You must log in to edit PetroWiki. Help with editing

Content of PetroWiki is intended for personal use only and to supplement, not replace, engineering judgment. SPE disclaims any and all liability for your use of such content. More information

PEH:Gas Treating and Processing

Jump to navigation Jump to search

Publication Information


Petroleum Engineering Handbook

Larry W. Lake, Editor-in-Chief

Volume III – Facilities and Construction Engineering

Kenneth E. Arnold, Editor

Chapter 5 – Gas Treating and Processing

Edward Wichert, Sogapro Engineering Ltd.

Pgs. 185-229

ISBN 978-1-55563-116-1
Get permission for reuse

Natural gas is a mixture of many compounds, with methane (CH4) being the main hydrocarbon constituent. When natural gas is produced from an underground reservoir, it is saturated with water vapor and might contain heavy hydrocarbon compounds as well as nonhydrocarbon impurities. In the raw state, natural gas cannot be marketed, and, therefore, it must be processed to meet certain specifications for sales gas. Additionally, it might be economical to extract liquefiable hydrocarbon components, which would have a higher market value on extraction as compared with their heating value if left in the gas.

Before the optimum design of any gas treating plant can be decided, at minimum, one must know the raw gas production capability to the plant; composition of separator inlet gas and condensate, and relative condensate/gas rates; specifications for the residue gas; and rate of gas sales.

The end user of natural gas needs to be assured of two conditions before committing to the use of gas in a home or factory: the gas must be of consistent quality, meeting sales gas specifications; and the supply of gas must be available at all times at the contracted rate.

Gas treating facilities, therefore, must be designed to convert a particular raw gas mixture into a sales gas that meets the sales-gas specifications, and such facilities must operate without interruption.

Typical Sales-Gas Specifications

Specifications for sales gas describe the required physical properties of the gas such that it can be transported under high pressure through long distance pipelines at ground temperature without forming liquids, which could cause corrosion, hydrates, or liquid slugs into downstream equipment. Limits on the content of certain nonhydrocarbon compounds are also specified. While the specific limits for each item might vary among transmission companies or customers, the overall specifications for sales gas generally include:

  • Maximum hydrocarbon dewpoint temperature at a pressure of 800 psig.
  • Maximum allowable CO2 content.
  • Maximum allowable H2S content and total organic sulfur content.
  • Maximum allowable water-vapor content.
  • Maximum allowable temperature of gas leaving the plant.
  • Minimum pressure to enter the gas transmission grid.
  • Minimum heating value.
  • Free of dust, treating chemicals, and other contaminants from the process plant.

In long distance transmission of sales gas by pipeline, the pressure is usually less than 1,000 psig. It is important that no liquids form in the line because of condensation of either hydrocarbons or water. Hydrocarbon liquids reduce the pipeline efficiency and might hold up in the line to form liquid slugs, which might damage downstream compression equipment. Condensed water can do the same. Additionally, water could form solid complexes (hydrates), which accumulate and block the line. The dewpoint temperature at any pressure is the temperature at which either hydrocarbons or water condense upon cooling of the gas. Thus, the specifications for sales gas include limits on the hydrocarbon dewpoint temperature, as well as limits on the water vapor content of the gas.

Knowing the specifications, and knowing the required sales gas flow rate and the composition of the raw gas and condensate entering the plant, the various process vessels can be designed, and the optimum process conditions of pressure and temperature can be specified.

Depending on the composition of the inlet fluids and the price at the plant gate, other sales products might be recovered in the plant as well. These could be ethane, propane, butane and pentanes-plus, which also must meet stringent specifications concerning purity. The possible processing steps, as illustrated in Fig. 5.1, are inlet separation, compression, gas sweetening, sulfur recovery or acid gas disposal, dehydration, hydrocarbon dewpoint control, fractionation and liquefied petroleum gas (LPG) recovery, and condensate stabilization.

Except for gas sweetening, the processing steps involve no chemical reactions. The gas/liquid product specifications are achieved by separating the compounds through changing the physical conditions of temperature and pressure to which the fluids are exposed. Contact with other compounds, such as glycol and absorption oil, affects the relative solubilities of certain compounds, thereby achieving separation from the main gas stream. Exposure to dry compounds, such as silica gel or molecular sieves, separate some compounds from the gas stream by physical adsorption. Distillation is used to separate the various hydrocarbon compounds into liquid fractions on the basis of differences in their volatilities.

Sour Gas Sweetening

Definition of Sour Gas

Sour gas is natural gas that contains hydrogen sulfide (H2S). A natural gas is "sour" when the H2S content of the gas mixture exceeds the limit imposed by the purchaser of the gas, usually a transmission company or the end user. Generally, the limit for H2S content is one grain of H2S per 100 scf of sales gas. The limit for H2S content in sales gas in some areas is 1/4 grain of H2S per 100 scf of gas. The mass specification of one grain per 100 scf converts a volumetric limit of 16 ppm. Sour natural gases can contain H2S in concentrations from several ppm to over 90%.

While the foregoing defines sour gas from a sales gas perspective, the standards and regulations applying to sour gas service and operations may use a different definition. In this respect, the National Association of Corrosion Engineers (NACE), which developed the standards for materials for use in sour service, defines sour gas service, to which their standards apply, on the basis of partial pressure of H2S and total pressure. NACE Standard MR0175 applies to natural gas systems having a partial pressure of H2S of 0.05 psia or greater, at an absolute pressure above 65 psia. If the partial pressure of H2S is at or above these limits, the steel and other equipment exposed to the sour gas must meet the conditions specified in NACE Standard MR0175 for sour service.

Other Sulfur Compounds

While H2S is the compound responsible for designating a natural gas as sour, there are other sulfur compounds, also present in sour gas, in much smaller concentrations. The sales gas specifications normally set a limit of 5 grains per 100 scf of gas for total sulfur content. Thus, sweetening solvents must be able to remove other sulfur compounds, as well as H2S, from the sour gas to meet the total sulfur limitation. Some of the other common sulfur compounds found in sour natural gas are mentioned next.

Mercaptans are compounds that occur naturally in sour gas. They are hydrocarbon compounds that have a sulfur atom inserted between a carbon atom and a hydrogen atom. Methyl, ethyl, propyl, and butyl mercaptan are some examples. Mercaptans have a strong offensive odor, and certain mixtures of manufactured mercaptan, such as tertiary butyl mercaptan and isopropyl mercaptan, as well as others, are added to sweet natural gas to odorize the gas prior to domestic or commercial consumption.

Carbon disulfide (CS2) and carbonyl sulfide (COS) might also be present in gases containing H2S, but usually only in small concentrations. These compounds also have a strong sour gas odor and are largely extracted from the sour gas in the sweetening operation.

CO2 Removal

Virtually all sour natural gases also contain CO2, but the converse is not true: many natural gas mixtures can contain CO2 without any H2S. CO2 and H2S are called "acid gases," or collectively "acid gas," as both gases are slightly soluble in water and form a mildly acidic solution. Most regenerative processes used for H2S removal from natural gas also remove CO2. If a natural gas contains amounts of CO2 in excess of the sales gas limit, but no H2S, there are specific processes available for CO2 removal.

Overview of Sweetening Process

Sour natural gas compositions can vary over a wide concentration of H2S and CO2, as well as a wide variation in concentration of the hydrocarbon components. If the H2S content exceeds the sales gas specification limit, the excess H2S must be separated from the sour gas. The removal of H2S from sour gas is called "sweetening." The process selected for sweetening a sour gas depends on the general conditions: H2S and mercaptan concentration in the sour gas, and sales gas H2S and total sulfur limits; maximum design flow rate; raw gas inlet pressure; requirement for sulfur recovery; and acceptable method of waste products disposal.

The selected process must be cost effective in meeting the various specifications and requirements. Throughout the world, regulations generally limit the flaring of H2S. Sweetening of gas streams containing very low concentrations of H2S can be done in many ways, depending on the general conditions. If the sour gas stream contains more than 70 to 100 pounds of sulfur/day in the form of H2S in the inlet gas to a sour plant, a regenerative chemical solvent is usually selected for the sweetening of the sour gas stream. For very low H2S content sour gas, a scavenger chemical is usually used. In such cases, the chemical is consumed, and the method for ultimate disposal of the spent chemical is a consideration.

Typical Process Equipment for Sweetening Sour Gas With a Regenerative Solvent

A schematic drawing of typical process equipment for sweetening sour gas with regenerative solvent is shown in Fig. 5.2. The first vessel is the inlet separator, which performs the important function of separating the fluid phases on the basis of density difference between the liquid and the gas. The sour gas flows from the separator into the lower part of the absorber or contactor. This vessel usually contains 20 to 24 trays, but for small units, it could be a column containing packing. Lean solution containing the sweetening solvent in water is pumped into the absorber near the top. As the solution flows down from tray to tray, it is in intimate contact with the sour gas as the gas flows upward through the liquid on each tray. When the gas reaches the top of the vessel, virtually all the H2S and, depending on the solvent used, all the CO2 have been removed from the gas stream. The gas is now sweet and meets the specifications for H2S, CO2, and total sulfur content.

