United States
Environmental Protection
Agency
Robert S. Kerr Environmental Research
Laboratory
Ada OK 74820
EPA-600/2-80-067
April 1980
Research and Development
&EPA
A New Correlation of
NH3, COa, and H2S
Volatility Data from
Aqueous Sour Water
Systems
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1 Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
' 4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-80-067
April 1980
A NEW CORRELATION OF NH,, C09, AND H9S
VOLATILITY DATA FROM AQUEOUS JSOUITWATER SYSTEMS
by
Grant M. Wilson
Thermochemical Institute
and Chemical Engineering Department
Brigham Young University
Provo, Utah 84602
EPA Grant no. R804364010
Project Officer
Fred fL Pfeffer
Source Management Branch
Robert S. Kerr Environmental Research Laboratory
Ada, Oklahoma 74820
Sponsored by the
American Petroleum Institute
Committee on Refinery Environmental Control
ROBERT S. KERR ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ADA, OKLAHOMA 74820
-------
DISCLAIMER
This report has been reviewed by the Robert S. Kerr Environmental
Research Laboratory, U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily reflect
the views and policies of the U.S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
-------
FOREWORD
The Environmental Protection Agency was established to coordinate admin-
istration of the major Federal programs designed to >protect the quality of our
environment.
An important part of the Agency's effort involves the search for infor-
mation about environmental problems, management techniques and new technologies
through which optimum use of the nation's land and water resources can be
assured and the threat pollution poses to the welfare of the American people
can be minimized.
EPA's Office of Research and Development conducts this search through a
nationwide network of research facilities.
As one of these facilities, the Robert S. Kerr-Environmental Research
Laboratory is responsible for the management of programs to: (a) investigate
the nature, transport, fate and management of pollutants in ground water;
(b) develop and demonstrate methods for treating wastewaters with soil and
other natural systems; (c) develop and demonstrate pollution control tech-
nologies for irrigation return flows; (d) develop and demonstrate pollution
control technologies for animal production wastes; (e) develop and demonstrate
technologies to prevent, control, or abate pollution from the petroleum re-
fining and petrochemical industries; and (f) develop and demonstrate technolo-
gies to manage pollution resulting from combinations of industrial wastewaters
or industrial/municipal wastewaters.
The use of inplant processes to remove undesirable components of a
wastewater stream prior to discharge to a wastewater treatment plant can often
effect significant improvements in treatment plant effluent quality. This
report contains the findings of a study to utilize new correlations between
sour water constituents so as to improve the ammonia removal efficiency of
sour water scrubbers in petroleum refineries.
W. C. Galegar
Director
Robert S. Kerr Environmental Research Laboratory
111
-------
ABSTRACT
A new correlation model has been developed for calculating sour water
equilibrium data at temperatures from 20 C to 140 C. The correlating
equations in this new sour water equilibrium model (SWEQ) have been used to
obtain a computer program capable of handling the various chemical and physical
equilibria of NH~, C0?, and H?S in sour water systems including the effects
of carboxylic acfds on ammonia (NhL), Carbon Dioxide (C02), and Hydrogen
Sulfide (HpS) in sour water systems including the effects of carboxylic acids
on ammonia fixation and release by caustic addition.
This new SNEQ correlation model has been used to evaluate published and
new vapor-liquid equilibrium data, and comparisons are made with the Van
Krevelen prediction equations as published by Van Krevelen. Average errors
between calculated and measured partial pressure data can be summarized.
Both models predict low temperature data quite well, but at high temp-
eratures the Van Krevelen model deviates considerably from measured data,
and errors between the SWEQ model and measured data increase from about 11%
to about 29%. Comparisons with variations of the Van Krevelen model as
published by other authors have not been made.
The basic NFL-FLS-H^O equilibrium program has been inserted into a tray
by tray program by CoNOCQ. Two brief example problems have been run to date.
The calculated stream requirements appear to be approximately 30 percent
greater for a refluxed tower and 20 percent more for a non-refluxed unit com-
pared to Van Krevelen - Beychok procedures. Definite conclusions cannot be
drawn until wider user experience is obtained.
Details of the SWEQ correlation model, correlating equations, the com-
puter program, and evaluations of experimental data are given in this report.
This report covers a period from March 15,"1976, to March 17, 1977, and work
was completed as of November 30, 1977.
IV
-------
CONTENTS
Foreword . . iii
Abstract 1y
Figures vi
Tables vii
Abbreviations and Symbols ...
1. Introduction 1
2. Project Objectives. , 2
3. The SWEQ Model 3
4. Computer Program Based on the SWEQ Model 20
5. Sample Problem Using the SWEQ Model 51
6. Comparisons and Evaluations Between Calculated and
Measured Data 59
Evaluation of Van Krevelen Model 59
Evaluation of SWEQ Model 75
Evaluation of New BYU Data 85
Ammonia Fixation by Acids and Release by Caustic
Addition 86
7. Accuracy of Correlation 91
8. Summary 92
References 94
Appendix 97
-------
FIGURES
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Sample plot of the volatility of NH3> C02, and H2S versus
Flow diagram of option 3 of SWEQ computer program ......
Ammonia mean ratio of measured over calculated partial
Carbon dioxide mean ratio of measured over calculated partial
pressures based on SWEQ correlation
Hydrogen sulfide mean ratio of measured over calculated
partial pressure based on SWEQ correlation
Ammonia mean ratio of measured over calculated partial
pressures based on Van Krevelen correlation
Carbon dioxide mean ratio of measured over calculated partial
pressures based on Van Krevelen correlation
Hydrogen sulfide mean ratio of measured over calculated
partial pressures based on Van Krevelen correlation ....
Free ammonia versus pH adjustment by caustic addition at 25°C
Free ammonia versus pH adjustment by caustic addition at 80°C
Sample plot of pH versus caustic addition showing variation of
oH at 25 C and at column temperature
Page
19
23
24
25
26
79
80
81
82
83
84
87
88
90
VI
-------
TABLES
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Summary of Equations Used to Calculate Temperature and
Composition Effects on Henry's Law Relations '
Summary of Chemical Equilibria Involved in Calculating
NH3-C02-H2S-H20 Vapor-liquid Equilibria
Effect of Composition and Ionic Strength on Chemical
Equilibrium Constants '
Summary of Temperature Parameters, Used to Calculate Chemical
Equilibrium Constants in Table 1
Subroutine KREAC
Subroutine HENRY
Subroutine YFX
Subroutine CFX
Subroutine PHST
Subroutine SPECV
Subroutine NPH ,
Subroutine SPECL
Subroutine PRESY
Subroutine NTEMP
Input data for Sample Problem with SWEQ Computer Program . . .
Computer Output from Data in Table 16 with Computer Program
Based on the SWEQ Model
Page
6
12
14
17
27
32
33
33
34
34
35
37
38
39
39
48
50
vii
-------
Number Page
IS H?S-hLO, NH^-H?S-H?0, and NH--H20 Systems, Comparison of
Calculated and Measured7Data of Miles and Wilson and
of Clifford and Hunter17 60
19 NH3-H?S-H?0 System, Comparison of Calculated and Measured
Data of Terres29 61
20 NH2-H?S-H?0 System, Comparison of Calculated and Measured
Data of Van Krevelen, et al.1 62
21 H?S in Aqueous Buffer Solutions, Comparison of Calculated
and Measured Data of Shih, et al.2B 63
22 NH3-C02-H20 and C02-H20 System, Comparison of Calculated
and Measured Data of Van Krevelen, et al J and Data From
Lange's Handbook20 64
1--CO?-H?0 System, Comparison of Calculated and Measured
Data of Otsaka, et al.22
23 NH^-C
65
24 NH^-COp-HpS-HpO System, Comparison of Calculated and
Measures Data of Cardon and Wilson 66
25 NHo-C02-H2S-H20 System, Comparison of Calculated and
Measurea Data of Badger and Silver^ 67
26 NH-j-hLO System, Comparison of Calculated and Measured
Data of Breitenbach and Permanl6,23 68
27 NH.,-C02-H2S-H20 System, Comparison of Calculated and
Measurea Data of Van Krevelen, et al.' 69
28 Summary of References of Experimental Data 70
29 Summary of Deviation Errors Between Calculated and
Measured Ammonia Partial Pressures 72
30 Summary of Deviation Errors Between Calculated and
Measured Carbon Dioxide Partial Pressures 73
31 Summary of Deviation Errors Between Calculated and
Measured Hydrogen Sulfide Partial Pressures 74
Al Computer Program Used for Calculating Vapor-Liquid
Equilibrium Data From the Van Krevelen Correlation . . 99
vm
-------
SECTION 1
INTRODUCTION
Previous design calculations of vapor-liquid equilibrium compositions
in sour water strippers have primarily been based on a correlation by Van
Krevelen (1) as outlined in Aqueous Wastes from Petroleum and Petrochemical
Plants by M.R. Beychok (2). The Van Krevelen correlation has proved suffi-
ciently reliable and many sour water strippers have been designed and built
using his correlation as a basis. New vapor-liquid equilibrium measurements
have been made since Van Krevelen's correlation published in 1949 including
new measurements at Brigham Young University sponsored by the API Technical
Data Committee. Although used considerably, the Van Krevelen correlation has
been previously recognized to be deficient in the following areas:
1. Only data to 60°C were correlated; thus the use of the correlation
at sour-water stripper temperatures of 100 to 120 C represented an
extrapolation of existing data.
2. The calculation method outlined by Van Krevelen did not allow for
mixtures containing ammonia over hydrogen sulfide ratios less than
1.5 in the liquid phase.
3. The calculation did not take into account reduced volatilities of
hydrogen sulfide and ammonia at low parts per million concentrations
due to the ionization constants of the two compounds in water.
Subsequent sections of this report give details of a new sour water
equilibrium model (SWEQ) which is based on new higher temperature data and
which avoids deficiencies mentioned above. This new correlation model also
permits the addition of caustic for release of NH3 held by carboxylic
acids or stronger acids.
-------
SECTION 2
PROJECT OBJECTIVES
The development of a new correlation for ammonia, carbon dioxide, and
hydrogen sulfide volatilities from aqueous sour water systems has required
the completion of the following project objectives:
1. Compare new NhL-HpS-HpO experimental vapor-liquid equilibrium data
developed by Bfignam Young University with previously published data
by Van Krevelen.
2. Check and "fine tune" (if necessary) the new vapor-liquid equilibrium
equations developed by Brigham Young University to the measured
experimental data.
3. Compare BYU equations to equilibrium expressions previously published
by Van Krevelen and Beychok.
4. Modify the BYU equilibrium equations to allow calculations with or
without external pH adjustment (i.e., using caustic).
5. Modify the existing BYU computer program to allow equilibrium
calculations with or without adjustment.
These objectives have been achieved by first developing a correlation
model in which literature data of Van Krevelen, new BYU data, and other liter-
ature data have been used to develop equations capable of predicting data
over wide ranges in concentration and temperature. Based on these equations
a new sour water equilibrium computer program has been developed which is
capable of handling the various chemical and physical equilibria of sour
water systems including the effects of carboxylic acids or stronger acids on
fixation and release by caustic addition.
This new sour water equilibrium correlation has now been used to evaluate
published and new vapor-liquid equilibrium data. Details of the correlating
equations, computer program, and data evaluations are given in subsequent
sections of this report.
-------
SECTION 3
THE SWEQ MODEL
The SWEQ correlation model developed from this project is very similar
to the model used by Van Krevelen (1) except that some of the limitations
imposed by that model have been removed. Van Krevelen assumed that FLS and
COo only exist in aqueous solutions as ionized species. This is virtually
true at concentrations where NHL is in excess, but such an assumption would
not be true when these acid gases are present in the absence of NH- or other
basic components. The method used here, therefore, avoids this problem by
considering the chemical equilibrium between ionic species of H?S or C02 and
undissociated H9S or C09 in the liquid as follows. ,
- + k - (HS"} (" } (1)
H2SU) "• HS +H (H2S) UJ
H2C03(£) + HC03- + H+ k - 3- (2)
The SWEQ model not not take into consideration the equilibrium between dissolved
C02 and carbonic acid (H2C03) according to the following reaction
C02 + H20 * HC03 (3)
because the presence of other acidic or basic component does not affect this
equilibrium. This reaction is apparently slow enough that the kinetics of
absorption of C02 into basic aqueous solutions is slower than for H2S. In
spite of this slower reaction rate, the assumption is made here that sufficient
contact time or catalyst is used to achieve chemical equilibrium. ' By this
method, the partial pressure of H2S or C02 in the vapor phase above a solution
can be calculated from the concentrations of the undissociated species as
follows.
'Because of the slower absorption of C02 into water and because of the
possibly slow conversion of bicarbonate ion to carbonate ion by excess ammonia,
a warning is given that actual plate efficiencies could be low compared to
expected efficiencies when C02 is present^
-------
PH2S = H^
PC02 = HC02 CH2C03 (5)
where
H2S> PC02 = partial pressure of H2$ or C02
HH2S, HCQ = Henry's constants for H2$ and C02
H2S> CH,,CO- = liquid phase concentrations of H2S and
6 H2C03' mo1es/K9 of solution
The Henry's constant used here must apply at finite concentrations as well
as infinitely dilute concentrations, so, in general, Hn2S and Hco2 become
dependent on the composition of the solution. This method of calculating
HpS and CCL partial pressures is analogous to Van Krevelen's method for
calculating ammonia partial pressures which a composition dependent Henry's
constant is used. The addition of Henry's constants and undissociated H2$
or H^CO., species concentrations makes possible the calculation of vapor-
liqufd equilibria at acid gas concentrations in excess of ammonia or of other
basic components; thus the Van Krevelen restriction to compositions with
excess ammonia is avoided.
This method for calculating vapor-liquid equilibrium data under condi-
tions of simultaneous chemical equilibrium requires two properties that
must be correlated in terms of analytical equations as follows.
1. Analytical equations for the effect of temperature and composition
on Henry's Law constants so that component partial pressures in
the vapor phase can be calculated from calculated concentrations
of undissociated NH3, C02, and H2S in the liquid phase.
2. Analytical equations for the effect of temperature and composition
on chemical equilibrium constants so that the concentrations of
undissociated NH3> C02, and H2$ in the liquid can be calculated.
Rather than do an exhaustive recorrelation of existing literature data
for these properties, an attempt has been made to use existing correlations
where possible. Modifications to these existing correlations have been made
when necessary to improve the representation of multicomponent data studied
in this project. Fortunately, the Henry's constants for NH3, C02, and H?S
can be based primarily on binary data in water. This simplifies the corre-
lation because these properties are fairly well known. Multi-component
vapor-liquid data thus serve primarily to establish the effects of high
concentrations of the various compounds in solution on these Henry's constants,
-------
By this method, the Henry's constants for ammonia and carbon dioxide at low
concentrations of each compound have been taken directly from the literature.
Van Krevelen's correlation was made in terms of component concentrations
in moles per liter (£") of solution or molarity. This method introduces an
unnecessary variable which is the density of the solution. This occurs
because the density is needed to calculate the molarity when the number of
moles or number of pounds or grams of each component in a mixture are speci-
fied. The SWEQ model avoids this problem by using concentrations in moles/Kg
of solution. At low concentrations of the solutes the density of the solu-
tion is about one, so the low concentration parameters of Van "Krevelen's
correlation still apply. However at conditions where the density deviates
significantly from unity, then parameters in the two correlations cannot be
directly compared. At these conditions, the parameters in the SWEQ model
have been determined by directly fitting available phase equilibrium data
using concentrations in moles/Kg of solution. By this method there is no
ambiguity in the correlation because concentrations in moles/Kg of solution
have only been used in the correlation, and the method avoids the need for
density at the various concentrations and temperatures of the correlation.
At low concentrations of the components, published Henry's constants and
chemical equilibrium constants have been used in units of moles/Kg of water
because the two sets of units are the same at the zero concentration unit.
Table 1 summarizes the various equations used in the SWEQ model for
calculating Henry's constants for NH3, C02, and H2S. The Henry's constant
for ammonia at low ammonia concentrations has been taken directly from the
equation of Edwards, Newman, and Prausnitz (3) rather than from Van Krevelen
because their correlation is more recent and includes data which was not
available to Van Krevelen. Exisiting literature data for the volatility of
ammonia from aqueous solutions scatters considerably, but the equation of
Edwards ert aj_ appears to correlate the data of greated precision.
The Van Krevelen model does not require Henry's constants for CCL and
HpS, so these have been obtained from another source. Kent and Eisenoerg
have recently published correlations (4) on H2S and CCL partial pressures
from aqueous monoethanol amine and diethanol amine solutions which appear
to correlate these systems quite well. In their correlation they adjusted
the amine equilibrium constant for reaction with hydrogen ions to obtain
agreement with published data on H2S and CCL partial pressures. By this
method they obtained a model capable of accurately predicting equilibrium
in HoS-CCL-amine systems. Their equations for the Henry's constant for CCL
has Been used without any changes as it is given in Table.!. Their Henry's
constant for H?S however was increased about 12% in order to improve the
represenation of multicomponent data by a change in the first constant as
noted at the bottom of Table 1.
The use of Henry's constants to correlate volatility data introduces
two methods for calculating concentration effects. One method is to assume
that the Henry's constant varies with the concentration of the various com-
pounds in solution, and the other method is to assume the various compounds
in solution. In some cases, the choice of a concentration parameter in the
Henry's constant or of using a concentration parameter in the equilibrium
5
-------
Water
TABLE 1. SUMMARY OF EQUATIONS USED TO CALCULATE TEMPERATURE
AND COMPOSITION EFFECTS ON HENRY'S LAW RELATIONS*)
Compound
Ammonia
Carbon
Dioxide
Hydrogen
Sulfide
Lit.
Ref.
3
4
4
Fortran
Symbol
HA
HC
HS
Equation '
ln(HA) « 178.339 - 15517. 91/T - 25.6767
+ (.06)(2CC + CS)
ln(HC) = 18.33 - 24895. 1/T + .223996 X
ln(HS) * 100.684* - 246254/T + 2.39029
ln(T) +
108/T2 -
X 108/T2
.019660T + (131.4/T
.090918 X 101]/T3 +
- 1.01898 X 1011/!3
- .1682) (CAS)
.12601 X 1013/T4
HW
ln(HW)
+ 1.59734 X 1013/T4 - .05(CAS) + (.965 - 486/T)(CC)
c^ = 14.466 - 6996.6/(T-77.67)
"' T » temperature in °R
CAS » free ammonia concentration, gram-moles/Kg of solution.
CC a total C02 1n solution, gram moles/Kg of solution.
CS • total HgS in solution, gram-moles/Kg of solution.
' Henry's constant in psia/(gram-moles/Kg of solution).
c' Water vapor pressure in psla; the partial pressure in water 1s calculated from Raoult's Law
Constant adjusted from 100.573 to 100.684 in order to fit new H2S solubility data; and
multicomponent NH,-CO«-H9S-H«0 data.
-------
constant has been arbitrary. A summary of methods used in the SWEQ model is
given in the following.
Compound
Affected
NH
C0
H2S
Concentration Effects on
Volatility Data Correlated by
Henry's Const. Equil. Const.
free NH
absorbed C02
and H2S
free NH
absorbed H2S
ionic strength
absorbed C02 absorbed C02
Principal Data
Correlated
NH3-H20
H2S-C02-NH3-H20
H2$-C02-NH3-H20
C02-NH3-H20
H2S-NH3-H20
H2S-C02-NH3-H20
Van Krevelen used a Henry's constant for ammomia which he assumed to
be only dependent on free ammonia concentration. Additional effects of ab-
sorbed H2S and CCL were found necessary in the SWEQ model in order to corre-
late more recent RpS-COp-NHo^O, so any concentration effects for these
compounds in his model correlated in the equilibrium constant.
The effects of free ammonia, and of absorbed H2S or CCL on the Henry's
constants used in the SWEQ model are given in Table 1. In this table, the
Henry's constant of ammonia is proportional to a constant times CAS, free
NH3, and to a constant times'(2 CC + CS), absorbed C02 and H2S. No concen-
tration effects were introduced in the SWEQ model on the Henry's constant of
C0?, but effects for free ammonia and absorbed C00 were introduced to corre-
late H9S volatility data as shown by terms proportional to CAS and CC in
Table f.
An equation for water is also given in Table 1. Water generally exists
as the principal component even in concentrated solutions of electrolytes so
that liquid-phase non-ideality effects on the partial pressure of water are
small. For this reason, the partial pressure of water in the vapor phase can
be calculated from Raoult's Law where the moles of each ionized and unionized
species in solution is considered in calculating the mole fraction of water.
The partial pressure of water is then calculated from its vapor pressure
according to the following equation.
= p
H20
H0
(6)
-------
where
o
p H20 = vapor pressure of water
XH20 = liquid phase mole fraction of water
By this method, the constants for water in Table 1 are simply the vapor
pressure of water fitted over the range from 25°C to 1500C from data in the
steam tables.
No attempt has been made in the SWEQ model to correct for non-ideal
behavior in the vapor phase. At low pressures, errors from assuming ideality
are probably less than +_ 5%; but at pressures of 50 psia or higher, the
errors will be greater than this and serious consideration should be made
to correct for non-ideal in the vapor phase.
Besides Henry's constants, one must correlate the chemical equilibria
of reactions occurring in the liquid phase as mentioned above. The Van Kre-
velen correlation is limited because the effects of other acidic or basic
components cannot be readily taken into account. This problem is avoided
in the SWEQ model by assuming that the various chemical equilibria are depen-
dent on the concentrations of either the ionized or undissociated species of
a component and the hydrogen ion concentration. For an acid, the general
form of the equilibrium equation is as follows:
AH + A" + H+ k = (
while for a base the equilibrium can be written as follows:
BOH + H+ -> H+ * B+ + H,0 . k = ID(^]U+, (8)
In principle, the assumption of equilibria according to these equations makes
possible the calculation of the equilibrium species concentration of each
component knowing only the total concentration of that component and the pH.
