5.040
77.0
158
386
832
79.9
163
402
869
75.6
152
370
791
These equations require a knowledge of the vapor pressure of the pure
organic and its solubility as a function of temperature. They then give
us H or y at one concentration. The concentration at which we know E and y
is the highest concentration obtainable, and we must assume that R and y
remain constant with respect to concentration at concentrations below •
saturation. Now it is experimentally true that if the concentration is low
enough H and y are independent of concentration. For example, for such
sparingly soluble gases as oxygen and nitrogen, Henry's law applies with
H constant. In a later section we extrapolate phenol-water equilibrium
data to show that Y is constant at low enough concentrations. In using
the above technique to obtain H or y we have no way of knowing if the
organic is sufficiently insoluble for accuracy; we just assume it is and
this technique is less reliable the more soluble the organic.
An example follows.
trichloroethane
At 25°C Hwang and Fahrenthold give, for 1, 1, 1 -
solubility (mg/1)
4,400
23
-------
Dilling et al6 give (at 25CC),
solubility (mq/1) vapor pressure (mm Hg)
1,300 123
The Chem. Eng. Handbook gives
•insoluble" 125
The two different solubilities are
X » 4,400 x 18 x 10"€/133.4 = 5.9 X 10~4
or
X " 1.75 x 10~4
The Henry's law constant is
H = 123/5.9 x 10~4 « 2.08 x 105 mm Hg = 274 atm
or
H «= 925 atm.
' Kavanaugh and Trussel , who did not give the base data, calculated H = 400 ana.
The numbers used illustrate not only the calculation but also the
discrepancies that frequently are found.
The use of this method will require a correlation of vapor pressure and
solubility (or activity coefficient) as a function of temperature.
4.3 Vapor Pressure as a Function of Temperature .
Reid, Prausnitz and Sherwood (Chapter 6) have considered the many
correlating equations available. We are only interested in temperatures
between about 0° and 100°C (273° to 373°K). We. are not interested in
compounds with very low vapor pressures because they will not be stripped.
For our purposes the preferred correlating equation is
Log p* » A + B/(C + T)
24
-------
which is called Antoine's equation. An alternative equation, less satisfactory
at low temperatures, is the Clapeyron equation
Log p* - A + B/T
where T is in °K.
4.4 Activity Coefficient as a Function of Temperature
Reid, Prausnitz and Sherwood3 (Page 307) state that the effect of
temperature on the activity coefficient is a particularly troublesome
question. They suggest using
Log Y (constant composition) = c + d/T
where T is in °K.
Since we are working at concentrations so low that Y is independent of
concentration, the limitation of constant concentration can be ignored.
If only one point is available, the constant, c, should be taken as
zero so
Log Y - d/T
This is the form of the equation that we most often use. It is the way
that Hwang and Farenthold extrapolate from 25°C to 100°C. Fortunately
the effect of temperature on y is very much less than the effect of
temperature on p* so the fact that we cannot correlate 7 very well is not
of extreme importance. Tsonopoulos and Prausnitz , working with aromatic
molecules in the range 0-50°C found it best to assume V independent of
temperature.
We return to our example on 1, 1, 1 - trichloroethane. As we have seen,
A
Hwang and Fahrenthold give, at 25°C, a solubility of 4,400 mg/1 or a mole
fraction of 5.9 x ID*4. Thus at 25°C (298°K) Y = 1/5.9 x 10~4 » 1,700. At
100°C (373"X) Y is given by
Log 1.700 273 ,_
Logv " 298 1*"
so Y (373eK) = 384.
25
-------
At 100°C Hwang and Fahrenthold give
p* - 213 kPa = 2.1 atm,
so, at 1 atm,
K = y/x = 7P*/P = 384 x 2.1 « 806 atm
The value given by Hwang and Fahrenthold is 796 and they nay have had
other solubility data. Note that from 25°C to 100°C we assumed that 7 ,
changed by 4.4 «•!«•* and p* changed by 13 times. The change in p* is much
the most important effect of temperature.
4.5 Extrapolation of Activity Coefficients Measured at High Concentration
Activity Coefficient as a Function of Concentration
The vapor-liquid equilibrium of phenol in water has been measured by
several authors. Gilliland2 quotes a 1933 thesis and the lowest concentration
at which he gives information is x = 0.001, y = 0.002 at 1 atmosphere. If
we assume that at this concentration 7 is independent of concentration we
would have
K » y/x « 2
H = Py/x - 2 atm
7 - *y/P* * = 37
where p* (100°C) = 0.054 atm (see below).
Note that 7 is very much greater than one and Raoult's law does not
apply. We will now suggest ways of extrapolating equilibrium data that
are preferrable to linear extrapolation from the lowest point and show
that the value of 7 obtained above is accurate.
Many relationships have been proposed between 7 and concentration for
binary mixtures. All the relationships must satisfy the Gibbs-Duhem equation
and summaries have been given by Gilliland , Reid, Prausnitz and Sherwood
(page 300, Table 8-3), Holmes and Van Winkle4 and in other texts.
It seems that van Laar's equations, which are among the simplest to
use, are adequate for our purposes. We have not made a detailed comparison
with other equations.
26
-------
The van Laar equations are
Log yx «
Log y2 - B/[l +
where 1 and 2 are the two components of a binary mixture. If we have any
one experimental point, we know Y , Y , x. and x2 (» 1 - x^) and the two
equations can be solved for A and B. The equations can then be used to
explore the variation of Y with x.
The following manipulations give expressions specific in A and B which
are useful when a single point is available.
Ax, I/Ax.
* "2
+ l/Bx2)
1/Bx,
*2 'L°q 2 " (I/Ax, + 1/Bx,)2
1 £
x. Log 5'^/3E2 Log V - Bx_/Ax.
From the first Van Laar equation
A = Logy, [1 + Ax,/Bx,]2
1 12 2
» Log yx [1 + x2 Log Y^^ Log Xj
and similarly
B - Log 72 II + x1 Log T^j Log Tj]2
As a first example consider again phenol (component 1 or A) and water
(component 2 or B). A single point that is often known is the azeotrope.
At 1 atmosphere the azeotrope is
x. = 0.0195, x, » 0.9805
27
-------
The boiling point is 99.8°C so
p* • 0.054 atm (see correlation below)
p* » 0.993 atm.
We calculate
TX • 1/0.054 • 18.5, Y2 = 1/0.993 - 1.007
A, phenol =1.59
B, water «• 0.262
(It is worth remembering the p* and p* are the vapor pressures of the pure
components and do not add to one atmosphere). If A and B as calculated
above are used to calculate Y when x. • 0.001, we obtain
Log Y • 1.S9/T1 + (0.001)(1.59)/(0.999)(0.262)]2
- 1.57
Y1 • 37.2
which is an unusually good agreement with value found by linear extrapolation
from the point x = 0.001, y » 0.002.
The van Laar equations are not useful when the single data point is at
a very low concentration. At a low concentration of organics Y_ (for water)
approaches one and Log Y is very small. But x (the organic) is also small
and x./Log y. tends to become indeterminate. An error of one tenth of a
degree centigrade in the boiling point can drastically alter the calculation.
When data is available at low concentrations, linear extrapolation is the
preferred extrapolation.
Any extrapolation will be better if many points are available. To
extrapolate, the equations may be linearized so that the data can be
conveniently plotted to obtain A and B. One way to plot the data is to
plot x./x. against I/(In Yj) . From the first van Laar equation
so
x /x = _,
VX2 —
28
-------
Xl X2
Another way is to plot ~—TT T~T—7T against x, . The following
1 YI T X2 2 x
manipulations explain this plot.
I/Ax, 1/Bx
x In Y + x in Y -
1122
Xl X2
X
2 Xl
In plotting to obtain van Laar coefficients it must be remembered that
A and B are functions of temperature and isothermal plots are required.
When, as frequently happens, the available data is at constant pressure
and variable temperature, van Laar's equations may be written
T Log ^ - A/(l + Ax1/Bac2)2
T Log Y2 «
and T log Y., T Log Y_ used instead of Log X and Log Y in all the equations.
T is in °K.
Hicks et al have fitted van Laar's equations to data given in
references 2 and 8 to 11 and found
A, phenol « 1.941 - 0.00352 t*C
B, water • 0.324 - 0.00052 t°C
55 < t°C < 240
The effect of temperature on A and B was found by simply putting the best
straight line through the available values. From these equations we have
calculated Y,» for phenol, as shown below (we have ignored the temperature
limitation on the correlation).
29
-------
Temp Xl y
°C _A_ B (x;= 1 - XT)
30 1.835 0.308 0.01 42^91
0.001 65.06
0.0001 68.05
10~5^ 68.36
10"6 68.39
100 1.589 0.272 10~2 26.11
10~3 37.20
10~4 38.65
10"5 38.80
10*6 38.81
At 100-C and x^ « 10"3 (5,200 ppm) YI (-37) is as calculated from the
low concentration data point and YI is independent of concentration within
about 3%. YI is independent of concentration within 1% when s^ < 10~4
(= 520 ppm) and is independent of concentration within 0.05% when x < -5
1 10
(52 ppm).
The vapor pressure of phenol has been correlated by the equation
Log p* (ion Hg) - A + B/(C + t°C)
where
A * £ Reference
!• 7.50 -1724 192 Hicks et al7
2. 7.14 -1518 175 Dean13
3. 7.13 -1516 174 Reid et al3
4. 6.93 -1383 159 Gteehling12
These equations give very similar results and numbers 2 and 3 seem
to be a good average set of constants.
At low concentrations, for phenol
Log E(mm Hg) *• Logy + Log p* = A + Log p* =
1.941 - 0.00 352 t + 7.13 - 1517/7174 + t) =
9.071 - 0.00352 t°C - 1517/(174 + t°C)
30
-------
4.6 '• Direct Measurement of Henry's Law Constant
The solubility of gases is usually expressed in the form of a Henry's
Law constant. It is quite easy to vary the partial pressure of a gas and
measure its solubility which gives H directly. Most toxic pollutants of
interest are not gases and solubility measurements of H are not usually
available.
However, Mackay, Shiu and Sutherland have devised a most useful
stripping technique to measure H which has been used extensively by Warner,
Cohen and Ireland20. Nitrogen is bubbled through a column of water is
which the organic was dissolved. The water column is held at constant
temperature and a record is kept of the concentration of the organic in
the water as a function of time. It is experimentally demonstratable
that the organic is in equilibrium between the water and the nitrogen
so that the rate of removal of organic by stripping,
- V dc/dt = (pp) G/RT
where
3
V is the volume of the water, m
c is the concentration of the organic, moles/m
t is time, hours
G is the nitrogen rate, m /hr
R is the gas constant, (m ) (atm)/(mole) (°K)
T is the temperature, DK
and
(pp) (the partial pressure of organic over the solution)
» yp s HX = HC X 18 X 10~6
if c « c at t » to
o
so,
if c
_
In (c/co) • (18 x 10 HG/VRT) t
•6
and a plot of In c against t has a slope of 18 x 10 HG/VRT
from which H can be determined.
We should mention here that we are not concerned with stripping
contaminants that react with water such as NH3 and CO2. Since it is HH3»
not HH*, which is volatile, and CO2/ not HCO^, which is volatile, stripping
calculations involve simultaneous calculations of chemical equilibrium in
solution. The basis for these calculations has been given by Hicks et al .
31
-------
Henry's Law constants have been given for the single contaminants
for NB , H S, and HCN, which are of interest. Some calculations from the
formulae given in the references are presented on Table 4-1. For single
18
contaminants we suggest using the newest reference which is Edwards et al .
The formulae are
In E (ammonia) - -149.006 - 157.552/T +
28.1001 In T - 0.049227 T
In H (H2S) » 342.595 - 13,236.8/T -
55.0551 In T -I- 0.0595651 T
In E (HCN) = 1446.005 - 49,068.8/T -
241.82 In T+ 0.315014 T
where T is in °K
-6-
and E is in (atm) (kg)/mole = E atm x IB x 10
19
Formulae given by the API show that H (ammonia) is altered by 100%
if the acid gases O>2 + E2S total about 160 mg/1 and that E (E2S) is altered
by 100% if the alkaline gas NE3 is present at about 80 mg/1. In most cases
the simple formulae for E given above are not useful and chemical equilibrium
must be taken into account.
TABLE 4-1. BENRX'S LAW CONSTANTS FOR GASEOOS CONTAMINANTS
•
E in (atm - kg/mole)
t°c
T*K
NH3
Ref.
Ref.
Ref.
H2S
Ref.
Ref.
Ref.
HCN
Ref.
Ref.