The rich solution leaves the contactor at the bottom and is flowed through a pressure letdown valve, allowing the pressure to drop to about 60 psig. In some major gas plants, the pressure reduction is accomplished through turbines recovering power.[1] Upon reduction of the pressure, the rich solution is flowed into a flash drum, where most dissolved hydrocarbon gas and some acid gas flash off. The solution then flows through a heat exchanger, picking up heat from the hot, regenerated lean solution stream. The rich solution then flows into the still, where the regeneration of the solvent occurs at a pressure of about 12 to 15 psig and at the solution boiling temperature. Heat is applied from an external source, such as a steam reboiler. The liberated acid gas and any hydrocarbon gas not flashed off in the flash drum leave the still at the top, together with some solvent and a lot of water vapor. This stream of vapors is flowed through a condenser, usually an aerial cooler, to condense the solvent and water vapors. The liquid and gas mixture is flowed into a separator, normally referred to as a reflux drum, where the acid gas is separated from the condensed liquids. The liquids are pumped back into the top of the still as reflux. The gas stream, consisting mainly of H2S and CO2, is generally piped to a sulfur recovery unit. The regenerated solution is flowed from the reboiler or the bottom of the still through the rich/lean solution heat exchanger to a surge tank. From here, the solution is pumped through a cooler to adjust the temperature to the appropriate treating temperature in the absorber. The stream is then pumped with a high-pressure pump back into the top of the absorber, to continue the sweetening of the sour gas.

Most solvent systems have a means of filtering the solution. This is accomplished by flowing a portion of the lean solution through a particle filter and sometimes a carbon filter as well. The purpose is to maintain a high degree of solution cleanliness to avoid solution foaming. Some solvent systems also have a means of removing degradation products, by having an additional reboiler for this purpose in the regeneration equipment hook-up. In some designs, the rich solution is filtered after it leaves the surge drum.

Sweetening Solvents

The desirable characteristics of a sweetening solvent, used in the process just discussed, are outlined next.

  • Required removal of H2S and other sulfur compounds must be achieved.
  • Pickup of hydrocarbons must be low.
  • Solvent vapor pressure must be low to minimize solvent losses.
  • Reactions between solvent and acid gases must be reversible to prevent solvent degradation.
  • Solvent must be thermally stable.
  • Removal of degradation products must be simple.
  • The acid gas pickup per unit of solvent circulated must be high.
  • Heat requirement for solvent regeneration or stripping must be low.
  • The solvent should be noncorrosive.
  • The solvent should not foam in the contactor or still.
  • Selective removal of acid gases is desirable.
  • The solvent should be cheap and readily available.

Unfortunately, there is no one solvent that has all the desirable characteristics. This makes it necessary to select the solvent that is best suited for treating the particular sour gas mixture from the various solvents that are available.[2] The sour natural gas mixtures vary in H2S and CO2 content and ratio; content of heavy or aromatic compounds; and content of COS, CS2, and mercaptans.

While most of the sour gas is sweetened with regenerative solvents, for slightly sour gas, it may be more economical to use scavenger solvents or solid agents.[3] In such processes, the compound reacts chemically with the H2S and is consumed in the sweetening process, requiring the sweetening agent to be periodically replaced.

Regenerative Chemical Solvents

Most of the regenerative chemical sweetening solvents are alkanolamines, which are compounds formed by replacing one, two, or three hydrogen atoms of the ammonia molecule with radicals of other compounds to form primary, secondary, or tertiary amines respectively.[4] Amines are weak organic bases that have been used for many years in gas treating to remove CO2 and H2S from natural gas as well as from synthesis gas. These compounds combine chemically with the acid gases in the contactor to form unstable salts. The salts break down under the elevated temperature and low pressure in the still.

Because the chemical reactions are reversible by changing the physical conditions of temperature and pressure between absorber and still, amines are highly suitable for removing the acid gases from the hydrocarbon gas stream. However, sometimes nonreversible reactions take place to a minor degree, forming degradation compounds. Such degradation compounds must be periodically removed by distillation. This occurs in a reclaimer vessel. Primary amines are more prone to forming degradation compounds than the other solvents.

Another chemical solvent, potassium carbonate (K2CO3), has also been used for removing H2S and CO2 from manufactured or natural gas. It is not widely accepted in the sour natural gas industry, but it is periodically mentioned in the literature as finding some application under specific conditions. This chemical solvent was originally developed by the U.S. Bureau of Mines for removing CO2 from manufactured gas. Table 5.1 lists the regenerative chemical solvents generally used for sweetening sour gas and gives the acronyms, chemical formulas, and molecular weights.

Primary Amines

Monoethanolamine (MEA).[5] MEA was the earliest amine used for sweetening sour gas. It is a stronger base than DEA and also has a higher vapor pressure than DEA; therefore, vapor losses are higher than for DEA. MEA forms nonregenerative (degradation) compounds with CO2, COS, and CS2, which is a disadvantage, as the degradation compounds must be removed periodically to lessen the corrosion rate. A reclaimer is usually incorporated in an MEA sweetening train to periodically remove the degradation products from the solution by distillation. MEA has been used for more than 60 years in process applications, and the process operation and problem areas are well understood. Solution strengths of MEA are usually in the range of 15 to 22% by weight MEA in water. Mol loadings (moles of acid gas picked up in the contactor per mole of solvent circulated) are generally in the range of 0.25 to 0.33 moles acid gas per mol of MEA.

Diglycolamine (DGA). The DGA process was developed by The Fluor Corp. in the 1950s, which called the process the Econamine Process. The advantage of DGA over MEA appears to be the lower solution circulation rate owing to the higher solvent concentration, resulting in higher acid gas pickup per volume of solution circulated.[6] This yields capital savings, as the regeneration equipment is smaller for DGA than for MEA. Disadvantages appear to be degradation of the chemical with CO2 and greater solubility of heavier hydrocarbons in the solution, as compared to MEA. This is a serious drawback if the acid gas stream is fed to a Claus plant, as additional air is required for the combustion of the hydrocarbons. Also, this dilutes the sulfur compounds in the sulfur recovery train. Solution strength is on the order of 50 to 70% by weight of DGA in water, with mol loadings in the range of 0.3 to 0.4 moles of acid gas per mole of DGA circulated. The DGA process train usually includes a reclaimer.

Secondary Armines

Diethanolamine (DEA).[7] DEA became a popular sour gas treating solvent in the 1960s after it was developed for such application in France. It can be used at higher concentrations than MEA. DEA has the advantage of picking up more acid gas per solution volume circulated, thus effecting some energy saving in circulation and regeneration. It does not form the nonregenerative products with COS and CS2 as is the case with MEA, which is another advantage over MEA. DEA is also generally less corrosive than MEA. Solution strength is usually in the 25 to 40% range, with mol loadings of 0.35 to 0.63.

Diisopropanolamine (DIPA). This secondary amine is not used by itself as a sweetening solvent but is part of the Sulfinol solvent formulation.

Tertiary Armines

Triethanolamine (TEA). TEA is not in general use for gas sweetening.

Methyldiethanolamine (MDEA).[8] MDEA reacts more slowly with CO2 than the previously described amines. It forms a slightly different salt with CO2 from those of the other amines, at a lower rate of reaction. The difference in the rates of reaction with H2S and CO2 gives MDEA a desirable feature over other amines, namely selectivity of H2S over CO2 . This is an attractive feature in cases where it is not necessary to remove all the CO2 from the gas stream. By leaving some of the CO2 in the natural gas, the circulation rate of the solution can be reduced, or the treating capacity of an existing unit can be increased with MDEA as compared with DEA.

MDEA concentrations are on the order of 30 to 50% by weight, with mol loadings of 0.40 to 0.55 moles acid gas per mol of amine. As a tertiary amine, MDEA is naturally a weaker base and is, therefore, less corrosive than the primary and secondary amines. The energy required for regeneration is also less than the requirement for the other amines.

Proprietary Armine Solvent Formulations

By making use of the difference in the rates of reaction between MDEA and the acid gases, several proprietary solvents have been developed that are suited for preferential extraction of H2S, with only partial removal of CO2 . These proprietary formulations usually contain MDEA plus other amines at various concentrations in aqueous solutions to tailor them for specific applications. Several chemical companies have developed such proprietary solvents.[3]

Hot Potassium Carbonate (K2CO3) (Hot Pot) [9]

The potassium carbonate process was developed for removing CO2 from manufactured gas. It reacts with both acid gases. Because the contacting of the sour gas occurs at very high temperatures, such as 195 to 230°F in this process, it is sometimes referred to as the "hot pot" process. It requires lower heat input for regeneration and, therefore, is somewhat less costly to operate than some amine processes. Also, no heat exchanger is required in the regeneration equipment. The process has difficulty in meeting the H2S specification of the treated gas, if the H2S/CO2 ratio is not extremely small. This process is significant for treating gas with a large concentration of CO2. The chemistry of the process can be enhanced by the addition of various catalysts, and this has resulted in the process being referred to by various trade names.

Computer Simulation of Sweetening Processes

The design and optimization of sweetening processes can be done by computer, using programs such as HYSIM from Hyprotech of Calgary, which contains AMSIM from D.B. Robinson and Associates Ltd. of Edmonton, Alberta; ProTreat from Optimized Gas Treating Inc. of Houston; TSWEET from Bryan Research and Engineering Inc. of Bryan, Texas; and proprietary programs from the chemical supplier.

Estimating Solution Circulation Rate

The circulation rate or the acid gas pickup by the solution can be estimated with the next two formulas.



The specific gravity of the amine solutions is shown in Fig. 5.3.[10] The specific gravity of potassium carbonate may be approximated by 1 + weight fraction K2CO3 in solution.

Types of Operating Problems

The main problems that can be encountered in the operation of sour gas treating facilities using chemical solvents are failure to meet H2S specification for sales gas; solution foaming in the contactor or regenerator; corrosion in pipes and vessels; and solvent losses.[11]

Failure to Meet H2S Sales-Gas Specifications. Treated gas that does not meet the H2S specifications is not admitted into the sales-gas transmission lines. Potential causes for "going sour" are a change in the acid gas concentration of feed gas; a change in the feed gas temperature; too hot lean amine solution; too low solvent concentration in solution; inadequate regeneration of solution; insufficient contact in absorber; too low amine circulation rate; too low absorber pressure; too high concentration of degradation products; too high inlet gas rate; mechanical damage or problems in absorber; and foaming.