If the pH is not known it can be calculated by trial and error until electrical
neutrality is achieved in a given mixture of compounds. This method of
calculation permits the development of generalized calculation methods, so
that new compounds can be added as necessary. In many respects the method
is similar to an equilibrium flash calculation where the feed composition and
equilibrium K-values of individual components are known. In a flash calcu-
lation, the concentrations of each component in the vapor and liquid phase
is known. But in general, the fraction as vapor or liquid is not kftown so
an iterative calculation is made until the concentrations in each phase add
to 100%. For acid-base equilibria, the problem is nearly as simple except
that the iteration parameter is pH instead of fraction as vapor or liquid.
The picture for H^S and COp is slightly more complicated because both com-
ponents have second ionization constants so that two chemical reactions must
be simultaneously solved at a given pH value. In this case, a calculation
example is given as follows.
8
-------
AH" + H+ k = AH (9)
AH -> A + H k2 = (AH-} (10)
To solve these equations, it is assumed that the total concentration
of both ionized plus undissociated species concentrations is known by
chemical analysis; but that the concentration of each individual species is
not known. In this case, the concentrations of individual species can be
related by the following equations.
(AH2) = nA -a-3 (11)
(AH~) = a (12)
(A=) = 6 (13)
From these equations, the following equations are obtained for k-, and k2.
k _ >)(H+j
kl ~ (nT-a-B) (14)
2 (a)
These simultaneous equations can be algebraically solved for a and 3 to
give the following equations.
k1k2/(H+) (16)
If no second ionization occurs, then k2 is zero and a becomes as follows*
a =
k] (18)
After a and 3 have been calculated from equations 16 and 17, then the con-
centration of undissociated species can be calculated from equation 11.
Because of computer round-off error due to subtracting two large numbers to
get a small number, it has been found better to calculate the undissociated
species concentration from equation 14 instead of equation 11. This is done
by rearranging equation 14 to the following equation in which the round-off
error ia avoided.
nA-a-3= (AH2) = (
-------
The calculation of chemical equilbria in mixtures containing both
ammonia and carbon dioxide requires allowance for the reaction of bicarbamate
ion with free ammonia to produce carbamatc ion as follows.
HC03" + NH3 -> H2NCOO" + H20
i, - (MCOQ")
K " (HC03 )("H3}
This introduces a third simultaneous reaction for CCL and a second simultan-
eous reaction for NH~. This added complexity makes necessary a second itera
tive calculation procedure to calculate individual species concentrations at
a specified pH value. This calculation is made by assuming various bicarbo-
ate concentrations from which the concentrations of the other species can be
algebraically calculated. The resulting concentrations of individual CCL
species are then added compared with the specified moles of C02 in the solu-
tion as follows.
n
CO,
= (C02) + (KC03~) + (C03~) + (H2NCOO~) (21)
'Z(calc)
The amount of carbonate is then adjusted up or down by the following ratio.
(nrn )
(HC03 ^ = (HC0~> 2 actua1
old
Fortunately, this iteration method appears to converge after only town or
three iterations.
This discussion of chemical equilibria involving HLS, C02, NH-, and
water outlines the details of various steps used in the SWEQ model to calcu-
late the concentration of each individual species in solution. Table 2 gives
a summary of the various reactions which are accounted for by the model.
There are a total of eight reactions listed in this table. First ionization
constants are involved in reactions 1, 3, 5, 7, and 8; and second ionization
constants are involved in reactions 2 and 6. In addition, bicarbonate ions
react with ammonia to produce carbanate ions in reaction 4. The corresponding
equilibrium equations based on the extent of each reaction occurring are
given to the right for each reaction in Table 2. Except for reaction number
4 for carbamate formation, the equilibrium concentrations of each species
are shown to be proportional to the pH of the solution. If the pH is known,
it becomes a rather easy matter to compute the equilibrium concentration of
each species in solution. If the pH is not known, an iterative method has
to be devised as discussed above whereby an initial pH is assumed. Then as
steps in the iteration loop, the concentration- of each species is calculated
by calculating the extent of each chemical reaction. From the calculated
species concentrations the sum of all electronic charges HT can then be cal-
10
-------
culated by the equation shown at the bottom of Table 2. Generally this sum
will not be zero, but the assumed pH can then be adjusted to bring the sum
closer to zero thus forming a closed iteration loop. Iteration can then be
formed until the two valves agree within a small tolerance. This calculation
method is very convenient and powerful because it can be readily expanded to
include other basic compounds as future needs arise.
A rigorous thermodynamic approach to the problem of calculating chemical
equilibria in electrolyte solutions involves the use of activity coefficients
for each species in solution requiring interaction parameters between each
species. These activity coefficients are then used to calculate the effect
of composition and ionic strength on the chemical equilibrium constants. Such
a method has been proposed by Edwards, Newman, and Prausnitz (3) for aqueous
solutions of volatile weak electrolytes. However, because of assumptions in
their model, their correlation is not suitable for concentrated solutions of
these compounds. ' To avoid this problem and to minimize computer time required
for calculating the activity coefficient of each individual species, a more
empirical method was used for the SWEQ model.
In the SWEQ model, the equilibrium constants in Table 2 are assumed to
be given by equations of the following form.
In K. = In K? + aCu <- + bCrn + C I0'4 (23)
1 1 HO \j\Jy
where K- = equilibrium constant
K? = equilibrium constant at infinite dilution of
all species
a,b,c = parameters
r
HpS = Total moles h^S absorbed/Kg of solution
CCQ = Total moles C02 absorbed/Kg of solution
2
I = ionic strength = 1/2 ^,-C^Z^ , Z.. = ionic charge
The constant a and b have been found to be independent of temperature while
c is found to be dependent on temperature. In many respects, this empirical
'A oaoer was aiven by Edwards* Newman, and Prausnitz at the 70th AICHE
Meeting, New York Session, 13-17 November 1977, on "Vapor-Liquid Equilibria
in Multicomponent Aqueous Solutions of Volatile Weak Electrolytes." They
report a new correlation similar to their first paper, 4) but the range of
application has been extended to temperatures from 0 to 170°C (32 to 338°F)
and total soHtte concentrations up to 10 molal. This new work was published
as the final report of this project was being written, so no comparisons with
the SWEQ model have been made.
11
-------
TABLE 2. SUMMARY OF CHEMICAL EQUILIBRIA INVOLVED IN CALCULATING
NH3-C02-H2S-H20 VAPOR-LIQUID EQUILIBRIA
Chemical Reaction Equilibrium Constant*
i . co, + H,O -»• HCO: + H+ k = (H+) a
„ „ 2 3 (np-o-B-e)
~ ~e
„ „
c~ ~e
2. HCO: - co" + H+
o J
3. NH« + H •*• NH* If _
3 4
n.-6-e 6
4. NH3 + HCOj •»• H2NCOO" + HgO k =
n^-6-e a e
5, H2S -^ HS" + H*
6. HS" -^ S" + H* k
Y $
7. H20 + H+ + OH" k • (H+)(o + CCAU)
(a + CCAU)
*
8. RCOOH •*• RCOO" + H+
"sA"c * nsA--c
4.
The sum of all electronic charges 1s given as follows: (HT) = a + 2e-6 +y +2i() +a + e +5 -CCAU - H
-------
method is similar to the method used by Van Krevelen. Van Krevelen found
that the equilibrium constants for reaction of hLS and CCL with NH-, were
proportional to ionic strength, so a single correlation parameter was intro-
duced to account for this effect. This has been changed slightly in the SWEQ
model in order to predict multi component equilibrium data at high concentra-
tions by introducing a and b as additional parameters for the separate effects
of absorbed C02 and H2S. Actual parameters used in the SWEQ model are sum-
marized in Table 3. This table shows that only three coefficients have been
introduced. A multiplying factor of -.278 times the concentration of absorbed
H2S and a temperature function time the ionic strength appear for the first
dissociation constant of C02 as given by reaction 1 in Table 3-. The effect
of ionic strength has been taken directly from Figure 3 of Van Krevelen's
papers by fitting the curves in his plot to an analytical equation of the
following form
(Effect of ionic strength on In K) = CIn (24)
where C = temperature dependent parameter
n = empirical exponent (a value of 0.4 was
found although a value of 0.5 would be more
correct from Debeye-Huckel considerations)
Van Krevelen's correlation was made in terms of the following reaction.
+
C0
2(g)
H2°(e)
NH
3(e)
HC0
NH
where
where
=
V.K. (pco ){NH3
This equilibrium constant can be rewritten in terms of a Henry's constant
for C02 as follows.
/.K.
)(H2C03)(NH3)
From Table 2, this represents the sum of reactions 1 and 3 as follows.
V.K.
In the SWEQ model, it is assumed that HCQ and ko are independent of compo-
sition; thus any effect of ionic strengtn2of ionfc strength on 1C, K becomes
a similar effect on K-.
The term of -0.278 CH2$ in Table 3 has resulted from fitting C0? partial
pressures data from quarternary FUS-COp-NHo-hLO mixtures measured at Brigham
Young University. (5) It does not affect ternary C02-NH3-H20 data and only
13
-------
TABLE: 3. EFFECT OF COMPOSITION AND IONIC STRENGTH
_ ON CHEMICAL EQUILIBRIUM CONSTANTSA) _
in K. = in Ki° + aC^ + bC^ + cl°'4
where K.° = equilibrium constant at infinite dilution of all
1 species
a,b,c = parameters
CM c» Crn = total moles of H9S or CO, absprbed in one Kg of
H2S C02 solution 2 d
?
I = 1/2 £.jC.Z. = ionic strength
Z- = ionic charge
Chemical
Reaction
in Table 2
1
2
3
4
5
6
7
8
a
-.278
0
0
0
0
0
0
0
b
0
0
0
0
.427
0
0
0
c
-1.32 + 1558.8/T°R
0
ob)
0
0
0
0
0
equation and constants given here are discussed in the section
on the SWEQ model, equation 23.
^No effect of ionic strength is required for NHo because its equili-
brium constant is used in combination with either h^S or C02«
14
-------
2.
becomes important when significant concentrations of both CCL and H2S are
present. The third coefficient in Table 3 appears as a mumplying factor
of 0.427 times the concentration of absorbed C0? 'which affects the first
dissociation constant of HLS as given by reaction 5 in Table 3. This effect
was also found necessary besides concentration terms in the Henry's constant
to correlate the quarternary H2S-C02-NH3-H?0 data. Van Krevelen found that
a multiplying factor of 08089 times ionic strength to be necessary in log-,0
K. In the SWEQ model, this is accounted for in the concentration-dependent
terms of the H2$ Henry's constant given in Table 1.
The concentration effects given in Tables 1 and 3 were developed in the
following steps:
1. Binary NH,-H,>0 data were correlated to obtain the effect of free
NH3 onitbi HSnry's constant of NHg.
Binary H2S-H20 and ternary H2S-NHo-H20 data were correlated to
obtain an adjusted zero concentration Henry's constant of H2S
as noted at the bottom of Table 1, and an additional concentration
parameter proportional to free NH3 concentration for the Henry's
constant of H2$ was introduced as shown in Table 1. The chemical
equilibrium constant of reaction 3 in Table 2 for the combination
of NH3 plus H to give (NH» ) was also adjusted as an empirical
parameter in order to fit the H2S-NH3-H20 "data. It was also found
necessary to introduce an effect of absorbed H2S on the Henry's
constant of NH, in order to correlate the ternary data. Thus
four effects wire correlated:
a) The zero concentration Henry's constant of H2$
b) The effect of free NH~ on the Henry's constant
of H2S J
c) The equilibrium constant of NH3 + H+ -> NH4+
d) The effect of absorbed H9S on the Henry's constant
of NH3. *•
3. Binary CCL-H20 and ternary C02-NH3-H20 data were correlated. The
effects of ionic strength on the first dissociation constant of
($2 was used directly from Van Krevelen's correlation. The
available data appear to be suitably correlated by this one
effect so no new additional concentration parameters were in-
troduced. However, the zero concentration dissociation con-
stant of C02 was adjusted slightly in order to obtain an im-
proved representation of the C02-NH3-H20 data. The equilibrium
constant for the reaction of HCO," and NH- to produce H2NCOO~
carbamate ions in reaction 4 of fable 1 was not changed from
Van Krevelen's correlation.
15
-------
4. Quarternary H?s-COo-NHq-H?0 data were correlated to obtain the
effect of absorbed H7S on the first dissociated constant of
H?CCL and the effect of absorbed C0? on the Henry's constant
of H^S and on the first dissociation constant of H2$.
Comparisons between measured and calculated data are given in a subse-
quent section of this report.
The chemical equilibrium constants for the reactions given in Table 2
are dependent on temperature. This effect is calculated in the SWEQ model
from equations of the same form given by Kent and Eisenberg (4) as follows.
In K.° = A + B/T + C/T2 + D/T3 + E/T4 (28)
where T is in degrees Rankine and concentrations are in gram moles or gram
ions/Kg of solution. Actual parameters used are given in Table 4. In many
cases the parameters are the same as the ones used by Kent and Eisenberg (4).
Various changes were made in these constants as noted at the bottom of Table 4.
These changes were as follows.
1. The reaction constant of NH, + H+ -> NhL+ was first adjusted
empirically using available H2S-NH3-H20 volatility data and
the equilibrium constant of HpS as puolished by Kent and
Eisenberg.
2. The reaction constants for the first and second ionization
constants of CCL were adjusted from Kent and Eisenberg's
equations to fit available C02-NHo-H?0 data. This was done
so as not to affect the H^-NH^-H^O correlation.
3. After Parts 1 and 2 were done it was found by detailed com-
parisons of measured and calculated data given in subsequent
tables of this report that both the H?S and CO, volatility
data could be adjusted slightly to improve thefr predicted
values. This was done by changing the first dissociation
constants of H2S and C02. The original constant for NHv
was left unchanged. The net effects of these various changes
are noted at the bottom of Table 4.
The equilibrium constant of NHL reacting with HCCL" to produce H?NCOO~
carbamate ion was used as published by Van Krevelen. The dissociation con-
stant of HLO was used as published by Kent and Eisenberg t4). The ionization
constant of carboxylic acids (RCOOH) in water (H20) are nearly independent
of temperature thus a single constant is used for RCOOH ionization according
to reaction 8 in Table 2. The value\of -11.28 is based on a pK of about
4.9 reported by Bomberger and Smith ' from potentiometric titrailons of actual
refinery sour water streams. This reaction has been introduced into the
calculation method so that the effect of carboxylic acids on the volatility
of NH3 can be calculated. A molecular weight of 60.05 is assumed in the SWEQ
calculation model, but another value could be entered if necessary. The
amount of carboxylic acid in a given sour water stream can be obtained from a
16
-------
TABLE 4. SUMMARY OF TEMPERATURE PARAMETERS, USED TO CALCULATE CHEMICAL
EQUILIBRIUM CONSTANTS IN TABLE 1
(Iniq = A + B/T + C/T2 + D/T3 + E/T4 for T 1n °R, and concentrations in gram
moles or gram ions per Kg of soln.)
Chemical
Reaction
In Table 2
1
2
3
4
5
6
7
8
Lit
Ref
Temperature Parameters
4,7,8 -241.79*
4,7,8,9 -295.60*
* 1.587*
1 -5.40
4,11,12 -293.88*
4,11,12 -657.965
4,.10
6
39.5554
-11.28
536256*
655893
11160*
3465
683858*
1649360
-177822
0
TT
-4.8123 X 108
-5.9667 X 108
0
0
-6.27125 X 10
-15.8964 X 101
1.843 X 101
0
1.94 X 10
.11
2.4249 X 10
0
0
,11
-2.96445 X 10
-3.7192 X 10
0
0
I13
,13
.8*
2.555 X 10
,11
Jl
-3.91757 X 101
,13
6.72472 X 10" -10.6043 X 10
,11 1 «sinSt u> 1 A I
-------
potentiometric titration of samples taken from the stream as performed by
Bomberger and Smith. This information can then be used to calculate the
amount of caustic to be added in order to release the NH3.
In the SWEQ model, the volatilities of H2S, C02, or NH3 in solution are
dependent on the H ion concentration or pH of the Solution. This effect
is shown in Figure 1 where the ratio of vapor over liquid concentrations on
a weights-basis are plotted at 120°C (or about 30 psia) versus pH measured
at 25 C. These data were calculated assuming a 0.01 weight-% concentration
in the liquid phase. From this plot, we see that H^S and CO- have greater
volatilities at low pH levels while NH-, has greater volatilities at high pH
levels. This means that a process for simultaneous stripping of all three
components from solution must operate at an intermediate pH where all three
have reasonable volatilities. From this plot, the optimum pH measured at
25 C appears to be around 10, but we find that it varies depending on the
mixture involved. The equilibrium-constant parameters in Tables 3 and 4 and
the Henry's law equations in Table 1 give all the parameters necessary for
predicting vapor-liquid data in NI-L-CO^HL-HpO systems. Other acidic or basic
components could be added to the correlation simply by adding parameters for
the added components to these tables and by incorporating them into the com-
puter program.
Details of a computer program based on the SWEQ model and comparisons
with literature data are given in the next sections of this report.
18
-------
in
c
O
O)
10,000
8,000
6,000
4,000
2,000 -
1,000
800
C O
O O
O O
CM
3 -M
O- 3
•r- O
_J .Q
(C
S ----
O)
> rd
O •r-
(/)
i- 0.
§.0
-------
SECTION 4
COMPUTER PROGRAM BASED ON THE SWEQ MODEL
A computer program for calculating NH-, C02, abd HgS volatility data
from aqueous sour water systems has been developed basea on the SWEQ model.
The program is written to handle a wide range of conditions and temperatures.
The estimated ranges of applicability are as follows:
Property Range
Temperature 20°C to 140°C
Pressure up to 50 psia*
Composition 1 ppm to about 30 weight %
dissolved NH3> carboxylic acid,
salts, and caustic
pH 2 to 14
Corrections for vapor phase non-ideality are recommended at
pressures above 50 psia.
As presently written, the program will handle NH3, CO^, H^S, and water plus
NHg fixation effects due to carboxylic or stronger acids and the effects of
caustic addition.
This computer program uses the new vapor-liquid equations presented
in the previous section of this report which were developed from both old
and new experimental data. This same program was used to develop data
comparisons given in the next section of this report.
The main features of the SWEQ model as it has been programmed are as
follows:
1. As shown in the next section, it is more precise than the Van
Krevelen method of prediction. This improvement is primarily
due to the use of actual data at the conditions of commercial
interest for development of the SWEQ model while the Van
Krevelen correlation is used at extrapolated conditions.
2. The program will take into account NH_ fixation effects
due to carboxylic acids in sour water systems,,
3. The program will also take into/account caustic addition
to release fixed NhL.
20
-------
4. The program can be readily converted to a subroutine for
equilibrium stage calculations for various separation pro-
cesses. Calculations can be made going either up or down
in a distillation process.
5. The SWEQ model can be expanded to additional acidic or basic
components with only minor changes to introduce new ioniza-
tion constants and Henry's constants.
Various ostions in the computer program are available to the user as
follows.9'
1. Option 1 allows the calculation of vapor-liquid equilibrium
data at a specified temperature and liquid composition. This
option would be used for circumstances in which the temperature
at liquid vapor equilibrium is known rather than the pressure.
This was the option used in correlating available experimental
data of this project. It may also be useful in some process
situations.
2. Option 2 allows the calculation of vapor-liquid equilibrium
data at a specified pressure and liquid composition. This
option would normally be used for equilibrium stage process
calculations going up a distillation tower. The program
calculates the temperature and vapor composition from a given
stage. The pressure change from stage to stage must be con-
trolled by the user in specifying the pressure of the equilib-
rium calculation.
3. Option 3 allows the calculation of vapor-liquid equilibrium
data at a specified pressure and vapor composition. This
option would normally be used for equilibrium stage process
calculations going down a distillation tower. The program
calculates the temperature and liquid composition from a
specified vapor composition and pressure. This option would
normally be used for sour water stripper calculations. The
pressure increment between stages must be controlled by the
user in each pressure specified to the program. Option 3
also calculates water in the condenser vapor at a specified
pressure, temperature, and vapor stream composition on a
water- free basis. In this case, a zero water content is
specified as input data for the calculation. This response
for zero water content only occurs with Option 3.
Ammonia fixation and caustic addition effects can be calculated with
all three options given above. Ammonia fixation effects can be calculated
by entering as data a specified wt. % of carboxylic acids in the liquid
the three options listed here, a. fourth, option for a flash
calculation has been completed. This was done after this report
was written, so the results are not in this report. Please contact
the author for the details.
21
-------
analysis. The amount to be entered may be determinate from a potent!ometnc
titration of the sour water under study. The method of titration could be
the same or similar to that used by Bomberger and Smith (6). A molecular
weight of 60.05 is assumed in the computer program. This number was assumed
without any real basis and can be changed in the program without affecting
other parts of the program. The effect of caustic addition can be calculated
in two ways as follows:
1. If a negative pH is specified as input data, then the program
ignores the entry and calculates the pH based in the amount of
caustic in wt % specified in the input data. The input concen-
tration refers to the liquid phasie even when option 3 is used
for calculating down a distillation tower.
2. If a positive pH is specified as input data then the program
computes the amount of caustic necessary to obtain the specified
pH. In this case, the concentration of caustic specified in
the input data is set to zero.
Both of thesie pH options use or compute pH data at the temperature of the
equilibrium stage. If the pH of the liquid at 25 C is desired, the user must
specify this temperature ;and the liquid composition obtained from a higher-
temperature equilibrium stage calculation. This would involve the use of
distillation option number 1.