17
18
19
17
18
19
17
18
15
288
0.0101
0.0100
0.0104
7.24
7.49
8.79
0.0571
0.0466
50
323
0.0472
0.0478
0.0478
16.55
15.87
17.95
0.258
0.269
100
373
0.246
0.248
0.248
28.9
27.45
30.13
2.93
0.997
32
-------
5. THE THEORY OF STRIPPING
5.1 Organization of the Section, Batch and Continuous Stripping,
Isothermal and Adiabatie.
If contaminated water is placed in a vessel, a pond or a lake and air
or other vapor is passed through it, the contaminants may be slowly
vaporized and, as time passes, the water is cleaned. This is a realistic
situation; con-tarnination nay have been due to a spill or other one-tine
discharge. Vapor bubbling through a vessel or pond is described by the
equation for a batch stripper given in Section 5-3.
A volatile contaminant trill vaporize from a lake even if air is not
bubbled through the water. In this case the rate of removal depends on
the rate of diffusion to the water surface, as well as on the volatility
(Henry's Law constant} of the contaminant. A description of this situation
has been given by Mackay and Leinonen and is discussed in the next
Section (5.2).
Now consider a continuous stripper; that is, the water as well as the
vapor is flowing. The passage of vapor through the vessel will normally
mix the water. The cleanest water we can hope to get from a simple vessel is
when vapor and water reach equilibrium and then the water cleanliness will
depend on the Henry's Law constant of the organic contaminant and the vapor
to water ratio. The only way to obtain cleaner water is to pass more vapor.
Usually increasing the vapor rate is not economical and the practice is to
pass water down a tower counter-current to the vapor flowing up. This means
that the cleanest vapor is put in contact with the cleanest water and that
as the vapor rises it comes into contact with dirtier and dirtier water until
vapor leaves the top of the tower with a high load of contaminant.
The tower is usually full of some sort of packing designed to encourage
contact between liquid and vapor. An empty tower, called a spray tower,
can also be used, but the results will be much the sane as a single sinple
vessel with vapor bubbled through. The choice of tower packing is discussed
in Section 6.
The designer of a stripping tower has the job of determining the height
of the tower and the vapor/liquid ratio required to reach the wanted degree
of stripping. (The diameter of the tower is controlled by the throughput
rate). There are two approaches to determining tower height. The first
approach, which is mostly used when the tower packing is continuous from
33
-------
bottom to top, is to write an equation for rate of transfer of contaminant
from the water to the vapor phase. The rate equation is integrated to
determine the tower height. This procedure needs the rate coefficient of
mass transfer which must be determined experimentally.
When the stripping tower contains discrete trays or plates rather than
a continous packing, it is usually more convenient to think in terms of
equilibrium stages "jsee Figure 5-1.). The vapor and liquid are assumed to
thoroughly mix and reach equilibrium on a plate. The liquid then falls to
the next equilibrium plate below and the vapor rises to the equilibrium plate
above. The height of the tower is expressed as the number of equilibrium
plates .to do the job.
Equilibrium is not reached on real plates and an experimentally
determined efficiency is needed to tell how many real plates are required
when the number of equilibrium (or theoretical) plates has been calculated.
We find it simpler to work in terms of equilibrium plates. (The words
"equilibrium" and "theoretical" are interchangeable and the words "plates"
and "stages" are interchangeable.) In Section 5.5 we briefly describe the
rate of mass transfer approach and show how the height of packing equivalent
to a theoretical plate (HET?) can be found by taking measurements on an
operating tower. In Section 5.6 we give the complete equations for the
equilibrium stages approach and use the results to classify organic pollutants
by the ease with which they can be stripped.
The simplest equations to write apply to isothermal stripping. In
isothermal stripping, when steam is the vapor, the water must enter the
tower preheated to the boiling point; when air is in the vapor, the water
must enter at the wet bulb temperature of the air. Air strippers are not
usually isothermal; they are adiabatic. Water is usually evaporated and
the liquid water is usually cooled.
Adiabatic stripping with air is considered in Sections 5.4 and 5.8.
Particularly in the case of a counter-current air stripping it can be difficult
to design equipment which will behave satisfactorily as the air conditions
change with the seasons. We suggest that the best control of such equipment
is to add a little steam to the air when the air is particularly dry. Steam-
air mixing is discussed in Section 5.9.
Finally, in Section 5.10, a brief analysis is given of a system in
which the water is circulated. This is particularly applicable to cooling
towers.
34
-------
DIRTY
VAPOR
DIRTY
WATER
U)
in
DIRTY
VAPOR
DIRTY
WATER
PACKING
SUPPORT
GRID
CLEAN
VAPOR
CLEANED
WATER
WATER
CLEAN
VAPOR
PACKED TOWER
PLATE TOWER
Figure 5-1. Pictorial of packed and plate towers.
-------
5.2 Evaporation from a Lake
Mackay and Leinonen have gi
volatile contaminant in a lake. Their formula is
Mackay and Leinonen have given a formula for the half life of a
where tQ _ - half life in hours
L » depth in meters at which the half life is measured
K_ e overall mass transfer liquid coefficient, m/hr.
also
1
*L kL (18 x ID*"6) H
where
k_ • liquid side mass transfer coefficient, = (approx) 0.2 m/hr
k « gas side mas transfer coefficient, = (approx) 30 m/hr
B = Henry's Law constant in atm
18 x ID*"6 - m3/TOl of water
R » gas constant » 8.2 x 10** (m ) (atm)/(mol) (°K)
T a temperature in °K.
At 70°F « 21DC « 294°K, at a depth of 1 meter,
1 1 + 1
KL °*2 (2.24 x 10"2) H
and tQ 5 = 3.45 + 30.8/H.
This equation is graphed on Figure 5-2. It seems reasonable to classify
organic molecules as
H(21°C) > 10, easily stripped
H(21°C) - 2 to 10, intermediate
E(21°C) < 2, slow to be stripped.
This classification is compatible with another classification given in
Section 5.7.
5.3 Isothermal Batch Stripping
Consider a solution of M moles of contaminant in W moles of water at
the boiling point. Saturated steam is passed through in such a way that
the vapor leaves in equilibrium with the solution. Water does not leave
the solution, but when an amount of steam, dv, is passed an amount of
contaminant, -dM , leaves the solution and appears in the vapor. The vapor
leaving is dV - dM moles (dM is a negative quantity) .
36
-------
1000
100
Figure 5-2. Half life in a lake as a function of Henry's Law constant.
37
-------
(The derivation applies to air as well as to steam so long as the stripper
is isothermal or approximately isothermal. The closer the air to being
saturated at the water temperature, the closer the system is to isothermal.)
-dM
1 dV - dM
is in equilibrium with the solution composition, i.e.
y = xH/P « Kx
where
K « H/P
Since x - Mc
W +• M
-dM KM
dV - dM W + M,.
c c
and KdV = \ K-l - -2- \ dM
<, Mc J C
If this equation is integrated between V = o and V «= V, and MC
(initial) and MC = Mef (final) one obtains
Mcf
K7 o (K-l)(Mrf - Mci) - W In —
This can be rewritten as
K ~
The preferred nomenclature is to use
Mci
X.
1 Mci +
38
-------
Mci XL
30 -5— « , „ - (approx) x. for low values of x.
™ i l
(and we are interested in x. less than 10~ in all cases).
AlSO -Si «s 1 _ p
"ci *
where FR is the fraction of contaminant removed.
The final equation is, for low initial concentrations,
S - - (K-i) x± PR - ln(l-FR>
In all the cases of interest to us x. is so low that the first term
on the righthand side can be neglected and
^ - (approx) - lna-FR)
Seme calculations using this formula have been made for benzene,
nitrotoluene and nitrobenzene and are presented on Table 5-1 for a total
pressure of 1 ata so that K = H. Graphs are presented later on Figures 5-3
and 5-5.
5.4 Adiabatie Batch Stripping with Air
»
The general situation is not isothermal and water evaporates into the fir
stream. If the tank is adiabatic (heat not supplied), which it will usually be,
the water will cool. It is necessary to calculate the rate at which the
temperature falls as the air is passed through the water. It is then possible -to
calculate the rate at which the organic is stripped, taking account of the
decrease in Henry's Law constant as the temperature falls.
A procedure for making the calculations is given in Appendix 1.
Calculations were made for benzene (R(20°C) «= 278 atm), o-nitrotoluene
(H(20°C) « 6 atm) and nitrobenzene (H{20eC) «= 0.9 atm). Benzene stripped
so fast that the calculations were of no interest. The results for
o-nitrotoluene are shown on Figures 5-3 and 5-4. The results for nitro-
benzene are shown on Figure 5-5. In each case the starting solution was
saturated at 20°C.
39
-------
TABLE 5-1. ISOTHERMAL BATCH STRIPPING
Vapor Passed Fraction Remaining
Compound °C H(atm) (V/W) Cl-Fn)
Benzene -- 10 213 0.011 0.1
0.022 0.01
0.032 0.001
20 278 -0.0083 0.1
0.017 0.01
0.025 0.001
o-Nitrotoluene 10 4.23 0.54 0.1
1.1 0.01
1.6 0.001
20 6.0 0.38 0.1
0.77 0.01
1.15 0.001
Nitrobenzene 10 0.53 1.3 0.5
3.0 0.2
20 0.91 0.76 0.5
1.8 0.2
2.5 ' 0..1
40
-------
Water enters at 20°C.
I
o
•H
-U
(0
U
JJ
i
o
U
10
• 1
0.1
Air
10°C
30% R.H.
Air
20°C
70% R.H.
Isothermal
at 108C
\
\
\
Isothermal
at 20°C
j ii iii
1.0
V/W, moles air per mole water
2.0
Figure 5-3. Affect of air quantity on residual concentration when stripping
o-nitrotoluene.
-------
20
Air 20°C, 70% R.H.
Wet bulb temp.
Air 10°C, 30% R.H
Wet bulb temp
20CC, 70% R.H
10°C, 30% R.H
0.98
2.0
V/W (moles air/moles water)
Figure 5-4. Water temperature and water vaporized in batch stripping.
-------
10'
I
.2 10-
*>
IB
M
I
10'
Water enters at 20"C.
- Isothermal, 10°C
Air
10°C, 30%R.H.
Isothermal, 20°C
Air
20°C, 70% R.H.
1.0
V/W, moles air per mole water
2.0
Figure 5-5. Effect of air quantity on residual concentration when stripping
nitrobenzene.
43
-------
o-Nitrotoluene (H(20°C) «= 6 atm, H(10°C) = 4.2 a tan) is moderately easy
to strip. The concentration is reduced to 1/100 Oth of the initial concen-
tration by 1.1 moles air per mole water when the air is warm and quite
saturated (20°C and 70% relative humidity) and by 1.7 moles air per mole
water when the air is colder and dryer (10°C and 30% RH) . For comparison
benzene (H(10°C) = 213 atm) is very easy to strip and the concentration is
reduced to 1/1, 000th of the initial concentration by about 0.^,3 moles »•**•
per mole water. The warm air case for o-nitrotoluene is satisfactorily
duplicated by the isothermal calculation at 20°C. However, the cold
case is not really well duplicated by an isothermal calculation at 20°C
(the water temperature) nor at 10°C (the air dry bulb temperature) . The
answer lies in between at about 15°C.
The reason for the 15°C average can be seen from the graph of water
temperature on Figure 5-4. Water starts at 20°C and, when the concentration
is reduced to 1/1, 000th after passing 1.7 moles air per mole water, the
water temperature has dropped to 9°C. A temperature of 15°C is a good
average.
Nitrobenzene (H(20°C) =0.9 atm and H(10°C) =0.5 atm) is -very
difficult to strip. The • concentration is not even reduced to 1/lOth
by 2 moles air per mole water (see Figure 5-5) . The warm air case is
reasonably calculated by a 20°C isothermal calculation and the cold air
case is reasonably calculated by a 10°C isothermal calculation. The
temperature curve which is independent of the organic pollutant (discussed
later) , shows why these 'average temperatures are reasonably satisfactory
(see Figure 5-4) .
The key to the calculation is to determine the temperature curve. With
the temperature known the stripping can be calculated using the isothermal
equation with E evaluated at a reasonable average temperature or H can be
allowed to vary stepwise as the temperature falls.
A general calculational procedure is given in Appendix 1. With a dilute
solution stripping the organic has a negligible effect on the temperature of
the solution which is overwhelmingly dependent on the amount of water which
vaporizes. It is, therefore, possible to illustrate what the computer does
by a simple hand calculation.
The one essential formula is a means of calculating the humidity of
satured air, H moles water per mole dry air.
44
-------
y
HCsatd.) = -
where
-— exp 21.158 - ^f°:!L - 6.977 x 10~3 (t°C + 273)
A / I fc^^T^/J I
•»• |_ —
Now consider water at 20°C. Air enters at 10°C and 30% relative humidity.
