GAS
`
`A .3
`
`PURIFICATION
`
`Akermin, Inc.
`Akermin, Inc.
`Exhibit 1008
`Exhibit 1008
`Page 1
`Page 1
`
`

`

`Akermin, Inc.
`Akermin, Inc.
`Exhibit 1008
`Exhibit 1008
`Page 2
`Page 2
`
`

`

`Akermin, Inc.
`Akermin, Inc.
`Exhibit 1008
`Exhibit 1008
`Page 3
`Page 3
`
`

`

`FIFTH EDITION
`
`Akermin, Inc.
`Exhibit 1008
`Page 4
`
`

`

`Gulf Publishing Company
`Houston, Texas
`
`Akermin, Inc.
`Exhibit 1008
`Page 5
`
`

`

`¯ II
`
`Arthur L. Kohl
`Richard B. Nielsen
`
`Akermin, Inc.
`Exhibit 1008
`Page 6
`
`

`

`FIFTH EDITION
`
`GAS
`PURIFICATION
`
`Copyright © 1960, 1974, 1979, 1985, 1997 by Gulf Publishing Company,
`Houston, Texas. All rights reserved. Printed in the United States of America.
`This book, or parts thereof, may not be reproduced in any form without
`permission of the publisher.
`
`Gulf Publishing Company
`Book Division
`P.O. Box 2608 [] Houston, Texas 77252-2608
`
`109876543
`
`Library of Congress Cataloging-in-Publication Data
`Kohl, Arthur L.
`Gas purification. -- 5th ed. / Arthur Kohl and Richard Nielsen.
`p. cm.
`Includes bibliographical references and index.
`ISBN 0-88415-220-0
`1. Gases---Purification. I. Nielsen, Richard (Richard B.) II. Title.
`TP754.K6 1997
`665.7---dc21
`
`96-52470
`CIP
`
`Akermin, Inc.
`Exhibit 1008
`Page 7
`
`

`

`Contents
`
`Preface, vii
`
`Chapter 1
`Introduction, 1
`
`Chapter 2
`Alkanolamines for Hydrogen Sulfide and Carbon Dioxide Removal, 40
`
`Chapter 3
`Mechanical Design and Operation of Alkanolamine Plants, 187
`
`Chapter 4
`Removal and Use of Ammonia in Gas Purification, 278
`
`Chapter 5
`Alkaline Salt Solutions for Acid Gas Removal, 3,30
`
`Chapter 6
`Water as an Absorbent for Gas Impurities, 415
`
`Chapter 7
`Sulfur Dioxide Removal, 466
`
`Chapter 8
`Sulfur Recovery Processes, 670
`
`Chapter 9
`Liquid Phase Oxidation Processes for Hydrogen Sulfide Removal, 731
`
`Akermin, Inc.
`Exhibit 1008
`Page 8
`
`

`

`Chapter 10
`Control of Nitrogen Oxides, 866
`
`Chapter 11
`Absorption of Water Vapor by Dehydrating Solutions, 946
`
`Chapter 12
`Gas Dehydration and Purification by Adsorption, 1022
`
`Chapter 13
`Thermal and Catalytic Conversion of Gas Impurities, 1136
`
`Chapter 14
`Physical Solvents for Acid Gas Removal, 1187
`
`Chapter 15
`Membrane Permeation Processes, 1238
`
`Chapter 16
`Miscellaneous Gas Purification Techniques, 1296
`
`Appendix, 1374
`
`Index, 1376
`
`Akermin, Inc.
`Exhibit 1008
`Page 9
`
`

`

`Chapter 5
`
`Alkaline Salt Solutions for
`Acid Gas Removal
`
`INTRODUCTION, 330
`
`ABSORPTION MECHANISMS, 331
`
`ABSORPTION AT ELEVATED TEMPERATURE, 334
`
`Hot Potassium Carbonate (Benfield) Process, 334
`Catacarb Process, 363
`Flexsorb HP Process, 369
`Giammarco-Vetrocoke Process, 371
`
`ABSORPTION AT AMBIENT TEMPERATURE, 378
`Carbon Dioxide Absorption in Alkali-Carbonate Solutions, 378
`Seaboard Process, 381
`Vacuum Carbonate Process, 383
`Vacasulf Process, 392
`Tripotassium Phosphate Process, 393
`Sodium Phenolate Process, 396
`All(acid Process, 397
`Caustic Wash Processes, 401
`
`REFERENCES, 410
`
`INTRODUCTION
`
`A prime requirement for absorptive solutions to be used in regenerative CO2 and H2S
`removal processes is that any compounds formed by reactions between the acid gas and the
`solution must be readily dissociated. This precludes the use of strong alkalies; however, the
`salts of these compounds with weak acids offer many possibilities, and a number of process-
`es have been developed which are based on such salts. Typically the processes employ an
`
`Akermin, Inc.
`Exhibit 1008
`Page 10
`
`