Solution Foaming. Solution foaming occurs when gas is mechanically entrained in liquid as bubbles.[12] The tendency to form bubbles increases with decreasing surface tension of the solution owing to interference of foreign substance at the surface of the solution on the tray. Foaming is thought to be caused by factors such as: liquid hydrocarbons entering the contactor with the sour gas; acidic amine-degradation products; treating chemicals from wells or gathering system; treating chemicals from makeup water; compressor oil; and fine solid suspensions, such as iron sulfide.

While solids suspended in the solution by themselves might not cause foaming, it is thought that they tend to stabilize the foam. The results from foaming can be severe upsets in the process tower, leading to carryover and loss of chemical, and possible damage to downstream process equipment or material. The best way to reduce the propensity for foaming is to ensure that the sour gas entering the contactor is clean, free of condensed liquids, and that the solution is cleaned up by mechanical and carbon filtration. The addition to the solution of antifoam agents is sometimes effective in controlling the foaming tendency of the solution. However, this does not solve the basic problem. Too much antifoam in the solution can actually add to the foaming problem.

Corrosion. Corrosion is common in most amine plants. It is necessary to control the corrosion rate by the addition of corrosion inhibitor and by use of stainless steel in certain pieces of process equipment. In the case of MEA solutions, corrosion rates tend to increase with increasing solution strengths beyond about 22% MEA, as well as with high levels of amine degradation products in the solution. Most of the process piping and vessels in amine plants are built with carbon steel, meeting NACE MR0175 guidelines.

It is not possible to predict with certainty where corrosive attack will take place. Experience has shown that the most likely areas for corrosive attack are those where the temperatures are high, such as in the top part of the still, the reboiler tubes, and heat exchangers, as well as some connecting piping. Hydrogen blisters are sometimes evident after many years of service in the shell of the contactor or still. Hydrogen-induced cracking can also occur in welds in the vessels or piping after many years of service. Corrosion/erosion can occur in areas where fluid velocities are high, such as in the return line from the reboiler, at the point of entry of the reboiler vapors into the still, and downstream of pressure letdown valves.

Corrosion rates in amine systems, especially MEA systems, generally increase with increasing temperature; increasing amine concentration; increasing mole loadings; and pure acid gas as compared with CO2 and H2S mixtures.[13] MEA is generally much more corrosive than DEA, and MDEA is only slightly corrosive.

Use of Corrosion Inhibitors. The use of corrosion inhibitors is a common practice to reduce the attack on steel by H2S and CO2 in aqueous environments. In most sour gas sweetening installations, a corrosion inhibitor is continuously injected into the sweetening solution.

Solvent Loss. In all regenerative solvent systems, it is necessary to periodically add pure solvent to the solution because of the loss of solvent during operation. Solvent losses in gas treating systems can occur because of vaporization, entrainment, degradation and removal of degradation products, and mechanical losses.

Solvents used in gas treating, like any other liquids, have a vapor pressure that increases with temperature. In a gas sweetening system, there are three vessels where gas and liquid streams separate: contactor, flash tank, and reflux drum. By far the largest gas stream is the one leaving the contactor. To reduce the solvent losses from this source, a water wash process is usually applied to the treated gas downstream of the contactor. Solvent losses from the flash tank are usually quite small, as the amount of gas leaving this vessel is usually small when compared to the total plant stream. When the solution is regenerated in the still, some solvent leaves the still overhead with the acid gas stream and the water vapor. Upon cooling the still overhead stream and condensing most of the water and amine, the liquid is returned to the top of the still as reflux, which also recovers most of the solvent. Nevertheless, some solvent vapor leaves the top of the reflux drum with the acid gas stream. Lower reflux drum temperatures reduce solvent losses at this point.

Entrainment of solvent occurs during foaming or under high gas velocity situations. By preventing foaming and by staying within design throughput, entrainment losses can be avoided.

In amine systems, some degradation of the solvent occurs. Primary amines are most susceptible to this problem, and such systems require special separation equipment to periodically remove the degradation products that contribute to corrosion. The degradation products are mainly caused by irreversible reactions between the solvent and CO2.

The most serious losses of solvent usually result from mechanical actions or problems. These include filter changeouts, drips from pumps or flanges, and vessel cleaning and draining.

Physical Solvents

In addition to the chemical solvents, there are also physical solvents available for extracting the acid gases from natural gas. Physical solvents do not react chemically with acid gases but have a high physical absorptive capacity. The amount of acid gas absorbed is proportional to the partial pressure of the solute, and no upper limit, owing to saturation, is evident, as is the case with chemical solvents. Hence, they are mainly suited for sour gases with high acid gas content at high contacting pressures. The physical absorption solvents have the advantage of regeneration by flashing upon reduction of pressure and, therefore, do not require much heat in the stripping column. This makes physical solvents useful as bulk-removal processes, followed by final cleanup using a chemical solvent because physical solvents have difficulty in achieving the H2S limit specified for sales gas. Unfortunately, they also tend to absorb heavier hydrocarbons, which is a disadvantage if the acid gas is fed to a Claus plant for sulfur recovery.

There are several physical solvents mentioned in the literature. Fig. 5.4 is a process schematic of a typical physical solvent process. A brief description of the more prominent physical solvent processes is discussed next.

Selexol Process[14]

The Selexol process was developed by Allied Chemical Corp. The solvent is dimethyl ether of polyethylene glycol and is usually used in its pure form. It has a preference for H2S over CO2, and, therefore, some CO2 remains in the gas stream, depending on the mole loading of the solvent. Several stages of flashing are provided for in the regeneration step, to allow the absorbed hydrocarbons to evolve from the solution. The flashed gases from the initial flash stages are compressed and returned to the inlet of the absorber. Selexol is noncorrosive and also removes water vapor from the gas stream.

Fluor Solvent Process[15]

The solvent in this process is propylene carbonate and was developed in the late 1950s by The Fluor Corp. This solvent also has a greater affinity for H2S than for CO2 and also dehydrates the feed gas. Absorptive capacity is highly temperature dependent, favoring the lower temperature.


The Purisol solvent process was developed in Germany by Lurgi. The solvent used is N-methylpyrrolidone, which has a high absorptive capacity for the acid gases.

Hybrid Process


The Shell Sulfinol process is a hybrid process using a combination of a physical solvent, sulfolane, and a chemical solvent, DIPA or MDEA. The physical solvent and one of the chemical solvents each make up about 35 to 45% of the solution with the balance being water. The sulfinol process is economically attractive for treating gases with a high partial pressure of the acid gases, and it also removes COS, CS2, and mercaptans. Other advantages are good heat economy, low losses because of low vapor pressure, and absence of corrosion. A disadvantage of this process is that the sulfolane absorbs heavier hydrocarbons from the gas, some of which are then contained in the acid gas feed stream to the sulfur plant. Thus, the Sulfinol process is best suited for very sour, lean gas.

Reduction/Oxidation (Redox) Process


Typical Sulferox process equipment is illustrated in Fig. 5.5. In this process, the sour gas is contacted in a small contactor in cocurrent flow with a water solution containing about 4% ferric iron ions, held in solution by proprietary chelate ligand. The H2S ionizes in the solution, and the ferric iron ions exchange electrons with the sulfur ions to form ferrous iron ions and elemental sulfur. The gas and the solution leave the contactor and are flowed into a separator. The gas out of the separator is sweet and requires further treating for dewpoint control. The solution is flowed into additional vessels for separating the elemental sulfur and for restoring the ferrous iron ions back to the active ferric iron state by contacting the solution with air. The chemistry is illustrated next.


Nonregenerative Chemical Solvent (Scavenger) Processes

When the gas is only slightly sour, that is to say, contains only a few ppm of H2S above the specification limit, a simpler sweetening process might have economic advantages over the typical processes described in the previous sections.[19] These processes scavenge the H2S from the sour gas, with the chemical being consumed in the process. It is, therefore, necessary to periodically replenish the chemical, as well as dispose of the end product of reaction containing the sulfur. Table 5.2 gives a summary of some common chemicals used for this purpose. The process equipment consists of a tower containing a solution of the chemical, or the chemical is in suspension in water. The sour gas is bubbled through the solution, and the chemical reacts with the H2S. The chemicals do not react with CO2.

The Sulfa-scrub process is a more recent development.[20] The chemical used in this process is triazine. The end product is beneficial as a corrosion inhibitor and is water soluble. As a result, disposal of the end product is convenient, as it is simply added to a water-disposal system. Sulfa-scrub can be injected into the flowline at the well, and it reacts with the H2S while the gas is flowing to the plant. Thus, there might not be a requirement for a treating tower.

Dry Sweetening Processes

While the sweetening of sour gas is predominantly done with regenerative solvents, there are also some dry processes that can be used for this purpose. Because these processes are batch processes, two or more towers are usually used, so that one tower can be taken out of service for chemical charge replacement without interruption of gas flow.

Iron Sponge (Iron Oxide) [21]

Iron sponge consists of wood chips that have been impregnated with a hydrated form of iron oxide. The material is placed in a pressure vessel through which the sour gas is flowed. Because this is a batch process, usually two vessels are installed—one in service and the other on standby. The H2S reacts with the iron oxide to form iron sulfide. In due course, the iron oxide is consumed. While it is possible to regenerate the iron sulfide with air to restore the iron oxide, in practice this is not done. Instead, the tower containing the spent iron sponge is taken out of service, and the standby tower is placed in service. The spent iron sponge is moistened with water, removed, and disposed of at an approved disposal site, and the tower is filled with a new charge of iron sponge. Care has to be exercised in handling the spent material in the dry state, as it is pyrophoric. When dry iron sulfide is exposed to air, a spontaneous chemical reaction between the iron sulfide and oxygen takes place—oxidizing the iron sulfide to iron oxide and emitting sulfur dioxide into the air.