For air saturated at 10°C
HCsatd.) = 0.0121 moles water/moles air
so H(30%) - 0.00363
Suppose we have 1 mole water and allow 0.2 moles air to pass through
it. For a first try assume the air leaves at 20°C.
H(satd.,20°C) » 0.0233
so 0.2 moles of air take out
0.2 (0.0233 - 0.00363) » 0.003934 moles water.
The latent heat of water at 20°C is 10,550 cals/mole and the specific
heat is 18 cals/(mole)(°C) so the water temperature falls
10,500 x 0.003934/18 - 2.31°C.
Since the air must leave at the average water temperature we cannot
assume that the air leaves at 20°C. We will assume that the air leaves
at 19 °C.
H(satd. 19°C) - 0.0219
so 0.2 moles of air take out
0.00365 moles water
and the water cools by 2.13°C. Since the water starts at 20°C and ends
at 17.9°C, our assumption of an average exit air temperature of 19°C is
satisfactory. The calculation must now be repeated for the next 0.2 moles
of air with the water starting at 18°C.
The calculation is tedious and may not be worthwhile. The discussion
of how much detail is needed is deferred to Section 5.8 in which we describe
continuous adiabatic stripping.
5.5 Isothermal, Counter-Current Stripping and Estimation of Height of
an Equivalent Theoretical Plate
The theory of packed towers is given in all chemical engineering texts
on mass transfer and only a brief summary is given here.
45
-------
Summary of the Theory
A packed tower is pictured on Figure 5-6. The tower is Z meters high
and S m in cross section. Dirty water enters at a rate W moles/hr and,
because we assume that water does not evaporate, water leaves at the same
rate. The vapor rate (steam or air) is'V moles/hr. x and y are mole
fractions of the contaminant in the liquid and vapor phases.
The overall rate of mass transfer of contaminant from water to vapor
at height Z of the tower can be described by the equation
dJ, (Flux in moles/hr) = 1^ a (x - x*) SdZ - - Wdx - Vdy
In this equation
SdZ is the differential volume of the tower, m ,
a is the interfacial area caused by the packing,
m /m of tower,
jr is the mass transfer coefficient, moles/(hr) (m )
(unit mole fraction driving force measured in the water)
x = mole fraction of contaminant in the water
x* ° mole fraction that would occur in the water if the water
were in equilibrium with the vapor
We consider a tower at atmospheric pressure so
x« - y/H
where
y = mole fraction of contaminant in the vapor at
height Z of the tower
H «* Henry's Law constant expressed in atmospheres
The differential equation for mass transfer can be written in integrated
form as
x
'in Ju KT a S Z
dx
_KL
x - y/H w
out
Two equations are possible, one based on water concentrations and water
flow, the other based on gas concentrations and gas flow. The choice is
arbitrary but water concentrations are the more convenient.
46
-------
DIRTY WATER
W moles/hr
xin
DIRTY VAPOR
V moles/hr
CLEANED WATER
Tower cross
section =
S m2
CLEAN VAPOR
W moles/hr
xout
V moles/hr
'in
= zero
Figure 5-6. Nomenclature for a packed tower.
47
-------
The integral on the left hand side of the above equation is called
the number of transfer units, NTD. The bigger the difference between x.
~"""^^*"""~ Ul
and xout/ the larger is the -number of transfer units needed to perform the
stripping. Evaluation of NTD can, in general, be made by numerical
integration of concentrations measured in the liquid and vapor at several
points in an operating tower. Only a few points are needed and the use of
Simpson's rule is usually adequate. However, in the particular case of
interest here, where Henry's Law applies, direct integration is possible
as discussed below.
The term w
v"5"
which has the dimensions of length, the same as Z, is called the height of
a transfer unit. The larger K_a the smaller is the height of packing
needed to do a given transfer. The height of the tower is, from our
definitions,
Z a HTD x NTD
The height of a transfer unit and hence the mass transfer coefficient,
K^a, is determined experimentally by finding the number of transfer units in
a tower of known height at constant water and vapor rates.
Direct Integration to Determine the Number of Transfer Dnits
Since Wdx » - Vdy » dJ
the flux, J, is a linear function of x and of y and, because H
is taken to be constant, of (x - y/H) . The slope of the line of plotting J
against driving force (see nomenclature on Figure 5-6) is
d(x - y/H) (xin " yout/H) " *out
" " Jtotal
(y^ has been taken as zero)
It is convenient to write
xin " W^ " *top
X • A
out bottom
x - y/H = A
Putting dJ B A s dz gives
48
-------
V A S Jtot
which can be integrated to give
in ^top
111 A" K a S Z
hot L .
bot- dZ
top ~^>ot tot
The lefthand side is the reciprocal of the logarithmic-mean driving force
(LMDF) and we can write
Jtot * V S Z (LMDF)
Since Jtot " W (*in * "out*
we can also write
W(x. - x J
in out
Z " l^a S (LMDF)
• ETD X
Xin " Xout
LMDF
from which it follows that
X. - X
NOT = in out
LMDF
Design of a Packed Tower
For a given job, W, x. , x „ are given. The vapor rate, V, is chosen
7 in out
and y calculated from the material balance
out
W(*in ' "out* ' %«t
It is then possible to calculate LMDF and NTD. A suitable packing is chosen
for which K_a is known for the water and vapor rates chosen. HTU and Z can
then be found. Repeated calculations will give the trade-off between
decreasing the vapor rate, V, and increasing the tower height, Z. Additional
calculations can show the advantage of different packings.
Height of an Ecuivalent Theoretical Plate
We find it easiest to visualize stripping towers in terms of equivalent
theoretical plates which are used in the following sections. The number of
49
-------
theoretical plates to do a given separation does not depend on the equipment.
The height of packing to do a given separation depends on the choice of
packing. We need, therefore, to find the height of packing equivalent to
a theoretical plate (HETP).
The height of a tower (Z = HETP) which is equivalent to a theoretical
plate is the height which causes the vapor leaving at the top to be in
equilibrum with liquid leaving at the bottom. That is
" Xo«t
In this case
Xin -
vi-
(xin - Xout> * (yout - yin)/H
X. - X .
becomes , in out
(x. - x .) - x . + y. /H
in out out in
Now the mass balance is
so
W(xin - xout> = * V (yin -
xout ' -H (xin- Xout}
If we write
R(range) = x -
In R
WR/VH
1/LMDF = R - WR/VH
and NTU = R/LMDF = J".^^ = W/VH - 1
HETP = H-TJ x ln W/VH
HETP - HTU X
50
-------
5.6 Number of Equilibrium Stages in Continuous Isothermal Stripper.
In this section we find the number of equilibrium, or theoretical,
stages to do a job of stripping. The theory will then be used to classify
pollutants by ease of stripping.
Theoretical stages are very useful for comparing pollutants, but they
are not, in any sense, realizable pieces of equipment. A real tower may
contain packing, in which case one has to know the height of an equivalent
theoretical plate. Or a real tower may contain plates. But a real plate is
not an equilibrium stage and one has to know the plate efficiency to find
the number of real plates when the number of theoretical plates is known.
The difference is large. Efficiencies may be as low as 0.5 and two real
plates may be required to do the work of one theoretical plate.
The General Formula
A simple stripping tower is shown on Figure 5-7. Dirty water at a rate
F moles/hr and having a mole fraction x. of contaminant enters at the top.
Clean stripping vapor at a rate S moles/hr enters at the bottom. Cleaned
water at a flow rate W moles/hr and having a mole fraction XQ of contaminant
leaves at the bottom, x ^ is specified. Dirty vapor at a flow rate V moles/hr
out
and having a mole fraction y of contaminant leaves at the top. For the
moment we consider only the cases where
F = W and V o S
That is, the cases in which water is neither evaporated nor condensed. If
steam is the stripping vapor the feed enters at the boiling point. If air
is the stripping vapor, it enters saturated at the water temperature. Our
assumption is an approximation but the concentrations of contaminants are
so low that removal of contaminant has negligible effect on the flow rate
of the total streams.
A mass balance for the contaminant around the bottom of the tower
including the first plate gives
Mass IN » Mass OUT
wx2 » wxx + vyx
A similar mass balance including the second plate gives
Wx • Wx- + Vy
If the (n-l)th plate is included
Wx = Wx, + vy .
n 1 Jn-l
51
-------
DIRTY VAPOR
DIRTY WATER'
F. X.
-,
n
n-1
CLEANED WATER
W, x
'out
' x.
STRIPPING VAPOR
(STEAM OR AIR)
Figure 5-7. Simple stripping tower showing equilibrium stages.
52
-------
This equation has two unknowns, x and y , and another equation is needed.
n n—l
We assume that the vapor and liquid leaving a plate are in equilibrun so
yn » (H/P) xn - Kxn
Equilibrium plates are theoretical plates. Actual stripping column plates
do not reach equilibrium. The approach to equilibrium is discussed in
Section 6.
The calculation of the number of theoretical plates can now be made.
x2 - XL + (V/W) y^ = x^ +• (VK/W) x^
X3 = Xl * (V/W) Y2 " Xl * (VK/W) *2
=» X1 + (VK/W) x1 + (VK/W)2 xx
and so on;
*n - x1tl +
To sum the series inside [], let
T - 1 + VK/W + (VK/W)2 + . . . + (VK/W)11""1
then (VK/W) T - VK/W + . . . + (VK/W)°
and ((VK/W) -1) T - (VK/W)n - 1.
so
Xl (VK/W) - 1
xn
If n is the number of plates to reduce the feed concentration,
x. = x , to the effluent concentration, x » x , then
xout m (VK/W) -
Xin =
n * * ' In (VK/W)
Sample Calculation
Suppose that epichlorohydrin is to be steam stripped at 1 atmosphere.
H a 20 atm and K « H/P ** 20. Let the feed concentration be 5,140 mg/1.
Since M.W. = 92.53,
x. = 5,140 x 18 x !0"6/92.53 - 10"3
in -6
we choose x ,=10 (5.14 mg/1)
out
53
-------
Let V/W - 0.15;
VK/W = 3
n + 1 - *** ! » " 5.9
Additional calculations give
V/W n
0.1 9
0.125 7
0.15 6
0.2 5
0.3 4
Note that the larger the value of V/W, that is the more steam used
for stripping, the smaller the number of plates. This is always true of
stripping columns; the designer has a tradeoff between a lot of plates or
• .
a lot of stripping vapor.
The Range of Practical Operabilitv
There are practical limits on the number of theoretical plates and
the mass of stripping vapor used. For the purposes of this report we will
make the following assumptions:
(1) the column is limited to 20 theoretical plates
(2) the desired ratio x ./x. is between 10~2 and 10~4
out in
(3) stripping vapor rates will lie on the range
0.1 < V/W < 0,3 for steam
0.75 ^ V/W < 2.5 for air
Assumption number 3 is discussed ia the following subsection.
The relationship between VK/W, n and xin/xout is graphed on Figure 5-8.
VK/W must be greater than one. Given the upper limit on the number of plates
of 20 plates a low value of VK/W -1.1 may just be useful. There is a
general lower limit of VK/W = 1.2. As VK/W is increased fewer plates are
needed. When VK/W = 10 about 3 plates will do the job. Increasing VK/W
above 10 does not give a large decrease in the number of plates and will
probably not be worth paying for the extra vapor. In general we will use
1.2 < VK/W< 10
with a probable extreme of
1.1 < VK/W < 100.
54
-------
in
ui
VK/W
100
50
20
10
I I I
5x10
-2
20 PLATES
• i •
I I I I I I I I I
10~2 5xlO~3
10~3 5xlO~4
xout/xin
10
-4
Figure 5-8. Relationship between stripping vapor rate and number of plates.
-------
Choice of Vapor Rate
To compare air and steam some discussion is necessary. The general
equations for the horsepower needed to drive a gas compressor are :
HP a WH/(33,OOOe)
,. (VI^/WL r-^*-1
\ ' \ w / L.
(n-l)/n - (k-l)Ae
HP is horsepower » 0.745 KW
W is gas flow in Ib/min
H is polytropic head (ft-lb)/lb
e is polytropic efficiency » 0.77
Z , Z. are compressibility factors for suction and discharge, - 1
s a
« is molecular weight
w '
T. is suction temperature, = 530 °R(°R = 460 + °P)
r is the compression ratio =1.3 for 4.5 psi pressure drop
k is ratio of specific heats a 1.40
(k-l)A for air - 0.286
From this equation, for a temperature of 70°F and compression ratio
of 1.3, the energy to supply S Ib moles of air for stripping is 0.11 S kw-hrs.
The energy to supply S Ib moles of steam for stripping is about 18,000 S
Btu. To compare steam to air one must convert kw-hrs to Btu. The simple
energy conversion for generating electricity is 10,000 BtuAw-hr which makes
1 Ib mole of steam equivalent in energy to 16.4 Ib moles of air. However,
the value of 1 kw-hr is usually closer to the value of 20,000 Btu which
makes 1 Ib mole of steam of equal value to about 8 Ib moles of air. (An
example of value is if electricity is worth 4$/kw-hr and low pressure
steam worth 52/lQ6 Btu).