Table 5 gives a flow chart of the main program. The format for entry of
data to the program is the same regardless of the options used. The basic
program involves the reading of input data which then converts the data so
it can be processed by options 1, 2, or 3 in the program. After these options,
the calculated equilibrium data are then printed by the program.
Flow charts for options 1, 2, and 3 are given in Figures 3, 4, and 5
respectively. These options primarily act as executive programs which call
various subroutines necessary to perform the calculations. Iteration loops
are involved in each of the options because of the problem of calculating
simultaneous chemical equilibria at each condition. The primary interation
of pH is done in each option by calculating equilibrium concentrations of
each species at assumed pH values. Initially chosen pH values are arbitrary
so a test is made to check for electrical neutrality of the solution. For an
arbitrary pH, neutrality will not occur; so then a new pH is chose in subse-
quent iterations until electrical neutrality within a small tolerance is
achieved.
A direct listing of the main program is given in Table §, and listings
of the various subroutines used by the main program are given in Tables 6
to 15. These subroutines and their functions are as follows:
22
-------
Read option and wt% carboxylic
acid and caustic in liquid
Read temperature, pressure, pH,
and either liquid or vapor
compositions in wt% depending
on option
Convert temperature iir°C to
absolute temperature in °R and
°K
I
Test for wt% NH. to be a
positive number
yes
Add a small value to the-com-
position to avoid calculation
problems at zero concentrations
Option 1
Calculate pressure
and vapor compn.
from temperature
and liquid compn
Option 2
Calculate temperature
and vapor compn. from
pressure and liquid compn.
Option 3
Calculate temperature
and liquid compn.
from pressure and
vapor compn.
Print Equil.
Data
Figure 2. Flow diagram of SWEQ main program.
23
-------
Convert input compositions to moles/Kg of
soln in liquid - subroutine CFX
Initialize iteration parameters for pH
calculation - subroutine PHST
Calculate chemical equilibrium constants -
subroutine KREAC
o
Calculate the concentration of each
species in the liquid - subroutine SPECL
Estimate a new pH value - subroutine NPH
Test if input pH is positive
Iterate ten times with a new estimated
value for caustic concentration each time
>
o
c
c
to
I/I
0)
Test to see if electrical neutrality is
computed within a small tolerance
yes
Calculate equilibrium pressure and vapor
compositions - subroutine PRESY
10
o
c
O)
&.
O)
Print results
Figure 3. flow diagram of option 1 of SWEQ computer program.
24
-------
o
Convert input composition to moles/Kg
of soln. in liquid - subroutine CFX
Initialize iteration parameters
Initialize pH parameters for each
temperature iteration - subroutine
PHST »
Calculate chemical equilibrium con-
stants - subroutine KREAC
no
Calculate the concentration of each
species in the liquid - subroutine
SPECL
Estimate a new pH value - subroutine
NPH
Test if input pH is positive
yes
Iterate ten times with a new estimated
value for caustic concentration each
time
Test to see if electrical neutrality
is computed within a small tolerance
yes
Calculate equil. pressure and
vapor compositions - subroutine
PRESY
Estimate new temp.
NTEMP
- subroutine
Test to see if calculated press-
ure is within a small tolerance
of specified pressure
X
Print results
r ten
ations
Figure 4. i?low diagram of option 2 of SWEQ computer program.
25
-------
Convert input compositions to mole
fraction in vapor - subroutine YFX
I
-------
TABLE 5. COMPUTER PROGRAM BASED ON THE SWEQ MODEL
C SWEQ COMPUTER PROGRAM
C THIS COMPUTE* PROGRAM HAS WRITTEN BY GRANT M. WILSON FOR THE API CREC
C COMMUTE, RONALD G. GANTZ SOUR WATER STRIPPER PROJECT MANAGER. QUESTIONS
C ABOUT THIS PROGRAM SHOULD BE DIRECTED TO EITHER GRANT M. WILSON OR RONALD
C G. GANTZ. THIS PROGRAM IS WRITTEN IN FORTRAN FOR OPERATION ON A TIME
C SHARE TERMINAL CONNECTED TO A DIGITAL EQUIPMENT CO.MODEL 10 COMPUTER.
C IT CAN BE CONVERTED FOR USE AS A SUBROUTINE OR FOR BATCH OPERATION. FOR
C OPERATION AS A SUBROUTINE THE ERROR MESSAGES NOW PRINTED ON THE TERMINAL
C WOULD HAVE TO BE CHANGED SO THAT THE EXECUTIVE PROGRAM WOULD TAKE
C CORRECTIVE ACTION. ERROR MESSAGES ARE NOW PRINTED BY STATEMENTS 320, 420,
C AND 515 IN THE MAIN PROGRAM; AND BY STATEMENTS 12 AND 35 IN SUBROUTINE
C SPECV AND PY STATEMENT 7341 IN SUBROUTINE SPECL.
C THIS PROGRAM CALCULATES EQUILIBRIUM VAPOR-LIQUID COMPOSITIONS FOR WEAK
C ELECTROLYTE MIXTURES CONTAINING NH3,C02,H2S,RCOOH,CAUSTIC, AND WATER.
C COMMENTS IN THE PROGRAM DESCRIBE VARIOUS OPTIONS POSSIBLE AND THE FUNCTION
C OF VARIOUS PARTS OF THE PROGRAM
COMMON TC,TK,TR,PSI,P,XA,XC,XS,XW,CA,CC,CS,CAS,CCS,CSS,
lYA,YC,YS,YW,Wft,fcC,WS,WW,PHO,PH,TOL,HT,DPH,PHA,PH8,AL,BT,GA,
2DE,SI,HP,SD,EPS,EKS,£KCA,EKCAO,£KCB,EKA,EKrt,EKCC,EKSB,EKSA,HTU,
3EL1,HA,HC,HS,MW,CCST,RHQ,XSA,XCAU,CSA,CCAU,ZET,WSA,«CAU,CCAUS
COMMON ICD
OPEN (UNI T = 20, DE V ICEs ' DSK', ACCES'Ss'SEQIN', FILEr'SWSD1)
DATA WA,WC,WS,WW,WSA,WCAU/17.03,«4.01,34.08,18.02,60.05,
-------
TABLE 5 (continued)
2 RtAD(20,100(>) TC,PSI,PHO,XA,XC,XS,XW
IF(XK) 10,10,20
10 ICO = 1
GO TO 30
20 ICO = 0
30 XCAU = XCAUO
C CONVERT TO ABSOLU1E TEMPERATURES IN DEC. K AND OEG. R.
TK = TC+273.15
C A SMALLKVALUE*ISBAOOED 10 THE COMPOSITIONS IN ORDER TO AVOID CALCULATION
C PROBLEMS AT ZERO CONCENTRATIONS.
XA = XAtlfc-12
XC - XC+1E-12
XS = XS-HE-12
Xw = XWtlt-12
C NEGATIVE AMMONIA CONCENTRATION SIGNALS NEW OPTION, RCOOH, OR CAUSTIC
C DATA.
IF(XA) 1,200,200
200 GO TO (300,400,500).NOOPT
C OPTION 1 CALCULATES PRESSURE AND VAPOR COMPOSITION FROM SPECIFIED
C TEMPERATURE AND LIQUID COMPOSITION. SUBROUTINE FUNCTIONS ARE LISTED WITH
C THE SUBROUTINES
300 CALL CFX
PH = 7
HTO = 1
CALL PHST
CALL KREAC
ELI = CA
C ITERATION LUOP TO CALCULATE EITHER EQUILIBRIUM PH OR EQUILIBRIA AT
C SPECIFIED PH.
DO 310 1C = 1,100
CALL SPECL
CALL NPH
IF(PHO) 305,305,302
302 CCAU = (CCAU+HT)/2
IF(IC-IO) 310,310,330
C TEST FOR PH CONVERGENCE
305 IF(ABS(HT/TOL)-.0001) 330,330,310
310 CONTINUE
320 WRITE(5,1010)
330 CALL PRESY
C EQUILIBRIA CALCULATED; TRANSFER TO PRINT OUT OF RESULTS.
GO 10 900
C OPTION 2 CALCULATES TEMPERATUKE AND VAPOR COMPOSITION FROM SPECIFIED
C PRESSURE AND LIUU1D COMPOSITION.
400 CALL CFX
P = PSI
ELI = CA
PH = 7
HTO = 1
C FIRST 00 LOOP ITERATES TO DETERMINE TEMPERATURE.
(continued)
28
-------
TABL'E 5 (continued)
DO 440 IEsl.100
CALL PHST
CALL KREAC
C SECOND 00 LOOP ITERATES TO FIND EQUIL. PH AT ITERATION TEMPERATURE.
DO 410 1C = 1,100
CALL SPECL
CALL NPH
IF(PHO) 405,405,402
402 CCAU = (HT+ccAu>/2
IF(IC-IO) 410,410,130
C TEST FOR PH CONVERGENCE
405 IF(ABS(HT/TOL)-.0001) 430,430,410
410 CONTINUE
420 WRITE(5,J010)
430 CALL PRESY
CALL NTEMP
C TEST FOR PRESSURE CONVERGENCE.
IF(ABS(PSI/P-1)-.001) 460,460,440
440 CONTINUE
WRITE(5,1011)
C EQUILIBRIA CALCULATED? TRANSFERS TO PRINT OUT OF RESULTS.
460 GO TO 900
C OPTION 3 CALCULATES TEMPERATURE AND LIQUID COMPOSITION FROM SPECIFIED
C PRESSURE, VAPOR COMPOSITION, PLUS WTZ OF RCOOH AND CAUSTIC IN LIQUID.
500 CALL YFX
P = PS1
CSA s 10*XSA*RHO/WSA
CCAU = 10*XCAU*RHO/WCA"
ELI = 0
CC = 0
CA = 0
CS = 0
CAS s 0
PH x S
HTO * 1
c FIRST po LOOP ITERATES TO (^TERMINE TEMERAIURE.
DO 530 IE s 1,100
CALL PHST
C SECOND DO LOOP ITERATES TU FIND EQUIL. PH AT ITERATION TEMPERATURE.
DO 510 IC=1,100
CALL KREAC
CALL SPECV
CALL NPH
IF(PHO) 505,505,502
502 CCAU = (CCAU+HT)/2
IFUC-10) 510,520,520
C TEST FOR PH CONVERGENCE.
505 IF(ABS(HT/TOL>-.0001> 520,520,510
510 CONTINUE
515 WK1TE(5,1010>
520 CALL NTEMP
(continued)
29
-------
TABLE 5. (continued)
C
C
C
C
C 3
900
C TEST FOR PESSURE CONVERGENCE.
1F(A6S(PSI/P-1)".001) 550,550,530
530 CON1INUE
hRITE(5,1011)
C EQUILIBRIA CALCULATED; TRANSFER TO PRINT OUT OF RESULTS.
550 GO TO 900
C COMPOSITION DATA IN THE LIQUID PHASE ARE USED IN THE PROGRAM IN TERMS
OF MOLES OF COMPOUND PER KILOGRAM OF SOLUTION. THE VAPOR PHSAE IS IN TERMS
OF MOLE FRACTION. THE NEXT TEN STATEMENTS CONVERT THESE BACK TO WEIGHT
PERCENT. THE ORIGINAL CONVERSION OF THE INPUT DATA TO MOLES PER KILOGRAM
AND VAPUR MOLE FRACTION IS DONE IN SUbROUTINES CFX, YFX, AND FOR OPTION
PARTLY IN THE MAIN PROGRAM.
XA = 100*CA*wA/(1000*RHO)
XC = 100*CC*WC/(1000*RHO)
XS = 100*CS*rtS/(1000*«HO)
XCAU = 100*CCAU*WCAU/(1000*RHO)
XW = 100-XA-XC-XS-XSA-XCAU
YT s YA*toA+YC*WC+YS*WS+YW*Wrt
YA = JOO*YA*wA/YT
YC = 100*YC*WC/YT
YS = 100*YS*WS/YT
YW s 100*YM*nw/YT
XMT = XA/WAfXC/wC+XS/WStXCAU/WCAUtXW/WW+XSA/WSA
XMT = 100/XMT
XMA = XA*XMJ/WA
XMC = XC*XMT/WC
XMS = XS*XMT/*S
XMCAU * XCAU*XMT/WCAU
XKW s XW*XMT/WW
XMSA s XSA*XM1/WSA
YMT = YA/WA+YC/WC*YS/rtS+YW/WW
YMT s 100/YMT
YMA = YA*YMT/wA
YMC s YC*YM1/«C
YMS = YS*YMT/wS
YMW = YW*YMT/WW
TF = TC*I.8+32
PKPA = P*6.895
ATM s P/14.696
C OUTPUT FROM THE FOLLOWING STATEMENTS IS MORE OR LESS SELF EXPLANATORY
C IN THE FORMAT STATEMENTS.
WRITE(5,1030) TCrTF,TK,TR,P,ATM,PKPA,PH
hRITE(5,10?0)
EK s YMA/XMA
XA,YA,XMA,YMA,EK
WR1TE(5,1040)
EK s YMC/XMC
WRITE(5,1050)
£K s YMS/XMS
I»RITE(5,1060)
EK s YMW/XMrt
WRITE(5,1070)
XC,YC,XMC,YMC,EK
XS,YS,XMS,YMS,EK
XH,Ytt,XMW,YMM,EK
(continued)
30
-------
TABLE 5 (continued)
WRITE(5,1072) XSA,XMSA
WR1TE(5,107«) XCAU.XMCAU
XTOT = XA+XC*XS+X«+XSAtXCAU
XMTOT s XMA+XMC+XMS*XMKtXMSA+XMCAU
YT01 s YA + YC + YS+Y*f
YMTOT s YMA+YMC+YM8+YMW
WRITE(5,1075) XTOT,YTOT,XMTOT,YMTOT
GO TO 2
1000 FORMAT(IOE)
1001 FORMAT(I,2E)
1010 FORMATC PH DID NOT CONVERGE IN 100 CYCLES')
1011 FORMATC TEMPERATURE DID NOT CONVERGE IN 100 CYCLES')
1020 FORMATC
1« WEIGHT PERCENT
2' COMPONENT LIQUID VAPOR
3-VALUE')
1030 FORMAH/X
I1 TEMPERATURE1,F8.2, ' C, SF8.2,' F, SFS.a,' K»',F8.2,' RV
2* PRESSURE ',Fe.2,' PSIA,',F8.3r' ATM,»,F9.2f' K-PASCALSV
31 PH',8X,F8.3//)
MOLE PERCENT1/
LIQUID VAPOR
1040
10SO
1060
1070
1072
1074
1075
1060
AMMONIA S5F10.5)
CARBON DIOXIDE *,5F10.5)
HYDROGEN SULFIDE',5F10.5)
WATER SSFiO.S)
CAR80XYL1C ACID • ,F10.5, IOX,F10.5)
FORMATC SODIUM HYDROXIDE1,F10.b,10X,F10.5>
FORMATC TOTAL V5F10.5)
FORMAT(X,F4.0»3F7.3.7F6.2,9F6.3»F6.2)
END
FORMATC
FORMATC
FORMATC
FORMATC
FORMATC
31
-------
WBLE 6. SUBROUTINE KREAC
SUBROUTINE KREAC
C THIS SUBROUTINE CALCULATES CHEMICAL EQUILIBRIUM CONSTANTS AS FOLLOWS.
C SYMBOL EQUILIBRIUM
C EKS H2S FIRST IONIZAT10N
C EKCAO C02 FIRST IONIZATIOM AT ZERO IONIC STRENGTH
C EKCB C02 SECOND IONIZATION
C EKA NH3 PLUS PROTON GOING TO AMMONIUM ION
C EKW WATER DISSOCIATION
C EKCC BICARBONATE PLUS AMMONIA GOING TO CARBAMATE
C EKSB H2S SECOND IONIZATION
C EKSA RCOOH IONIZATION
C THE EFFECT OF IONIC STRENGTH ON EKCAO IS CALCULATED BY THE CALLING PROGRAM.
COMMON IC,TK,TR,PSI,P,XA,XC,XS,XK,CA,CCrCS,CAS,CCS,CSS,
lYA,YC,YS,YW,WA,ViC,WS,Ww,PHO,PH,TOL,Hr,OPH,PHA,PHB,AL,BT,GA,
2DE,SI,HP,SD,EPS,EKS,EKCA,tKCAO,EKCB,EKA,EK*l,EKCC,EKSB,EKSA,HTO/
3£LI,HA,HC,HS,HW,CCST,RHO,XSA,XCAU,CSA,CCAU,ZET,WSAr*CAU,CCAUS
EKS = EXP(-293.88+68385tt/TR-6.27125E8/UR*TR)+2.5551Ell/(TR**3)
l-3.91757El3/(TR**«)+.
EKCAO * EXP(-2/J1.79+536256/TR-/».8123E8/(TR*TR) + l,
EKCB = EXP(-295.64655893/TR-5.9667E8/(TR*TR)+2.4249EH/(TR**3)
1-3.7192E13/(TR**«J)
EKA * EXPU.587+11160/TR)
EKW = EXP(39.5S5«-177822/TR+1.8«3E8/(TR*TR)-.85«)E11/(TR**3)
1+1.0292E13/(TR**0))
EKCC = EXP(-5. 40+1925*1. 8/TR)
EKSB = EXP(-657.965+1649360/TR-15.e964E8/
-------
TABLE 7; SUBROUTINE.HENRY
SUBROUTINE HENRY
C THIS SUBROUTINE CALCULATES HENRY'S CONSTANTS FOR NH3,C02, H2S, AND H20
C RESPECTIVELY AS HA, HC, HS, AND Hh. HA OF AMMONIA IS DEPENDENT ON THE
C CONCENTRATIONS OF SPECIES NH3, C02, AND H2S RESPECTIVELY BY THE SYMBOLS
C CASr CC, AND CS. HS OF H2S IS DEPENDENT ON CAS AND CC. HW FOR *ATER IS
C THE VAPOR PRESSURE OF WATER.
COMMON TC,TK,TR,PSI,P,XA,XC,XS,XW,CA,CC,CS,CASrCCS»CSS,
1YA,YC,YS,YW,WA,KC,WS,WK,PHO,PH,TOL,HT,DPH,PHA,PHB,AL,6T,GA,
2DE,SI,HP,SD,EPS,EKS,EKCA,EKCAO,EKCB,EKA,£K*,EKCC,EKSR,EKSA,HTO,
3ELl,HA,HC,HS,HW,CCST,RHO,XSA,XCAU,CSA,CCAU,ZET,AiSA,*lCAU,CCAUS
TK = TR/1.8
HA a EXP(178.339-15S17.91/TR-25.6767*ALOG(TR)
|».01966*TR+(131.4/TR-.1682)*CAS
lt.06*(2*CC+CS))
HC = EXPU8.33-24895.1/1R + .223996E8/(TR*TR)-.090918E11/(TR**3)
H.12601E13/(TR**4))
HS = EXPUOO.684-24.6254E4/TRt2.39029E8/(TR*TR)-1.0l898EU/(TR
l**3)+l.S973tE13/(TR**a)-.05*CAS+(.965-a86/TR)*CC)
HW - EXP(1«.466-6996.6/(TR-77.
RETURN
END
TABLE, 8. SUBROUTINE YFX
SUBROUTINE YFX
C THIS SUBROUTINE CONVERTS COMPOSITIONS IN WTX TO VAPOR CONCENTRATIONS
C IN MOLE FRACTION. VAPOR COMPOSITIONS FOR NH3.C02, H2S, AND KATER RESPECTIVE
C ARE GIVEN BY THE SYMBOLS YA, YC, YS, AND Y*.
COMMON TC,TK,TR,PSI,PfXA,XC,XSrXw,CA,CC,CS,CAS.CCS,CSS,
lYA,YC,YS»YW,^AfwC,WS,WW,PHO,PH,TOLTHT,OPH,PHA,PHB,AL»BT,GAr
20E,SI,HP,SD,EPS,EKS,EKCA,EKCAO,£KCB,EKA,EK*,EKCC,EKSH,EKSA,HTO,
3ELI,HA,HC,HS,hW,CC-ST,RHO,XSA,XCAU,CSA,CCAU,ZET,^SA,rtCAU,CCAUS
XT =
YA = XA/(WA*XT)
YC s XC/(rtC*XT)
YS = XS/(wS*XT)
Yw = XW/(WW*XT)
CA = 1
RETURN
END
33
-------
TABLE.9. SUBROUIINE CFX
SUBROUTINE CFX
C THIS SUB«OUI1NE CONVERTS COMPOSITIONS IN WTX TO LIQUID CONCENTRATIONS
C IN MOLES PEH KG OF SOLUTION. LIQUID COMPOSITIONS FOR NH3, C02, H2S, RCOOH,
C AND CAUSTIC RtSPECTlVELY ARE GIVEN BY THE SYMBOLS CA, CC, CS, CSA, AND CCAU
COMMON TC,TK,TR,PSI,P,XA,XC,XS,Xrt,CA,CC,CS,CAS,CCS,CSS,
lYA,YC,YS,Yw,KA>«C,*S,rtH,PHO,PH,TOL,HT,DPH,PHA,PHB,AL,8T,GA,
20E,SI,HP,SD,EPS,EKS,EKCA,EKCAO,EKCB,EKA,EKW,EKCC,EKSB,EKSA,HTO,
3ELI,HA,HC,HS,Hrt,CCST,KHO,XSA,XCAU,CSA,CCAU,2ET,«SA,WCAU,CCAUS
F s 1000*RHO/(XA+XC+XS+XW+XSAtXCAU)
CA = XA*F/wA
CC = XC*F/rtC
CS = XS*F/WS
CSA = XSA*F/WSA
CCAU a XCAU*F/WCA(J
RETURN
ENO
TABLE 10. PHST
SUBROUTINE PHST
C THIS SUBROUTINE , INITIALIZES PARAMETER VALUES FOR ITERATIVE CALCULATION
C OF PH. THIS SUBROUTINE DETERMINES THE VALUE OF THE TOLERANCE TOL TO BE USED
C IN TESTING FOR PH CONVERGENCE, AND INITIALIZES PH AND OTHER PARAMETERS FOH
C THE ITERATION
COMMON TC,TK,TR,PSI,P,XA,XC,XS,XW,CA,CC,CS,CAS,CCS,CSS,
2De,SI,HP,SD,£PS,EKS,EKCA,EKCAO,EKCB,£KA,EK«,EKCC,EKSB,EXSA,HTO,
5ELI,HA,HC»HSrHw,CCST,«H(l,XSA,XCAU,CSA,CCAU,Z£T,WSA,hCAU,CCAUS
IF(PHO) 40,30,30
30 PH = PKO
40 PHA = 0
PHB = 14
HTA = -CA
HTB s 2*CC+CS
IFCHTB+HTA) 50, $0,60
50 TOL = -HTA
GO TO 70
60 TOL = HTB
70 *L = 0
TOL s TOLtlE-f
DPH s 1
EPS s 0
HIO s 1
RETURN
END
34
-------
TABLE IV. SUBROUTINE SPECV
SUBROUTINE SPECV
C THIS SUBROUTINE CALCULATES EQUILIBRIUM SPECIES CONCENTRATIONS IN THE
C LIQUID PHASE FROM A SPECIFIED VAPOR COMPOSITION, TEMPERATURE, AND AN
C ASSUMED PH. SPECIES CONCENTRATIONS IN THE LIQUID ARE GIVEN BY THE FOLLOWING
C SYMBOLS.