`

`Alkaline Salt Solutions for Acid Gas Removal 331
`
`aqueous solution of a salt containing sodium or potassium as the cation with an anion so
`selected that the resulting solution is buffered at a pH of about 9 to 11. Such a solution, being
`alkaline in nature, will absorb H2S and CO2 (and other acid gases), and, because of the
`buffering action of the weak acid present in the original solution, the pH will not change
`rapidly as the acid gases are absorbed. Salts that have been proposed for processes of this
`type include sodium and potassium carbonate, phosphate, borate, arsenite, and phenolate, as
`well as salts of weak organic acids.
`The major commercial processes that have been developed for hydrogen sulfide, carbon
`dioxide, and mercaptan absorption using aqueous solutions of sodium or potassium com-
`pounds axe discussed in this chapter. Several of the processes described are no longer com-
`mercial, but brief descriptions are included because of their historical significance. The priaci-
`pal technologies which are still employed are (1) processes based on hot potassium carbonate
`solutions, which are used for the removal of carbon dioxide (and sometimes hydrogen sulfide)
`from high pressure gas streams (the solutions normally contain a promoter to enhance the rate
`of carbon dioxide absorption); (2) processes based on absorption in ambient temperature sodi-
`um or potassium carbonate solutions with vacuum regeneration, which are used primarily for
`the removal of hydrogen sulfide from coke-oven gas; and (3) processes based on ambient
`temperature absorption into solutions containing free caustic, which are used to remove mer-
`captans and small traces of carbon dioxide or hydrogen sulfide from gases.
`
`ABSORPTION MECHANISMS
`
`As point~ed out in Chapter 1, the occurrence of a chemical reaction in the solution has the
`effect of increasing the liquid phase absorption coefficient over that which would be
`observed with simple physical absorption. This increase can be quantified in terms of an
`enhancement factor: the ratio of the actual absorption coefficient with reaction to the absorp-
`tion coefficient which would be expected under the identical conditions if no reaction
`occurred. The prediction of enhancement factors for various classes of reactions is quite
`complex (see Astarita et al,, 1983) and requires a knowledge of the reaction path and rate as
`well as liquid phase physical properties. As the reaction rate increases, the enhancement fac-
`tor, and thus the liquid phase absorption coefficient, also increases. When the reaction rate is
`extremely fast, the liquid phase absorption coefficient can be high enough to make the gas
`phase mass transfer resistance controlling.
`When hydrogen sulfide is absorbed into an alkaline solution, it can react directly with
`hydroxyl ions by a proton Wansfer reaction:
`
`H2S + OH- = HS- + I-I20
`
`(5-1)
`
`This reaction is extremely rapid and can be considered instantaneous in comparison with
`diffusion phenomena (Savage et al., 1980).
`Since hydrogen sulfide is absorbed more rapidly than carbon dioxide by aqueous alkaline
`solutions, partial selectivity can be attained when both gases are present. The data of Garner
`et al. (1958) indicate that selectivity is favored by short gas-liquid contact times and low
`temperatures. Commercial applications of selective absorption based on short residence time
`contact are described by Hohlfield (1979) and Kent and Abid (1985).
`Carbon dioxide is a slightly stronger acid in solution than hydrogen sulfide. Its ionization
`constant for the first step ionization to H÷ and HCO3- is approximately 4 x 10-7 at 25°C
`compared to 1 x 10-7 for the corresponding hydrogen sulfide ionization. As a result, under
`
`Akermin, Inc.
`Exhibit 1008
`Page 11
`
`