Note that for steam the vapor/liquid ratio is the same for weight
rates of flow as for molar rates of flow. For air
vapor/liquid weight ratio • 1.6 vapor/liquid molar ratio
56
-------
' In practice steam strippers (such as ammonia strippers) usually operate
with steam rates between 0.1 and 0.3 times the feed rate, that is
V/W » 0.1 to 0.3
just as we have chosen.
Air strippers, in the form of cooling towers usually operate with
gas/liquid weight rates in the range 0.75 to 2, that is molar rates in
the range 0.47 to 1.25, so our chosen molar ratios are on the high side
and have been chosen to .give equal cost between steam and air.
5.7 Pollutants Classified by Ease of Stripping
At this point it is possible to classify compounds by the ease with
which they can be stripped. The classification is given on Table 5-2.
Consider compounds listed as very easily stripped. These are all compounds
for which it is practical and simple to obtain VK/W > 10. For steam we
chose V/W • 0.1 so K(100°C) "z. 100. For air we chose V/W » 0.75 so
K(20°) > 13. • Since K = H/P and we usually are concerned with atmospheric
pressure strippers, H - K if H is expressed in atmospheres.
The "difficult to strip" compounds have VK/W » 1.2; V/W for steam =
0.3 and K (steam) > 4. A few compounds with H(100°C) between 2 and 4 atm.
have been classified as "very difficult to strip". These compounds require
more than 20 plates. Compounds with H(1004C)< 2 atm have been classified
"cannot be stripped*.
Of a total of 185 listed toxic pollutants Table 5-2 shows that
68 are very easily stripped or decompose in water.
36 are strippable with various degrees of difficult.
48 cannot be stripped or probably cannot.
33 lade information.
For most compounds the classification was independent of the choice
of stripping vapor (air or steam). A few compounds, marked *, are
classified for air stripping and will be more easily stripped by one
category higher if steam stripping is used.
Touhill (Ref. 37 page 235) has compiled a large list of reports on
stripping. Our classification is compatible with this list.
5.8 Adiabatic Continuous Stripping with Air
If the water does not enter the top of the column at the wet bulb
temperature of the air, then water is stripped and, in the usual case
57
-------
Table 5-2 TOXIC POLLUTANTS CLASSIFIED BY EASE OF STRIPPING
Very easily stripped
H(100°C)>100 atm
H(20°C) > 13 atm
5-4+ 2-Chloroethyl vinyl ether
7-13 Triethyl amine
9-1 Benzene
9-2 Chlorobenzene
9-3 1,2-DiChlorobenzene
9-4 1,3-OiChlorobenzene
9-5 1,4-Dichlorobenze
9-6 1,2,4-Trichlorobenzene
9-7 Hexachlorobenzene
9-8 Ethyl benzene
9-TO Toluene
9-15 Benzyl chloride
9-16 Styrene
9-T8 Xylenes
10-1 2-Chloronaphthalene
** 10-3 Benzo(b)fluoranthene
** 10-4 Benzo(k)fluoranthene
** 10-5 Benzo(a)pyrene
10-9 Acenaphthene
** 10-12 Chrysene
** 10-13 Fluoranthene
10-15 Naphthalene \
10-17 Ryrene
+ the left hand column are the code numbers used in the Treatability
Manual and in Appendix 3.
** probably, but data is poor
58
-------
(very easily stripped - continued)
11-1 Aroclor 1016
11-2 Aroclor 1221
11-3 Aroclor 1232
11-4 Aroclor 1242
11-5 Aroclor 1248
11-6 Aroclor 1254
11-7 Aroclor 1260
12-1 Methyl chloride
12-2 Methylene chloride
12-3 Chloroform
12-4 Carbon tetrachloride
12-5 Chloroethane
12-6 1,1-Di Chloroethane
12-7 1,2-DiChloroethane
12-8 1,1,1-TriChloroethane
12-9 1,1,2-Tri Chloroethane
12-10 1,1,2,2-Tetrachloroethane
12-11 Hexachloroethane
12-12 Vinyl chloride
12-13 1,2-Oh'chloropropane
12-14 1,3-Dichloropropene
1*2-15 Hexachl orobutadi ene
12-16 Hexachlorocyclopentadiene
12-17 Methyl bromide
12-18 Oichlorobromomethane
12-19 Chiorodibromomethane
12-20 Bromoforra
12-21 Dichlorodifluoromethane
12-22 Trichlorofluoromethane
12-23 Trichloroethylene
59
-------
(very easily stripped - continued)
12-24 1,1-Dichloroethylene
12-25 1,2-Trans-di chl oroethylene
12-26 Tetrachl oroethylene
12-27 Ally! chloride
12-30 Ethylene dibromide
13-20 Heptachlor
13-25 Toxaphene
13-37 Isoprene
13-46 Carbon disulfide
14-4 Amy! acetate
14-5 n-Butyl acid
14-13 Vinyl acetate
15-1 Methyl mercaptan
15-3 Cyclohexane
Easily stripped
H(100°C) 20 to 100 atm.
H(20°C) 2 to 13 atm.
5-3 Bis(2-chloroisopropyl) ether
7-7 Acrylonltrile
9-19 Nitrotoluene
10-10 Acenaphthylene
10-14 Fluorene
10-16 Phenanthrene
13-8 Aldrin
13-9 Dieldrin
13-24 Chlordane
** 13-26 Captan
14-1 Acetaldehyde
** 14-16 Acrolein
14-18 Propylene oxide
probably, but data is poor
60
-------
Intermediate
H(100°C) 8 to 20 atm.
H(20°C) 1 to 2 atm.
5-2 Bis(2-chloroethyl) ether
8-5 Pentachlorophenol
8-6 2-Nitrophenol
9-9 Nitrobenzene
9-11 2,4-Dinitrotoluene
9-12 2,6-Dinitrotoluene
10-11 Anthracene
12-31 Epichlorohydrin
13-12 4,4'-DDD
13-21 Heptachlor epoxide
Difficult to strip
H(100°C) 4 to 8 atm.
H(20°C) 0.5 to 1 atm.
8-2 2-chlorophenol
8-3 2,4-D1chlorophenol
8-4 2,4,6-Trichlorophenol
14-15 Crotonaldehyde
15-4 Isophorone
* this is for air; a higher category for steam
** probably, but data is poor
61
-------
Very difficult to strip
H(100°C) 2 to 4 atm.
6-2 Diethyl phthai ate
8-1 Phenol
* 8-10 2,4-Dimethy!phenol
8-13 4,6-Dinitro-o-cresol
8-14 Cresol
* 9-17 Quinoline
* 14-3 Ally! alcohol
Cannot be stripped
5-7 Bis(2-chloroethoxy) methane
6-1 Dimethyl phthalate
7-10 Ethylenediamine
8-7 4-Nitrophenol
8-8 2,4-Dinitrophenol
8-9 Resorcinol
9-13 Aniline
9-14 Benzole acid
10-8 Benzo(ghi) perylene
13-2 Endosulfan sulfate
13-13 Endrin
** 13-18 Diurone
13-22 Carbofuran
** 13-28 Coumaphos
13-29 Diazinon
13-30 Dicamba
1.3-31 Dichlobenil
13-32 Malathion
13-33 Methyl parathion
13-34 Parathion
* this is for air; a higher category for steam
** probably, but data is poor
62
-------
(cannot be stripped - continued)
13-35 Guthion
13-38 Chlorpyrifos
13-39 Dichlorvos
13-41 Disulfoton
13-43 Mexacarbate
13-44 Trichlorfon
14-2 Acetic acid
14-6 Butyric acid
14-7 Formaldehyde
14-8 Formic acid
14-12 Propionic acid
14-14 Adipic acid
15-2 Oodecyl benzenesulfonic acid
15-5 Strychnine
15-7 Zinc phenol sulfonate
Poor data, but probably difficult to strip
5-5 4-Chlorophenyl phenyl ether
6-3 Di-n-butyl phthalate
6-5 Bis(2-ethylhexyl) phthalate
7-1 N-nitrosodimethylamine
7-2 N-nitrosodiphenylamine
7-5 3,3'-Diphenylhydrazine
7-6 1,2-Oiphenylhydrazine
13-11 4,4'-DDT
13-17 Kepone
13-40 Oiquat
13-42 Mevinphos
14-9 Fumaric acid
14-10 Maleic acid
63
-------
Poor data; better data worth obtaining as may be strippable
5-6 4-Bromophenyl phenyl ether
6-4 Di-n-octyl phthalate
6-6 Butyl Benzyl phthalate
7-3 N-nitrosodi-n-propylamine
7-4 Benzldine
7-9 D1ethyl amine
8-12 p-Chloro-m-cresol
10-2 Benz(a}anthracene
10-6 Indeno(l,2,3-cd)pyrene
10-7 01benzo(ah)anthracene
13-10 4,4'-DDE
13-27 Carbaryl
14-17 Furfural
15-6 2,3,7,8-Tetrachl orodibenzo-p-dloxln
Data inadequate for comment
7-8 Butyl amine
7-11 Monoethylamine
7-12 Monomethylami ne
7-14 Trimethylaraine
12-28 2,2-Dichloropropionic acid
13-1 a-Endosulfan
13-3 8-Endosulfan
13-4 o-BHC
13-5 8-BHC
13-6 6-BHC
13-7 Y-BHC
13-14 Kelthane
13-15 Naled
13-16 Dichlone
13-19 Endrin aldehyde
64
-------
(data inadequate for comment - continued)
13-23 Mercaptodimethur
13-36 Ethion
13-45 Propargite
14-11 Methyl methacrylate
Decompose in water
5-1 Bis(chloromethyl) ether
12-29 Phosgene
65
-------
which is adiabatic, the water cools. The Henry's Law constant varies from
plate to plate as the temperature varies. It is a simple matter to estimate
the fraction stripped on each plate once the temperature is known on
each plate. The method for calculating the temperature on each plate is
given in Appendix 2 and some results are discussed below. But a word of
caution is required.
We calculate the temperature, and hence the fraction of organic
stripped, on a theoretical or equilibrium plate. When a real tower is
built, it is important to remember that real plate efficiencies may .not
be_ the same for water and for the organic. Also, the height of packing
equivalent to a theoretical plate is not the same for water and for the
organic. The rate of stripping of water is controlled by the rate of
diffusion of water vapor from the interface through the air layer and into
the main body of the air. The rate of stripping of volatile organics is
controlled by the rate of diffusion from the main body of the water to
the interface; this is always an order of magnitude slower than the diffusion
rate in the gas if the organic is easily strippable. The liquid side
resistance for organic stripping has been illustrated in Section 5.2 where
evaporation from a lake was described.
Figure 5-9 shows a typical warm water such as might be fed to a
cooling tower with typical mid-season air (not the height of summer nor
the middle of winter). The feed rates are 0.37 moles air/mole water. In
one theoretical plate the water temperature falls by 13°C (24°F) which is
about the way that cooling towers are designed. In fact our calculations
suggest that the usual cooling tower is close to one theoretical plate.
In the next 5 plates the temperature only falls an additional 10°C.
For this situation stripping of an organic can reasonably be estimated
using the isothermal stripping equation with H evaluated in the temperature
range 30 to 35°C. It is worth pointing out that in this particular
application, air is heated as it rises. The hot air near the top of the
tower has a much greater capacity for water vapor than the entering air.
Thus most of the water which evaporates does so on the top plate and the
biggest temperature drop for the water occurs on the top plate.
On Figure 5-10 is shown the effect of stripping a cold water with
a rather dry air. Note, as before, that on the top one or two plates the
temperature of the water falls rapidly, but that on the lower plates the
66
-------
, 50
40 —
u
o
-------
a*
co
U
a
01
rt
M
0)
10 -
Water feed temperature = 20°C (68°F)
Air feed: 10°C (50°F) , 30% R.I1.
0.47 moles air/mole water
0.62 "
II II
0.93
Wet bulb temperature
^v
I I I
I I I I I I 1 1
2 4 6 8 17 18 19
Number of theoretical plates
Figure 5-1O. Water temperature when stripping cold water with air.
-------
water temperature changes very little. The approach to the air wet bulb
temperature is very slow. The vapor pressure of water falls off rapidly
as the temperature falls and so, therefore, does the driving force for
mass transfer. The fractional approach to the wet bulb temperature falls
rapidly as the water temperature falls.
While it is not difficult to calculate the temperature on each plate
with a computer, it is very tedious to do by hand. We have not made
enough estimates to find any reliable simplification. The best we can
suggest, and it is not accurate, is that isothermal stripping be assumed
to occur at a temperature which is the arithmetic mean of the feed water
temperature and the air wet bulb temperature.