C SYMBOL SPECIES
C CAS NH3
C CCS C02
C CSS H2S
C AL HC03-
C BT C03--
C DE NH4+
C EPS CARBAMATE ION
C GA HS-
C SO S—
C SI OH-
C 2ET RCOO-
C CCAUS NAt
C THIS CALCULATION IS PERFORMED AT AN ASSUMED PH SO THAT SPECIES CONCENTRATIO
C CAN BE SOLVED FROM DIRECT ALGEBRAIC EQUATIONS. SOME ITERATION IS REQUIRED
C BECAUSE OF THE EFFECT OF IONIC STRENGTH ON THE FIRST IONIZATION OF C02
C AND THE EFFECT OF SPECIES CONCENTRATIONS ON THE VOLATILITY OF NH3 AND H2S.
C ELI = IONIC STRENGTH.
COMMON TC,TK,TR,PSI,P,XA,XC,XS,XW,CA,CC,CS,CAS,CCS,CSS,
lYA,YC,YS,Y«,KA,WC,WS,WW,PHO,PH,TOL,HT,t>PH,PHA,PHB,AL,8T,GA,
2DE,SI,HP,SD,EPS,EKS,EKCA,EKCAO,EKCB,EKA,EM,EKCC,EKSB,EKSA,HTO,
3ELI,HA,HC,HS,HW,CCST,RHO,XSA,XCAU,CSA,CCAU,ZET,«SA,wCAU,CCAUS
COMMON ICO
CALL HENRV
XW x .9
PW s 1E-19
PA s 0
PC s 0
PS s 0
EKCA s EKCAO
DO 30 1=1,100
EKAP s EKCA
EKCA s EKCAO*EXP((-1.32tl558.8/TR)*ELI**.
-------
TABLE 11. (continued)
P a PW
ICO s 0
GO TO 6
5 Yw = Pw/PSl
YTOT = U-Yw)/(YAtYC + YS)
YA s. YA*YTOT
YC = YOYTOT
YS = YS*YTOT
PA = PSI*YA
PC = PSI*YC
PS = PSI*YS
60 TO 7
6 PA s (YA*Prt/YW+PA)/2
PC = (YC*Prt/Yrt+PC)/2
PS s (YS*PW/YW+PS)/2
PBA s p
P = PW+PAtPO+PS
7 CAS = PA/HA
CCS = PC/HC
CSS = PS/HS
HP = tXP(-?.30259*PH)
Ai. = EKCA*CCS/HP
8T s £KCB*AL/HP
OE s EKA*HP*CAS
EPS s EKCC*CAS*AL
GA * EKS*CSS/HP
SO = £KSB*GA/HP
SI s EKW/HP-CCAU
2ET = EKSA*CSA/(HP+EKSA)
CCAUS = CCAU-SI
CA = CAS*OE+EPS
CC ~ CCS+ALtBTtEPS
CS = CSS+GA+SD
ELI s ((AU+4*6T*OE*EPS+GA*fl*SDtSl4HPtZET*CCAUStCCAU)/2+ELI)/2
TN1" s (IOOO*RHO-CA*«A-CC*«C-CS*WS-CSA*WSA-CCAU*WCAU)/WW
1-ALtEPS-SI
TNM s CA+CCtCS+BT-OE+GA+SD+SI+TNrt+ZET42*CCAU+CSA
XiMO v- XW
XW = TNrt/TNM
XH s (XMtXwO)/2
8 JF(XW) 80,80,9
80 XW s XW+.l
GO TO 8
C TEST FOR ITERATION CONVERGENCE
9 IF(ABS(P«/PwO-l.)-.OOl) 10,10,30
10 IF(ABS(PA/PAO-1.)-.001) 15,15,30
15 IF(ABS(PC/PCO-1.)-.001) 20,20>30
20 IF(ABS(PS/PSO-l.)-.OOl) 50,50,30
30 CONTINUE
35 WRITE(5,«0)
40 FOKMATC OIDNT CALCULATE LIQUID IN 100 CYCLES')
50 RETURN
END
36
-------
TABLE 12. SUBROUTINE NRH
SUBROUTINE: NPH
C THIS SUBROUTINE CALCULATES A NEW ESTIMATED PH FROM A PREVIOUS PH.
C CRITERIA USED ARE THE SUM OF IONIC CHARGES 10 ELECTRICAL NEUTRALITY, HT;
C AND ANY CHANGES IN SIGN OF HT FROM A PREVIOUS ITERATION.
COMMON TC,TK,TR,PSI,P,XA,XC,XS,Xrt,CA,CC,CS,CAS,CCS,CSS,
!YA,YC,YS,YW,WA,WC,i«S,WW,PHOfPH,TOL,HT,DPH,PHA,PHB,AL,BT,GA,
2D£,SI,HP,SD,EPS,EKS,EKCA,EKCAO*EKCB,EKA,EK'Al,EKCCrEKSB,EKSA,HTO,
3ELI,HA,HC,HS,HW,CCST,«HO,XSA,XCAU,CSA,CCAU,ZET,«SA,WCAU,CCAUS
HT - AL+2*BT+GA-DE+S1-HP+2*SD+EPS+ZET-CCAU
TOL s 2*CCtCS
IF(TOL-CA) 60,70/70
60 TOL = CA
70 IF(PHO) 81,81,80
80 HT = HT+CCAU
GO TO 88
81 IF(HTXDPH) 8«,8«,82
82 HTO 8 ,5*HTO
84 DPH s. -HT*HTO/(ABS(HT)+1E-19)
PH = PHtOPH
GO TO (8fi,88re5),NOOPT
85 REF = HP*(HT+EKA*CAS*HP)
PH = SQRT(REF/(EKA*CAS))
PH = -,5*ALOGlPH*HP)/2.30259
88 RETURN
END
37
-------
TABLE 13. SUBROUTINE SPECL
c THIS suoRouHNE CALCULATES EQUILIBRIUM species CONCENTRATIONS IN THL LI«UIO
C PHASE FROf. A SPECIFIED LIUUID COMPOSITION, TEMPERATURE, AND AN ASSUMED
C Ph. SYMBOLS USEU FOk SPECIES CONCENTRATIONS ARE THE SAME AS FOR SPECV.
C ITERATION IS NECESSARY BECAUSE THE CARBAMATE CONCENTRATION CANNOT BE
C SOLVED DIRECTLY, AND BECAUSE EKCA IS DEPENDENT ON IONIC STRENGTH.
COMMON TC,TK,TR,PSI,PfXA,XCfX5,XW,CA,CC,CS,CAS,CCSrCSS,
lYA,YC,YS,Yi»,WA,wC,WS,WW,PHO,PH,TOL,HT,DPH,PHA,PHB,AL/BT,GA»
2DE,SIrHP,Sr),EPS,EKS,fcKCA,EKCAO,EKCB,EKA,EKrt,EKCC,EKSH,EKSAfHTO,
•*ELI,HA,HC,HSfHW,CCST,*HO,XSA,XCAU,CSA,CCAU,2ET,«SA,WCAU,CCAUS
HP = EXP(-2.30259*PH)
AL = CC
DO 734 1AL = 1,100
EKCA = EKCAO*EXP((-1.32+155a.8/TR)*ELI**.«)
CCS = HP*AL/EKCA
BT = EKCB*AL/HP
DE = EKA*hP*CA/((H-AL*EKCC)*U+EKA*HP/(ltAL*EKCC))>
EPS = AL*EKCC*(CA-DE)/(ltAL*EKCC)
CCSI = CCS^AL+BTfEPS
C TEST FOR ITERATION CONVERGENCE
IF(ABS(CCST/CC-1)-.0001) 736,736,733
732 IF(CCST-1E-16) 733,733»73«0
733 AL = CC/2
CCS = CC/2
BT = IE-19
DE = EKA*HP*CA/(1+EKA*HP)
EPS = IE-19
60 TO 736
73<»0 AL s. AL*CC/CCST
6A s. CS*EKS/(HP*(1 + (1*EKSB/HP)*EKS/HP))
SD = GA*EKS6/HP
SI s EKW/HP-CCAU
ZET = EKSA*CSA/(HPtEKSA)
CCAUS = CCAU-SI
7>fl ELI = (ALt1*BT+DE+EPStGA+1*SD+SItHP+ZET+CCAUStCCAU)/2
7341 WRITE(5,735)
735 FORMAT(• CARBAMATE 01DN1 CONVERGE IN 100 CYCLES')
736 RETURN
END
38
-------
TABLE 14. SUBROUTINE PRESY
SUBROUTINE PRESY
C THIS SUBROUTINE CALCULATES EQUILIBRIUM VAPOR COMPOSITION AND PRESSURE
C FROM TEMPERATURE AND CALCULATE!) SPECIES CONCENTRATIONS OF NH3, C02, AND H2S
C IN THE LIQUID
COMMON TC,TK,TR,PSI,P,XA,XC,XS,XW,CA,CC,CS,CAS,CCS,CSS,
lYA,YC,YS,Yrt,WA,WC.«S,WW,PHO,PH,TOL,HT,DPH,PHA,PHB,ALr8T,GAr
2DE,Sl,HP,SD,EPS,EKS,E*CA,EKCAO,EKCB,EKA,EKrt,EKCCffcKSB,EKSA,HTO,
3ELl,HA,HC,HS,HW,CCST,RHO,XSA,XCAU,CSA,CCAU,ZET,rtSA,WCAU,CCAUS
CCS = HP*AL/EKCA
CSS s HP*GA/EKS
CAS s DE/(EKA*HP)
CALL HENRY
PA s CAS*HA
PC s CCS*HC
PS s CSS*HS
TNW s UOOO*RHO-CA*WA-CC*WC-CS*WS-CSA*WSA-CCAU*WCAU)/KW
1-AL+EPS-SI
TNM s CA+CC+CS+BT-DE+GAtSD+SI+TNW+ZETt2*CCAU+CSA
P« a TNW*HW/TNM
P * PA+PCtPStPW
YA = PA/P
YC a PC/P
YS s PS/P
YW = PW/P
RETURN
END
TABLE IS. SUBROUTINE NTEMP
NTEMP
C THIS SUBROUTINE ESTIMATES A NEW TEMPERATURE IN AN ITERATIVE CALCULATION
C TO AGREfhITH A SPECIFIED PRESSURE. THE ONLY CRITERION USED IS THE CALCULAT
C PRESSURE OF A PREVIOUS ITERATION VERSUS THE SPECIFIED PRESSURE. AN ASSUMEl
C EFFECTIVE HEAT OF VAPORI/ATION OF 9000X1.987 BTU PER POUND MOLE IS USED
C TO ESTIMATE A NEW TEMPERATURE.
COMMON TC,TK,ltf,PSI,P»XA/XC,XS,XW,CA,CC,CS,CAS/CCSfCSS,
IYA,YC,YS.YW,WA,WC,WS,W«»PHOfPH,TOL,HT,DPH,PHA,PHB,AL.BT,GA,
2DE,SI,HP,SD,EPS,EKS,EKCA,EKCAO,EKCB,EKA,EKrt,EKCCftKSB,EKSA,HTO,
3ELI,HA,HC,HS,HW,CCST,RHO,XSA,XCAU,CSA,CCAU,Z£T,rtSA,WCAU,CCAUS
TRi r TR
TR s -ALOG(PSI/P)/9000+1/TR1
TR s 1/TR
1C = TR/1.8-273.15
TK s TR/1.8
RETURN
END
39
-------
Table Function
Subroutine No. of Subroutine
KREC 6 Calculates chemical equilibrium constants from
parameters in Tables 3 and 4
HENRY 7 Calculates Henry's constants from parameters in
Table 1
YFX 8 Converts vapor compositions in wt % to vapor
concentrations in mole fraction
CFX 9 Converts liquid compositions in wt % to
liquid concentrations in moles/Kg of solution
PHST 10 Initializes pH iteration parameters
SPECV 11 Calculates pressure and equilibrium species con-
centrations in the liquid phase from a specified
vapor composition, temperature, and asssumed pH
NPH 12 Calculates a new estimated pH from a previous pH
SPECL 13 Calculates equilibrium species concentrations in
the liquid phase from a specified liquid composi-
tion, temperature, and an assumed pH
PRESY 14 Calculates equilibrium vapor concentrations and
pressure from temperature and calculated species
concentrations in the liquid
NTEMP 15 Estimates a new temperature in an iterative cal-
culation so that the calculated pressure will
agree with a specified pressure
A discussion of each of these subroutines in the order listed above is given
in the following text of this report.
KREAC
Equations used in KREAC come from Tables 3 and 4. The symbols used in
the subroutine relate to the various chemical reactions in Table 2 as follows.
Fortran
Symbol
EKS
EKCAO
EKCB
EKA
EKW
EKCC
EKSB
EKSA
Chemical Reaction
in Table 2
5
1
2
3
7
4
6
8
40
-------
The effect of ionic strength of EKCAO of reaction 1 is not computed in the
subroutine because it changes each pH iteration. This effect is therefore
computed in subroutine SPECL for each pH iteration cycle where the ionic
strength from the previously computed cycle is used for the next iteration.
HENRY
Equations used in HENRY come from Table 1. The symbols used in the
subroutine relate to the Henry's constant parameters given in Table 1 as
follows.
Fortran Henry's
Symbol Constant for
HA free NH,
HC free CO^ (or H,CO-)
HS free H^S * J
HW vapor pressure of water
YFX
The conversion of wt.% in the vapor to mole % in the vapor from sub-
routine YFX is fairly straight forward. The fortran symbols and associated
molecular weights entered by means of a data statement at the beginning of
the main program are as follows.
Component Symbol Molecular Height
NH, WA 17.03
CO- WC 44.01
H9S WS 34.08
HfO WW 18.02
RCOOH WSA 60.05
NaOH WCAU 40
CFX
Subroutine CFX is similar to YFX except that the concentrations in wt.
% are converted to liquid concentrations in moles/Kg of solution. To do this,
the sum of all wt.% given as input to the program are summed and divided into
1000 x RHO to obtain the normalizing factor. A value of RHO = 1 has to be
used or the concentrations will not come out in moles/Kg of solution. This
assignment is made in the main program as the first executable statement. The
number of moles of each component is then computed from the normalizing factor
times its concentration on a weight basis divided by its molecular weight.
PHST
Subroutine PHST initializes parameters used in the pH iteration pro-
cedure. If a positive PHO (for pH) is specified to the subroutine, then the
subroutine assigns PH = PHO and the other parameters have no effect. If a
negative PHO is specified then it means that the program must compute the pH.
41
-------
In this case, it assigns the limits over which the pH can be varied which
are from 0 to 14 and assigns a tolerance to be used by subroutine NPH to
test for convergence. The tolerance variable is assigned to either the sum
of acid gas concentrations if they are in excess or to the Nl-L concentration
if it is in excess. Carbon dioxide reacts with two moles of NH~, so its con-
centration is multiplied by two in computing the acid gas^oncentration. If
the tolerance assigned by this method is less than 1 x 10 then a tolerance
of 1 x 10 moles/Kg of solution is assigned. The variables DPH and HTO are
iteration parameters used by NPH. For their use, see subroutine NPH.
SPECV
Subroutine SPECV is the main subroutine used in option 3 to calculate
temperature and liquid composition from specified pressure, vapor composition,
plus RCOOH and/or caustic in the liquid. The steps of this subroutine are not
too obvious, so details of the calculation procedure will be discussed here.
Temperature iteration and pH iteration are done outside of the subroutine,
so the subroutine calculates pressure and liquid composition from temperature,
pH, and a specified vapor composition. This is d ne by first estimating the
partial pressure of water in the vapor phase using the vapor pressure of
water and Raoult's law as follows.
PW = (HW) x (XW) (29)
where PW = water partial pressure
HW = vapor pressure of water
XW = mole fraction water in liquid phase; initially
assumed to be 1.0
The partial pressures of the other components are then calculated from the
mole ratio of the components over water times the partial pressure of water.
(YA, YC. or YS)
(PA, PC, and PS) = [YW]x
where PA, PC, and PS = partial pressures of NH-, C0?,
and H2S, respectively
YA, YC, or YS = vapor mole fractions
YW = water mole fraction
PW = water partial pressure
The total pressure P is then calculated as the sum of the partial pressures;
and the concentrations of free NH3 (CAS), C0? (CSS), and H?S (CSS) are calcu-
lated from their partial pressure divided by the Henry's constant of each
component. These Henry's constants depend on the composition of the liquid
phase so this computation involves an iterative procedure where Henry's con-
stants computed from the liquid composition. This procedure could diverge
instead of converge, so each new partial pressure is assumed to be the average
of the new computed partial pressure and the old computed partial pressure.
This technique requires a minimum of ten iterations to achieve an accuracy of
j^..0.1%; so it uses more computer time in order to avoid possibility of
42
-------
diverging instead of converging. A maximum of 100 cycles is specified in
the subroutine for convergence; if this number is specified in the subroutine
for convergence; if this number is exceeded, the subroutine writes to unit 5
a warning signal that 100 cycles are exceeded. If this occurs, one may want
to give the old partial pressure more weight than the new one so as to improve
convergence.
Once the concentrations of free NH-, C02, and H?S in the liquid have
been calculated for each iteration cycle, then the concentrations of all
species concentrations in the liquid phase can be computed according to the
chemical equilibria summarized in Table 2. Symbols used by the subroutine
for each species present are summarized as comment statements at the beginning
of the subroutine in Table 11. Once these concentrations have been computed,
then the mole fraction of water can be recomputed and then the iteration cycle
is repeated, iterations are continued until the new computed partial pressure
of each component equals the old computed partial pressure within a tolerance
of +_ 0.1%. In each iteration cycle, the subroutine allows for any RCOOH or
caustic specified to be in the liquid phase as input data to the subroutine.
NPH
Subroutine NPH estimates new pH values based on information gained from
previous pH iterations* This subroutine uses the requirement of electrical
neutrality as the determining equation for either "increasing or decreasing
the pH. The equation for electrical neutrality can be written as follows.
I C. 2,. = 0
where C. = concentration of component i in moles/Kg
of solution
Z. = electronic charge
In general, for a randomly selected input pH, the electrical neutrality
summation will not equal zero. In this subroutine, this summation is repre-
sented by the symbol HT. In order to bring to zero, the step length for a
new pH value is computed from the following equation:
ApH = 0(k) x (HT)/|HT| (32)
where pH = computed pH increment, DPH
k = a proportionality constant, HTO
HT = electrical neutrality summation
|HT| = absolute value of HT
If HT changes sign compared to a previous iteration, then the proportionality
constant k is increased by a factor of two, and the iteration is continued.
By this procedure, the pH increments are only determined by the algebraic
sign of HT compared with previous iterations. Thus, when HT changes sign,
then the increments are reduced by a factor of two. DPH and HTP are intially
set to unity by subroutine PHST. This convergence method is slow, but
43
-------
dependable. Other faster methods could probably be devised to speed up this
calculation.
SPECL
Subroutine SPECL is similar to SPECV in that the pH, temperature, and
composition of one of the phases is given and the pressure and composition of
the other phase is calculated. In the case of SPECL, the total amounts of
NH3, CCL, H S, H?0, RCOQH, and caustic in the liquid phase are given; and the
composition of tne vapor phase is calculated.
This subroutine is used for both options 1 and 2 of the main program.
The method of computing the concentrations of each individual species in the
liquid requires a knowledge of the chemical equilibrium constants which in
turn are dependent on the concentrations of the individual species present.
Thus, an iterative calculation procedure is required where the ionic strength
ELI is initially set to equal the total NH- concentration CA in the main pro-
gram. Subsequent iterations then give better values for the ionic strength.