SulfaTreat (Iron Oxide) [22]

Several years ago, a new iron-oxide-based dry product with the trade name of SulfaTreat was introduced for sweetening sour gas. The product is placed in towers, as illustrated in Fig. 5.6, through which the sour gas is flowed. The gas stream should have a superficial gas velocity of no more than 10 ft/minute, and the temperature of the gas should be between 70 and 110°F. The gas must be water saturated at the tower conditions of temperature and pressure. SulfaTreat has a different molecular structure from that of iron sponge and, upon reaction with H2S, forms iron pyrite instead of iron sulfide. The charge of SulfaTreat is replaced when consumed.

This process is usually installed in a two-tower configuration, as shown in Fig. 5.6. The slightly sour gas is flowed through both towers in series. The H2S content is monitored in the gas between the two towers. When the concentration of H2S starts to increase in this gas, it is an indication that the SulfaTreat chemical in the first tower is consumed. This tower is then temporarily bypassed, and a fresh charge of chemical is installed. The tower containing the new chemical charge becomes the second tower in the continuation of the operation.

Molecular Sieves[23]

Molecular sieves are crystalline compounds created from alumina silicates, with controlled and precise structures, which contain pores of uniform size. These compounds have an affinity for various molecules, especially for polar compounds such as water, H2S, and CO2. The pore size can be controlled during the manufacturing process and can be tailor-made for specific molecules, such as H2S. Molecular sieves can, therefore, be used for removing water from sour gas, or they can also be used for sweetening sour gas that exceeds the H2S specification by a few ppm. The process requires two or three towers filled with molecular sieves, one of which is used for adsorption, while the others are being regenerated by the application of a hot gas stream. Sweetening with molecular sieves is suitable for large volume, very low H2S concentration gas.

Screening Program for Optimum Process Selection

The Gas Research Institute (GRI) of Chicago, now called the Gas Technology Institute (GTI), has performed large scale investigations of redox and scavenger sweetening processes. The results have been compiled in reports and papers, which are available from GTI. Furthermore, computer screening programs have been prepared, and these are also available for a nominal fee. The two programs dealing with scavenger process selection are CalcBase™ and SeleXpert™. Information on these programs can be accessed on the GTI website:, then search for "H2S scavenger."

Dehydration of Natural Gas

Dehydration of natural gas means extracting water vapor from the gas to a specified maximum limit for residual water content. There are various processes available for dehydration, such as absorption with glycol; adsorption with dry desiccant; absorption with a deliquescent salt; and refrigeration and hydrate suppression with a chemical.

Dehydration With Glycol

All raw natural gas is fully saturated with water vapor when produced from an underground reservoir. Because most of the water vapor has to be removed from all natural gas before it can be commercially marketed, all natural gas is subjected to a dehydration process. The water vapor content of natural gas at equilibrium saturation is shown in Fig. 5.7, which is based on the well-known McKetta and Wehe chart and expanded to 400°F on the basis of data of Olds, Sage, and Lacy.[24][25] As can be seen, the water content increases with increasing temperature and decreasing pressure. For gas sales in colder areas of North America, the specification limit for water content in the sales gas is 4 lbm/MMscf. In warmer, southern areas, the limit is generally 7 lbm/MMscf in sales gas. When natural gas is a feedstock to a turboexpander plant for high natural gas liquids (NGL) recovery, virtually all the water must be removed before chilling the gas to very low temperatures.

There are four glycols that are used in removing water vapor from natural gas or in depressing the hydrate formation temperature. Table 5.3 lists these glycols and shows some of the properties of the pure material.[26] Ethylene glycol (EG) is not used in a conventional glycol dehydrator, as described in Sec. 5.12. The main use of EG in the dehydration of natural gas is in depressing the hydrate temperature in refrigeration units. Of the other three glycols, triethylene glycol (TEG) is the most commonly used glycol for dehydration of natural gas because of the advantages relative to diethylene glycol (DEG): TEG is more easily regenerated to a higher degree of purity; vapor losses are lower; and operating costs are lower.

Tetraethylene glycol would have to be regenerated at higher temperatures than TEG to reach the required purity for application in a glycol dehydration unit.[25] Thus, of the four glycols, TEG is the best suited for dehydration of natural gas. In glycol dehydration, TEG is usually referred to only as "glycol." Unless otherwise specified, that convention is used in the rest of this chapter.

Process Description

Fig. 5.8 is a schematic drawing of the typical process equipment for glycol dehydration. While the overall process equipment is similar for all glycol dehydration units, there can be considerable variation among installations.

The gas flows through a separator to remove condensed liquids or any solids that might be in the gas.[27] Some absorbers incorporate the separator in a bottom section of the vessel, in which case the gas then flows upward through a chimney tray into the glycol absorber portion of the vessel. The glycol contactor or absorber can contain trays, random packing, or structured packing. If it is a trayed vessel, it will contain several bubble-cap trays. Lean glycol is pumped into the upper portion of the contactor, above the top tray but below the mist eliminator. The trays are flooded with glycol that flows down from tray to tray in downcomer sections. The gas rises through the bubble caps and is dispersed as bubbles through the glycol on the trays. This provides the intimate contact between the gas and the glycol. The glycol is highly hygroscopic, and most of the water vapor in the gas is absorbed by the glycol. The rich glycol, containing the absorbed water, is withdrawn from the contactor near the bottom of the vessel above the chimney tray through a liquid level control valve and passes to the regeneration section. The treated gas leaves the contactor at the top through a mist eliminator and usually meets the specified water content.

The rich glycol can be routed through a heat exchange coil in the top of the reboiler column called the still. The heat exchange generates some reflux for the separation of the water from the glycol in the top of the still and also heats the rich glycol somewhat. In some installations, the rich solution passes to a flash tank operating at about 15 to 50 psig, which allows absorbed hydrocarbon gas to separate from the glycol. The glycol then flows into the still through a filter and a heat exchanger, exchanging heat with the regenerated glycol. It drops through a packed section in the still into the glycol reboiler vessel, where it is heated to the necessary high regeneration temperature at near atmospheric pressure. At the high temperature, the glycol loses its ability to hold water; the water is vaporized and leaves through the top of the still. The regenerated glycol flows to the surge tank, from where it is routed through the lean/rich heat exchanger to the glycol pump. The pump boosts the pressure of the lean glycol to the contactor pressure. Prior to entering the contactor, it exchanges heat with the dry gas leaving the contactor or some other heat exchange medium.

Function of the Inlet Separator

The first and foremost piece of equipment that the gas flows through is the inlet separator. This vessel can be a separate, detached vessel, or, on smaller units, built in to the bottom of the contactor. Its function is to separate any condensed liquid from the gas before the gas enters the contactor. If the gas does not contain condensate (liquid hydrocarbon), the vessel is a two-phase separator. If the gas is a rich gas, with some condensate as well as liquid water forming at the inlet conditions of pressure and temperature in the separator, then a three-phase separator is installed. It is absolutely essential that no liquid, condensate, or water enters the absorber section. The separator is usually equipped with a mist eliminator section in the top of the vessel. As the gas moves through the mist eliminator section, small droplets that might be in the gas will coalesce on the fine wire mesh and form larger droplets that drop down through the gas into the liquid section below.

The inlet separator is equipped with liquid level controls, allowing the accumulated liquids to exit the vessel through a level control valve. If for some reason the liquid level in the vessel should rise above a certain limit, a high-level alarm or shutdown automatically occurs.

Function of the Contractor or Absorber

The contactor is the vessel in which the mass transfer of the water occurs from the gas to the glycol. Most of the water vapor is extracted from the gas phase into the liquid glycol phase. For this to occur, it is necessary to create a large surface area between the gas and the liquid glycol. This is accomplished with specific internal equipment configurations, such as through the installation of trays, structured packing, or random packing.

The most common trays used in this application are bubble cap trays, as illustrated in Fig. 5.9. The gas flows from below each tray through the bubble cap and forms small bubbles of gas in the glycol liquid that flows across and on top of each tray. After flowing across one tray, the glycol flows down to the next tray below through a downcomer, which ensures that the gas cannot bypass any tray. The gas bubbles provide the large surface area needed to effect the transfer of the water from the gas to the glycol. Because of the short contacting time on each tray, equilibrium in mass transfer is not reached—several trays are needed in a contactor to bring about the necessary dehydration of the gas. In the design of dehydrators, the calculations make use of theoretical equilibrium stages for determining how many times the gas and glycol must be contacted. Because of the dynamic conditions, equilibrium in mass transfer is not reached—approximately four actual trays are used for each theoretical equilibrium stage. In practice, about 6 to 10 trays are installed in a contactor, usually spaced 24 in. apart. In more recent designs, 12 to 14 trays are installed in glycol absorbers to minimize glycol circulation.

Structured packing consists of arrangements of corrugated steel internals, over which the glycol flows downwards as a thin film. Elements of structured packing are illustrated in Fig. 5.10. The gas flows upward through the structured packing and is in intimate contact with the large surface area of the glycol that flows downward as a film on the packing. This creates a very efficient model for mass transfer to occur.[28] The design for the height of the packing required is related to the number of theoretical stages required. The suppliers of structured packing have developed the relationship of packing height per theoretical equilibrium stage. When using structured packing, it is essential that the glycol be evenly distributed across the top of the packing. To ensure that the large surface area is provided by the downward flowing glycol, it is also essential that the steel be thoroughly clean, so that all of the steel is wetted by the glycol.[29]

Similarly, random packing of various types can also be used in glycol contactors to create the surface area for mass transfer. Various types of random packing are illustrated in Fig. 5.11. Again, the total height of the packing in the vessel is related to the number of theoretical stages used in the design and the height of packing per theoretical stage. The suppliers of the packing have correlations for packing height per theoretical stage.