When the air rate is about doubled from 0.47 moles/mole water to
0.93 moles/mole water, the steady temperature on all but about the top
three plates is reduced from about 13.2°C to about 7.5°C. The stripping
of an organic depends on the group VVW » (at one atmosphere) VH/W.
Returning to a previous example of nitrobenzene we find that
H(13.2°C) = 0.632
and HC7.S°C) » 0.454
This means that doubling the air rate multiplies the fraction stripped by
(0.454 x 2)/0.632 = 1.4. There is a positive advantage to increasing the
air flow, but it is a much smaller advantage than night have been thought
to occur if the effect of temperature had been neglected. Each case will
have to be estimated but our preliminary conclusion is that varying the
air rate is not an efficient way to control a stripping tower.
But some control is necessary. The temperature and humidity of the
air vary from hour to hour, day to night and season to season. If the
organic is very easily stripped, the variations in the air conditions
will matter very little. A small overdesign will guarantee satisfactory
stripping under all likely conditions. For example, Recon Systems, Inc.,
of Somerville, NJ, have designed and installed very reliable stripping
columns for such extremely volatile contaminants as trichloroethane and
trichloroethylene. There is, however, a. lot of difficulty in designing
a column for stripping a less volatile compound. , Consider o-nitrotoluene
as an example. Take 99% stripping at V/W - 1.
69
-------
Temp (°C) Number of Theoretical Plates
20 2.6
15 2.8
10 3.2
5 6.6
The column can be designed for 5°C water on the coldest winter days and
the column wiill be twice as high as needed in summer time, or the column
can contain about 3 plates and satisfactory stripping will not be obtained
on the coldest days.
One alternative is to use steam stripping, which is not dependent on
the weather, for difficult-to-strip compounds. The energy cost will not be
higher, but the capital cost is usually higher than for air strippers
because the influent and effluent water must be heat exchanged for reasonable
energy efficiency.
A possible alternative, briefly explored below, is to mix steam and air.
5.9 The Use of Air-Steam Mixtures
If, to air at 10°C, 30% RH, is added 0.0085 moles steam/mole air, which
is the quantity of steam needed to saturate the air at 10°C, the dry bulk
temperature rises to 10.9°C and the wet bulb temperature is raised from
3.4°C to 10°C.
Now suppose that a stripping column does a good job when the air and
water enter at about 20°C, the air has a high humidity and the air to water
rate is 0.93 moles air/mole water. As is shown on Figure 5-10, if the
air temperature and humidity drop to 10°C, 30% RH, the mean water temperature
approaches 7.5°C. However, adding a little steam (less than 1% of the air)
raises the mean water temperature to about 11°C. In our example for
o-nitrotoluene we have seen that an increase in temperature from 7.5 to 118C
can divide the plates necessary to do-a given stripping job by about 1.7°.
The addition of a little steam when the air is dry seems to be a good
way of controlling a stripping tower.
5.10 The Effect of Recireulating the Water
Water can be recirculated to obtain more net stripping in a tower
of lower height or less plates. The penalty is a bigger tower cross
section and pumping costs. In a cooling tower, as pictured on Figure 5-11,
recirculation of the water is the normal practice.
70
-------
AIR
F.
HEAT
EXCHAN-
GER
1
STRIPPING
TOWER
t
F moles/hr
1
«-*•
MAKEUP
M moles/hr
XM
SLOWDOWN
B moles/hr
XB
AIR
V moles/hr
y s 0
Figure 5-11. Stripping in a cooling tower with circulated water.
71
-------
Suppose that in a single pass down the stripping tower the feed
concentration at the top, x , is reduced to a bottoms concentration, x ,
F B
and we write „
*B
—-— =» r (the removal)
r is a function of the water rate, air rate and Henry's Law constant, but
does not depend on x . It is important to remember that if water is
evaporated, as it usually is, the concentration of organic may increase
in a pass through the tower if the organic is not volatile; that is r can
be greater than 1.
Now let suffix, n, denote the conditions of the circulating water
entering or leaving the tower for the n'th time. Water leaving the tower
fron the n'th pass is circulated and enters the tower for the (n+1) 'th pass.
Fx,, , - WX,, - Bxa + MX..
F, n+1 B, n B, n M
so
***. n+l/r - <» - B) XB, n * ""M
but,
P « W - B •*• M
or W - B = F - M-
so
FXB, n+l/r " {F " M) XB,n *
If, now, one puts
F/M » R (the recycle ratio)
_ (R-l)r r
*3, n •»• 1 R * ^, n R M
The average number of passes made by the water is, in fact, equal to the
recycle ratio, that is
n » R
72
-------
1.0
0)
c
•H
u
c
o
u
1
•a
*
c
-H
U
c
o
o
o.o -
0.6
0.4
0.2
20
1.0
curves are labelled by the value of
B,l , the concentration
M per pass.
40
Recycle ratio, R
60
80
Figure 5-12. The effect of recycle when the organic is stripped.
-------
When the system is first started up the concentration everywhere is
the makeuD concentration x ; that is
"" n
XF,1 " XM
S.l - r XM
It follows that
"*
„ -„ + x-+x
•» 5 Xu 5 M H M
,3 _2 M _2 H R M
R K
Rn-2
(R-l)r fB^ (R-l)nrn+1
R ^M ° Rn
^ • • • T-
_
A
(R-l) r *B
""
Rn-l Rn Rn
74
-------
Putting
n = R
and consolidating gives
• j ^ ^ \ ** ** t * %
r +• (R-1) g (1-g)
Tf O^ 1
*B,n _ R8 •*"
xu " R(l-r) + r
A numerical investigation of the above equation shows that for nearly
any recycle (R > 2) if an important amount of stripping occurs per pass so r
is small (rs approx 0.2) it is sufficiently accurate to write
XB,R
for
*M
** ^^^ •
r^ 0.2
For larger values of r the equation is graphed on Figure 5-12. So long
as r< 1, the concentration in the blowdown falls rapidly as the recycle
ratio increases to 10 or 20 and then falls much less with additional recycling.
(A cooling tower has a recycle ratio on the order of 50).
If r » 1, recycle has no effect. It must be remembered that when r = 1,
the concentration of the organic caused by evaporation of water is exactly
offset by the stripping of the organic.
When r > 1> Figure 5-13 shows that the concentration increases rapidly
with the recycle ratio.
75
-------
2.0
o>
curves labelled by the value of
XB,1 , the concentration
per pass.
Recycle ratio, R
Figure 5-13. The effect of recycle when the organic Is concentrated.
-------
6. THE EFFICIENCY OF STRIPPING EQUIPMENT
6.1 Introduction and Summary
The development of the theory of stripping was based on the number of
theoretical plates required to do a given job of stripping. The number
of theoretical plates depends on the Henry's Law constant (and therefore on
the temperature) and on the ratio of water to vapor rates. The number of
real plates depends on their efficiency. We have found no data on the
efficiency of distillation trays or of packed towers for stripping toxic
pollutants, and very little data on stripping of any sort. All available
data comes from distillation and absorption. The efficiency for distillative
separation of miscible organic liquids is usually in the range 60% to 100%.
But the efficiency for absorption of C02 into water can be as low as 1%.
Stripping of an insoluble organic is probably more related to adsorbing an
insoluble gas than to a distillation, and we cannot assume a high efficiency.
A theoretical (equilibrium) stage can be obtained in practice. It is
the usual method for measuring the Henry's Law constant experimentally.
•
However, the liquid must be at least 60 cm deep, the vapor must be well
distributed and in fine bubbles and the flow rates must be low. It is
never economical to use multiple theoretical stages and cheaper, less
efficient equipment is always used.
The factors which affect the efficiency of a bubble-cap tray and of
a packed bed are listed in Sections 6.2 and 6.3. It is reported that
efficiency decreases as the Henry's Law constant increases. This
phenomenon is explained on the basis of the two film theory of mass transfer.
For most stripping jobs of interest here the liquid film will be the
controlling resistance. This is shown in Section 6.4. Finally a few
reasons for choosing particular equipment are given in Section 6.5.
The design of any particular piece of equipment is not given here;
in particular, the sizing of equipment to the correct capacity is not
discussed, nor do we describe the determination of optimum liquid and
vapor rates and of pressure drop. Designs are best made from manufacturers'
design manuals after the type of equipment has been chosen.
6.2 Factors Affecting the Efficiency of a Bubble Cap Tray
Real trays do not reach equilibrium and a real tower requires more
trays to perform a given degree of stripping than the calculated number
of theoretical stages. One way to define tray efficiency is to use an
overall efficiency, S°, defined as
77
-------
0 Number of theoretical stages required
* Number of actual trays required
The overall efficiency has been found to depend on the following factors
(as summarized by Gilliland and the Chemical Engineers Handbook ).
Viscosity. Efficiency increases as the liquid viscosity decreases. Different
authors have found different expressions for the dependence but the variation
is approximately
E" is proportional to (viscosity)~n
where n is 0.7 to 0.9
so the effect is large. Since the viscosity of water decreases from 0.89
centipoise at 25°C to 0.28 at 100°C there will be an important increase
in efficiency with temperature.
Liquid Depth. The efficiency increases as the depth of liquid on the tray
is increased. The price is an increased pressure drop.
Vapor Rate. The efficiency is not much dependent on vapor rate 'up to the
point where frothing and entrainment occur. Entrainment causes liquid to
be mixed backwards up the tower and the efficiency•to decrease.
Liquid Rate. Since liquid on a bubble-cap tray flows across the vapor
(see Figure 6-2 below), it is possible for there to exist more than one
equilibrium stage on a tray. Insofar as an increased liquid rate causes
more back mixing in the liquid, the efficiency falls somewhat as the
liquid rate increases. The effect is dependent on tray design.
Henry's Law Constant. The efficiency decreases as the Henry's Law constant
increases. This is a most important consideration in the design of
strippers because we are most often concerned with compounds having a high
Henry's Law constant. The effect of Henry's Law constant cannot be
understood in terms of intimacy of contact between vapor and liquid, as
the preceding factors have been. The effect is caused by the controlling
resistance to mass transfer being increasingly due to the rate of diffusion
of the organic through the liquid as the Henry's Law constant increases. A
preliminary understanding was given in Section 5-2 where evaporation from
a lake was considered.
78
-------
The overall efficiency is not a useful tool for understanding
resistances to mass transfer and it is customary to use an efficiency
called a Murphree efficiency, E,_ which is defined as
MV
where
y. • vapor composition entering the plate
y » vapor composition leaving the plate
y* a vapor composition in equilibrium with the
liquid leaving the plate
Since, on a theoretical plate, y = y*, the Murphree efficiency is a
o e
measure of
depth of liquid on the plate
• depth of liquid on a theoretical plate
that is, the Murphree efficiency compares the depth of liquid on the plate
to the height of liquid equivalent to a theoretical plate.
The mathematical model for defining a Murphree efficiency is just the
same as the model used in Section 5.5 where the height of packing equivalent
to a theoretical plate was determined. We will, therefore, next list the
factors affecting the efficiency of a packed bed and then describe the two
film theory for rate of mass transfer and show how the HETP increases and
tray efficiency decreases as the Henry's Law constant gets larger.
6.3 Factors Affecting the Efficiency of a Packed Bed.
A packed bed is more efficient the smaller is the height of an
equivalent theoretical plate (HETP) or the height of a transfer unit (HTU).
HETP and HTU are affected by the following factors:
HETP and HTU Factor
Decrease Viscosity decreases
Liquid flow rate decreases
Little change Vapor flow rate alters below
the flooding rate
Increases Henry's Law constant or relative
volatility increases
79
-------
The increase in HETP with relative volatility has been given by
Coulson and Richardson (Ref . 30 page 649) who quote both Murch and Ellis
to show that HETP is directly proportional to relative volatility. An
understanding of this phenomenon comes from the two film theory for mass
transfer.
6.4 Two Film Theory of Mass Transfer.
We assume that for an organic molecule to be stripped it must diffuse
through a Liquid film to the liquid-vapor interface wherp it transfers to
the vapor phase. Then the organic molecule* must diffuse through the
vapor film to the main body of the vapor where it is swept away. The rate
of mass transfer can, therefore, be written
J (moles/hr) (ft of packing)
where
- x.) = koatyi " y> " *!,*** ~ **}
a - interfacial area, ft2/ft3 of packing
k_ » transfer rate through the liquid film, moles/ (hr)
*• . 2
(ft of interface) (mole fraction driving force)
k_ » transfer rate through the vapor film, moles/ (hr)
« -
(ft of interface) (mole fraction driving force)
K, a overall mass transfer rate for a liquid concentration
it
driving force
The mole fractions x, x, x*, y, y are defined on Figure 6-1.
at equilibrium
y - Kx*
y.-Kx.