The calculation method used in this subroutine is based on the calculation
method discussed in the previous section of this report on the SWEQ model;
equations 7 to 22. Because of H2NCOO~ formation, equations 16 and 17 are not
used to solve for a and &; instead a is used as an iteration parameter along
with ionic strength. Initially a (Fortran symbol AL) is assumed to be the
total C02 concentration in the liquid; equations 14 and 15 are then used to
calculate the concentration of free C02 (CSS) and of hLNCOO" ions (BT). These
are also listed as equations 1 and 2 in Table 2. The equilibrium NH. con-
centration in solution is obtained by simultaneously solving equations 3 and
4 in Table 2 by algebraic methods to obtain the equations for NHd~ concen-
tration (DE) and H2NCOO~ concentration (EPS) used in this subroutine. Iter-
ation is continued until the sum of all C0? species equals the amount of CCL
in the liquid from the starting composition. If the sum of the species
concentrations is higher or lower than the starting composition, then AL is
proportionately changed by multiplying the old AL by the ratio of starting
composition over the sum of the species concentration as follows.
(33)
CC = C02 starting concentration
CCST = sum of CC"2 species concentrations
After this calculation, the concentrations of (HS~) and (S=) ions are cal-
culated using equations similar to equations 16 and 17 in the section on
the SWEQ model. The actual equations involved are equations 5 and 6 in
Table 2. These can be solved to give the following:
44
-------
=
(34)
+) + k5 + W(H+0
(35)
In the subroutine, these have the following symbols:
Fortran Symbol
Y GA
4- SD
k5 EKS
ke EKSB
Mr CS
H* HP
After this calculation, the only species left are from RCOOH and from
water dissociation; these are calculated from equations 7 and 8 in Table 2
where SI represents the extent of water dissociation and AET represents the
extent of RCOOH dissociation. From the calculated species concentrations,
the ionic strength can be calculated and iteration is then continued until
the sum of CO- species equals the COp in the starting composition within
+_ 0.01%. When this test is satisfiea, the subroutine returns to the main
program.
PRESY
This subroutine computes the partial pressure of NHo, COp, H2S, and
water from equations 4, 5, and 6 given in the section on the 5WEQ model.
To do this, the individual species concentrations of C02, H2S, and NH3 repre-
sented by CCS, CSS, and CAS are computed from equations similar to equation
19 in the section on the SWEQ model. In order to calculate the partial
pressure of water, two quantities are first calculated in the subroutine.
These are the total number of moles of water, TNW, present in 1 Kg of solu-
tion (RHO = 1) and the total moles of all components, TNM, in 1 Kg of solu-
tion. TNW is computed from the residual weight left after subtracting the
weight of NHV C0?, H?S, RCOOH, and caustic respectively from the 1000 grams
of solution divided by the molecular weight of water. TNM is calculated by
summing the moles of all species present including water in 1000 grams of
solution. The partial pressure of water is then computed from the vapor
pressure of water, HW, times the moles of water over the total moles. The
total pressure is then calculated as the sum of the partial pressures, and
the vapor mole fraction of each component is calculated from its partial
pressure divided by the total pressure.
NTEMP
Subroutine NTEMP is used to estimate the correct temperature for an
equilibrium calculation where the total pressure is specified. This occurs
45
-------
in options 2 and 3. For this purpose a simple equation is used as follows
(36)
1n/PSpecified V _gooo / 1_
l^calculatedy \J Rnew
Nold
where -9000 corresponds approximately to the heat of vaporization of water.
-9000 = AHvap = 18,000 Btu/lb mole (37)
1.987 1.987
The above equation can be solved for T°Rneu) to give the following:
I_ „ _I_ -lAsM /9000 (38)
pcalc
In the subroutine, these have the following symbols:
TR - f Rnew (39)
PSI ' P (40)
TRI = T°RQld (42)
This subroutine also computes the temperature in °C from TR before returning
to the main program.
Tables 5 to 15 represent a total of ten subroutines used by the main
computer program. A large bumber of subroutines are used in order to make
it possible to devise various options in the main program. Many options are
possible; an attempt was not made to develop programming for each possible
option because of the large amount of programming required. Instead, three
options were programmed which demonstrate the use of the subroutines. Thus,
flow charts for options 1, 2, and 3, given in Figupeg 2, 3, and 4, primarily
involve the use of subroutines with some programming done in between to satis-
fy the requirements of the option. In option 1, the main iteration is to
calculate the pH. When the pH is specified, then the iteration changes
slightly to calculate the amount of caustic necessary to achieve the speci-
fied pH. This method of calculation occurs in all three options. Distilla-
tion options 2 and 3 involve a second iteration loop besides the pH iteration
loop. This is necessary to find the correct temperature at a specified
pressure. An example of input and output data for the computer program
listed in Table 5 is given in Tables 16 and 17.
46
-------
Table 16 explains the data format to be used in entering data to the
computer program. The information in this table must be studied carefully
before using the computer program. Table 17 gives an example of computer
output from data specified in Table 16. This listing is self explanatory.
The next section of this report gives a numerical example of calculations
necessary for an actual design problem and a subsequent section gives data
comparisons and evaluations between calculated and measured data.
47
-------
TABLE 16. INPUT DATA FOR SAMPLE PROBLEM
WITH SWEQ COMPUTER PROGRAM
Parameter
Option numbers for calculating
liquid composition and temp-
erature from a specified vapor
composition and pressure
Weight percent carboxylic acid
in liquid
Weight percent caustic in
liquid
Temperature, °C (For Option 3
this is used as a starting
temperature)
Pressure, psia; specified
pressure
pH, a positive entry specifies
the pH for the calculation.
The computer program will
adjust the amount of caustic
in the liquid independent, of
the concentration entered
above when a positive pH is
entered
Weight percent concentrations
in the vapor phase
NH
co
Symbol
NDOPT
XSA
XCAUO
TC
PS I
PHO
Value
Entered
.05
.05
100
20
8.5
XA .01
XC .01
XS .01
XW 100
(continued)
48
-------
TABLE 16. (continued)
Format for data <
3
NDOW
100
sntry:
.05
XSA
20
PST
.05
XCAUO
8.5 .01
PHO" "XT
.01 .01 100
"XT ~XS~ "XT
Additional lines of temp., pressure, etc. can follow
111—1 111 This entry will signal a new option line to
follow this one.
NDOTT XSA" XCAUO
Then lines of temp., pressure, etc.
49
-------
TABLE 17. COMPUTER OUTPUT FROM DATA IN TABLE 16 WITH
COMPUTER PROGRAM BASED ON THE SWEO MODEL
TEMPERAIURE
PRESSURE
PH
lOfl.88
20.00
8.500
c,
PSIA,
287.99
1.361
f,
ATM,
382
137
.03
.90
K,
687.66
R
K-PASCALS
COMPONENT
AMMONIA
CARBON DIOXIDE
HYDROGEN SULflDE
WATER
CARBOXYL1C ACID
SODIUM HYDROXIDE
TOTAL
WEIGHT
LIQUID
0.00091
0.00017
0.00073
99.93077
0.05000
0.01742
100.00000
PERCENT
VAPOR
0.01000
0.01000
0.01000
99.97001
100.00000
MOLE PERCENT
LIQUID
0.00096
0.00007
0.00039
99.97572
0.01501
0.00785
100.00000
VAPOR
0.01058
0.00409
0.00529
99.98004
100.00000
K-VALUE
11.02807
57.78305
13.62491
1.00004
50
-------
SECTION 5
SAMPLE PROBLEM USING THE SWEO MODEL
Information given in the two prior sections of this report on the SWEQ
model and on the computer program based on the Sk'EQ model can probably be
better understood by giving a numerical example which shows the calculations
necessary in an actual sour water stripper design case. For this purpose
the following sample problem ' is given.
A refluxed sour water stripper operates at a condenser temperature of
212 F at a pressure of 8.7 psig (23.4 psia). To achieve the desired removal
of HpS and NH3 from the stripper feed, the overhead gas from the condenser
must contain 48 Ib/hr of NH3 and 49.7 Ib/hr of H?S. Determine the amount of
water in the exit gas, and the reflux composition.
Gas rates
NH3 = 17.03
= 2'82 mole/hr
mole
HS = 49.7/34.08 = 1.46 mole/hr
From Raoult's Law, the partial pressure of water in the vapor phase is:
p.p.(H20) - (V.P.iyj) . X^
Assume that X = ^ At ^ v.p _
p.p.(H20) = 0.9 (14.7) = 13.2 psia
The partial pressure of (H9S = NH-) = 23.4 - 13.2 5,10.2 psia, therefore
the total moles of overhead gas =J(2.82 + 1.46) x irr = 9.82 moles/hr.
Assumed water rate is 9.82 - (2.82 + 1.46) = 5.54
the assumed vapor composition is
le/hr. In summary,
a^This sample calculation is given through the courtesy of Ron Gantz
and co-workers of CONOCO who did the numerical calculations and wrote this
sample problem. It has been checked at Brigham Young University and found
to be correct.
51
-------
1 b/hr mole/hr mole/fr. p_.p. , psia
NH3 48 2.82 .287 6.72
H2S 49.7 1.46 .149 3.49
H00 99.8 5.54 .564 13.19
* "9782 T7DT50
Calculation of the liquid composition in equilibrium with the assumed
vapor composition involves simultaneous solution of the appropriate chemical
equilibria and phase equilibrium equations. The chemical reactions (Tables
2 to 4) are:
NH, + H+ t NH.+ kNH
3 * 4 NH3
H+
+ JP J
HS" ->• S + H k - =
Mi j 5 f H KHS-
The chemical equilibrium constants are correlated in the general form (Table
4).
In ki = A + B/T + C/T2 + D/T3 + E/T4
where T is the temperature in °R.
At 100°C (= 671. 76R),
In k =39 5554 - ^ 77822 1..843-108 .8541 -IP11, 1.4292 • IP13
in k 39.5554 -r- + ~ (671 .7)3 + (671.7)4
In kNH = 1.587 + 11160/671.7
»J
k = 8.032 • 107
Ink - -293 88 + ^83§5g 6!?7T25/J°8+ 2-5551 ' ^^ 3.91757 • IP13
in ic ^3.88 + + (671.7)3 --- (671.7)4
kH $ = 2.805 • 10"7
52
-------
In k - fi<57 QfiR + IlilM 15.8964-IP8.6.72472- 1011 10.6043-I O13
in KHS ot>/.yb& + -^7T-7 (671.7)2 (671~7)J (67i;7)4
kHS" = 9.06 • 10"13
For phase equilibria, the Henry's Law coefficients from Table 1 are used:
In (HNH ) = 178.339 - 1591 - 25.6767 In (TR) + .01966 • (TR)
-------
From the Henry's Law coefficients for H?S and NFL, and the assumed vapor
partial pressures, the free H2S and NH3 concentrations in the liquid can
be calculated:
3
C$s(free H2S) = ^^ = 7.89 • 10~3 gm-mole/Kg
A pH must now be assumed - use 8.5
A pH must now be assumed - use 8.5
pH = -log10[H+]
[H+] . e-2-303 - PH . 3J5 . 1Q-9
The chemical equilibrium equations can now be solved for the concentrations
of all other species in solution:
. .
H3 [NH3JLH*J
] = [NH3][H+]
[NH4+] = (1.85) (3. 15 • 10"9) (8.032 • 107) = 0.468 gm-ions/Kg
k
K
H2S
[HS"] = kH2SH2S [H9S] = C
IJFT
-i - (2.805- 10"7)(7.89- 10"3) n 7n
] = - 3.15 .10-9 - = °-70 gm-ions/Kg
HS
54
-------
[S=] =
LHJ
)= 2.02.10-4gm.1ons/Kg
If the assumed pH and values of C.., (L- are correct, the solution should
be electrically neutral, that is: b5
[H+] + [NH4+] = [MS'] + 2[S=] + [OH
3.15 • 10"9 + 0.468 = 0.68 + 2(2.02 - 10'4} + 1.6 • 10"4
0.468 * 0.68
A trial and error procedure for pH, with successive substitution for species
concentrations at each pH level, must be used to reach a converged solution.
First, adjust the Henry's Law coefficients for the current values of H«S
and NH3 concentrations:
""'i*. J682 (CAS) + .06 Cs
In (HNH ) = In (3.640) +
where GS = total H2S = C$s + [HS~] + [S=]
» .00789 + 0.68 + 1.96 - 10'4 - 0.688 9
In (HMM ) = 1.292 + N!TT - .16821(1.85) + .06(.688)
wi, 671.7 I
'3
HM,, = 3.99
gm-mole/Kg
In (HH s) = In (442.5) =-.05 CA$
H = 403.4 psia
HH2S gm-mole/Kg
Calculate new free H2$, NH3 concentrations
c = P-P-(NH3l = |^| = 1 .68 gm-mole/Kg
• H S J-yy
c = fiiEiiifei) = 3.49 = 8i65 . 1Q-3 g
55 H,i q H-UO.f
55
-------
The total anmonia concentration, C^, is
CA = CA$ + [NH4+] = 1.68 + 0.468 = 2.15 gm-moles/Kg
The total H^S concentration, C<., is
CS = CSS + tHS~J + CSJ = 8'65 ' 10" + °'70 + 1<96 ' 10~ = -709
A new water mole fraction in the liquid should be calculated for use with
Raoult's Law to provide a new vapor composition.
A free water concentration, Cu n
C
H20 f - CS ' ^ • wt- H2S> - CA '
C 0 = [1000 - 0.709(34 08) - 2.15(17.03)] _^
£• lo.Od
r - co o gm-mol e
LH20 ~ ^-- K^
The free water mole fraction in the liquid is
XH90 = H2°
L ZCi
SCi = [NH3] + [NH4+] + [H+] + [H2S] + [HS~] + [S=] + [HgO]
= 1.68 + .468 + 3.15 • 10"9 + 8.65 • 10"3 + 0.70 + 1.96 • 10"4
+ 52.2 + 1.6- 10"4
= 55.06
p.p.(H20) = (0.948)(14.7) = 13.9 psia
p.p..(H2S + NH3) = 23.4 - 13.9 = 9.5 psia
Total moles in the vapor = 4.28-^t= 10.54
56
-------
Moles of H20 vapor = 10.54 =-4.28= 6.26
The new vapor partial pressures are:
P-P.(NH3) = 1^- • 23.4 = 6.26 psia
P.P.(H2S) = i4jj_ - 23.4 = 3.24 psia
t— \ \J •
-------
increasing the value of H would increase NH. and decrease HS .
Thus, the correct pH must be lower than the initial assumption of 8.5.
If [H+] = 4.15 • 10"9, and [NH3], [H2S] are assumed constant, then
This is close enough to use for the next guess
nH = ln (4.15-IP"9) _ R
PH -2.303 8'38
Using this pH and the current values of free H?S, NH~, begin again the
successive substitution procedure for species concentrations and continue
until a final solution is reached (achieving electrical neutrality). In
most cases, several more pH trials may be required.
The final solution is
pH = 8,38
Vapor Composition Liquid Composition
Ib/hrmole/hr wt. fr. mole fr.
NH3 48 2.82 .036 .0384
H2S 49.7 1.46 .018 .0096
H20 114 6.33 .946 .9520
58
-------
SECTION 6
COMPARISONS AND EVALUATIONS BETWEEN
CALCULATED AMD MEASURED DATA
Information in this section will be discussed in the following order:
a) Evaluation of Van Krevelen prediction model
b) Evaluation of SWEQ prediction model
c) Evaluation of new NH3-H2S-H20 and NH3-CO?-H2S-H?0 data
d) Ammonia fixation by carDoxyfic acids and release of
NH3 by addition of caustic
These subjects will be discussed by frequent referral to data summarized in
Tables 18 to 27 which contain comparisons between calculated and measured
vapor-liquid equilibrium data. Not all literature data were examined in
this project because of the limited scope of the project. However, an attempt
was made to examine as much data as possible. Table 28 summarizes various
references collected during the project. This table also indicates the type
found in each reference and whether the data were used for modeling purposes.
Tables 29, 30, and 31 give summaries of deviation errors between calculated
and measured partial pressures in Tables 13 to 27 for NH,, C0?, and HpS
respectively.
In developing the SWEQ model, some individual experimental points have
been ignored and some entire data sets have been ignored. As a general policy,
individual experimental points in a given set of measurements have been ig-
nored in developing the correlation model when deviation errors from these
points appeared to be radically difference from the main set of data. Entire
sets of published data were ignored when deviations appeared to have little
or no definite pattern and were also very large. When this has occurred, the
data ignored and reasons for ignoring are noted at the bottom of the table.
Evaluation of Van Krevelen Model
The Van Krevelen prediction model which applies when NhL/hLS ratios are
greater than 1.5 was derived by Van Krevelen et al'' from low temperature
data. These are compared with Tables 20 to 27 where columns headed VK repre-
sent predicted partial pressures from the Van Krevelen model ' and columns
headed MEAS represent measured partial pressure data. The following is a
summary of the various comparisons.
listing of the Van Krevelen computer program is given in the Appendix.
59
-------
TABLE 18. H2S-H20, NH3-H2S-H20, AND NH3-H20 SYSTEMS, COMPARISON OF CALCULATED AND
MEASURED DATA OF MILES AND WILSON21 AND OF CLIFFORD AND HUNTER17
Temo.
•c
89,
88,
8B,
80,
120,
120,
120,
150.
150,
150,
80,
80,
*e,
•»!
Holes/Kq of Soln.
HH,
— 3-
.900
,VB0
.000
.HUB
\Hl
.009
.900
,009
,010
.994
»339
I"'
!s2s
*«', 7.399
80, 4,535
C$2.
9,008
0,000
0.000
9,000
0,000
8,000
0.000
!00o
,000
,000
,000
,0t>0
.000
0,000
0.000
9,1100
0.000
80, 13,7*5 0,fl(i0
80, 14.358
120, (
120, 1
1,515
,B'2
120. 0.959
120. 0.415
120. «
120, 1
,519
.933
120, 8.H99
120. 0.011
60, 3
,452
69, 8.B55
60. 13.155
60, IS
• 3*2
60, 23.*00
60, 27.066
ice, 3
,452
100, 8.675
tee. 12.358
100. 17,386
140, 0,555
l«0, 3.285
140, !
.692
I8B. 7,829
.000
,000
,000
.000
,009
,{100
,000
,000
,«0a
,c 00
0,000
0,000
8,000
0,000
8,000
0,000
0,009
0)008
8,000
0,000
0,000
0,000
H£
9,054
»,297
0,5119
0.515
0,047
9,2190
9,341
0,044
0.136
9.232
1,358
9.925
4,157
1,066
5,553
4,250
4,340
1,015
2,831
3,528
0,696
0,946
0.077
0,178
0,108
0,413
1.698
0.001
0,000
8.000
0.000
B.flBfl
0,000
0.000
0,000
0,800
0,4)00
0,000
0,080
0,000
0,000
0,000
VK API MEAS
Partial
YK
t
Pressure
H,S
API
IB68.6
5932,6
K1M.I
10297^7
0,0 5,
lit
85,
137,1 |27,
200.
296,
342,
384.
776,
775,
170,
91 ,
145,
505,
1605,
3.
181,
495'
367)
365,
6611,
1229,
HI6.
1491 ,
1530|
399,
479,
442,
2023,
2099,
24,
106,
174,
87,
135.
521.
2583.
3,
214.
634,
1342,
2705,
4178,
3610,
711,
2065,
3255,
5258,
294,
r 1788,
2389,1 3166,
2808,2 4435,
12.1
72,4
136,5
208,9
465,3
597,7
574,4
2373.9
2285,1
23,8
18215
81,2
177,3
568,7
3179,5
2.
241.
627.
1442.
2957.
4559,
6079,
770,
2311,
3867,
6979,
299,
1825,
3376,
«92T,9
349,0
1225,
1056,
99,
172,
233,
1024,
285,
65,
246,
762,
8,
1166,7
4932,2
8422,4
1177,5
3637,5
6223,5
6160,6
2399,0
8670,7
3««>.8
7652,1
2244, T
19b
-------
TABLE 19. NH3-H2S-H20 SYSTEM, COMPARISON OF CALCULATED AND MEASURED DATA OF TERRES
29
Temp.
IS'
20,
20,
20.
20,
40.
40.
40.
60,
60,
60,
Holes/Kg of Soln.
NH3_
0.M0
1.574
3.975
4.T27
5,526
1.356
3.453
4,234
5,484
1,891
3.253
3.991
5.205
C0j_
*,0f)0
B.PVU5
0,ian0
0.000
0.000
e|0s>0
0.000
0,000
H2i
0.411
0,783
1.998
2,356
2,758
1.719
2,139
2,746
0,954
1,631
1.998
NH3
VK API.
a
8,
22!
25.
52)
4,2
a. a
26,1
33.0
40,7
52,4
66.7
65, 93,9
57,0 55. 1
94, 8 1B2.6
114,1 131,7
145,0 186.7
Partial
HEAS
0.0
0,0
0.0
5,6
13.0
8.2
22,2
26,9
30,4
25,5
48,5
76,1
145,0
Pressure,
VK
4,9
8,3
16,8
18.1
19,6
59,7
69,4
123,'z
181,1
205,1
240,2
mm Hg
H,S
API
5,6
10.1
24,6
28,0
32.3
67,8
84,2
104,2
130,4
213.2
255,4
131,2
WAS
5J
65,
UK.
190,
till'.
293,
134,
2«5,
250.
365,
-------
TABLE 20. •NH,-H«S-0,SYSTEM, COMPARISON UF CALCULATED^.MEASURED
n3""2
DATA Of VAN KkEVELEN, ET AL.