`

`Gas Purification
`
`conditions of extended gas-liquid contact where equilibrium is approached, carbon dioxide
`can displace previously absorbed hydrogen sulfide. This phenomenon is used commercially
`in the processing of Kraft paper mill liquors.
`The chemical reaction of absorbed carbon dioxide with alkaline carbonate solutions takes
`place through two parallel mechanisms: (1) direct formation of HCO3- by reaction of CO~
`with the hydroxyl ion and (2) reaction of CO2 with water followed by dissociation of carbon-
`ic acid. According to Astarita et al. (1981), the predominant mechanism at pH > 10 involves
`the direct reaction of dissolved CO2 with OH-:
`
`CO~ + OH- = HCO3- (fast) (5-2)
`
`HCO3- + OH- = CO3= + H20 (instantaneous)
`
`(5-3)
`
`At pH < 8 the principal mechanism is based on the hydration of dissolved CO2 to form
`carbonic acid followed by reaction of the carbonic acid with OH-:
`
`CO2 + H20 = H2CO3 (slow)
`
`H2CO3 + OH- = HCO3- + HzO (instantaneous)
`
`(5-4)
`
`(5-5)
`
`In the pH range of interest for commercial operations, pH > 8, the mechanism involving
`the direct reaction of carbon dioxide to form bicarbonate ions (reaction 5-2) predominates
`(Astarita et al., 1981).
`Although the overall reaction of carbon dioxide with solution components results in the
`conversion of carbonate to bicarbonate, the local reaction rate is determined by the concen-
`tration of hydroxyl ions as indicated by reaction 5-2 (Tseng et al., 1988):
`
`reaction rate [g mol/(liter)(sec)] = koa(CO~)(OH-)
`
`(5 -6)
`
`The value of the second order rate constant, kon, can be estimated by the following equa-
`tion suggesl~l by Astarita et al. (1983):
`
`1og~0koa = 13.635 -- 2.895/T + 0.08 I
`
`Where: T = °K
`I = Ionic strength of the solution
`
`(5-7)
`
`Equations 5-6 and 5-7 indicate that the rate of reaction of carbon dioxide can be increased
`by increasing the CO2 concentration, the hydroxyl ion concentration, or the temperature.
`According to Astarita (1967), the reaction rate of carbon dioxide in carbonate-bicarbonate
`solution is not fast enough at room temperature to enhance the absorption rate appreciably
`over that of physical mass transfer. At temperatures above about 318°K (113°F) the reaction
`rate is sufficiently high to enhance the mass transfer rate significantly, but even at tempera-
`tures as high as 378°K (221°F) the reaction rate is not high enough to be considered instanta~
`neous (Savage et al., 1980).
`
`Akermin, Inc.
`Exhibit 1008
`Page 12
`
`

`

`Alkaline Salt Solutions for Acid Gas Removal
`
`The relatively low rote of absorption of carbon dioxide in carbonate-bicarbonate solutions
`has been an incentive for research on rate-increasing additives. Many such materials have
`been discovered and are usually referred to as promoters, activators, or catalysts. Ferrell et
`al. (1987) provide the following list of materials found to increase the rate of carbon dioxide
`absorption and the original references: formaldehyde, methanol, phenol, ethanolamines
`(Killeffer, 1937), arsenious acid (Roughton and Booth, 1938), glycine (Jeffreys and Bull,
`1964), and the enzyme carbonic anhydrase (Alper et al., 1980). Promoters known to be used
`in commercial processes include diethanolamine, sterically hindered amines, glycine, and
`arsenious oxide.
`In early theories to explain the rate enhancement effects of promoters it was assumed that
`arsenious acid and organic amines operated through different mechanisms. Roughton and
`Booth (1938) concluded that arsenious acid acts as a homogeneous catalyst that increases the
`rate of the key carbon dioxide reaction (reaction 5-2). A "shuttle" mechanism was proposed
`by Sb_rier and Danckwerts (1969) for amine type promoters at low temperatures. In this
`mechanism the carbon dioxide reacts rapidly with dissolved amine by a second order reac-
`tion of the following type:
`
`CO2 + RR’NH = RR’NCOO- + H+
`
`(5-8)
`
`Even with small amine concentrations in the solution, reaction 5-8 is significantly faster
`than reaction 5-2 causing reaction 5-8 to proceed near the gas-liquid interface. The carba-
`mate ion then diffuses into the bulk of the liquid where equilibrium is re-established by
`reversal of reaction 5-8. Carbon dioxide released by the reverse reaction is consumed by
`reaction 5-2, and the resulting free amine can diffuse back to the interface to react with addi-
`tional carbon dioxide.
`Astarita et al. (1981) proposed a general mechanism to cover both arsenious acid and
`amine promoters involving the following two steps:
`
`CO2 + promoter = intermediate
`
`intermediate + OH- = HCO3- + promoter
`
`(5 -9)
`
`(5-10)
`
`The relative rate of equation 5-10 represents the key difference between the two types of
`promoters. With arsenious acid, the step indicated by equation 5-10 is very rapid and takes
`place immediately after equation 5-9 at the gas-liquid interface. With amines at moderate
`temperatures, reaction 5-10 is believed to take place primarily in the bulk of the liquid. At
`the higher temperatures of desorption, the reaction may be fast enough for amines to act as
`homogeneous catalysts (Astarita et al., 1981). Savage et al. (1984) indicate that the rate-pro-
`motion effect of amine in carbonate solutions can be described quite well in terms of homo-
`geneous catalysis.
`Sartori and Savage (1983) compare the rate promotion capabilities and the effects on
`vapor liquid equilibria of a conventional amine (DEA) and a sterically hindered amine. In
`the sterically hindered amine, bulky substitaent groups are placed adjacent to the amino
`nitrogen site. This causes the hindered amines to form carbamates of intermediate to low sta-
`bility. The low stability of the hindered amine carbamate is credited with the observed faster
`absorption rate and higher CO2 loading at commercial operating conditions for carbonate
`solutions promoted with hindered amines compared to solutions promoted with DEA.
`
`Akermin, Inc.
`Exhibit 1008
`Page 13
`
`