Algebraic manipulation gives
Also
r- _ „ -, j~x _ x*~|
_±_ j j 1 + JL j L -i. +
-x k x-x k
80
-------
b
a
a
a
fi
a
o»
u
0
§
u
a
Vi
, at the interface
y, main body of
the vaoor
Equilibrium,
y = kx
(x - x^)-
I
x. , at the
l interface
•(x - x*)
x , in equilibrium
with the vapor
I
I
x. Bain body of
the water
mole fraction of the organic in water, x
Figure 6-1. Two film theory of mass transfer.
81
-------
or, at 1 atrosphere total pressure when
K = H measured in atmospheres
In Section 5.5 it was shown that the height of a packed tower
necessary to perform a desired job of stripping is given by
ZCheight) » NTD x HTU
In this equation NTD (the number of transfer units) is a function of
Inlet water concentration
Outlet water concentration
Henry's Law constant
NTU is a measure of the degree of stripping required. It does not depend
on the tower paclcing or the flow rates. The higher the Henry's Law constant,
the lower the NTU for the same reduction in water concentration.
HTU, the height of a transfer unit, is given by
W
HTU = £
K]
where
W/S is the moles of water flowing per hour per unit
cross section of paclcing (moles/ (hr) (f t ),
and
K_a is the overall mass transfer rate for a liquid
concentration driving force defined above in the units
moles/(hr)(ft of packing)
HTU **a" be written
W / 1 _l_
HTU = Sa I k Hk
V L G
and is seen to depend on the resistance to mass transfer in both the
liquid and gas films.
82
-------
The liquid film resistance is found by measuring the overall rate
coefficient, K a, when absorbing a slightly soluble gas such as carbon
L
dioxide or oxygen. The gas side resistance is small because pure gas is
used and all the resistance is in the liquid film. It is found that
k a is a proportional to D and W
where D is the molecular diffusity (so transfer is more
rapid with lower molecular weight molecules than
with large molecules)
W is the water rater
n is 0.5 to 0.8
Although k.a increases as the water rate increases, HTU » W/SJc a also
increases as W increases; that is, more packing is needed to deal with the
increased load even though transfer becomes faster.
The values of k_a and HTU depend on the packing as well as on the
L
properties of the gas being absorbed and on the flow rates, but the order
of magnitude can be given. Water rates vary from a low of
5 gallons/(min)(ft of cross section)
for very open slatted towers like cooling towers, to more than
60 gallons/ (min)(ft2).
That is, water rates are in the range
2,500 lb/(hr)(ft2) to more than 30,000
or.
140 moles/(hr)(ft2) to more than 1,700
HTU is in the range 1 to 5 ft with the larger values applying to higher
flow rates16'34'35'36. The order of magnitude of k^a is 100 to 400
moles/(hr)(ft )(mole fraction), k a is sometimes expressed in dimensioned
3 3
concentration units; since 1 ft of water » 3.47 Ib moles, 1 mole/(hr)(ft )
3.47 moles/(mole fraction) « (hr)(ft3)(moles/ft ).
33
-------
Trulsson has measured k a for adsorption and desorption of CO from
ii 2
air to deionized water in three plastic packings, Intalox saddles,
Tellerettes and Plasdek (Hunter's) Corp. He found k a to lie in the range
-1 3
0.01 to 0.05 sec (125 to 625 moles/(hr) (ft ) (mole fraction)} with the
value depending mainly on liquid load. The liquid load varied from 6 to
60 Kg/(m2)(sec)(10 to 100 gpm/ft2, 280 to 2,800 moles/(hr)(ft2)} and the
higher loadings gave the higher values of k a.
The gas film resistance is found by evaporating water into an air
stream or absorbing the very soluble gas, ammonia, from a dilute gas
stream into water, k a varies with the gas rate to about the 0.8 power,
G
and HTTT is the sane order of magnitude as for liquid film controlled systems.
Thus- k a will be the same order of magnitude as k a when the liquid and
vapor rates are the same (as in air strippers) and k a may be as low as
0.1 k a when the vapor rate is much lower than the liquid rate, as in a
steam stripper. Since k a and k a are the same order of magnitude in air
strippers, BTU for evaporating water is close to HTU for stripping which
means that the difficulties in calculating adiabatic strippers mentioned
in Section 5.8 may not be large.
It also follows that the overall rate of mass transfer, K a, is
predominantly controlled by the liquid film resistance when the Henry's Law
constant, H, is greater than about 10 a-dn for air stripping or greater than
about 100 atm for steam stripping. All the compounds shown on Table 5-2
as "very easily stripped", and many of those shown as "easily stripped" will
be liquid side controlled.
6.5 Choice of Equipment
The advantages and disadvantages of various water-vapor contacting
devices are listed in many tests. We summarize here the suggestions given
by King and by Morris fi Jackson , and by other sources referred to below.
The simplest equipment is a simple spray tower. Small water drops
should be helpful when the resistance is on the liquid side; however, spray
towers are prone to entrainment which causes internal circulation of the
liquid. Thus, true counter-current flow is not obtained and spray towers
do not do more than one theoretical stage. A spray tower is only useful for
easily stripped compounds when only one theoretical stage is needed*.
84
-------
Similarly, in an agitated vessel the liquid is mixed and only one stage
is obtained. It is quite simple to obtain a full equilibrium plate at the
price of a deep vessel and a larger pressure drop, but not more. Agitated
vessels are used for batch stripping but not continuous stripping.
When more than one theoretical plate is required, packed towers or
tray towers are used.
6.6 Choice of Packing or Trays
The choice of a packed tower or a tray tower, and the choice of type
of packing or type of tray, will be made on the basis of cost. A few rules
can be given:
(1) For steam stripping with only a moderate number of theoretical
plates (5 to 20) required, it is not certain whether packing or trays will
be cheaper. When many stages are needed, trays are usually cheaper. With
air stripping, when plastic packing can be used, packing is usually cheaper
than trays.
(2) The pressure drop is less through packing than through trays and
packing is the first choice for air stripping.
(3) The efficiency of a packed tower decreases steadily as the liquid
flow rate'is decreased below the design rate. This is because the liquid
tends to channel and not wet all of the surface of the packing. If very
variable liquid flows, or very low liquid flows are expected, bubble-cap
trays should be considered. A bubble-cap tray is pictured in Figure 6.2.
The liquid level on the tray is controlled by the overflow weir height.
The liquid flows across the tray and the vapor-liquid contact area is
independent of the flow rates. A valve tray is similar to a bubble-cap
tray but sieve trays and perforated trays are not the same and cannot be
used with variable liquid rate.
(4) Trays redistribute the liquid repeatedly and tall towers do
not result in the channeling usually found in tall packed towers. Thus,
when a close approach to equilibrium is required with many theoretical
plates, a tray tower is preferred.
(5) Tray towers are easier to clean than packed towers and are used
when suspended solids are present or precipitation may occur (such as when
lime is added to release ammonia for stripping). Very open packings can
also be used on turbid streams.
85
-------
Bubble cap
Plate
Gas riser
Liquid
s&f
$ >;•?•* x
**^lvM, /,
•V*"-* ^/
^^
r**^<^
l~x~ /,
WIA. .-.^ f.
Figure 6-2. Biisble-cap tray.
36
-------
Efficiency
It appeared to us that packing, which distributes the liquid in a
thin film, should be more efficient than trays when the liquid side is
controlling. We had very brief consultation with makers of packings (Mass
Transfer, Inc., Houston, Texas, and the Hunters Corporation, Fort Myers,
Florida) and with firms who make both packings and trays (Glitsch, Inc.,
Dallas, Texas and Koch Engineering Company, Ltd., Wichita, Kansas). The
manufacturers gave the opinion that overall cost was the deciding factor
and that trays need not be less efficient than packing.
For steam stripping both types of equipment are used. For air stripping
it is so important to maintain a low pressure drop that plastic (cheap)
packing is the first choice. All the manufacturers provide correlations
and manuals for design. Pressure drops are usually in the range 0.2 to 1
inch water per foot of packing, but can be as low as 0.05 inch water/ft for
very open grid packings. Open grid packings, such as used in small cooling
towers, should be given first consideration.
87
-------
7. COMPARISON OF PREDICTIVE CALCULATIONS TO STRIPPING DATA AT LOW
CONCENTRATIONS
In this section the results obtained front strippers in operation are
compared to those from theoretical predictive calculations. The applicability
of theoretical models were assessed by:
• Developing a methodology of comparison
0 Identifying limitations of the existing predictive calculation
methods, if any, and
0 Performing sensitivity calculations to determine the effect of
identified parameters on tower sizing and effluent contaminant
concentrations. • N
7.1 Methodology of Comparison.
Organic removal by air/steam stripping is a function of numerous
interrelated parameters, as discussed in previous sections of this report.
In light of the variety of toxic organic pollutants being considered in
this study, it is not feasible to incorporate the effect of each design
and operating parameter on the theoretical predictive calculations. Thus,
a one-to-one correspondence between the theoretical and experimental data
for all the organic pollutants is unlikely.
The simplified methodology of comparison, as developed and used in
the analysis, is more of a qualitative nature and consists of the following
steps:
1. The predictive calculations are based on sound engineering
principles and the results should be compared only with reasonably
accurate and precise experimental data. Thus, all data were first
checked for consistency. This was especially necessary when the sampling
data was gathered over a period of time and/or the analysis was performed
by two or more laboratories.
2. If inconsistencies were found, original sources of the data were
contacted to determine whether problems were encountered during sampling
and analysis.
3. A comparative analysis is more meaningful and conclusive if enough
data is available. Since the quantity of data available in the beginning
of this study was judged inadequate, a limited review of current literature
was made to obtain additonal experimental stripping data at low concentrations.
88
-------
4. The ease with which a particular compound is stripped depends on
the volatility, and thus Henry's Law Constant is one of the most important
parameters in determining the feasibility of organic removal by stripping.
For a given set of experimental data, the relative degree of removal was,
therefore, compared to the relative magnitude of Henry's Law Constant for
all the components. Calculations were made using predictive methods
whenever necessary production and equipment data, such as stripping agent,
type of contact, and the configuration of: the contacting device were
available. The differences and disagreements were outlined.
5. Further investigation was made to identify design parameters or
unreasonable assumptions in the theoretical models that could be
responsible for the disagreements in the results.
6. A hypothetical system of compounds with different characteristics
was utilized to demonstrate the degree of influence of the identified
parameters on the strippability of a compound.
7.2 Stripping Data at Low Concentration.
Table 7.1 provides a summary of sources and characteristics of stripping
data that were analyzed in this study. This section provides additional
information on the sources and the quality of data.
A. Commercial Scale Units
In the EPA organic data base, commercial wastewater strippers
used in petrochemicals and plastics plants were selected for sampling of
Influent and effluent organic concentrations.. The waste streams contained
phenolic compounds and chlorinated hydrocarbons. Sampling was conducted
for about thirty days period and part of the samples were analyzed by two
laboratories.
water Factory 21 is a 0.66 a /s plant that employs several of the
wastewater treatment processes, including air stripping, to treat waste-
water for trace organic contaminants. The data presented is for cooling
towers that use polypropylene spash-bar packing with an air-to-water ratio
of 3000 m /m at design capacity. The important information in these data
is that the towers were originally designed for ammonia removal and the
type of packing used has very low contact area per m of packing volume
than most other packing materials.
89
-------
TABLE 7-1. SUMMARY OF AVAILABLE STRIPPING DATA AT LOW CONCENTRATION
Source
Scale
Type of
Stripper
Stripping
Agent
Organic Compounds in Aqueous
Waste Stream
1. EPA Organic Commercial
Data Base
Plant #1290-010
2. EPA Organic Commercial
Data Base
Plant #2930-035
3. EPA Organic Commercial
Data Base
•Plant #3390-005
4. Water Commercial
Factory Data
5. Robert S.
Kerr
Environmental
Research Lab.
Laboratory
6. Water General Laboratory
Corporation
7. Treatability Laboratory
Manual
Multistage
Tray Tower
with
Total Reflux
Multistage
Tray Tower
Information
on refluxing
Not Known
Steam Benzene, Nitrobenzene,
2-Nitrophenol, 4-Nitrophenol,
2,4 Dinitrophenol, phenol
Not Benzene, Carbon Tetrachlor.ide,
Known Chlorobenzene, 1,2 Dichloroethane,
1,1,1-Trichloroethane,
1,1-Dichloroethane, Chloroethane,
1,1,2-Trichloroethane, Chloroform,
Ethylbenzene, 1,1-Dichloroethylerie,
Methylene Chloride, Toluene, Vinyl
Chloride, Tetrachloroethylene,
Trichloroethylene, Cis-1,2-Dichloro
ethylene
Not
Known
Air
Packed- Tower
Information
on refluxing
Not Known
Cooling Tower
Polypropylene
Spash-bar
Packing
Packed Tower Steam
Berl Saddles
No Reflux
Multistage
Tray Tower
No Reflux
Packed Tower
Polypropylene
Pall Rings .