VIC API WAS « API fCAS
21, *,}fll 8,01)8
24, 8,5*5 4.0P1
21, l.7»8 », t»a
24, 2.358 D.reo
21, 2.121 8. nee
21, 1.215 t.atiB
21, 1,565 8. ewe
28, l.|18 9, line
21, 1.788 4.01)8
21, 2,1'B 8, nee
28, 2,110 0,8(19
21, 0.268 0,008
21, 1.548 0,888
28, 1,175 8.0(10
21, 1,781 1.000
21, 2,268 1,880
21, 2,661 1,101
28, 1.260 8,888
28. 0.575 1.808
26, |,|58 1,888
21, 1,730 8,8ft8
II, 2,318 0,880 I
20, 2,120 8,080
21, 1,231 1,0(10
21, 1,545 1.888 f
28, 1,068 8.080
21, 1,720 8,001
21, 2,271 1,188 <
21, 2,870 6,*B0 I
28, 1,468 8,000 I
40, 1,198 0.001 1
41, 1,515 8. Bill f
41, 1,718 1.888
48r 2,350 0,000
41, 2,120 0,000
48, 1,215 0,000 t
41, 0,595 0.0B0 i
41, 1,111 0,881 i
48. 1,748 1,880 <
41, 2.4B0 B,8B1
41, 2,131 0,000
41, 1,241 8,000 1
41, 1,561 1,000 1
40, 1,175 1,100 1
10, 1,710 0,080 I
41, 2,261 1,000 1
41, 2,460 1,300
10, 0.260 0,000 1
II, 0,575 8,008 1
48, 1,150 0,800 1
41, 1,138 0,880 1
40. 2,310 1,080 )
40, 2,120 0.000 (
40, 1.260 1,888 I
41, 1,595 1,800 1
40, 1,161 1,881 I
41. 1,728 1,008 1
44, 2,270 0,000 1
10, 2.471 1,0m 1
10, 1,468 0.000 1
tl, 1,388 1,103 (
tl, 1,515 1,180 I
II, 1,710 1,880
tl, 2.351 1.080
t8, 2.128 8,001
68, 1,215 0,«B0 (
t8. 1,565 1,0110 1
tl, 1.1'" 1,8(8 1
10, 1,768 l,Bjigl 1
tl, 2,4*8 0,808
tl, 2,13* 0.8110
68, 1,26* 0,001 !
10. 1,548 0.RB8
II, 1.175 8.001 I
tl, 1,710 1,808
tl, 2.261 a.naa t
tl, 2,468 0,804
tl, 1,290 4, ana 1
tl, 1.575 *.9e8
tl, 1.158 0.8311
t8, 1.718 0.»«0
tl, 2.318 .Kxa
tl, 2,128 ,8fl«
11, 1,248 .""'
64, 1,545 ,•»*
tl, 1,168 ,tet
it, l.»2* ,841
tl. l.2*« API
tt, i.«'i ;«»»
tl, !.•«• «.««
1.145 1.
8,575 2,
1.165 1,
1.508 |,
,»«8 10,
1.148 1,
1,215 3,
1,541 6,
1,861 1,
1.158 12.
l.«48 15.
1. 118 1.
1.221 3.
>,455 7,
1.618 It,
).675 14,
1,118 17,
1.878 2.
«.145 4.
1,218 9,
>.4JS t3,
1. 565 17,
>.738 21.
1.868 2,
1,128 S,
1,220 1,
1.355 13,
l.«48 16,
1.558 22,
1.518 26,
>. 145 3.
1,375 t.
.165 16,
,548 21,
.618 27,
,148 4,
1,245 1,
1,568 16,
1,868 24,
.150 32.
.448 34,
1,118 4,
1,228 9,
1.455 19,
1,619 21,
1.475 35,
.118 44,
1.870 5,
1.145 It,
1.210 23.
1.415 11,
1.565 41.
1,710 54,
1.868 6,
1.128 12,
.228 22,
1,355 35,
1.448 16.
1,550 56,
1,510 6*.
I.US 7,
>.37S 13,
.165 38,
,540 49,
.410 62.
1,140 9,
1.245 18,
1.560 17.
1,460 56.
,15.0 74.
,448 17,
1,110 10,
1.220 22.
1,455 44,
1,610 65.
1.175 42,
I. 110 101.
1,870 11,
1.145 26.
1,218 52,
l,4!5 77,
1,565 |0*.
I.7I0 12).
>.«l«,8 1).
>.I28 21,
'.228 51.
1.155 18,
1 • *•)• 1 uC
• ™ !•* » *
••**• 118,
».'•"• 156,
1.2 0, 3,6 1.1 4.1
2,1 1, 7,4 1.5 1.
7,1 0, 21.5 21.1 11.
4,5 1. 26.1 37.6 27,
12.5 8. 24,6 44,| it.
1,4 0, 1,5 1,1 1.
3,1 8, 1,2 l.t 1.
4,6 8, 6,1 7,2
18, 0, 1,5 10,4
14. 0, 10. S 13,4
11, 0. 11.0 17,4
1. 1. 0,9 1.0
3, 0, 1,6 1,1
7, 8. l.S 1,7
12, 0, 4,1 5,1
16, 1, S.I 7,1
21. 1. 7.8 1,7
2, 0. 1,1 8,3
4, 0, 1,6 0,1
9, «. 1.2 1,3
14, 0 1,7 1,9
2«. 0 2.2 2.5
27,
2.
4.
9,
15.
21,1
21,1
37,'
2.1
5,1
17, i
23, (
31.3
s,<
7.<
16,1
25,
35,
43,
4,
9,
18.
29,
38,
58.
5,
11.
22.
35,
4l!
64,
5,
11.1
-I.I
J6,'
51,!
67,1
46,1
t,
12, <
37, <
50,
t5.
1.
16. '
34. <
53.1
75.
92.
9,
19.'
40, <
61, ,
12.'
107.1
II, i
23.
41. «
74,"
102,1
134,,
11.
25,
46.1
'4l«
la. •
'•4.1
»«l.1
14*, <
• 2,6 1,1
t.
1.
12.
IS.
1.
1.
1.
4.
6,
t.
1.
1.
1.
1.
2,
2,
8 8,2 1,2 1.
* 1.1 0,4 0.
8 0,7 0,7 1,
8 1,1 1,2 1,
1 1.2 1,1 1.
1 1 1,5 l.t 1,
I 1 1,1 1,1 I,
8 11,1 12.7 13.
> • 29,5 27,1 26,
! 8 15,2 11,1 1.
1 1 106,1 111,9 1,
0 117.2 140.7 125,
1 8 1,1 5,5 0,
» 12,7 11,6 12,
> • 24.4 23,2 21,
8 i;,6 11,3 32,
0 41,6 41,2 43.
8 51,6 56,1 54,
0 1,5 1,1 1,
8 6,4 5,1 6,
8 11,8 12,1 12
8 11,9 11,8 11
0 21.8 22,5 22.
1 27.9 21,1 11
8 1.1 1,8 1
8 2,4 2,1 I
8 4,6 4,1 4
0 k-7 6.8 6
r, • •, • ™
0 s.f, 0,f 1
0 10,4 9,5 19
0 1,1 8,7 1
* ',5 1,1 1
0 ;,7 2,4 2
8 1.1 1.1 4
> 0 4,8 4,2 5
' * 5,1 5,1 5
0 5,4 4,6 1
I • 44,2 40,1 19
• * 11.3 a*> T •
* • t * •*» « f 9
8 261,9 246,4 182
• 129,5 181,7 0
' 8 163,1 449,6 11*
1 • 11,8 11,1 IS
8 11,4 17,7 15
1 < 7S.6 75,6 66
' « 104,8 186,5 14
• 129,0 140,5 US
' • 159,9 112,1 151
• !•,' ll.l 11
! • 19.1 11.9 21
1 • 40,4 19,6 14
' • 51,5 59.8 54
1 • 71,1 71,6 64
» 0 16.5 92 • ••
• "», * ^« .
0 l.S 1
•* 31
10 714
_ »3 ••
" 14,2 11,
0 >fl Jb t •
• «»!,» It,
• 27.2 26.
0 la 5 ,,
*«•* 91,
1 > < 9
„ 2,5 2,
0 4 f t
* . * 4 .
0 • « •
° 1.5 1.
* 13.2 12.
• 14.9 11.
1. 16, « 14,
*. 16.6 |5,
Jm
m
981
I
5 M
1* M
Z 2%
Ji
3 *
• •
*
_ *
9 '
< i?
« It
* u
* tr
62
-------
TABLE 21. H2S IN AQUEOUS BUFFER SOLUTIONS, COMPARISON OF CALCULATED AND MEASURED DATA OF SHIH, ET AL.
26
Co
Partial Pressure, mm Hq
„ _ Holes/Kg of Soln.
pH
7.00
7 00
/ " on
7.00
7.00
7.00
7,90
7.83
7.80
7.77
•c NH,
60
S00.
IZ0,
140.
tes.
•0.
140'
.l>00
.HUB
,080
,B0B
,000
,1100
,000
|fe«. 0,000
115, 0.M00
CO,
.000
,000
,0P0
.000
,000
0,000
0,000
0.000
0,000
0,000
0,000
0,0(10
H,S
0,010
0,010
0,010
0.010
0,010
0,010
0.010
e,ei0
8,010
0.010
0.010
0,010
H,S
VK API
31.3
60,1
61, B
6a, a
86,4
136.7
10.1
11,5
12,4
15,0
0.0 42.2
MEAS
84. e
98. 0
104. e
ur.0
125.
133,
16.
22!
25,
32.1
-------
TABLE 22. NH,-C09-H90 AND C09-H,0 SYSTEMS, COMPARISON OF CALCULATED AND MEASURED
3 C. C C. f. -i yn
DATA OF VAN KREVELEN, ET AL. AND DATA FROM LANGE'S HANDBOOK^
PtrtUl Pressure, m Hi
Tump MoUi/Kfl of Soln. NH3 CO,
*C NH3_ COj, MjS n API. MEAS VK API HEAS
II, 0,410
20. 1.960
20, 1,990
20, 2, «il>
«•, a, tie
48, 0,*R0
B0 0,liU0 B,
,440 0,B«I0 II,
,543 0,1100 |0,
,)I6 0,0*0 3,
,257 0,000 4,
,t«9 0,000 5,
,623 0,000 S,
,5U 0,000 I,
,026 0,000 |2,
,S«0 0,000 2»,
,0«8 0,000 22,
,319 0.0U0 »,
,19! 0.000 43,
,3b4 0,000 10,
,670 0.0U0 tl,
,J7B 0,000 ai,
,696 0,000 54,
.998 0,000 34,
k8, 2.000 1.3J6 0,008 i7,
0, 0,000 0,076 0,000 0,
20, 0,000 e,Bi» 0,000 0,
40, 0,I>U0 0,024 0,000 0,
J.
9,
12,
12,
2,
4,
S,
4,
7,
13,
30,
!«•
»,
39,
27,
10.
61,
54,
33.
•»,
a,
a.
• .
60, 0,000 0,0i6 0,000 0,0 0,
0, 0,
'. 0.
n. 0,
12, 0,
e. fa.
4, a.
5. 4t,
0, 45,
9,0 Jfl,
II, S 25,
0,0 2,
21.5 3.
7,0 101,
4«,0 «.
29,2 23,
12.0 220,
95.0 4.
61,0 25.
36,0 94,
16,0 379,
a. a 0,
0,0 0.
0,0 0,1
B..
*,
0.
0,
2'i
9,
33,
37,
!«•
21,
2,
3,
95,
4.
22.
203,
4,
23,
90,
355,
»95,
791,
» 718.
0,0 0,« 660,
I 0,
0.
«.
a.
2T.
10.
e.
46,5
JJ;s Van Krevelen
2!
3,
93.
4.
20.
215.
5 .
2'.
66.
394,
T60,
T60,
760.
T60.
Data
Lange's Handbook
-------
TABLE 23. NH3-C02-H20 SYSTEM, COMPARISON OF CALCULATED AND MEASURED DATA OF OTSAKE, ET AL.
22
Temp Holei/Kg of Soln.
J£_ 113_ £22_ HjS VK
20, 0,990 0,820 B.PBB 0,
20. 2.111 W.298 0,1100 |7,
20, 1.120 1,411 0,000 0,
20. 4,|S7 0.SB0 0.000 29.
20, 2,877 1,94} 0.MB0 2,
20, 5,590 2,836 0,0(10 7,
20. 7,1187 1.9|9 0,000 8,
40, 8,712 0,761 0,000 111,
40, 10.714 2.579 0,000 ||6,
•B, 2.719 1,134 B. 000 16,
68. |,al4 1,261 0,000 14,
68, 2,866 0.918 0,000 77,
6«, 5,790 2.200 0.000 111,
60, 6.060 2.004 0.HH0 117,
60, 6,794 3,611 0,000 60,
60, 7.8M 1,495 0.0H0 |B4,
»B, 8.291 2.818 0,000 166,
•0, 2,824 1,218 0.U0I 121,
10. 3.8«6 0.816 0,000 2H2,
IB, 4,a9| 0.405 0.0UI 416,
10, 4,486 1,561 0,000 224,
II, 7,152 0.564 0.000 585,
111, 2.601 1,102 0.000 176.
100, 4,891 0,407 0,000 765,
til, 4,474 1,563 0,000 407,
100, *.1M «.916 0,000 121,
NH}
API
B.
21.
B,
41,
2,
II.
30s!
277.
M!
>2.
mi
87.
159,
269.
122,
305.
527.
252,
885.
185.
924,
473,
H99,
Partial
HEAS
B.1
20,
B,
48.
3,
M,
15,
285,
322.
19.
22,
177,
351,
364,
101,
311.
485,
108,
297,
490,
21 4,
802.
157.
921,
427,
1022,
Pressure, pro
VK
> 96,
>
f
3 ',
,
,
1,
21,
447,
3'!
21.
207,
74,
22,
223.
27.
3,
115,
2,
1203,
12,
409|
«*.
Hq
CO,
API
r 45,
I B.
86.
0.
28.
5,
B,
B,
20.
410,
22,
16,
21,
198.
22!
267,
33,
118,'
3.
1757.
j9
625,
•T,J
ME/
55.
«.
IH5,
a.
51,
7.
13.
B.
1.
22.
442,
3,
4,
1*.
2T1.
31,
20.
256.
41.
10.
180,
19,
1211.
43,
645.
1 107.
«
t
S
-------
TABLE 24.
SYSTEM, COMPARISON OF CALCULATED AND MEASURED DATA OF CARDON AND WILSON
ON
1 CHIP .
"C
50.
50.
50,
80,
88.
80,
80.
80,
80,
8B,
80.
80,
80,
110,
110,
110,
110,
110.
120.
120,
120,
128,
128,
Ma_
5, 445
10,142
1,511
5,606
1,963
4,543
4,101
1,048
0,542
Pi, 752
0.582
2,692
15,101
2.124
0,552
0,597
5.676
2,645
2.152
0,259
0,853
0.105
0,083
C0,_
1.706
2.665
8.PI55
1.485
0,027
2, IP9
1,999
0,204
0,460
C.137
0,384
0,995
2,7«0
0,754
0.541
0,072
2.511
1,566
0,094
0,231
0,426
0.011
0,828
Mjfl
2,689
3,711
0,172
3.000
0.165
1,774
1,487
0.083
0,093
0.499
0,049
1,117
3,769
1,155
0,115
0,166
1,281
0,786
0,166
0,138
0,306
0,022
0,B83
Partial Pressure, mm Hq
VK
15,5
67,8
49,8
72,4
211,7
30,7
29.5
86,6
0,8
13,5
14,3
42,2
579,9
41,6
0,0
98,4
249,6
«2.2
600,6
0,0
34,7
25,9
e,e
NHj
API
22.3
122,3
49,2
108,5
195,9
52,2
47,4
73,8
4,3
15,3
13.1
49,5
1425,6
112.3
17.7
90,5
409,7
118,8
630,1
12.0
76,6
23, a
«•*
CO, H,S
MEAS VK AP£ MEAS VK A£I_ MEAS
30,
263.
•«,
164,
164.
»5,
127,
1U2,
B.
T,
29,
102,
2176,
285,
15,
10P>,
353,
563,
734,
l«.
93,
1*.
195,1 309.0 239,9 594. 1286,2 1132,2
v 25,2 62,6 54.3 > l|4, 442,3 540,8 ,
aj 0,3 0,3 2,33; 2, 2,4 . s.sa;
, 975,4 1871,1 1613,0 x 1370. 4740,7 3960,2 N
aj 0,6 0,8 7.6aJ 7, 6,6 20,43;
4152.7 4706,4 3453,6 1215, 4267,5 2585,
4024,3 4362,3 3458,7 1117. 3519,8 2517.
26,0 31,9 65,5 12. 13,2 18,
0,0 2656,3 2436,6 0, 351,2 246,
312,5 275.7 346,9 1040, 767,1 589,
640,2 647,7 453,4 57, 60.3 47,
1090,7 1319,0 987.5 968, 1720,9 1354,
35,1 121,0 149,9 60, 626,6 847.
11867.4 7471.6 5976,5 7736, 5470,7 4616,
0,0 10589,9 9|45,7 0, 843,8 8ft.
103,4 176,2 169,1 160, 155,5 172.
» 5724.5 10211,210040.1 320, 3741.9 4425,
3; 26855,9 15998,6 9988,4 2594,0 4186,2 72(07.
25,1 54,6 103,4 48,2 49,2 67,
0,0 7296,7 6860,6 0,0 1269,3 1623,
\ 9596,0 6208,4 5619,8 3036,6 1502,6 1483,
d' 26. a 50,2 SB, 3 24,3 22, a 22.
46,6 e.B 899,5 752. a 0,0 787,3 563,
a) These points ignored in computing averages because they represent extreme deviations which
are probably due to major error in the measured values. In the third and fifth runs, the
entire run was suspected, so the entire run was ignored.
-------
TABLE 25.
SYSTEM, COMPARISON OF CALCULATED AND MEASURED DATA OF BADGER AND SILVER
" -
Temp.
28,
20,
28,
28,
20,
20,
20,
28,
28,
20,
28.
28.
28,
28.
28.
Moles/Kq
ffis-
.189 ft
.191 0
,390 0
,B«5 |i
,380 0
,W97 VI
.192 0
.196 0
.192 0
,193 0
,«92 0
,088 0
.1)95 0
,088 0
.095 0
of Soln
iz_
,410
.'90
,195
.500
.638
,660
.670
,686
.700
,T45
,770
,790
,800
.815
.816
t
HgS
0.189
B.194
0,390
0.D«5
0.360
0,097
0,192
0,196
0,192
0.193
0,092
0,066
0.095
0.066
0,095
NH3
VK API.
3, 4,0
2. 3,0
?, 3.0
2, 2.8
1. 1.7
1, 1.5
1, l.«
I, 1.3
1, 1.2
0, 1,0
0, 0,8
8, 0.7
8, 0,7
8. 0,6
8, 0.6
MEAS
4.1
2,9
2j7
r,3
U3
0,9
0,9
0,7
0,0
0,0
0,5
0,b
Partial
VK
l.«
j »
42
3,6
13,2
I •, 1
16,0
18,6
20,6
30,2
• 3,4
47,6
55,1
56',9
Pressure.
CO,
API
I,T
3,*2
3,9
3.0
12,1
It,*
13,6
15,0
IT.9
25,7
31,3
36,1
00,3
46^4
no Hq
MEAS
1.5
3,5
37
3.7
13,1
12.1
19*0
20,5
29,2
35,2
42,4
45,1
0,0
0.0
VK
3.6
I3*,7
1,1
26,9
6.1
13.9
15.8
16,6
22,1
11.4
12,5
15,2
16,9
H,S
API
3,6
5,4
1«,1
1,1
26,7
5,7
13,1
15,0
15,7
20,'S
10,
11,
13,
13,
i«;
MEAS
3, .
ie'.
" 0.
27,
9,
II,
15,
15,
27,
12,
8,
0.
13,
16.
-------
TABLE 26.
SYSTEM, COMPARISON OF CALCULATED AND MEASURED DATA OF BREITENBACH AND PERMAN16'23
PtrtUl Prestur«. an Hg
oo
Temp.
21.
20,
2«,
20'
201
20,
20|
20,
20,
20,
29,
40,
40,'
68,
68,
60,
60,
*»,
"I
Holei/Kq of Soln.
1.174
1,468
1.762
2.149
2.916
4,«04
S.872
a.a«a
11,744
14.680
IT.616
29|}40
15,212
41,104
46,976
1.762
2,916
5,872
11,744
IT,616
29.160
0!s87
0,940
1.468
2,916
4.404
5,872
8,801
11,744
.W08
,»Dfl
!e«0
,000
,000
,000
,000
,000
,000
.000
,000
.000
,f)R0
IHBB
.(100
tvnv
,000
,000
,000
|ae0
,000
,000
.800
,000
.001
,000
!o00
,000
,000
,000
.000
,000
,000
,000
,000
.000
.0U0
,000
.000
1080
.VH0
,0110
,000
,080
.000
,000
,000
,000
|000
,0tJ8
,000
,000
VK
12,
14,
l>,
22,
21,
17,
•5,
»».
65,
69,
79.
66,
5',
50,
42,
]4 ,
44,
69,
II',
167,
(79,
151,
108,
16,
51,
86.
159,
219.
268,
1«0.
181,
API
12.
H,
20,
28.
It.
*2,
91,
177,
100,
476,
724,
1551,
1117,
6011,
11274,
20101,
•T,
20?)
608,
1115,
4769,
14108,
Jl ,
51,
82,
177,
287,
411,
720,
1119,
MEAS
12,
15!
18,
24,
11,
58,
114 ,
166,
227,
298,
478,
6»6,
945,
1178,
1458,
fl 5 •
76,
167,
195,
612,9
1520,0
2760,0
10,2
«8,7
11,
I6S,
261.
161,
581.