`

`Gas Purification
`
`ABSORPTION AT ELEVATED TEMPERATURE
`
`Hot Potassium Carbonate (Benfield) Process
`
`This process was developed by the U.S. Bureau of Mines, at Bruceton, Pennsylvania, as
`part of a program on the synthesis of liquid fuel from coal. Research on CO2 removal was
`conducted with the objective of reducing the cost of synthesis-gas purification by designing
`a process that would take maximum advantage of the synthesis-gas conditions; i.e., high CO2
`partial pressure and high temperature. A flow sheet of the basic process is shown in Figure
`5-1, and a photograph of a large commercial plant is shown in Figure 5-2. The original
`process was described in considerable detail by publications of Benson and coworkers (Ben-
`son et al., 1954; Benson et al., 1956; The Benfield Corp., 1971).
`The hot tmtassium carbonate process was developed further during the 1970s by Benson
`and Field, who conducted much of the original work at the U.S. Bureau of Mines, and many
`improvements were made. Among these, the development of an amine activator (DEA) for
`the potassium carbonate solution, resulting in substantial lowering of capital and operating
`costs and higher treated gas purity, is probably the most important (The Berdield Corporation,
`1971). Major improvements were also made in energy economy through the recovery of inter-
`nal heat (Clayman and Clark, 1980; Baker and McCrea, 1981; Grover and Holmes, 1987;
`Bartoo and Ruzicka, 1983), and the process has been demonstrated to be suitable for partially
`selective removal of hydrogen sulfide in the presence of carbon dioxide (Astarita, 1967).
`Recent improvements to the process include the use of high efficiency packing in both the
`absorber and regenerator columns and the development of further improved activators there-
`by reducing capital and operating costs and resulting in higher treated gas purities when
`
`LEAN SOLUTION I.P/DRAULIC POWER ACID GAS
`RECOVERY TUREINE SEPARATOR
`COOLER
`
`ACID GAS
`REGI~IERATOR
`CONDENSER
`
`LEAN SOLI~ION
`PUMP
`
`REI:LUX
`WATER PUMP
`
`CARBGi~LATE
`REBOILER
`
`NOTES:
`1. COOLERS CAN BE ~IR OR WATER COOLEIX
`;l. HYDRAUUC ~IJRBINE IS USUAllY USED
`MODERATE TO HIGH PItESSUR~ FAIIIUTIES
`
`Figure 5-1. Flow diagram of hot potassium carbonate process for the absorption of
`C02, split-stream configuration. A = cooled lean solution; B = main lean solution
`stream; C = rich solution; 1 = feed gas; 2 = purified gas; 3 = acid gas. (UOP, 1993)
`
`Akermin, Inc.
`Exhibit 1008
`Page 14
`
`