Chloroform, Methylene, Chloride,
Toluene, Trichloroethylene,
Vinyl Chloride
Chloroform, Bromodichloromethane,
Dibromochloromethane, Chlorobenzene,
1,2; 1,3? and 1,4-Dichlorobenzene,
Tetrachloroethylene, 1,1,1-
Trichloroethane
Benzene, Chloroform, 1,1,2,2-
Tetrachloroethane, Chlorobenzene,
Ethyl Benzene, Tetrachloroethylene
Steam Ammonia, Phenol
Steam Same as in 5 +
1,2-Dichloroethane, 1,2-Trans
Dichloroethylene, 1,1,1 and 1,1,2
Trichloroethane, Trichlorofluoro
Methane
-------
In the EPA organic data base, information on the type of stripping
agent and the mode of refluxing is not available for Plant Nos. 2930-035
and 3390-005. In all cases, the influent and effluent concentrations of
each organic is available over a period of time.
B. Laboratory Scale Units
Nineteen bench-scale steam stripping runs were carried out in the
study by the Robert S. Kerr Laboratories. Three of these runs involved
acetone, methanol and 2-propanol to evaluate the performance of the equipment,
and the rest involved some single and some mixtures of priority pollutants
listed in the table. The experimental results were compared to those calculated
using stripping factor design models. The draft report of the study concludes
that the performance of the bench-scale stripper for priority pollutants was
unexpectedly low and further states that the use of estimated K-values instead
of the experimentally determined values may be the cause of poor performance.
It recommends that before future evaluations of steam stripping are performed,
the K-values be determined experimentally.
The data of WGC were obtained from a 20 tray continuous steam stripper
of laboratory scale. These data are limited in terms of the number and type
of compounds that are of interest in this study.
7.3 Data Analysis and Results.
All stripping data were first checked for consistency. A simple and
effective parameter that was used in this test is the removal efficiency
defined as:
F concentration in - concentration out ~\ .__
n (%) " \_ concentration in J X XU°
•
since it requires a minimum amount of information. However, for plants for
which additional information was available, a suitable theoretical model for
removal efficiency was used. For example, a continuous plate tower model
with total reflux was used to analyze the results of plant No. 1290-010.
Inconsistencies were identified in the EPA organic data base,
especially in the data obtained from Plant No. 1290-010. The samples
obtained from the plant were analyzed by two laboratories and the results
were found to vary significantly. The data sources were contacted and it
was learnt that significant problems were encountered during sampling and
91
-------
analysis. Table 7.2 tabulates all efficiency calculations for data collected
over a two months period. Similar calculations were performed for plant
Mo. 2930-035 which are tabulated in Table 7.3. Table 7.4 presents results
of the Water Factory study. The ease with which a particular compound is
stripped depends on the volatility, and thus Henry's Law Constant is one of
the primary determinants of feasibility of organic removal by stripping.
For a given set of experimental data, the removal efficiency was, therefore,
compared .to the magnitude of Henry's Law Constant for all compounds. As
expected and supported by predictive calculations, most stripping data
indicated significant removal of compounds with high Henry's Law Constant
(H @ 100°C > 20 atm).
Deviations from this behavior were observed for certain compounds.
There was not necessarily a direct correlation between a compound's Henry's
Law Constant and the removal efficiency. This means the removal efficiency
of compound A with Henry's Law Constant higher than that of compound B,
may not be higher than the removal efficiency of compound B. In addition,
the removal efficiencies of some compounds with high Henry's Law Constant
were not consistently high as would be expected. In many cases this may
be due to analytical and sampling errors. Nevertheless, we spent some
tin"* searching for alternative explanations. High liquid-phase resistance
would be one reason for low removal efficiencies over with high Henry's Law
Constant. The problem was investigated from theoretical aspects using pure
component properties. It was confirmed that liquid diffusivity does affect
the degree of removal and consequently is important in the design and sizing
of stripping equipment. This is discussed in further details with specific
examples in the next section.
In case of packed towers stripping a waste stream consisting of compounds
with Henry's Law Constant higher than 100 atm, the height of packing on the
removal efficiency was found to be nearly independent of the magnitude of
Henry's Constant. Liquid-phase resistance (diffusivity) was a factor affecting
the performance. Again, experimental data on liquid-diffusivity is not
abundant and empirical methods to estimate diffusivity may not provide
reliable values for many organic compounds. Developing better contacting
devices in terms of reducing liquid-phase resistance through more efficient
packing and better liquid distribution in plate columns seems to be the
probable solution to improving the performance of strippers.
92
-------
TABLE 7-2. ORGANIC REMOVAL EFFICIENCY-CALCULATIONS FOR PLANT
NO. 1290-010 OF THE EPA ORGANIC DATA BASE
Bats of
Sampling
1979
nl»
Benzene
(M - 1190) Z
Nitxulionzene
(M - 17)
2-Nitropnenal
(H - 14.81
4-Nitropnenl
(H - 4KlO~3)
2.4-Nitraphenol
(H < 1.5)
Phenol
(H - 2)
10/2
10/3
10/5
10/9
10/10
10/11
10/13
10/16
10/17
10/18
.
10/19
10/23
10/24
10/25
10/26
10/30
10/31
11/1
11/2
11/6
11/7
11/8
11/9
11/13
11/15
11/16
11/20
11/21
11/22
11/23
11/27
11/29
11/30
» 99.7
> 99.8
> 99.8
> 99.5
(99. 9) 3
> 99.7
87.6
(97.2)
> 99.8
> 98.9
(>99.7)
> 99.2
> 99.5
(99.9)
. > 99.7
» 99.7
> 99.8
> 99.5
> 99.7
(>99.6)
> 99.4
> 98.7
(99.9)
48.6
> 99.3
(H)
> 99.4
> 99.0
(>98.S)
> 99.6
* 99.2
> 99.1
> 98.3
> 97.5
(>99.0)
> 99.0
-
(>24.4)
> 99
> 99.5
(>98.5)
-
( - )
> 98.8
> 99.3
99.9
99.9
99.9
85.0
(99.9)
98.8
99.9
(99.7)
99.8
99.9
(99.8)
99.9
66.7
(99.9)
99.9
99.0
99.9
99.9
99.9
(99.6)
99.9
98.8
(99.9)
85.6
90.7
(99.9)
99.7
99.8
(99.4)
99.9
99.9
99.9
99.9
99.9
(99.9)
99.9
99.9
(99.9)
-
99.9
(99.9)
99.9
(99.7)
99.4
52.4
89.9
91.4
87.5
80.0
(35.5)
- 86.5
69.9
(15.8)
- 15.9
21.7
I- 6.4)
44.1
44.6
(10.4)
82.1
89.4
87.2
22.2
7.1
(-86.6)
70.3
90.8
(-72.8)
34.8
77.7
(-42.9)
98.6
66.1
(-402.6)
44.9
87.0
54.5
99.8
67.7
(-120.8)
10.2
S2.0
(48.8)
-
85.7
(92.8)
59.8
(38.1)
47.3
79.1
> 99.2
9.2
43
86.4
> - 31.5
70
(-139.3)
- 4.2
28.1
(-101.1)
50.8
68.9
(78.2)
49.5
62.7
45.2
8.2
48
(46.9)
30.9
2.7
(-56.1)
- 40.7
81.0
(63.3)
17.0
38.7
(»>
81.3
44.7
27.6
81.0
91.5
(-250.0)
57.1
63.5
(69.1)
-
52.3
(-55.2)
89.7
(-146.7)
89.7
33.6
78.4
90.7
99.3
58.3
(99.9)
- 48.9
89
(90.9)
- 34.6
64.3
(-40.8)
14.8
30.0
(38.7)
45.9
17.9
7.1
33.0
92.4
(79.0)
16.0
- 8.7
(-38.3)
15.0
60.9
(29.6)
43.7
42.8
(89.0)
43.9
89.9
20.5
4.3
. 84.1
(50.6)
22.2
48.4
(99.0)
-
90.3
(-4647)
29.1
(-533.0)
55.8
67.3
96.3
60.8
99.9
94.5
(93.8)
71.2
> 99.9
O99.8)
57.1
- 40.7
(41.0)
- 145.4
85.5
(-46.3)
84.7
84.1
80.4
68.1
51.1
(-12.9)
69.3
89.7
(-595.4)
68.9
66.0
(-292.3)
> 9.1
> SO
099.5)
> 98.8
21.2
9.1
18.8
> 99.7
099.6)
99.9
> 99.9
(95.4)
-
> 99.6
(>74.2)
98.8
(79.8)
99.8
99.9
Notes t
1. efficiency n%
'Cone (in) - Cone (out)
Cone (in)
x 100
2. Henry Law Constant H « 100°C.
3. Values tn brackets represent analysis of the sane sanples by another laboratory
«. Neqaciva values indicate outflow cone, oreaeer Chan inflov cone.
4. Hien iteaa/feed ratio (25 or greater).
-------
TABLE 7-3 .< ORGANIC REMOVAL EFFICIENCY CALCULATIONS FOR PLANT NO. 2930-035
OF THE SPA ORGANIC DATA BASE
Data o*
34ffl&LmQ
1979
Unit
Ceapound
A.
' 8
c
0
z
r
e
B
I
J
K
I
H
•
0
V
Q
March
1
-134
88.2
iao
80
Hi
97.9
78.8
98.9
31.2
80.1
JO.O
96.4
-653
< 83
-S2.4
99.4
H10
HOtaat
1. CrUciancy n
2. Onie 1 • OHC
4
2
99.9
99.3
100
99.2
100
> 99.6
97.6
100
65.8
100
iao
100
100
98.2
99.3
100
100
% PCD*:
% L
n »
lurch
1
60. S
99.2
M10
83.4
(U
99.1
94.8
97.6
to.
94.6
810
99.6
-U96
100
97. S
99
99
(in) - Cone
Cone (u)
10
2
99.3
99. S
100
92.4
99.1
> 99.6
98.9
100
66.9
99.5
H10
100
87.5
HI
96
100
100
(cue) -I
J
March
1
-JS.7
HI
100
76.8
Ml
100
95.7
100
93.2
96
M10
100
-1722
H10
75.6
100
98.9
BO
11
2
99.8
100
100
93.2
99.2
100
99.8
100
64.2
99.7
H10
100
97.8
-
98.fi
100
100
natch
1
72.6
810
H10
HI
81
100
96.6
99.9
99.9
91.7
810
99.9
HI
H10
HI
96.6
100
12
2
99
100
100
-158
100
100
95
100
88
99
810
100
91
99
97
100
N10
.7
.3
.3
.9
.8
.2
Scrippar (pH rang* 1-2)
Onie 2 - Win Strippar 2000
continued
94
-------
Date of
Sapling
1979
Unit
Caapouad
B
0
E
r
G
B
I
J
L
H
B
0
P
g
DM* Of
Saopllag
1979
Unit
B
0
B
r
e
H
z
j.
L
H
0
P
0
No teat
1. £
nay 5
1 2
Nl HI
Ml 61.6
Nl 99.8
99.8 100
M10 92.2
99.8 99.9
MI as.i
98.5 99.9
100 100
-127 99.7
-
97.6 99.9
98.6 100
99.9 99.9
Kirch 17
1 2
HI 100
78.8 91.0
99.9 100
99.9 99.9
100 84.2
99.9 99.9
65.5 99.2
97.} 99.9
99.9 99.9
Ml 97.4
HI 99.9
99.9 99.6
100
nay 12
1 2
Nl Nl
Nl 66
Nl 99.3
> 99.7 100
N10 99.6
» 98.9 100
Nl HI
» 99.6 100
> 99.9 100
» -1.9 98.2
> 98.9 99.6
» 78.6 99.9
98.9 99.9
-
"•«* 24 Apri
121
MI MI mo
Ml Ml HI
HI 100 HI
99 100 99.5
HI Ml
99.4 HI 99
Nl U HI
96.7 99.9 99.6
99.9 99.9 99.2
1.68 99 HI
88.8 99.8 99.3
98.1 100 91
99.9 N10 100
1 2 April 8
212
HID H10 100
HI > 99.7 99.4
100 HI 100
100 100 100
98.9 ' HI 99.8
100 > 99.9 100
Nl > 99.9 99.4
100 100 100
99.9 100 100
99.9 HI 99.9
99.9 99.7 99.9
100 100 100
100 10O 10O
CritLinj-u n % • I"0"* "* '"
Cone (in) j - «~
2. Unit 1 • QHC Stripper (pH rang* 1-2)
Unit 2 • WRI Stripper (pH range 10-12)
3. HI • Not detected ia influent but detected in affluent
H10 • Hoc detected in influent and affluent
4. Negative value* indicate higher concentration in effluent than in influent
5. Naaa of the compound*: (« 100°e) Henry'« taw Conatant
1190
1167
798
265
801
862
250
1257
1244
901
1812
3596
1118
1600
> 2000
A •
B • Carbon Tetrachloride
C • Chloreaanxane
0 • 1.2-DicUonethana
t • 1.1.1-fricnlocoothene
P • 1.1-OichloraataaiM
6 • 1.1.2-firicaloroethane
H - Calaroetbene
Z • Chlorofon
J - 1.1-Oichloroetfaylane
K • Ethyl Benzene
L - Hethylene Cbloride
H » Tatraenloroetbylena
H - Toluene
O - Trichloroetaylene
P • Vinyl Chloride
Q • Cia-1.2-0ichloroectiylena
-------
TABLE 7-4. ORGANIC REMOVAL EFFICIENCY RESULTS OF THE WATER FACTORY DATA
Compound % Range
Chloroform 79-83
Dichlorobromomethane -
Chlorodibromomethane 82
Chlorobenzene 96
1-2 Dichlorbenzene 88
1-3 Dichlorbenzene 83
1-4 Dichlorbenzene 92-97
1-1-1 Trichloroethane 91
Tetrachlorethylene 95
96
-------
7.4 Development of a Correlation for the Effect of Liguid-diffusivity
on Tower Sizing and Organic Removal Efficiency.