114,
-------
TABLE 27. NH--C00-H0S-H,0 SYSTEM, COMPARISON OF CALCULATED AND MEASURED DATA OF VAN KREVELEN, ET AL
ON
Partial Pressure, mm Hq
Temp Moles/Kq of Soln. NH3 CO, H,S
*C NH3_ <%_ HjS VK API MEAS VK API. HEAS VK APJ HEAS
20, 1.140 0,4t0 a. 188 5,5 3.* 0
20. i,2«e 0,750 0.29t 0,7 B.B n
20, 2,160 0,910 0,160 3,T 4,0 0
20. 2.158 0,4M0 0,600 8,9 9.9 0
20* 2.250 i,4na d.210 1.7 1,9 0
20, 0,790 0,250 0,180 2,6 2,6 0
• 0, 1,178 0,1110 0,184 10, 6 |0,3 0
40, 1.140 0,410 0,180 IB, 0 9.7 0
40, 1,130 0,210 0,290 14,1 IJ,
»0, 2,160 0.940 B,3hB 11,
48, 2,1)0 0,400 0,bt)0 23,
40, 0,700 0,104 0.3S0 5,
40, 0,790 0,250 0,180 7,
40, 0.740 0,380 0,160 3,
40, 2,230 1,4100 8.210 5,
60. 1,170 0.4i<| 0,184 26.
60, 1,140 fl.4|0 0,180 24,
60, 1,130 0.210 0,290 33.
60, 2.160 0,950 0,360 28,
60, 2.1*0 0,400 0,6(90 15,
60, 2.250 1.340 0,200 18.
60, 0,700 0,104 0,3)0 13,
60, 0.790 0.250 0,160 17,
60. 0,740 0,360 0,150 10,
60, 1,020 0,620 0.140 9,
60, 1.2«i 0.6S6 0,124 19,
69. 1,2»9 0,643 0,214 IS.
Hi
24.
5,
6.
3.
6,
24,
zz.
30.
29.
55,
1«.
12.
l«,
*.
>.
1«,
"•
0
0
0
n
0
0
«
0
0
0
0
0
0
0
0
0
0
0
0
0 2.1 t.
0 44.7 38,
0 6.4 5,
0 8,5 0,
0 30,7 26,
0 1,7 I,
0 10,2 8,
0 11,0 9,
0 3.1 2.
0 30,
0 3,
0 3.
0 a.
0 42,
0 128,
0 43,
0 4fc.
0 14,
0 131.
»T,
S60,
*>.
36,
»»«,
265.
2^.
^ *
4,
7.
35,
109,
«J.
46,
!"•
137,
1"».
352,
22,
36,
114,
265.
0 121,3 116,
0 177,2 174,
0
0
0
0
0
0
e
0
0
0
0
0
0
0
0
0
0
0
0
0
0 3,7 3,7 4.2
0 48.2 44.0 49.0
0 13,4 12,
« 7,3 7,
0 19,7 16,
0 3,7 3,
0 11,3 10,
0 11,6 ID,
0 12,0 11,
0 33,7 35,
0 24,8 24,
0 33,1 30,
0 12,5 11,
0 29,3 26,
0 40,5 43,
0 29.3 32,
0 30,0 33,
0 34,1 35,
0 74,7 107,
0 67,2 79,
0 59,3 95,
96.6 94,
14,8
8.4
10.11
4,5
11.7
12,3
12,1
36,4
27,0
32,3
13.3
27,8
37,6
29,5
30,0
32,8
64,4
77,5
59,5
80,4
34,| 35,3 32.0
55, « 58,3 53,5
70,0 79,2 63.7
29.7 35,9 25.8
80,3 96,4 77,0
-------
TABLE 28. SUMMARY OF REFERENCES TO EXPERIMENTAL DATA
Reference
15
16
17
18
19
20
21
22
23
24
25
26
Type of Data
NH3-C02-H2S-H20
50°C to 120°C
NH3-C02-H2S-H20
20°C
NH3-H20
0°C to 60°C
97°C to 147°C
NH3-C02-H2S-H20
20°C
NH3-C02-H20
20°C to 40°C
C02-H20
0°C to 60°C
80°C to 150°C
NH3-H2S-H20
80°C to 120°C
NH3-C02-H20
20°C to 100°C
NH3-H20
0°C to 60°C
NH3-H20
0°C to 60°C
NH3-C02-H20
20°C to 40°C
H2S-H20
(Buffered)
80°C to 185°C
Used in Correlation
yes
yes
yes
yes
no
no
yes
yes
yes
yes
yes
yes
no
yes
70
(continued)
-------
TABLE 28. (continued)
27 NH3-C02-H20 no
70°C to 120°C
28 NH3-C02-H20 no
Phase diagrams
60°C to 170°C
29 NH3-H2S-H20 yes
20°C to 60°C
1 NH3-C02-H20 yes
NH3-H2S-H20 yes
NH3-C02-H2S-H20 yes
20°C to 60°C
30 NH3-H20 no
114°C to 317°C
31 C09-H20 no
270feC to 550°C
32 H2S-H20 no
160°C to 330°C
33 NH3-H2S-H20 no
70°C to 90°C
34 H2S-H20 no
25°C
35 H2S-H20-Salt no
150°C to 330°C
36 H2S-H20 no
H2S-CH4-H20
71°C to 140°C
37 NH3-C02-H20 no
60°C to 150°C
More references are given in I. Wichterle, J. Linek, and E. Hala, Vapor-
Liquid Equilibrium Data Bibliography, Elsevier (1973). The references listed
above are ones for which copies of the data have been obtained.
71
-------
TABLE 29. SUMMARY UF DEVIATION ERRORS BETWEEN CALCULATED AND
MEASURED AMMO?!IA PARTIAL PRESSURES
Partial
Pressures
in Table no.
18
19*)
22
23
24
25
26 b)
Temp
°C
Van Krevelen
vs Meas. Data
No. pts Ave. Error
80 6
120 6
60 6
100 4
140 4
20 5
40 4
60 4
20 4
40 7
60 9
20 7
40 3
60 7
80 5
100 4
50 3
80 10
110 5
120 5
20 13
20 16
40 7
overall average
ignoring data in
Tables 23 & 30
60
29
387
130
33
184 J]
107a
60a'
15
8
8
45
90
139
14
15
98
126
115
73
8
2425!
223.b)
72
API SWEQ
vs Meas. Data
No. pts Ave. Error
10
8
6
4
4
5
4
4
4
7
9
7
3
7
5
4
3
10
5
5
13
16
7
14
17
8
18
5
329 a!
15°a
79a}
3
8
8
11
10
81
10
9
43
77
35
30
12
128^1
Ji'
24
a)
b)
These data appear to be of rather low quality and should be given very
little weight.
These data are in disagreement with data for NH^-H^O in Table 22 by
Clifford and Hunter. From our evaluation we believe the data of Clifford
and Hunter to be more nearly correct, and more weight has been given to
their data.
72
-------
TABLE 30. SUMMARY OF DEVIATION ERRORS BETWEEN CALCULATED AND
MEASURED CARBON DIOXIDE PARTIAL PRESSURES
Partial
Pressure
in Table no.
Van Krevelen
Temp.
°C
vs.
No. pts.
Meas. Data
Ave. Error%
vs.
No. pts
API SWEQ
Meas. Data
Ave. Error%
22 20 4
40 8
60 9
0
20
40
60
23
a)
20 5
40 3
60 7
80 5
100 4
24 50 3
80 9
110 4
120 3
25 20 15
overall average
ignoring data in
Table 27
13
8
9
102
1288
92
63
63
98
148
8
35
4
8
9
1
1
1
1
7
3
7
5
4
3
10
5
5
15
5
13
10
18
4
6
15
117*
66a\
23
29
20
25
J2
17
a) These data appear to be of rather low quality and should be given
very little weight because deviations are large and have little
apparent pattern.
73
-------
TABLE 31. SUMMARY OF DEVIATION ERRORS BETWEEN CALCULATED AND
MEASURED HYDROGEN SULFIDE PARTIAL PRESSURES
Partial
Pressures Temp.
in Table no. °C
18 80
120
150
80
120
19a) 2Q
40
60
20 20
40
60
21 80
100
120
140
160
x 185
24 50
80
no
120
25 20
27 20
40
60
No.
__
—
6
6
5
4
4
30
30
30
__
—
__
—
—
—
3
9
4
3
15
6
9
12
overall average
ignoring data
Table 23.
in
Van Krevelen
vs. Meas. Data
pts. Ave. Error%
__
—
22
44
787"!
225a
23a)
10
6
9
_ _
--
—
--
-_
—
211
139
188
46
10
12
6
8
24
vs.
No. pts
4
3
3
10
8
5
4
4
30
30
30
2
2
2
2
2
2
3
10
5
5
15
6
9
12
API SWEQ
Meas. Data
Ave. Error%
5
2
13
13
44
539!)
182a{
5a)
14
8
12
50
67
75
70
44
16
13
47
23
18
11
12
10
20
18
,
a) These data appear to be of rather low quality. Large and unreasonable
adjustments would have to be made in the correlation model to correlate
these data; therefore they were ignored.
74
-------
Table Temp. Ave. Error. %
No- System Range. °C NH^ CO: - Jg[
20 NH3-H2S-H20 20°C to 60°C ..... - 8
22 NH3-C02-H20 20°C to 60°C 9 9
25 NH3-C02-H2S-H20 20°C 8 8 10
27 NH3-C02-H2S-H20 20°C to 60°C ...... 8
From this comparison, average errors are -about 10% or less in these tables-
however only data from 20U to 60UC are compared, llhen higher temperature
data and other literature data are compared the agreement is not as good be-
cause of extrapolation errors. These comparisons are given in Tables 18, 19
24, and 25 where deviation errors can be summarized as follows.
Temp. Ave. Error. %
System Range, °C NH^
18 NH3-H2S-H20 80°C to 120°C 139 — 58
19 NH3-H2S-H20 20°C to 60°C 122b) — 379b)
23 NH3-C02-H20 20°C to 100°C 65b) 105b) —
24 NH3-C02-H2S-H20 50°C to 120°C 108 84 146
This comparison shows that deviation errors are baout 3 to 15 times higher
than for the systems from which the correlation was derived. This conclusion
doesn't change significantly even when suspected data noted at the bottom of
this summary are ignored. Thus, it is concluded that the Van Krevelen model
does well at temperatures from 20 C to 60 C which is the region from which i£
was derived, but its accuracy is much poorer at temperatures from 60 to 120_
which is the range of commercial interest for sour water strippers. Beychok
has recently proposed an NH, Henry's constant published by Edwards et al
which improves predicted MM, volatilities. By this method, the average error
for NH3 at 80°C in Table 29Jis reduced from 60% to 23%.
Evaluation of SWEQ Model
The SWEQ model has the advantage that both low- temperature and high-
temperature data were used in developing the model, thus it would be expected
to give better results than the Van Krevelen model. A comparison of the API
SWEQ model with the Van Krevelen model and with experimental data is given in
Tables 18 to 27 under the heading API. At some of the conditions in these
tables, direct comparison with the Van Krevelen model is not possible because
the Van Krevelen modes does not permit the calculation of equilibrium data at
NHo/H9S ratios less than 1.5 or for NFL/total acid gas ratios less than unity.
This condition occurred in the following cases.
b 'Measured data in Tables 19 and 23 are believed by the author to be un-
reliable. Data in Table 19 deviate radically. Large and unreasonable adjust-
ments would have to be made in the model to correlate these data. Data in
Table 23 exhibit large deviations with little apparent pattern.
75
-------
Table Temp.
No. System Range, °C
]8 H2S-H20 80°C to 150°C
H2S-NH3-H20 80°C and 120°C
21 H2S-bufferrH20 80°C to 185°C
22' C02-H20 0°C to 60°C
24 NH3-C02-H2S-H20 80°C to 120°C
(4 points)
A comparison of the SWEO model with the Van Krevelen model at low
temperatures where the Van Krevelen model was derived gives the following
results.
lable
No.
20
22
25
27
System
NH -H S-H 0
WM _rn _H n
NH3-CO?-H2S-H20
NH-,-CO -H S-H 0
Temp.
Range
°C
?n fin
20-fiO
20
20-60
A\
NH^
VK SWEQ
Q 7
8 12
te. Error "/,
CO?
VK SWEQ
Q in
8 12
f
9
H'
VK
0
10
0
?$
SWEQ
n
1 1
n
i R
1 U
Overall Ave. Error 9 10 9 11 9 12
This comparison shows that the SWEq model does about as well as the Van
Krevelen model except that the Van Krevelen model appears to be slightly
better. The overall average error from the Van Krevelen model is about
9% while the SWEQ model gives about 11%.
The picture changes considerably when high temperature data are compared
as follows.
76
-------
Temp. Ave. Error %
Range ~lfi^ CO^ [O
™-J™ _^_ __2
18 NH3-H2S-H20 80-120 45 13 — — 33 27
19 NH3-H2S-H20 20-60 122a) 197a) ...... 379a) 265a)
23 NH3-C02-H20 20-199 65s } 29a) ^ 9#> — —
24 NH3-C02-H2S-H20 50-120 108 53 84 24 146 31
Overall Ave. Error 77 36 84 24 90 29
(ignoring Tables 23 & 27)
a Measured data in Tables 23 and 27 are believed by the author to be
unreliable. Data in Table 23 deviate radically. Larqe and unreasonable
adjustments would have to be made to correlate these data. Data in
Table 27 have large deviations with little apparent pattern.
This comparison shows the SWEQ model to be superior to the Van Krevelen
model at high temperatures with deviations averaging about 84% for the Van
Krevelen model compared to about 29% for the SWEQ model. However, the pre-
dicition accuracy is still not as good as at lower temperatures. This can be
partly explained by the fact that much higher concentrations of the components
were studied in Tables 18 and 24 at high temperatures compared with concen-
trations in Tables 20, 22, 25, and 27 at low temperatures. Concentrations
up to 14 moles/Kg of solution are covered in Tables 18 and 24 while concen-
trations to only about 3.5 moles/Kg of solution are covered in Tables 20, 22,
25, and 27. These higher concentrations place very high demands on the SWEQ
model and makes actual correlation of the data more difficult. Besides this
problem there is the normal scatter expected from the data due to measurement
errors. It is not possible at this point to say which errors are correlation
erros and which errors are measurement errors; however, it is possible to con-
clude that the sum of both errors is on the order of about 29% for the SWEQ
model compared to 84% for the Van Krevelen model.
Another test for accuracy of the SWEQ model can be made by comparing
mean deviations in p /p lc where the mean value is calculated as follows.
P
meas
calc/mean value
i = 1,N
P
meas
calc
1/N
(43)
where N = number of points averaged
This comparison will show any bias errors that may exist between calculated
and measured data. As in the case of average errors, the bias errors can
result from either correlation bias or from bias in the experimental data.
These comparisons for the SWEQ model are given in Figures 6, 7, and 8 for
77
-------
NH-, C0«, and H9S respectively. Figure 6 shows that mean ratios of Pmeas/Pcalc
fof NH. lie primarily above unity with only the data of Badger and Silver
and of Breitenbach and Perman lying below unity. The amount of steam required
in a sour water stripper is primarily determined by the volatility of NH~.
For this reason, available NH3 volatility data from the literature from vari-
ous authors are compared in Figure 6 in an attempt to obtain a reliable
volatility correlation. The author suspects that the points below unity are
in error and that the true NH3 partial pressures are slightly above unity.
The scatter between various authors primarily represents bias in their measured
data, but a line of unity which falls below all of these points probably
represents correlation bias. This correlation bias in the case of ammonia can
be easily adjusted so that measured data will scatter symmetrically both above
and below unity; however, the author is hesitant to do this without further
justification from other data.
Figure 7 shows that mean ratios of Pmeas/P aic for H^S appear to fall
nearly symmetrically both above and below OnTty, so againf the scatter between
the points probably represents bias in measured data between the various
authors.
Figure 9, 10, and 11 show similar plots comparing mean ratios of
pcneas/pcalc from the Van Krevelen model for NH~, C0?, and H2S respectively.
These plots show wider scatter than it obtained from the SWEQ model. The
difference has to be due to correlation bias. This result would tend to infer
that there could still be correlation bias in the SWEQ model which has not
been identified. If such bias exists, it has to be on the order of the
deviations appearing in Figures 6 to 8 or less.
78
-------
o
re
c
re
O)
o
re
o
re
O)
10.0
9.0
8.0
7.0-
6.0
5.0
4.0
3.0
2.0
i.oJ
0.9
0.8
0.7
0.6
0.5
0.4*
0.3
0.2
0.1
_ 1 1 1 1 1 1 1
I O Miles & Wilson
- D Clifford
- A Van Krevelen NH3-C02
- Vcardon & Wilson
• Badger & Silver
Ofireitenback & Perman
-
V V
k A zn o
• -
-
: o
?
—
• i i i i _
—
-
_
D v X
_
—
-
-
1 1 1 1 1
20 30 40 50 60 70 80 90 100 110 120 130 140 150
Temperature, °C
Figure €. Ammonia mean ratio of measured over calculated partial pressures
based on SWEQ correlation.
79
-------
o
fO
ro
O)
CM
O
0
ft
O
to
o
o_
CO
(T3
Ol
Q_
10.0
9.0
8.0
7.0
6.0
5.0
4.0
3.0
2.0
1:3
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
I I I I 1 1 1 1 1 1 1 I
_ -
_ -
m ***
_
ltm •„
-
b A $ v ^
t " v v i
^ «
-
-
- A Van Krevelen NH3-C02
OLange's Handbook
~ V Cardon and Wilson
Q Badger and Silver
i i i i i i i i i i i i
20 30 40 50 60 70 80 90 100 110 120 130 140 150
Temperature, °C
Figure 7. Carbon dioxide mean ratio of measured over calculated partial
pressures based on SWEQ correlation.
80
-------
o
•r~
DC
c
03
0)
s:
CM
„
0
ro
u
Q.
10
(U
Q-E
10. n
.0
8.0
7.0
6.0
5.0
4.0
3.0
2.0
-
T-Ofl
0.8|
0.7
0.6
0.5
0.4
0.3
0.2
0.1
-
—
-
-
O
- A
V
- D
O
1
20 30
^^ V A o
O ^
Miles and Wilson
Van Krevelen NH3-H2S
Cardon and Wilson
Badger and Silver
Van Krevelen NH3-H2S-C02
i i i i i i
40 50 60 70 80 90
O
V *7
I
_
_
^
-
-
1 1 1 1 1
100 110 120 130 140 15
Temperature, °C
Figure 8. Hydrogen sulfide mean ratio of measured over calculated partial
pressure based on SWEQ correlation.
81
-------
o
•»™
4-*
c
Ol
i^1-
z
3E
•"• ""•
P.
LJ x/
O
v o a
A A ^
t~± ~
0.8|- -1
0.7
0.6
0.5
0.4
0.3
0.2
0.1
_ —
_ _
- -
_
O Miles and Wilson
D Clifford
A Van Krevelen NH3-C02
V Cardon and Wilson
' Badger & Silver
O Breitenback & Perman
I i i J I i i i I i i i
20 30 40 50 60 70 80 90 100 110 120 130 140 150
Temperature, °C
Figure 9. Ammonia mean ratio of measured over calculated partial pressures
based on Van Krevelen correlation.
82
-------
10.0
9.0
8.0
7.0
6.0
5.0
4.0
3.0
o
2 2-0
IO
s:
CM
8 i,rJ
0.9)
o 0.81
Ts 0.7
^° 0.6
£ O.J
0)
ex5 0.4
0.3
0.2
0.1
m
-
_
_
^,
-
_
o o
o
3 A ^ :
•B
•• _
A Van Krevelen NH3-C02
~ O Cardon & Wilson
D Badger and Silver
i i i i i i i i i i i i
20 30 40 50 60 70 80 90 100 110 120 130 140 15
Temperature, °C
Figure 10.
Carbon dioxide mean ratio of measured over calculated partial
pressures based on Van Krevelen correlation.
83
-------
o
to
OL
c
fO
2
CO
u
(O
u
o.
trt
(O
0)
o.
10.0
9.0
8.0
7.0
6.0
5.0
4.0
3.0
2.0
H
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
i i i i I I I I I i i i
. _
-
-
-
V
V
o
_ v :
-
— _
mm «
w •
- O Miles and Wilson
A Van Krevelen NH3-H2$
- V Cardon and Wilson
O Badger and Silver
O Van Krevelen NH3-H2S-C02
iii i i i i i i i i i
20 30 40 50 60 70 80 90 100 110 120 130 140 150
Temperature, °C
Figure 17. Hydrogen sulfide mean ratio of measured over calculated partial
pressures based on Van Krevelen correlation.
84
-------
Evaluation of New BYU Data
An evaluation of new NH.-l-LS-hLO and NH.-CO,-H9S-H90 data measured
at Brigham Young University Can^proDably be BestSiafle b§ comparison of
measured data with predicted data from the SWEQ model. These comparisons
are made in Table 18 for the hLS-H90 and NH~-H9S-H90 systems and in Table
24 for the NH--CO -H2S-H20 system/ The SWEQ m&del^predicts low temperature
data on which the Van Krivelen model is based with about the same accuracy
as the Van Krevelen model, and at the higher temperatures from 50°C to 120°C
in Tables 18 and 24, the accuracy is much better than the Van Krevelen model.
As discussed above, it is not possible to separate correlation errors from
measurement errors; so margins of error have to include both effects. The
following gives a summary of the average errors between predicted and measured
data.
Table Temp. No. of
No. *C Points
Ave. Error %
18
19
80
120
50
80
110
120
10
8
3
10
5
5
14
17
43
77
35
30
Overall Ave. Error % 36
23
29
20
25
24
13
44
13
47
23
18
29
It is concluded from this comparison that the new BYU data in Tables 18 and
24 are consistent with literature data correlated by Van Krevelen et al.,
with average scatter between measured and correlated data being on the order
of 36%, 24%, and 29% respectively for NH3, C02, and H2S. Two experimental
runs given in Table 24 were ignored in computing these averages and ammonia
analyses on two additional runs were ignored. The points ignored are noted
at the bottom of Table 24. The reason for ignoring these points is that the
deviations are so large that the experimental points appear to be unreasonable
and probably in serious error.