Function of the Reboiler

The rich glycol leaving the absorber must be regenerated to a high purity so that it can be recirculated to the absorber to continue its dehydration function. The regeneration is accomplished in the reboiler and the still column above the reboiler.
The rich glycol is preheated through heat exchange with the regenerated glycol and enters the top of the still at atmospheric pressure. By heating the glycol in the still and reboiler to near its boiling point, the glycol releases virtually all of the absorbed water and any other compounds and is then cooled for reuse. The heat is usually supplied through a fire tube in the reboiler in which natural gas is burned. The fire tube is always submerged in the glycol by having the glycol flowing from the reboiler over a weir or pipe, which is higher than the top of the fire tube.

Water Dewpoint Depression

Water dewpoint is the temperature at any given pressure at which the natural gas is saturated with water. Because sales gas generally has a maximum water content specification in mass per unit volume (e.g., 4 lbm/MMscf), it is necessary to determine from Fig. 5.7 the dewpoint temperature at the contactor pressure. For example, if the treating pressure were 1,000 psia, then the dewpoint temperature for 4 lbm/MMscf would be about 18°F. If the treating pressure were 500 psia, the dewpoint temperature would be about 5°F.

The overall objective of dehydration is to remove a sufficient amount of water from the natural gas so that the specification for maximum allowable water content in the treated gas is met. If the gas has not been processed, then the gas entering the glycol contactor is fully saturated with water at the pressure and temperature of the separator ahead of the contactor. The dewpoint depression to achieve is the lowering of the dewpoint temperature from the separator temperature to a temperature at the separator pressure where the water content meets the specified limit for the dehydrated gas. The next example shows the difference in dewpoint depression and the amount of water to be removed per MMscf, if the gas is entering the glycol contactor at 90°F and at 1,000 psia or at 500 psia. The water content specification in this example is 4 lbm per MMscf for the treated gas.

Examples of water dewpoint depression for two operating pressures are shown at 90°F.

Water content at 1,000 psia and 90°F = 45 lbm/MMscf (Fig. 5.7).
Dewpoint temperature for 4 lbm/MMscf and 1,000 psia = 18°F (Fig. 5.7).
Water to remove = 45 – 4 = 41 lbm/MMscf.
Dewpoint depression = 90°F – 18°F = 72°F.
Water content at 500 psia and 90°F = 78 lbm/MMscf.
Dewpoint temperature for 4 lbm/MMscf and 500 psia = 5°F.
Water to be removed = 78 – 4 = 74 lbm/MMscf.
Dewpoint depression = 90°F – 5°F = 85°F.

Fig. 5.12 provides a correlation of the equilibrium water dewpoint when gas is in contact with triethylene glycol of various purities at different contacting temperatures.[29] In actual operations, equilibrium is not achieved on any tray. To enhance the transfer rate of the water from the gas to the glycol, an approach temperature to equilibrium of 10 to 20°F is used in designing glycol dehydration units. For example, assuming a 10°F approach is realistic, to achieve the dewpoint temperature of 18°F at 1,000 psia contacting pressure, an equilibrium dewpoint of 8°F (18°F minus 10°F approach) is used in the design of the glycol dehydration unit. At a 90°F contacting temperature, the glycol purity must be about 98.8% to achieve this equilibrium dewpoint, according to Fig. 5.12. At a contacting pressure of 500 psia and 90°F on the other hand, the glycol purity must be about 99.3% to depress the water dewpoint to –5°F with an approach of 10°F. The lower the water dewpoint temperature requirement is for the treated gas, the higher the glycol purity has to be.

Glycol Regeneration

The degree of water removal from natural gas by glycol, or the depression of the water dewpoint of the gas, depends on certain conditions: glycol purity; glycol circulation rate (up to a certain limit); number of contacting stages (trays) or packing height; and amount of water in the inlet gas, which depends on the pressure and temperature of the gas.

These parameters must be considered at the design stage of the contactor, in addition to the maximum rate, pressure, and temperature of the gas. The higher the glycol purity, the more effective is the dewpoint temperature depression. If the glycol purity is insufficient, increasing the glycol circulation rate will not necessarily achieve the desired dehydration of the gas.

Glycol Purity Enhancement Methods

In some operating situations, a high glycol purity is required that cannot be achieved by the temperature in the glycol reboiler alone. There are several ways of enhancing the purity of the glycol beyond what is achieved in the reboiler.[10] One such method is the application of a small amount of stripping gas in the regenerating section. Fig. 5.13 shows the effect of using stripping gas to enhance the purity of the lean glycol solution.[10] Stripping gas is simply a small stream of natural gas that is flowed into the hot glycol. The flow of this gas is usually regulated manually with a small needle valve and is measured by means of a small rotameter.

There are a couple of ways of flowing this gas into the hot glycol. One way is simply to flow the gas into the glycol below the overflow line from the reboiler to the surge drum or directly into the glycol in the reboiler vessel through a perforated pipe below the fire tube. The other way is to install a small packed column between the reboiler and the surge drum and admit the gas at the base of this column. By contacting the hot glycol with natural gas, an additional small amount of water is "stripped" from the glycol into the gas, increasing the purity of the lean glycol. If a packed column is used as a contacting means between the glycol and the stripping gas, the stripping efficiency is considerably improved, as seen in Fig. 5.13.

The stripping gas that is added to the glycol in the regeneration section is emitted into the atmosphere with the released water vapor, unless the vapors from the still columns are routed to a heater or an incinerator, or are captured by a compressor and recompressed.

Components of the Glycol Circulating System

The Glycol Circulating Pump

The circulation of glycol is done with a reciprocating pump. The pump is driven by an electric motor, natural gas pressure, or the high-pressure, rich glycol returning from the contactor.

Electric-motor-driven pumps are usually employed in central dehydration facilities where electric power is available. In field installations, a natural gas powered pump or a glycol powered pump can be used. In the latter case, the high pressure, rich glycol, with the assistance of a small amount of high pressure gas, flowing out of the contactor, is used to provide the power needed to stroke the pump. The required pump rate in field dehydration units is usually small, and, therefore, a single plunger pump is normally used. When the gas rate is large, duplex or triplex pumps are used.

The main problems with glycol pumps are leaks through the packing around the plunger, as well as sticking check valves. If the packing gland nut is tightened too much, the rod may get scored. Usually, a small pan is placed under the pump or the plunger portion of the pump to contain the leaked glycol.

Heat Exchange

Because glycol must be relatively cool when entering the absorber and is heated to near its boiling temperature for regeneration, the liquid is continuously subjected to heating and cooling. To minimize the use of energy in the regeneration of the glycol at high temperature, several heat-exchange opportunities are built into most glycol circulating systems. Heat exchange usually occurs in certain locations: coil in the top of the reboiler still; coil in the surge tank or separate heat exchanger between rich and lean glycol; and pipe-in-pipe heat exchanger ahead of the glycol absorber or a coil in the top of the absorber.

In some cases, an additional heat exchanger is necessary to exchange heat with air, to cool the lean glycol ahead of the contactor. It is necessary to limit the temperature of the lean glycol to only a few degrees above the temperature of the gas to increase the absorption of the water by glycol. Too high a glycol temperature reduces the transfer of water from the gas to the glycol, and the water dewpoint may not be met. This is frequently the problem in summer operations, in which the gas is dehydrated after compression. On hot days, the glycol, as well as the gas, might be above normal temperatures. Usually, by exchanging heat with the dried gas through a double pipe exchanger or through a coil in the top of the contactor in small units, the temperature of the glycol is adjusted to a few degrees above the temperature of the gas leaving the contactor.


It is very important to maintain the glycol in as clean a condition as possible. For this reason, filters are always incorporated in a glycol circulating system. These filters are usually particulate filters and carbon filters.

The particulate filters are intended to remove solids down to a 5-μm diameter. Solids can occur from corrosion in the glycol system. Carbon filters are designed to remove dissolved impurities, such as compressor oil or condensate from the glycol solution. Particulate filters are usually installed on the rich glycol side and are in operation all the time. Carbon filters may be bypassed most of the time, if there is no dissolved hydrocarbon in the glycol. Impurities in the glycol solution might promote foaming in the contactor or still.

Surge Drum

Because the glycol that is being circulated might not always flow evenly at the same rate throughout the system, a vessel, the surge drum, is required that can handle any surges in the circulation rate. The reboiler always contains a liquid level above the fire tube. Glycol levels in the absorber or flash tank are essentially constant but might fluctuate slightly. Thus, there is a need for a vessel that can absorb slight temporary differences in circulation flow between the various vessels, as well as the thermal expansion of the glycol upon startup.

The surge drum is usually located below the reboiler, or at least at a level below the glycol in the reboiler. The glycol level in the surge drum is important because in some instances there is a heat exchange coil in the surge drum, as explained earlier. The level of glycol in the surge drum should be about at the two-thirds full level. The liquid level in the surge drum is an item that is usually checked by the operators. If the level is lower than normal, this might be the first indication of trouble, such as high glycol losses with the treated gas, losses with the vapors leaving the reboiler still, holdup in one of the vessels, or leaks in the piping.


A strainer should always be installed upstream of the suction of the glycol pump. The glycol strainer ensures that no solid particles enter into the glycol pump. The main problem with solids entering the pump is that they might lodge in the suction or discharge valves and prevent the pump from pumping at maximum efficiency.