`

`Alkaline Salt Solutions for Acid Gas Removal 335
`
`Figure 5-2. Two Benfield trains for natural gas purification. Courtesy of UOP and
`Natural Gas Corporation of New Zealand
`
`compared with the original amine activator (Bartoo et al., 1991). To date, the Benfield
`process is practiced in more than 600 plants worldwide for the removal of carbon dioxide
`and hydrogen sulfide from ammonia synthesis gas, crude hydrogen, natural gas, town gas,
`and others (UOP, 1992B). The process is licensed under the name of "The Benfield Process"
`by UOP, Tarrytown, New York.
`
`Akermin, Inc.
`Exhibit 1008
`Page 15
`
`

`

`Gas Purification
`
`The applicability of potassium carbonate to CO2 removal has been known for many years.
`A German patent, granted as early as 1904, described a process for absorbing COx in a hot
`solution of potassium carbonate and then stripping the solution by pressure reduction without
`additional heating (Behrens and Behrens, 1904). Williamson and Mathews (1924) studied
`the rote of absorption of COx in potassium carbonate solution and found that increasing the
`absorption temperature to 75°C (167°F) from 25°C (77°F) greatly increased the rate of
`absorption. The work of the U.S. Bureau of Mines, however, constitutes a major contribu-
`tion, in that it resulted in the development of an economically commercial process. A patent
`covering one aspect of this work was issued to Benson and Field in Great Britain (1955).
`This patent describes the use of potassium carbonate solution as an absorbent at temperatures
`near its amaospheric boiling point and its regeneration by flashing and steam stripping.
`As a result of the high-absorber temperature, the steam, which other regenerative process-
`es require to heat the solution to slripping temperature, is not required in the hot potassium
`carbonate system. In addition, the need for heat-exchange equipment between the absorber
`and slripper is eliminated. The high temperature also increases the solubility of potassium
`bicarbonate, thus permitting operation with a highly concentrated solution.
`
`Process Description
`
`As can be seen from the towdiagram, Figure 5-1, the process is extremely simple. In the
`split-stream process shown, a portion of the lean solution from the regenerator is cooled and
`fed into the top of the absorber, while the major portion is added hot at a point some distance
`below the top. This simple modification improves the purity of the product gas by decreasing
`the equilibrium vapor pressure of CO2 over the portion of solution last contacted by the gas.
`A somewhat more complex scheme termed "two-stage," has also been used for applications
`in which more complete CO2 removal is required (see Figure 5-3). In this modification, the
`main solution-stream is withdrawn from the stripping column at a point above the reboiler so
`that only a portion of the solution passes down through the bottom of the stripping column to
`the reboiler. Since this portion of the solution is regenerated by the total steam supply to the
`slripping column, it is thoroughly regenerated and is capable of reducing the CO2 content of
`the gas to a low value. The main solution-stream is fed into the midpoint of the absorber,
`while the more completely regenerated portion is fed at the top.
`Packing is used almost exclusively today in Benfield units in preference to trays. Metal
`packing is the standard, although ceramic and plastic rings are found in some units. Neither
`ceramic nor plastic packings are recommended for hot potassium carbonate service because
`plastic rings have a tendency to melt or deform at the high operating temperatmes, and the
`ceramic packing can become brittle and deteriorate with time causing small pieces to break
`off and damage other areas of the unit. Both carbon steel and stainless steel types are com-
`mon depending on the particular processing conditions. Choice of an appropriate packing
`can lead to optimization in grassroots facilities and inoreased capacity or debottlenecking in
`revamped units. See Chapter 1 for details on packing types and sizes.
`The basic Benfield flow scheme, without any heat conservation features, typically has a
`net heat duty in the range of 45,000 to 50,000 Btu/lb-mole COx. Several modifications of the
`basic flow patterns have been used, aimed primarily at greater heat economy and higher
`product gas purity. Examples of modified flow schemes are shown in Figures 5-3, 5-.4, 5-5,
`and 5-6. The Benfield LoHeat process, illustrated in Figure 5-3, uses low level heat, which
`would otherwise be lost in a solution cooler or in an overhead condenser, to satisfy part of
`the regeneration heat requirement. Hot lean solution exiting the regenerator is reduced in
`
`Akermin, Inc.
`Exhibit 1008
`Page 16
`
`