Mean ratios of Pmp,,/Pr,lr plotted in Figures 6, 7, and 8 for NH-, CO-,
and H2S respectively snow some bias between data in Tables 18 and 24 as
follows. Table
M«Y Comments
24
H2
appears okay
appears okay
NH3 at 50°C and 80°C appears about
40% too high
CO? appears okay
H2§ appears okay
85
-------
Based on this comparison, it is concluded that any bias in the measured
data is small except for NH, at 50°C and 8QOC in Table 24. If these data
points were ignored in ccmpdting the average error above for NH-, then the
overall average error would be reduced from 36% to 24% which is comparable
to deviation erros for CC0 and H.-.S based on the SWEQ model.
C. (L
Ammonia Fixation by Acids and Release by Caustic Addition
/
Little direct data appear in the literature on the volatilities of NH-,
C0?, and H9S9f-irom aqueous solutions as a function of pH» One study made by
Shfh et air, ' is given in Table 21 for the volatility of H2$ from buffered
solutions. In this table, the predicted H2S pressures are consistently lower
than measured values by a factor of about 6.7. This prediction error could
be the result of the salt concentration in the buffer solution which is not
accounted for by the SWEQ model.
In addition to the data by Shin et a!., new measurements of pH versus
caustic addition have been made at BYL). These results are shown as the
plotted curves in Figures 12 and 13. These comparisons show that predicted
free NH- concentrations are also lower than measured values as occurs in the
case of HpS. These data tend to indicate that the SWEQ model might be pre-
dicting both too low H2S and too low NHL partial pressures, but we doubt
this based on the measured volatility data of NFL and hLS examined in this
report. With these discrepancies, calculated pH levels coul:d be in error by
+_ 0.5 unit; this is a rather large error so more work should be done to re-
solve this question.
Ammonia fixation effects due to carboxylic acids and the release of
NH- by caustic addition are predicted by the SWEQ model as given in Table
32. This table gives a comparison of calculated tray to tray NH_, C0?, and
HpS volatilities going down a separation column at total reflux at 20 psia
column pressure. The initial vapor phase concentrations of NH,, C0?, and
HpS were 100 ppm on a weight basis for each component. The first set gives
calculated vapor and liquid compositions for three trays under conditions of
no carboyxlic acid or caustic present. In this example, the liquid concen-
tration of all three components drops to 0.1 ppm or less on the third tray.
When 500 ppm by weight or carboxylic acid is added to the liquid on each tray,
then the ammonia concentration goes up to 142.4 ppm on the third tray indi-
cating that the ammonia is fixed and is unstrippable. If caustic is then
added to a level of 172.5 ppm in the liquid on each tray, then the ammonia is
released from the carboxylic acid and concentrations less than 0.1 ppm are
predicted for NH,, C02, or H^S in the liquid of the third tray. If too much
caustic is added; then HpS wfll be fixed in the liquid phase; thus 500 ppm of
caustic produces an HpS concentration in the liquid phase of the third tray
of 310.1 ppm H2S. From this table, it appears that the optimum pH for equal
volatility of NH, and H?S is about 8.5. This pH corresponds to the hydrogen
ion concentration in the liquid phase at the temperature of the tray in the
separation column. In actual practice, samples of liquid would probably be
taken for pH determination at room temperature. The effect of temperature
on pH can be calculated from the SWEQ model. Figure 14 gives a plot of pH
at 25°C and at 120°C for the addition of caustic to the mixture shown in
Figure 1. The effect of temperature will be different depending on the
mixture, but this plot can give some idea of the effect.
86
-------
00
—I
11
Figure 12. Free ammonia versus pH adjustment by caustic addition at 25°C.
-------
125
TOO
75
(D
fD
50
25
8
PH
10
Figure 13. Free ammonia versus pH adjustment by caustic addition at 80°C,
-------
TABLE 32. COMPARISON OF CALCULATED NH,, C09, and H«S VOLATILITES VERSUS
EFFECTS FROM CARBOXYLIC ACID AND^CAUSTKTADDITION, TRAY TO TRAY
FROM THE COLUMN AT TOTAL REFLUX AND 20 PSIA COLUMN PRESSURE
RCOOH
oo
Initial Basis for Each Set
or OH"
1n Liquid
ppm wt
Tray
1
2
3
1
2
3
1
2
3
1
2
3
RCOOH
0
0
0
500
500
500
500
500
500
500
500
500
OH
0
0
0
0
0
0
172.5*
172.5
172.5
too
500
500
1s 100 ppm
ppm by wt.
in Vapor
NH,
100
5.8
0.6
100
138.8
142.2
100
4.5
0.2
100
3.9
0.2
co2
100
0.3
0.0
100
0.0
0.0
100
0.9
0.0
100
60.0
29.5
of NH,, COo, & HoS 1n Vapor
H,S
L
100
1.3
0.0
100
0.1
0.0
100
3.7
0.2
100
171.2
250.0
"" bpm by wt.
in Liquid
NH,
J
5.8
0.6
0.1
138.8
142.2
142.4
4.5
0.2
0.0
3.9
0.2
0.0
COg
3.0
0.0
0.0
9.0
0.0
0.0
0.9
0.0
0.0
60,0
29,5
11.8
H2S
1.3
0.0
0.0
0.1
0.0
0.0
3.7
0.2
0.0
171.2
250.0
310.1
Vapor/Liquid
wt Basis
NH.
•*
17
9.7
4.2
.72
.98
1.00
22*
23
23
26
26
26
C0?
370
1100
2590
5600
5100
5090
116
83
82
1.7
2.0
2.5
Ratio
H2S
78
241
620
1750
1570
1550
27*
20
19
0.6
0.7
0.8
pH
at col .
Temp.
in Liquid
8.035
7.515
7.026
6.195
6.332
6.341
8.500
8.646
8.652
10.165
10.097
10.026
Appears close to optimum caustic addition for best NH3 and H2S volatility.
-------
12
11 -
Initial Concentrations
NH0
ppm, wt
100
co2
H2S
RCCOH
I
100
100
500
100 200 300 400
Caustic Added, ppm by wt
500
Figure- 14. Sample plot of pH versus caustic addition showing variation of
pH at 25°C and at column temperature.
90
-------
SECTION 7
ACCURACY OF CORRELATION
The overall accuracy of the SWEQ model can be assessed by examination
of the error summaries in Tables 29 to 31 for NH-, C0?, and H?S respectively.
From these tables, the overall average error between measured and predicted
partial pressures can be summarized as follows.
Temperature Overall Ave. Error %
Compound Range, °C VK SWEQ
Ammonia 20 to 140°C 72 24
Carbon dioxide 20 to 120°C 35 17
Hydrogen sulfide 20 to 185°C 24 18
This comparison shows that SWEQ module is superiorio the Van Krevelen
model.9)
Data at low temperatures are represented better by both models than data
at high temperatures as shown in the following comparison taken from the
previous section of this report.
Ave. Error %
up to 60*^ above 60°C
Compound VK SWEPT VK SWEQ
Ammonia 9 10 77 36
Carbon dioxide 9 11 84 24
Hydrogen sulfide 9 12 90 29
This comparison shows that both models predict the low temperature data
quite well; but at high temperature, the Van Krevelen model deviates consider-
ably from measured data, and errors between the SWEQ model and measured data
increase from about 11% to about 24%.
Users of theSWEQ model must be aware that the errors summarized above
are average errors and that there might be regions where the correlation is
less accurate. More experimental data is required before a better assessement
can be made.
a^This is the model published by Van Krevelen without any modifications.
91
-------
SECTION 8
SUMMARY
A new correlation model has been developed for calculating sour water
equilibrium data at temperatures from 20°C to 140°C. The correlating equations
in this new SWEQ have been used to obtain a computer program capable of handling
the various chemical and physical equilibria of NHV CCL, and HUS in sour water
systems including the effects of carboxylic acids on NH~ fixation and release
by caustic addition.
This new SWEQ correlation model has been used to evaluate published and
new vapor-liquid equilibrium data and comparisons are made with the Van
Krevelen prediction equations as published by Van Krevelen. Average errors
between calculated and measured partial pressure data can be summarized as
follows.
Ave. Error %
U£
V*
to 60°C above 60°C
Compound VK SWEQ V_K SWEQ
Ammonia 9 10 77 36
Carbon dioxide 9 11 84 24
Hydrogen sulfide 9 12 90 29
This comparison shows that both models predict low temperature data
quite well; but at high temperatures, the Van Krevelen model deviates con-
siderably from measured data, and errors between the SWEQ model and measured
data increase from about 11% to about 29%. Comparisons with variations of
the Van Krevelen model as published by other authors have not been made.
Vapor-liquid equilibrium measurements made at Brigham Young University
are predicted by the SWEQ model with the following average errors.
Compound Ave. Error %
Ammonia 36
Carbon dioxide 24
Hydrogen sulfide 29
Data on measured NH~ partial pressures from NHg-CO^^S^O mixtures appear
too high by about 40% at 50 C and 80 C. If these points are ignored, then
the average ammonia error is reduced from 36% to 24%.
92
-------
Details of the SWEQ correlation model, correlating equations, the
computer program, and evaluations of experimental data are given in this
report.
93
-------
REFERENCES
1. D. W. Van Krevelen, P. J. Hoftijzerv and F. J. Huntjens, Recueil
Des Travaux Chimiques Des Pays-Bas, 68, 191-216 (1949).
2. M. R. Beychok, -Aqueous Wastes from Petroleum and Petrochemical
Plants. John Wiley & Sons, New York, N.Y., (1967).
3. T. J. Edwards, J. Newman, and J. M. Prausnitz, AIChE Journal. 2]_, 248
(1975).
4. R. L. Kent and B. Eisenberg, Hydrocarbon Processing. Feb., 87-90
(1976).
5. D. L. Cardon and R. M. Wilson, "Ammonia-Carbon Dioxide-Hydrogen Sulfide-
Water Vapor-Liquid Study", Final Project Report in progress for the API.
6. D. C. Bomberger and J. H. Smith, Report on "Evaluation of Ammonia
'Fixation1 Components in Actual Refinery Sour Waters", Stanford Re-
search Institute, December 10, 1977.
7. H. S. Harned and S. R. Scholes, J. of Amer. Chem. Soc., 63, 1706
(1941). ~~
8. Cuta and Stratfelda, Chem. Li sty, 48_, 1308 (1954).
9. B. N. Ryzhenko, Geokhemiya. 137-138 (1963).
10. Handbook of Chemistry and Physics. 51st ed., The Chemical Rubber Co.,
D-122.
11. Handbook of Physical Constants, Revised ed., The Geological Society
of America, Memoir 97, Sec. 18 (1966).
12. A. L. Ellis and N. B. Milestone, Geochim. Cosmochim. Acta, 31, 615
(1967).
13. Handbook of Chemistry and Physics, 51st ed., The Chemical Rubber Co.,
D-143-D-144 (1970-71).
14. American Petroleum Institute, Publication No. 946, "Sour Water Strip-
ping Project Committee on Refinery Environmental Control, American
Petroleum Institute", June 1975.
15. E. H. M. Badger and L. Silver, J_. Soc. Chem. Jjid_., 5_7_, 110-12 (1938).
94
-------
16. Breitenbach, Bull. Univ. Wis. Eng. Exp. Sta. Ser. 68 (as given in
Perry's Chemical Engineers' Handbook. Fouth ed., 1963).
17. I. L. Clifford and E. Hunter, J_. Phys. Chem. , 37,101 (1933).
18. I. 6. C. Dryden, J_. Soc. Chen. Ind. , 66_, 59 (1947).
19. S. Ikenko, Kogyci Kagaku Zasshi , 64, 627 (1961).
20- Lunge's Handbook of Chemistry, Eighth ed., 1952 (original source not
given).
21. D. H. Miles and G. M. Wilson, "Vapor-Liquid Equilibrium Data for De-
sign of Sour Water Strippers", Annual Report to API for 1974, October
1975.
22. E. Otsaka, S. Yoshimura, M. Yokabe, S. Inque, Kogyo Kagaku Zasshi.
63_, 1214-1218 (1960). -- " --
23. E. P. Perman, J_. Chem. Soc. London, 83, 1168 (1903).
24. Sherwood, ing. Eng. Chem.. 1_7, 745 (1925). (as given in Perry's Chemi-
cal Engineers' Handbook, Fouth ed., 1963).
25. S. Pexton and E. H. M. Badger, J_. Soc. Chem. Ind. . 57^, 106 (1938).
i
26. T. T. C. Shih, B. F. Hrutfiord, K. V. Sarkanen, and L. N. Johansen,
TAPPI, Technical Assoc. of_ the Pulp and Paper Industry, 50, (No. 12)
27. T. Takahashi, Kogyo Kagaku Zasshi. 65_, 837-843 (1962).
28. T. Takahashi, S. Yoshimura, K. Fukii, and E. Otsaka, Kogyo Kagaku
Zasshi. 65_, 743-745 (1962).
29. V. E. Terres, W. Attig, and F. Tscherter, Gas, u.. Wasserfach.- 98,
512-516 (1957).
30. M. E. Jones, J_. Phys. Chem. 67, 1113-1115 (1963).
31. S. D. Malinin, Geochem. Int.. 11 , 1060-75 (1975).
32. T. N. Kozintseva.x Geochem. Int.. No. 4, 750-756 (1964).
33. V. I. Oratovskii, A. M. Gamolskii, and N. N. Klimenko, J_. Appl.Chem.
USSR*. 37,, 2363-2367 (1964).
34. H. Gamsjager and P. Schindler, Helv. Chim. Acta. 52, 1395-1402 (1969).
35. T. N. Kozintseva, Geokhemiya. 121-134 (1965).
95
-------
36. J. L. Vogel, M.S. Thesis, The University of Tulsa, 1971.
37. R. J. Frohlich, Ph.D. Thesis, Polytechnic Institute of Brooklyn Uni
versity, Microfilms No. 60-3497, 1957.
38. M. R. Beychok, Hydrocarbon Processing. Sept., 261-263 (1976).
96
-------
APPENDIX
COMPUTER PROGRAM FOR CALCULATING SOUR WATER EQUILIBRIA
BASED ON THE VAN KREVELEN EQUATIONS
Table 33 gives a listing of the computer program used for calculating
NH,, C02, HpS and.kLO partial pressure data for comparing the SWEQ model with
the Van Krevelen " equations. The input and output of this program is very
similar to the SWEQ model.
The main calculations are done starting with the following statement.
DO 2030 I = 1,100
This is the start of an iteration loop which extends to statement 2030. This
iteration loop calculates the amount of C03= (BT) and H?NCOO" (EPS) in
solution for various assumed concentrations of HCO~~ (Ac). The concentration
of HCO?~ is adjusted in each iteration so that the total of HCO-" + C02~ +
HpNCOO* concentrations add to the COp content of the mixture. The folTowing
Fortran symbols are used for the chemical equilibrium constants.
EK1 C02(g) + NH3 + H20 + NH4+ + HC03" (A-l)
EK2 NH4+ + HC03- -» H2NCOO~ + H20 (A-2)
EK3 NH3 + HC03- -> NH4+ + C03= (A-3)
EK4 H2S(g) + NH3 "* HS~ + NH4+ ^A"4^
Other symbols have the same meaning as symbols in the SWEQ computer program.
After correct values of HCOo", C0~~, and H?NCOQ- concentrations are found,
the program proceeds to calculate NH-, COp, H-S and H^O partial pressures and
vapor concentrations. The results are then printed out.
This Van Krevelen computer program only computes vapor composition and
pressure from a specified liquid composition and temperature. No other
options were programmed. The equations for the chemical equilibrium constants
and ammonia Henry's constant were obtained by fitting tabular values given
by Van Krevelen. The Henry's constant of ammonia above 90 C is based on
Beychok's2) graphical extrapolation. Because of the tabular and graphical
nature of Van Krevelen's correlation and the graphical nature of the Henry's
constant for ammonia given by Beychok, there is some arbitrariness in the
computer program because another person using different equations to fit the
tabular data and graphs would obtain slightly different results.
97
-------
BeycholoRQ longer recommends the ammonia Henry's constant published
in his book, ' but comparison is made with the book because it represents
a basis for comparing any changes or variations.
Further discussion of the Van Krevelen-correlation can be obtained by
referring to either Beychok ' or Van Krevelen» '
98
-------
TABLE A-l. COMPUTER PROGRAM USED FOR CALCULATING VAPOR-LIQUID
EQUILIBRIUM DATA FROM THE VAN KREVELEN CORRELATION
QPEN(UNIT«20,DEVICE>'DSK', ACCESS- 'SE9IN',FIUE«'SWSD')
DATA WA, WC, WS,WW/17t03,44. 81,34.08,18. 02/
RHQ • \
REAO(20,1000)
250
1994
I99S
1996
2000
2010
2020
2030
2040
3000
3010
3020
3030
TC*273,15
TK*i,8
XA+1E-19
XCMC-19
X5+1E-19
XW+JEM9
1000*RHO/(XA+XC+X3*XW)
XA*F/WA
XC'F/WC
CC
TK
TR
XA
XC
XS
XW
f «
CA
CC
C5
AC
EK2
EK3
CCS
6A i
CSS
SO i
00 2030 X»l,100
CAS * CA*AL« 1.5
IF(CAS) 1994,1994,1996
PA « 0
PC • 0
PS « 0
60 TO 1501
FORHATC M2S AND C02 JN EXCESS')
60 TO 1501
EPS • E*2*CAS*AC
IF(Dfc) 2000,2000,2010
BT « 0
60 TO 2020
BT f EK3*CAS*AWDE
CCE * AU+EPS+BT
0
CS
0
0
This is a test to see if the ratio
of NHj/acid gas is greater than 1.5
If not so, then the computation is
skipped.
i
r
Iteration loop to calculate amounts
of HCO", COj,
and HgNCOtt"
At « AL*(.5*,5*CC/'CCE)
3000,3000,2030
FORHATC ITERATION 010 NOT CONVERGE IN 100 CYCUES'J
HW • EXP(14tfl66*6996,6XtTR.77t67))
ZFCTC-90) 3020,3010,3010
HAD • EXP(^3,17..022**C)
60 TO 3030
HAD « EXPC*17.03+4315/TK)
HA * ,92*HAO*EXPCi0576*CAS)
HA » I/(HAt51,7l}
PA « S1,71*HA*CAS
TNW • C1000*RHC
TNM « CA+CC*BT*OE+6A+SO*TNW*CS
PM « HW*TNW/TNM
(continued)
99
-------
TABLE A-1, (continued)
1000
1861
1010
1011
1020
1030
1040
1050
1060
1070
1089
EM • -»EXP(»l,764*ffc*7*AUOG(TC))*.0e9*CS«U,929..539/TK}*CC
5lf7i*DE*C3/(EK46CA8*51.7t)
PA/P
PC/P
PS/P
PW/P
PC
PS
P
YA
YC
YS
YW
1501 XA
XC
XS
xw
YT
YA
YC
YS
YW
HRITE(5,1030)
WRITE(5,1080)
gK « YA/XA
WRITEC5,|040) XA(YA,CK
EK * YC/XC
EK « YS/X3
EK n YW/XW
XS|YS,EK
GO TO Z
fORMATtSl)
FORMAT(• PH DID NOT CONVERGt IN 100 CYCLES*)
PCRMATC TEMPERATURE DID NOT CONVERSE IW 100 CYCLES')
FORMAT(» COMPONENT LIQUID VAPOR K-VALUE*)
FORMAT(//' TEMPERATURE, C»,F6f2,/« PRESSURE, PSIA«,F8,2,/
FORHATC* AMMONIA *,3F10t5)
FORMAT(» CARBON DIOXIDE *,3Fl0,5)
FORMATC HYDROGEN suLFiDE»,3Fi0,5)
FORMAT(» WATER »,3F10,5)
ffND
100
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing}
REPORT NO
EPA-600/2-80-067
riTLE AND SUBTITLE
2.
A New Correlation of NH3, C02, and H2S Volatility Data
from Aqueous Sour Water Systems
3. RECIPIENT'S ACCESSION-NO.
5. REPORT DATE
April 1980 issuing date
6. PERFORMING ORGANIZATION CODE
. AUTHOH(S)
Grant M. Wilson: Brigham Young University
a. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
American Petroleum Institute
2101 L Street Northwest
Washington, DC 20037
10. PROGRAM ELEMENT NO.
C33B1B
11. CONTRACT/GRANT NO.
R804364010
12. SPONSORING AGENCY NAME AND ADDRESS
Robert S. Kerr Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Ada. Oklahoma 74ft?ft ;
13. TYPE OF REPORT AND PERIOD COVERED
Final 1976-1977
14. SPONSORING AGENCY CODE
EPA/600/15
15. SUPPLEMENTARY NOTES
American Petroleum Institute project officer: Ron Gantz
16. ABSTRACT
A new correlation model has been developed for calculating sour water
equilibrium data at temperatures from 20°C to 140°C. The correlating
equations in this new sour water equilibrium model have been used to
obtain a computer program capable of handling various chemical and
physical equilibria of NHs, CO?, and H2S in sour water systems in-
cluding the effects of carboxylic acids on ammonia fixation and re-
lease by caustic addition. A bibliography of related literature data
is included in the report.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a.
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Ammonia
Hydrogen sulfide
Carbon dioxide
Volatility
Aqueous systems
Carboxylic acids
Caustic
Henry's Constants
Chemical equilibrium
constants
Sour water equilibria
Sour water stripping
Correlation of volatility
data
Computer program
Ammonia fixation
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
Unclassified
;ES
109
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
101
4 U.S. GOVERNMENT PRINTING OFFICE. 1980-657 -146/5655
------- |