Glycol Flash Tank

Whenever gas is in contact with a liquid at elevated pressures, such as natural gas and glycol in the contactor, some of the gas physically dissolves in the liquid. The greater the contacting pressure, the more gas dissolves in the liquid. Thus, some natural gas dissolves in the glycol in the absorber in addition to the water vapor. When the glycol reaches the flash tank, its temperature has been raised through the coil in the reboiler still, and the pressure in the flash tank is at a much lower level, generally between 15 to 50 psig, than the pressure in the contactor. In light of these changed conditions of pressure and temperature between the absorber and flash tank, most of the dissolved gases evolve from the glycol in the flash tank.

On larger systems, the glycol flash tank can be designed as a three-phase separator to help remove any condensate that becomes entrained in the glycol. This bulk separator increases the operating life of the downstream filters.

Glycol Piping

All the vessels through which glycol is circulated are interconnected with steel piping. Glycol is a substance that is prone to leak through threaded connections in piping, as well as through the packing on glycol pump plungers. For this reason, some operators prefer welded piping rather than threaded piping for the glycol system. There are, however, many threaded glycol systems that have provided leak-free service.

Because the pump rate is usually small, the piping in most field installations is of small diameter. It is important to check the piping for leaks and to repair them as soon as possible.

Instrumentation and Controls

Most glycol dehydration units are sufficiently automated that they can operate unattended. The degree of automatic control of the equipment can vary considerably and depends largely on the specifications by the owner company. The discussion in this section highlights the main control points, which may be considered as the minimum control of a dehydration unit. The controls relate mainly to gas flow, temperatures, pressures, glycol circulation, and lean glycol concentration. Pressure gauges should be installed on all vessels, including the reboiler, and on the discharge side of the pump. Similarly, thermometers should be installed on all vessels, as well as ahead and after all heat-exchange equipment on both cold and hot lines.

Control of Gas Flow

Gas flow is usually controlled with a flow control valve upstream of the inlet separator. The operator can set the flow to a certain rate. If the set rate is not met, then the valve opens fully and allows the available gas flow to enter the separator and the contactor. Downstream of the contactor, there may be a meter, which meters the gas flow, or the meter may also be located upstream of the separator. At some point downstream of the contactor, there usually is a back-pressure valve. This valve ensures that the pressure in the contactor is steady without abrupt changes. The pressure is set above the downstream line pressure to ensure steady operation of the contactor.

Lean Glycol Circulation Rate

To achieve the required water dewpoint depression, it is necessary to circulate a certain amount of lean glycol per pound of water to be removed from the gas. The rate of glycol circulation depends on several conditions, which are all interrelated. These conditions are lean glycol purity, after regeneration, which depends on the reboiler temperature and whether or not stripping gas is used, with zero or one stage contacting for the stripping gas; water content of the gas, which depends on gas temperature and pressure in the inlet separator; number of actual trays (or equivalent packing height) in the contactor; and the design approach temperature in the contactor.

In general, a circulation rate of 3 to 5 gal of lean glycol per pound of water to be removed from the gas is required. If the glycol purity is not sufficiently high, any larger circulation rate might not give the necessary dewpoint depression.

Usually there is an attempt to match the circulation rate to near the minimum required rate to achieve the necessary drying. Overcirculation has disadvantages: the heat load on the regenerator is increased, requiring more fuel gas consumption; the lean glycol returning to the contactor is at a higher temperature because of less efficient heat transfer; more hydrocarbons are absorbed, especially compounds such as benzene, toluene, ethyl benzene, and xylene (BETX), if these compounds are present in the gas; and additional acid gas is absorbed, if sour gas is being dehydrated.

Because there is concern about the absorption of other compounds besides water, as well as for energy efficiency, the glycol circulation rate should be set to remove the required water only. In field installations, using gas driven pumps, the pumps are set to the required pump rate by a gas control valve. This is usually a manually operated needle valve. The manufacturer of the pump provides a chart for the pump that shows the pump rate in volumetric units per time vs. the number of strokes of the plunger per minute.

Reboiler Temperature

The temperature of the glycol in the reboiler determines largely the purity to which the glycol is regenerated. However, there is a limit on the temperature to which the glycol can be heated. This limit is a few degrees below the decomposition temperature, as shown in Table 5.3, because above this temperature, the glycol molecule breaks down. In light of this, the normal temperature in which TEG is heated in the reboiler is about 380 to 390°F. This temperature range results in a lean glycol purity of just under 99% on a mass basis, the other 1% being water.

Thus, it is very important to control the reboiler temperature to the range of 380 to 390°F or some other range that provides adequate regeneration of the rich glycol. In most glycol dehydration units, the heat for regeneration is supplied by burning a small amount of the gas in a fire tube in the reboiler vessel. The size of this vessel is determined by the maximum design rate of glycol circulation, and the size of the fire tube itself is designed for a limit on the heat flux from the fire through the steel tube to the glycol on the shell side of the fire tube. The larger the fire tube, the lower is the heat transfer rate per unit area. The flame should be burning along most of the tube, as opposed to an intense flame at the front of the burner. The fire tube should be designed for a heat transfer rate per square foot of fire tube no greater than 7,000 Btu/h.[30]

A thermowell located in the shell of the reboiler and immersed in the glycol is equipped with a temperature regulator that controls the instrument gas supply to a control valve on the fuel gas supply line to the burner. By setting the regulator at the desired temperature, the gas flow to the burner is automatically controlled, resulting in a narrow operating temperature range for the reboiler. A pilot light ignites the gas to the main burner when the controller allows the gas to flow. The reboiler controls also include a high-temperature shutdown and a shutdown of the fuel supply in case of pilot-light failure.

Most glycol reboilers are equipped with a flame arrestor at the air inlet to the burner. The flame arrestor consists of a tightly wound metal sheet, with sufficient space between the wound metal to allow sufficient air through the arrestor into the burner. If an external source of flammable vapors is sucked in with the air through the flame arrestor, such vapors will not ignite outside of the flame arrestor, as the temperature of the gas is cooled below the ignition point, thus preventing a backflash or explosion.

Liquid Level Controls

The main liquid level of concern is the level of the condensed liquids in the inlet separator. This vessel can be a two-phase or a three-phase separator. It is very important that no condensed liquid flows with the gas into the contactor. If condensate or salt water gets into the contactor, the result could be foaming or deposits of salt occurring on the fire tube. Heavy hydrocarbons will eventually gum up the packing in the reboiler column or plug the filter. Flashing of hydrocarbons in the still could damage the packing in the still column. In light of this, most glycol units are equipped with high-level alarms and shutdowns, which activate when the liquid level is exceeded in the inlet separator.

The glycol level in the contactor is also important, as any increase in level above the gas inlet pipe can result in interruption of circulation of glycol because of insufficient glycol returns. Both liquid levels, in the inlet separator and the contactor, are controlled by conventional liquid level floats and outlet valves.

Where a flash tank is employed, it is again important to ensure that the level of the glycol be maintained at a set level. The liquid outlet valve must be the throttling type, as opposed to snap acting, to ensure a smooth and steady flow of rich glycol to the regenerator. The glycol level in the surge drum should be at about 2/3 to 3/4 full. In small units, the heat-exchanger coil is in the surge drum and has to be totally submersed to be effective in heat transfer.

Pressure and Temperature Indicators

All vessels are usually equipped with pressure gauges, as well as pressure relief valves. An operator checking the operation of the glycol unit can quickly see the pressure at which each vessel is operating. The same cannot, in many instances, be said about the temperature, especially in the glycol lines upstream and downstream of each heat exchanger. Ample installation of dial thermometers on the glycol lines is helpful but lacking in many cases. Thermometers are usually installed in the inlet separator and the reboiler. As a minimum, additional thermometers should be installed on the glycol line ahead of the contactor after the heat exchanger and ahead of the reboiler still. Ideally, dial thermometers are installed on all lines entering and leaving the heat exchange equipment.

Contractor Design Considerations

While the contactor provides the intimate contact between the gas and the glycol, it also must have a diameter sufficiently large that there is separation between the gas and liquid phases. The gas flows upward, and any liquid glycol droplets that might form must be able to fall down through the gas stream. The basic equation requires the determination of the maximum superficial gas velocity in the tower. The contactor towers are equipped with a mist extractor installed at the top of the tower. The suppliers of packing have the correlations for packing height per theoretical stage.

The allowable superficial gas velocity in the tower using bubble cap trays is determined by[10]


The density of the gas at contactor condition of pressure and temperature is determined by


The internal diameter of the vessel is approximated by Eq. 5.5, on the basis of the allowable gas velocity in the vessel, as determined by Eq. 5.3.


Design engineers make use of these formulas, with proprietary adjustments, to determine the diameter for maximum gas flow through the absorber. The height of the vessel is determined by the number of theoretical stages used in the design, the spacing between the trays, and the relationship between the theoretical stages and the number of actual trays or packing height. Additionally, there is space provided below the bottom tray and above the top tray for disengaging between the gas and the glycol. A stainless steel mesh mist eliminator is installed near the top of the vessel. The contactors are also designed to very stringent codes, with respect to shell thickness requirement related to the maximum operating pressure rating. These items are stamped on a plate attached to each contactor.

Water Dewpoint Determination

It is always of interest to know that the dehydration unit is performing as required and that the exit gas meets the necessary dewpoint specification. If the gas has to meet sales gas requirements for water content, the dewpoint temperature that has to be met at the operating pressure of the contactor can be obtained from Fig. 5.7. This merely allows the operator to determine the appropriate dewpoint temperature. To determine the actual dewpoint temperature, a dewpoint tester instrument is usually used.