The effect of diffusivity was demonstrated by considering the example
of a continuous packed bed tower stripping organics from aqueous waste
stream.
The height of a packed bed (Z) can be determined from the knowledge
of two factors: the height of a transfer unit (HTU) and the number of
transfer units (NTU)
Z » HTU y NTU (1)
The MTU is defined in terms of an integral of influent and effluent
concentrations and this integral equation can be solved analytically for
dilute solutions and solutes obeying Henry's Law to give the following
expression
NTO - - t-r— a I -=*- u - ±* + T I (2)
where S = stripping factor
- Ha G/PtL
X. = mole fraction in influent
X » mole fraction in effluent
o
We have plotted the NTU as a function of stripping factor for
various removal efficiencies ranging from 60 percent to 99.99 percent.
The results are illustrated in Figure 7.1. An important observation
made from this figure is that for compounds with high Henry's Law
Constant and consequently high stripping factor ( > 50), the number of
transfer units is nearly independent of the magnitude of Henry's Law
Constant for any given removal efficiency. Thus, it can be stated that
for compounds with high Henry's Law Constant,
NTU * (removal efficiency alone) (3)
97
-------
15
10 D
I*]
NTU = No. of transfer units
S » Stripping factor
n = Removal efficiency
(Note: At S>50, NTU / f(S) for a given n)
H=99.99%
a n Q n = 99.9%
A
A
— A
A A A n = 99%
A
AA
A A A n = 90%
o o o o n = 80%
• • • • • n = 60%
i = LVI 1
0 . 50 100 1000 5000
Figure 7-1. Dependency of NTU on stripping factor and removal efficiency.
-------
and therefore, for a given removal efficiency
Z « HTD (4)
The height of a transfer unit is defined as
HOT = L/T^ a CQ (5)
When a solute has a large Henry's Law Constant, as do slightly soluble gases
or volatile liquids, the overall transfer coefficient K. depends primarily
• •*
on the local transfer coefficient (k_), and therefore
HOT = L/k, a C_
L o
For a given packing and loading rate L, a and CQ are the same for any compound
in the mixture, and therefore
HOT « 1AL t6)
A typical empirical correlation used for liquid-phase mass transfer
coefficients in towers containing randomly packed materials is given by
Perry' kLa / L \ 1~n / WL \ °*5
All the factors in this expression, except the diffusivity D^, are the same
for any organic in the mixture which means that
B^»
k (7)
Combining Equations 4, 6 and 7 we arrive at the final correlation:
In other words, for a given removal efficiency and for compounds with high
Henry's Law Constant, the height of a packed bed is inversely proportional
99
-------
to the square root of the component's liquid diffusivity in dilute aqueous
solutions.
For example, if there are two compounds in a solution with diffusivities
0 and D , and the height of packed bed required is Z , and Z respectively,
then the relationship between the ratios D../D- and Z_/Z- will be as shown in
Figure 7.2.
100
-------
DR(1:2) = Ratio of Diffusivitioa (Compound I/Compound 2)
ZR(2:1) = Ratio of Heights for any removal'efficiency and
stripping factor greater than 50
(Compound 2/Compound 1)
2.0
n 1.5
1.0
i.o
_L
2.0 3.0
DR(1:2)
4.0
Figure 7-2. The effect of liquid diffusivity on the height of a packed bed.
-------
REFERENCES
1. Lewis, G.N. and Randall, M. , Thermodynamics, McGraw-Hill 1923.
2. Gilliland, E.R., Robinson and Gilliland's Elements of Fractional
Distillation, McGraw-Hill 1950.
3. Reid, R.C., Prausnitz, J.M. and Sherwood, T.K., The Properties of
Gases and Liquids, 3rd Edition, McGraw-Hill Book Company, New York,
1977. -v
4. Hwang, S.T. and Fahrenthold, P., "Treatability of the Organic
Priority Pollutants by Steam Stripping," pages 37-60 in Water-1979,
AICHE Symposium Series 197, Vol. 76, A1CHE, N.Y. 1980.
5. Kavanaugh, M.C. and Trussell, R.R.., "Design of Aeration Towers to
Strip volatile Contaminants from Drinking Water," Journal AWWA, p.
684-692, December 1980.
6. Dilling, W.L., Tefortiller, N.B. and Kallos, G.J., "Evaporation
Rates and Reactivities of Methylene Chloride, Chloroform, 1,1,1 -
Trichlorethylene, Tetrachloroethylene and Other Chlorinated Compounds
in Dilute Aqueous Solutions," Environmental Sci. and Tech. 9_ (9)
833-838, September 1975.
7. Hicks, R.E., et al, "Wastewater Treatment in Coal Conversion," U.S.
EPA Report 600/7-79-133, NTXS Catalog PB-297-587/8WE June 1979.'
8. Weller, R., Schuberth, H. and Leibnitz, £., "The Vapor LiquidoPhase
Equilibria for the System Phenol/n-Butylacetate/Water at 44.4 C,
"Journal fur Praktisehe Chemie, 21, 234, 1963.
9. Horsley, L.H., "Azeotrope Data III," Advances in Chemistry Series
116. American Chemical Society, Washington, D.C. 1973.
10. Rhodes, F.H., Wells, J.H.* and Murray, G.W., "Vapor Composition
Relationships in the Systems Phenol-water and Phenol-cresol," Ind.
Snq. Chem. 17 (11), 1199, 1925.
11. Bogart, M.J.P. and Brunjes, A.S., "Distillation of Phenolic Brines,"
Chem. Eng. Prog. 44 (2), 95, 1948.
12. Gaehling, J. and Onken, a., "Vapor-Liquid Equilibrium Data Collection,'
Dechema Ch
-------
15. U.S. EPA "Treatability Manual, Volume 1, Treatability Data," EPA
600/8-80-042a, July 1980.
16. Perry, R.H. Chilton, C.H. and Kirkpatrick, S.P., "Chemical Engineer's
Handbook," 4th ed. McGraw-Hill, 1950.
17. Edwards, T.J., Newman, J. and Prausnitz, J.M., "Thermodynamics of
Aqueous Solutions Containing Volatile Weak Electrolytes," AICHE Journal
21.(2) 248-259, March 1975.
18. Edwards, T.J., et al "Vapor Liquid Equilibria in Multicomponent
Aqueous Solutions of Volatile Weak Electrolytes," AICHE Journal
24(6) 966-976, November 1978.
19. American Petroleum Institute, "A New Correlation of NH, CO. and
H_S Volatility Data from Aqueous Sour Water Systems, API Pub. 955,
API, Washington, D.C., March 1978.
20. Warner, H.P., Cohen/ J.M. and Ireland, J.C., "Determination of
Henry's Law Constants of Selected Priority Pollutants" Municipal
Environmental Research Laboratory, EPA, Cincinnati, Ohio 45268.
21. Tsonopoulos, C. and Prausnitz, J.M., "Activity Coefficients of
Aromatic Solutes in Dilute Aqueous Solutions," Ind. Eng. Chem. Fundamentals
10 (No. 4) 593-600, 1971.
22. Verschueren, K., Handbook of Environmental Data On Organic Chemicals,
Van Nostrand Reinhold 1977.
23. Handbook of Lab Safety, 2nd ed. Chemical Rubber Company 1971.
24. Handbook of Chemistry and Physics 55th ed. 1974 CRC Press, p. D2-
D36.
• _
25. Glasstone, S. Physical Chemistry, MacMillan, 1948.
26. National Academy of Science "Polychlorinated Biphenyls," Washington,
D.C. 1979.
27. Rossi, S.S. and Thomas, W.H., "Solubility Behavior of Three Aromatic
Hydrocarbons in Distilled Water and Natural Sea Water," Environmental
Sci. and Tech. 15(6) 715 to 716, June 1981.
28. Weber, R.C., Parker, P.A. and Bowser, M., "Vapor Pressure Distribu-
tion of Selected Organic Chemicals" Report EPA-600/2-81-021, EPA
IERL Cincinnati, February 1981.
103
-------
29. U.S. EPA "Aldrin/Dieldrin - Ambient Water Quality Criteria," 1979,
NTIS Catalog PB 297-616 also - similar,
"Chlordane, PB 292-425
"DDT," PB 297-923
"Endosulfan," PB 296-783
"Endrin," PB 246-785
"Heptachlor," PB 292-434
"Hexachlorocyclohexane," (BHC) PB 297-924
"Toxaphene," PB 296-806
"Haloethers." PB 296-796
"Phathalate Esters," PB 296-804
"Benzidene," PB 297-918
30. Mackay, D., Shiu, W.x. and Sutherland, R.P., "Determination of Air-
Water Henry's Law Constants for Hydrophobia Pollutants," Environmental Sci.
and Tech. 13(3), 333-337, March 1979.
31. Mackay, D. and Leinonen, P.J. "Rate of Evaporation of Low-Solubility
Contaminants from Hater Bodies to Atmosphere". Environmental
Sci. and Tech. 9.U-3) 1178-1180, December 1975.
32. Neerbon, R.F., "Compressor Selection for the Chemical Process
Industry," Chemical Engineering 78-94, January 20, 1975.
33. Wescott, J.W., Simon, C.G. and Bidleman, T.F., "Determination of
Polychlorinated Biphenyl Vapor Pressure by a Semi-micro Gas Saturation
Method," Environmental Sci. and Tech. 15_(11) 1375-1378, November
1981.
34. King, C.J. Separation Processes McGraw-Hill 1971.
35. Morris, G.A. and Jackson, J. Absorption Towers 1st ed. 1953,
reprinted 1974 by Krofchak Ltd., Cambridge, Ontario.
36. Coulson, J.M. and Richardson, J.F., Chemical Engineering, McGraw-
Hill, 1955.
37. Touhill, Shuckrow & Associates, Inc. "Concentration Technologies
for Hazardous Aqueous Waste Treatment," U.S. EPA-600/2-81-019,
February 1981, NTIS Catalog PB 81-150583.
38. Trulsson, S.G., "Design of Adsorption/Desorption Towers, Mass.
Transfer and Pressure Drop Data for Plastic Packings," Dissertation,
Dept. of Chemical Engineering, Lund Institute of technology, Lund,
Sweden, March 1979 (lent to the authors by Munters Corp.) . ..
&
39. Smith, J.K. et al "Laboratory Studies of Priority Pollutant ^reatability,"
Project Summary of EPA-600/52-81-129 Sept. 1981.
40. Hsu, Jackson, Matouk, Rudorfer and Zirps, "Determination of Henry's
Law Constant for Toxic Pollutants," Report for Chemical Engineering
Laboratory Project 10-26, Massachusetts Institute of Technology,
December 1980.
104
-------
41. Stephen, H. and Stephen, T., eds. Solubility of inorganic and Organic
Compounds, Pergamon Press, London, 1963.
42. Leighton, D.T., and Calo, J.M., "Distribution Coefficients of Chlorinated
Hydrocarbons in Dilute Air-Water Systems for Groundwater Contamination
Applications", J. Chem. Sng. Data 26 382-385, 1981.
43. Leinonen, P.J. and Mackay, D., "The Multi-component Solubility of
Hydrocarbons in Water", Canadian J. of Chem. Enq. 51 230-233, 1973.
44. Smith, J.M. and Van Ness, H.C., Introduction to Chemical Engineering
Thermodynamics, McGraw-Hill, 197TI
105
------- |