`

`Alkaline Salt Solutions for Acid Gas Removal
`
`REFLUX ~JMP
`
`A~D G~
`
`AriD GAS
`CO6E)F.NSi~
`HYDRAUMC POWE~
`
`REGENERAIOR
`
`CONDENSA~ COND£NSA~
`PUMP
`TANK
`
`LF.~N SOLUTION lEAN S~4.LmON RE~OILER
`S~IVE-LEAN
`S~AI.LF.R~ S4~.BllQN
`C.~qgONA’I~
`COOLER
`SOLUllQN PUMP
`FLASH T~K
`RE~OILER
`
`~JMP
`
`FEEDGAS
`
`Figure 5-3. Flow diagram of two-stage Benfield LoHeat process with internal steam
`generation. A = lean solution; B = ~mi-lean solution; C = rich solution; 1 = feed gas;
`2 = purified gas; 3 = acid gas. (UOP, 1993)
`
`1. CO(ORS C~N BE AIR C~ WA~ COOLB).
`2. HYDRAULIC TUR911~ IS USUALLY USF.D IN
`MO~EI~AI~ TO HIGH PRESS~J~.E FACR.ITI~S
`
`HEA11NG MEDIUM
`
`Figure 5-4. Flow diagram of Benfield LoHeat process wib~ bob~ steam ejectors and
`mechanical vapor recompression. A = lean solution; B = rich solution; 1 = feed gas;
`2 = purified gas; 3 = acid gas. (UOP, 1993)
`
`Akermin, Inc.
`Exhibit 1008
`Page 17
`
`

`

`Gas Purification
`
`1. CQ4)&~LS CA~I BE A~R OR WATER COOL.El),
`2, H’~)RAUMC ~ IS USUALLY USED IN
`
`Figure 5-5. Flow diagram of Benfield Hi-Pure process with LoHeat system, A = cooled
`lean solution; B = main solution stream; C = rich solution; 9 = lean amine; E = rich
`amine; 1 = feed gas; 3 = acid gas. (IJOP, 1993)
`
`PURE GAS
`
`CARBoNA’I~
`
`LF.N4 SOLUTION
`
`ACID GAS
`
`C&RBONA~£
`REGENERATOR
`
`REBOILE~
`
`FEED GAS ~
`
`Figure 5-6. Flow diagram of Enhanced LoHeat Benfield process with one-stage rich
`solution flash. (UOP, 1993)
`
`SOLUTION PUMP
`
`Akermin, Inc.
`Exhibit 1008
`Page 18
`
`

`

`Alkaline Salt Solutions for Acid Gas Removal
`
`pressure to produce steam. Most LoHeat units use multiple stages to increase the overall
`energy efficiency by 10 to 15% (Bartoo et al., 1991). Each stage operates about 2 psi below
`the preceding stage. The flashed steam is then recompressed using a steam ejector or
`mechanical compressor and re-injected into the base of the regenerator (Grover and Holmes,
`1987). As described by Baker and McCrea (1981), usable heat recovered from gas and liquid
`streams may reduce outside energy requirements by as much as 60% and, in some cases,
`result in the export of low pressure steam. It should be noted, as shown in Figure 5-3, that
`waste heat reboiling of internal reflux water to generate motive steam for the LoHeat ejector
`system via a condensate reboiler can further reduce the net energy requirements.
`Whether the LoHeat process is used or not, the amount of stripping steam or gross regen-
`eration energy required for solution regeneration is approximately the same. However, in a
`Ben field unit that utilizes the LoHeat process, part of the gross energy requirement is gener-
`ated internally through the flash heat recovery. Hence, the total external energy requirement,
`or net regeneration energy, is reduced. Typical LoHeat energy requirements depend on the
`process configuration, but usually range from 30,000 to 35,000 Btu/lb-mole CO2 when steam
`ejectors are used.
`One important LoHeat option is the Ben field hybrid LoHeat scheme. As shown in Figure
`5-4, this LoHeat system utilizes a combination of steam ejectors and a mechanical vapor
`recompressor (MVR) (Grover and Holmes, 1987). Typically, a multi-stage flash tank is
`employed with the first few stages operating on steam ejectors and a final stage using an
`MVR. The MVR allows for larger recompression ratios, thereby allowing a deeper flash and
`hence increased energy savings. Typical heating requirements for a hybrid LoHeat system
`will range from 25,000 to 28,000 Btu/lb-mole CO2.
`Another modification described by Benson and Parrish (1974) is the HiPure process, which
`is capable of producing a treated gas containing less than one part per million of H2S and less
`than 50 parts per million of CO~. The ability to remove these compounds to such low levels
`makes it an excellent choice for purification of natural gas to pipeline purity. The Benfield
`HiPure process has also been used in large liquefied natural gas (LNG) facilities where
`extremely low product specifications for CO~, H~S, COS, and mercaptans are required.
`This process, as shown in Figure 5-5, uses two independent, but compatible circulating
`solutions in series to achieve high purity combined with high efficiency. The process gas is
`first contacted with normal hot potassium carbonate followed by contact with aqueous amine
`solution. The hot potassium carbonate serves to remove the bulk of the acid gases, while
`final purification is achieved with the second solution. The two solutions are regenerated
`separately in two sections of a regenerator with the stripping steam leaving the lower section
`of the regenerator being re-used in the upper section. The two systems are thermally integrat-
`ed by using waste heat from the amine circuit to provide a portion of the regeneration heat in
`the carbonate circuit. The combined heat required for the two solutions is generally lower
`than that for a conventional hot carbonate system. Although the capital cost of the HiPure
`unit is somewhat higher than that of a normal Benfield unit due to the additional equipment
`required, the savings in heat energy and the ability to produce high purity product gas make
`this process quite attractive.
`An extension of the conventional Benfield process, termed the "Enhanced LoHeat Ben-
`field Process," has been developed (G-rover, 1987). As shown in Figure 5-6, the thermal
`energy required for regeneration is substantially reduced as a result of a one or two-stage
`rich solution flash step, which serves as the driving force for a major portion of the regenera-
`tion. The balance of the regeneration energy is supplied via conventional thermal stripping.
`The bulk of the rich solution is regenerated in the flash tank and recycled back to the bottom
`
`Akermin, Inc.
`Exhibit 1008
`Page 19
`
`