Before electronic instruments were commonplace, the water dewpoint of natural gas was usually determined with a U.S. Bureau of Mines dewpoint tester. Currently, there are several different types of electronic instruments on the market that determine the amount of water in the gas, as opposed to determining the condensation temperature of the water. Such instruments need very clean gas to function properly. A slipstream is taken of the gas to be tested and is usually filtered to remove any impurities.

One brand of electronic dewpoint testers makes use of a metal oxide layer that adsorbs water molecules. By measuring the electrical impedance across the adsorbed metal oxide surface, a reading of the amount of adsorbed water is obtained. The amount of adsorbed water is in balance with the small amount of water in the surrounding gas. Thus, a direct readout is obtained of the moisture content of the gas being tested.

Normal Operation Checklist

While glycol dehydration units are designed to operate unattended, periodic inspection of the equipment and its operation is necessary. All items listed next should be checked.

  • Check the still column vent. Water vapors should be visible. There should be no pressure on the reboiler. Ensure that there is no ice buildup in winter.
  • Check the lean glycol temperatures across all heat exchangers and in the reboiler.
  • Check the pump operation, strokes per minute, and lubrication oil.
  • Check the operation of the glycol filter for pressure drop. Change the filter if necessary.
  • Check the glycol level in the surge drum. Add makeup glycol if necessary.
  • Check the stripping gas rate. Adjust rate as required.
  • Check the liquid levels in the inlet separator, contactor, and flash tank.
  • Drain any fluid from the fuel gas scrubber.
  • Check the operation of the burner in the fire tube. Check sight glass to ensure it is not broken and gasket is in good shape; clean if necessary.
  • Check equipment for liquid leaks, and repair if required.

Trouble Diagnosis

Glycol dehydrators usually operate trouble free. However, there are some problem areas that can occasionally occur.


One of the more serious problems is foaming. The cause of foaming is usually difficult to determine. However, if the solution is not continuously cleaned by filtration, certain materials can cause foaming. One of the more common causes of foaming is entrained hydrocarbon liquids. It is essential that the inlet separator provide good separation between condensed liquids and gas going to the contactor. Antifoam agent is usually a temporary solution, and the real problem must be identified and corrected.


Corrosion is usually caused by degradation products in the glycol, which can be generated by too high a skin temperature of the fire tube in the reboiler.

Not Meeting Water Dewpoint

There can be many reasons for not meeting the required water dewpoint depression. The first step is to check the water dewpoint temperature with a dewpoint tester. A high water dewpoint can be caused by:

  • Gas inlet temperature higher than design.
  • Gas inlet pressure lower than design, combined with normal or higher temperature.
  • Insufficient glycol circulation owing to too low a pump rate or a low glycol level in the surge drum, check valves on suction or discharge of pump not holding, or suction strainer plugged.
  • Insufficient glycol regeneration because of a too low reboiler temperature, high water in inlet separator carrying water into absorber, a leak in the rich/lean glycol exchanger, insufficient stripping gas, or fouled stripping column packing.
  • Foaming in absorber: check liquid level in inlet separator, place charcoal filter in service, or temporarily cut back throughput, if necessary.

Environmental Concerns

Upon the regeneration of the rich glycol to lean glycol, the water that was absorbed in the contactor is released in the regenerator, and in the past was vented into the atmosphere. Unfortunately, the glycol not only absorbs water in the contactor, it also physically absorbs hydrocarbons and acid gas. The absorptivity of paraffinic hydrocarbons, such as methane, ethane, etc. is not great. However, the aromatic hydrocarbons such as benzene, toluene, ethylbenzene, and xylene (BTEX) are more easily absorbed. The problem with these vapors is that they are considered carcinogenic, and all contribute to atmospheric pollution.

Glycol Dehydrator BTEX and VOC Emission Control

The quantity of BTEX and volatile organic compounds (VOC) emitted from natural gas processing facilities has become a major environmental concern over the past decade.[31] With passage of the 1990 amendments to the Clean Air Act in the United States and similar regulations in the rest of the world, several regulatory programs were established to control BTEX and VOC emissions from the still vent of glycol dehydration systems.

Most environmental regulations that were promulgated have focused on two main areas that affected the oil and gas industry. The first area was the reduction of VOC and nitrous oxide (NOx) emissions, which were known to react together in the presence of sunlight and create what is commonly referred to as smog. The second area was a list of hazardous air pollutants (HAP) that were to be controlled, most of which were found to be carcinogenic compounds. Among the list of HAP were the four BTEX compounds, which are commonly found in natural gas streams.

To minimize emissions of these vapors to the atmosphere, several emission control processes have been developed by the gas industry.[32] One such method is to recover condensable compounds and to use the remaining vapors as part of the fuel gas supply to the reboiler. Fig. 5.14 illustrates the process equipment. The vapors emitted from the glycol reboiler still column are cooled in a natural or forced draft air cooler to temperatures below 120°F. The condensed liquids are collected in a small two-phase separator and pumped back into the process system for recovery of saleable hydrocarbon liquids. Noncondensable gas from the two-phase separator is burned in the glycol reboiler firebox to reduce fuel gas consumption and to achieve an overall minimum destruction efficiency of 99.7%.

It is important that all emission control devices be engineered with proper controls for safe operation. Safety controls should be used as a minimum. A level safety high (LSH) should be installed on the two-phase separator to prevent condensed hydrocarbon liquids from going into the reboiler firebox, if the condensed liquid pump fails. An in-line flash arrestor should be installed in the noncondensable gas piping to prevent flame propagation from the reboiler firebox back into the BTEX system. A pressure safety valve (PSV) should be installed on the glycol reboiler or emission control device to protect the system from overpressure. Block and bleed valves should be installed in the noncondensable gas piping to protect the glycol reboiler during a high-temperature situation.

GTI has sponsored research in the area of emission controls from glycol units.[33][34] A computer program has been developed that is designed in part to estimate the amount of BTEX absorbed in a conventional glycol unit. The program is called GLYCalc™ and is available from GTI at a nominal cost.

Dry Dessicant Dehydration

While dehydration with glycol is the most common process used to meet the water dew- point specification for sales gas, under certain conditions solid adsorbents are also used for this purpose. Materials with a great affinity for water are silica gel, Sorbead, alumina, and molecular sieves.

These solid compounds are prepared as round or slightly elliptical beads having a diameter of about 4 to 6 mm, and in the case of molecular sieves, are also manufactured in small cylindrical forms, about 2 to 3 mm in diameter and about 6 to 8 mm long. Each of these compounds has its own characteristic affinity and adsorptive capacity for water, and the designer has to chose the material best suited for the purpose.

Silica gel is silicon dioxide (SiO2), manufactured as small round beads with a large pore surface area onto which the water, contained in the vapor phase in the gas, is adsorbed by the desiccant at relatively low temperatures. The affinity for water is temperature dependent, and the affinity for water is broken at high temperatures, such as 390°F. It is necessary to avoid liquid water droplets from contacting the silica gel, as liquid water damages the desiccant. Thus, it is important to have effective gas/liquid separation ahead of a dry desiccant unit. Sorbead is also a silica gel product that contains about 3% aluminum oxide (Al2O3) that makes the gel more resistant to damage from water droplets. Certain manufacturers of Sorbead desiccants provide different grades of Sorbead, such as grade R for regular natural gas drying and grade WS for additional damage resistance from water droplets. A tower filled with Sorbead contains mainly Sorbead R and is topped off with a one-foot-thick layer of Sorbead WS to offer degradation protection from water droplets.

Alumina is used mainly for drying air. Molecular sieves are artificial zeolites that are manufactured so that they result in solids with uniform small pore spaces. They have a great affinity for water. Molecular sieves are usually installed in applications in which very low residual water content is required, such as ahead of a low temperature hydrocarbon extraction process. Molecular sieves are suitable for drying very sour natural gas that also contains aromatic compounds.[35] The heavier hydrocarbons might be difficult to remove from the silica gel during the regeneration step.

Adsorptivities and design water contents are best obtained from the vendors of the desiccant. In general, design water adsorption is on the order of 8 to 12 lbm of water per 100 lbm of silica and Sorbead desiccants. When new, the adsorption capacity is higher than these numbers. However, when desiccant ages with a number of regeneration cycles, a good design value is about 8 to 10 lbm of water per 100 lbm of desiccant.

Typical Process Equipment

Fig. 5.15 illustrates the typical equipment for a dry desiccant process for dehydrating natural gas. Two towers are required for gas dehydration. As the gas flows through one tower containing the solid desiccant, water is adsorbed onto the surface of the material. After adsorbing for several hours, the material becomes fully saturated with water. For this reason, the process incorporates two towers. While one tower is on adsorption, the other tower is being regenerated by the application of heat through a small slipstream of gas and subsequent cooling. The regeneration time is slightly less than the adsorption time of the first tower. Thus, when one tower is fully saturated, the second tower has been fully regenerated. The towers are then switched, and the process of drying the gas continues uninterrupted.

To design a silica gel dehydration system, it is necessary to select the duration of the onstream time per tower before switching towers. This is usually 8 to 12 hours for fully water-saturated gas. After establishing the cycle time and the amount of water to be removed from the gas during one cycle, the bulk volume of desiccant can be estimated. The diameter of the tower is designed on the basis of the superficial gas velocity, and the height of the tower is determined by dividing the cross-sectional area of the tower by the determined bulk volume of the desiccant. The vendors of the desiccant usually provide the design criteria and perform the design.