`

`340
`
`Gas Purification
`
`section of the absorber where about two-thirds of the CO2 contained in the feed gas is
`removed. The top section of the absorber receives thermally regenerated solution, which is
`used to remove the remaining CO2 and produces a treated gas having the desired low level of
`impurities. Low level waste heat within the process is used to enhance the rich solution flash.
`Typical heating requirements depend on whether a one- or two-stage rich solution flash is
`chosen, but usually range from 18,000 to 25,000 Btu/lb-mole CO2. The approximate heat
`requirement for several versions of the Benfield process is shown in Figure 5-7 as a function
`of the acid gas partial pressure.
`
`A large amount of comprehensive physical data on the potassium carbonate-potassium
`bicarbonate-carbon dioxide-water and potassium carbonate-potassium bicarbonate-potassi-
`um bisulfide-carbon dioxide-hydrogen sulfide-water systems is available in literature (Ben-
`son et al., 1954; Benson et al., 1956; Tosh et al., 1959; Tosh et al., 1960; Allied Chemical
`Corp., 1961; Bocard and Mayland, 1962). Some typical data are given below; however, for
`complete information, the reader is referred to the original sources.
`The effect of temperature and percentage conversion to bicarbonate on the solubility of the
`salts in the system, potassium carbonate bicarbonate, has been determined by Benson et al.
`(1954), and their data are presented in Figure 5-8, together with data from other literature
`
`0
`O
`
`I
`40
`
`I
`80
`
`PSi
`
`I
`120
`
`160
`
`ADD 20~ FOR HII:~RE PROCF~q~
`
`FE~’D GA~ PARTIAL PR~’__~u_ JRE OF (CC~ + H2,S}
`
`Figure 5-7, Approximate heat requirements of Benfield process systems as a function
`of acid gas partial pressure in the feed gas. (Bartoo, 1984)
`
`Akermin, Inc.
`Exhibit 1008
`Page 20
`
`

`

`Alkaline Salt Solutions for Acid Gas Removal 341
`
`(cid:128).~1 I
`
`8 Bo
`
` .00
`
`/ /
`
`/ ,, ,/ /
`///
`
`80
`
`100 I~0 140 160 IBO ~ ~0 ~40 ~60 ~BO 300
`TEMPERA~RE, *F
`Figure 5-8. Effect of temperature and percentage conversion to bicarbonate on
`solubility of K2C03 plus KHC03. Lines represent solubility limits for given salt
`concentrations (measured as equivalents of I~CO~). Forty, 50, and 60% lines based on
`data of Bensan et al. (1954); other lines based on data of Perry (1950), Seidell (1940),
`and Hill and Hill (1927)
`
`sources (Perry, 1950; Hill and Hill, 1927; Seidell, 1940). Lines on the chart represent the con-
`ditions under which crystals of potassium bicarbonate begin to precipitate for varying potassi-
`um carbonate concentrations. At 240°F, for example, the 60% solution can be converted to
`only about 30% bicarbonate without the formation of a precipitate. A 50% solution can be
`50% converted; and a 40% solution can theoretically be converted 100%. On the basis of
`these data, it is concluded that a 40% equivalent concentration of potassium carbonate is
`about the maximum that can be used for the treating operation without precipitation occur-
`ring, and a 30% solution is considered to be a reasonable design value for most applications.
`If cooling of tire solution occurs at any point in the system, even 30% potassium carbonate
`may be too great a concentration. On the basis of commercial-plant experience with units
`treating natural gas, Buck and Leitch (1958) recommend 30% potassium carbonate equiva-
`lent as a maximum concentration. They found no appreciable effect on the absorptive capac-
`ity of the solution when its concentration was reduced to as low as 20%.
`The equilibrium vapor pressure of COz over a solution containing the equivalent of 20%
`potassium carbonate as a function of conversion to bicarbonate, based on the data of Tosh et
`al. (1959), is presented in Figure 5-9. These authors, who investigated vapor/liquid equilib-
`ria for 20, 30, and 40% equivalent potassium carbonate solutions, found that the COz equi-
`librium vapor pressure remains practically the same for the range of 20 to 30%. This in
`effect confirms the observation of Buck and Leitch (1958) for commercial operations. The
`experimental CO2 vapor pressure data were used by Tosh el al. (1959) as the basis for calcu-
`
`Akermin, Inc.
`Exhibit 1008
`Page 21
`
`

`

`342
`
`Gas Purification
`
`.01
`0 20 40 60 80 100
`PERCENT OF K2C03 CONVERTED TO KHC03
`Figure 5-9. Equilibrium vapor pressure of C02 over 20 percent equivalent potassium
`carbonate solution. (Tosh et aL, 195g~
`
`lating the equilibrium constant, K, for the three solution concentrations according to the
`expression:
`
`K = (KHCO3)2/(K2CO3)Pco2
`
`(5-11)
`
`with KHCO3 and K2CO3 expressed in gram moles per liter and Pco2 in mm Hg. K was found
`to be constant at a given temperature for each degree of conversion for the 20 and 30%
`
`Akermin, Inc.
`Exhibit 1008
`Page 22
`
`

`

`Alkaline Salt Solutions for Acid Gas Removal 343
`
`potassium carbonate solutions. From the values of K, the equilibrium vapor pressure can be
`calculated for any conversion within the given range of solution concentrations. Table 5-1
`shows average K values (arithmetic mean of all experimental data points) for 20 and 30%
`solutions (Tosh et al., 1959).
`
`Table 5-1
`Average Values of K for 20 and 30 percent K2COa Solutions
`
`Temperature °C
`
`K, 20% Solution
`
`K, 30% Solution
`
`70
`90
`110
`130
`
`0.042
`0.022
`0.013
`0.0086
`
`0.058
`0.030
`0.017
`0.011
`
`Source: Data of Tosh et al., 1959
`
`The equilibrium vapor pressure of water over a solution containing the equivalent of 20%
`potassium carbonate as a function of conversion to bicarbonate is shown in Figure 5-10.
`This chart is also based on data obtained by Tosh et al. (1959). Here, again, there is not much
`difference in the vapor pressure of water for 20 and 30% equivalent potassium carbonate
`concentrations.
`Additional vapor-liquid equilibrium data based on published information and experimen-
`tal work have been reported by Bocard and Mayland (1962). Other physical data on the
`potassium carbonate-potassium bicarbonate-carbon dioxide system are shown in Figures 5-
`11 to 5-14. The data of Bocard and Mayland (1962) have been converted to a series of
`homographs by Mapstone (1966).
`The potassium carbonate-potassium bicarbonate-potassium bisulfide-carbon dioxide-
`hydrogen sulfide-water system has been studied extensively by Tosh et al. (1960) and Field
`et al. (1960) o