Dehydration With Deliquescing Dessicants

Deliquescing desiccants are salts that adsorb water vapor; the water then condenses and dissolves the salt. The water drops down as brine and is removed from the vessel. In the past, the common deliquescing desiccant was calcium chloride (CaCl2).[36]

Deliquescing desiccants have been used to dehydrate natural gas since the early 1920s. Early desiccant systems were plagued with numerous operational problems, such as channeled gas flow and bridging, which caused washouts and velocity problems within the desiccant bed and ultimately resulted in inadequate dehydration of the gas stream. Almost all the operational problems associated with the early systems can be attributed to the poor quality of desiccant that was used. Early desiccants were made in briquette form and were very soft, crumbly, and porous. The shape of the briquettes was irregular and, when partially consumed, became more irregular. As gas flowed through the desiccant bed, it found the easiest path and bypassed the rest of the bed. Once started, this process accelerated desiccant consumption in the flow zone, and a channel would form through the desiccant bed. The drying process stopped because wet gas no longer contacted the desiccant. Also because of its porous nature, gas could penetrate through the briquette, causing it to hydrate internally. This caused the briquette to expand and bloom, making the desiccant fuse together into a hard mass. The operator had to mechanically break up the bed, which was not easy.

Significant advancements have been made over the past 10 years to greatly improve deliquescing desiccant dehydration. New dry-forming technology produces a hard, nonporous, low permeability desiccant. Hydration can only occur on the outside of the desiccant, which helps to maintain its general shape as it is consumed. Flow efficiency remains relatively constant as the desiccant bed is consumed.

Fig. 5.16 shows a typical improved gas dehydration system using deliquescing desiccants. Gas containing water vapor enters the bottom of the vertical dehydration vessel and flows up through a series of full diameter diffusion baffles, which evenly distribute the gas stream across the diameter of the vessel. As the wet inlet gas stream flows up through the desiccant bed, the desiccant starts to hydrate, removing water vapor from the gas stream. This water accumulates on the desiccant surface and drips off the desiccant into the vessel sump as the hygroscopic brine on the desiccant surface continues to remove water vapor from the gas stream. This process, known as "deliquescing," causes desiccant salts to dissolve into the fresh water accumulating on the desiccant. Desiccant is consumed at a rate based on the dilution factor of each desiccant formulation.

The ability of each desiccant to remove water vapor is based on the vapor pressure difference between the hydrate of the metal halide salt used in the desiccant and the partial pressure of water vapor in the inlet gas stream. Gas exiting the vessel top has been dried to a point consistent with the equilibrium point of each desiccant. There are several different hygroscopic grades of desiccant available. Selection of the correct desiccant is based on the inlet gas conditions and the required outlet moisture content. If more hygroscopic grades are needed, it is normally more economical to use several grades in series, flowing from lowest grade to highest in separate vessels, rather than using a high desiccant grade in one vessel. An exception is for very low flow rates such as instrument or fuel gas in which operating-cost savings might not offset additional equipment costs incurred by using multiple vessels.

As desiccant is consumed, new desiccant must be added periodically by isolating and depressurizing the vessel, removing the top service closure and pouring desiccant into the vessel. This interval is predictable, and if necessary, the vessel is simply oversized to provide a longer interval between service operations.

One of the most hazardous operations is the opening of vessels that have been previously under pressure. Even with a safety interlock, a small amount of pressure could remain trapped in the vessel. Small "telltale" vents can become plugged, giving the operator a false sense that there is no internal pressure. When opening such a system, bolts should be loosened but not removed until the pressure seal is broken and a knife can be inserted into the pressure-containing cavity.

Water removed from the gas combines with salts in the desiccant to form brine water, which accumulates in the sump. This brine is removed (typically with automatic controllers) to brine storage, where it can normally be disposed of as common oilfield brine. There are no other products or emissions. Desiccant is not typically affected by high Btu gas; however, inlet gas should flow through standard filters or separators to remove liquids, as required in any dehydration process design.

Because deliquescing desiccant systems are closed, there are no VOC or BTEX emissions. Because equipment is relatively simple compared to glycol systems, capital cost is normally less than that of TEG systems. This is especially true if emission control systems are required with the TEG unit. Deliquescing desiccant systems are very simple to operate and require minimal maintenance. There are no moving parts other than a diaphragm-operated dump valve used to discharge brine water, and 100% turndown is possible with desiccant dehydration. However, the cost associated with manually refilling the desiccant can be substantial, as can be the maintenance cost associated with very corrosive brine.

The operating cost associated with a deliquescing desiccant system is directly proportional to the amount of water vapor removed from the inlet gas stream. Operating cost is lower for high-pressure and low-temperature gas streams in which small quantities of water vapor exist. Compared to conventional TEG dehydration, more gas can be sold with desiccant dehydration because there is no fuel or pump gas consumption.

Water Dewpoint and Hydrocarbon Dewpoint Control

Short-Cycle Dry Dessicant Process

Some processes can be used to control the water content as well as remove heavier hydrocarbons from natural gas. The short-cycle dry desiccant process can be used in this way.

This process is similar to the dry desiccant adsorption process covered in dehydration. It is usually selected for a gas stream that is fairly lean in propane and butane but contains sufficient amounts of C5+ so that the hydrocarbon dewpoint is not met. Silica gel is generally the preferred adsorbent.

There are basically two differences between using this process for water dewpoint control and for hydrocarbon dewpoint control, namely that three towers are required for hydrocarbon dewpoint control, and the cycle time for the towers is short, usually 20 to 30 minutes as compared to 8 hours or longer for dehydration. The desirable feature of the short cycle adsorption unit is that it recovers about 60 to 70% of the pentanes-plus in the gas stream and meets both the hydrocarbon and water dewpoints.

Fig. 5.17 shows the arrangement of a typical three-tower, short-cycle hydrocarbon recovery unit. While one tower is on adsorption, the other two towers are on their heating and cooling cycles. Usually, the towers switch every 20 to 30 minutes. The length of the cycle time depends on the design of the facility.

There are many factors that influence the performance of a short-cycle unit. First and foremost, the separator ahead of the unit must remove all liquids and solids in the gas. A slug of liquid into the bed would damage the adsorbent and perhaps necessitate a changeout. The adsorption temperature should be as low as possible without causing hydrates ahead of the unit, to adsorb more of the heavier hydrocarbons. Proper regeneration of the bed is very important. A portion of the process stream is used for this purpose. Regeneration gas temperatures are in the order of 480 to 530°F. The tower should be heated to an outlet gas temperature of 390°F.

While adsorbing and desorbing are important, the adequate condensing of the desorbed hydrocarbons is also very important. If the condensing is inefficient, the heavier hydrocarbons have no means of exiting from the gas stream. In warm climates, this is a major problem, requiring that the condenser be conservatively sized. Upon condensing, the liquids must be separated from the lighter hydrocarbons in the regeneration stream. Conservatively-sized vertical separators are usually used in this service.

The design of the short-cycle dry desiccant units is usually performed by the seller of the desiccant. The seller develops the computer software and tests the simulation results in actual applications. Thus, the design results from the vendor are quite reliable.

Dewpoint Control by Refrigeration

The refrigeration process is used in gas plants to remove heat from certain process streams. Refrigeration in natural gas treating is a process that serves a dual dewpoint control function—namely, it is used to meet the hydrocarbon dewpoint as well as the water dewpoint specification for residue or sales gas. The temperature to which the gas is cooled depends first on meeting these dewpoint specifications. This is the minimum cooling requirement. Cooling the gas to lower temperatures than the minimum temperature for dewpoint control has to be justified by the economics of liquefied petroleum gas (LPG) recovery. This requires a cost analysis of the value of additional LPG recovery versus increased capital and operating costs. Additional recovery of LPG is achieved by chilling the gas to colder temperatures, such as –20 to –30°F, or by contacting the gas stream with lean oil in an absorption tower.

Refrigeration is basically pumping heat from one medium to another. Heat by itself can only flow from a higher temperature medium to a lower temperature medium. Thus, refrigeration is a process that provides the cooling medium to which the gas is exposed. Refrigeration systems generally operate trouble free but can drop in efficiency, which requires investigation.[37]

Fig. 5.18 shows the typical refrigeration equipment for natural gas cooling. The heat exchanger cools the incoming gas to the refrigeration unit by exchanging heat with the cold gas, which has been chilled to the design cold temperature in the propane chiller.

Because the gas entering the refrigeration unit is normally saturated with water vapor and the temperature to which the gas is cooled is substantially below the hydrate point of the gas, some means of preventing hydrate formation must be instituted. The formation temperature of hydrates at a given pressure can be suppressed by the addition of chemicals, such as methanol or glycol. In conventional refrigeration units, the common chemical used for hydrate suppression in the process gas is monoethylene glycol, usually referred to as ethylene glycol, EG, or simply glycol. Glycol must be added to the natural gas being cooled at two points: the inlet to the gas/gas heat exchanger and the inlet to the propane chiller. It is important to evenly distribute the glycol in the gas stream so that all gas is protected from freezing. This requires spraying the regenerated glycol evenly on to the tube sheet in these two vessels so that some glycol flows through each tube with the gas.

Propane, or some other refrigerant, boils in the chiller at a very low, controlled temperature, removing heat from the gas stream and, thereby, condensing a portion of the gas. The cold gas, condensate, and glycol flow from the chiller to a three-phase separator. The condensate goes to a fractionation unit. The gas is sufficiently cooled so it meets both the hydrocarbon and water dewpoints. It exchanges heat with the incoming gas to the refrigeration process.

The rich glycol is separated from the hydrocarbon gas stream in a three-phase separator and is routed to a regenerator. The concentration of the regenerated glycol is usually on the order of 75 to 80% glycol, with the balance being water. Sufficient glycol is injected at the two injection points to result in a mixture of water and glycol to depress the hydrate temperature to the required level, which for design purposes is the refrigerant boiling temperature.

The amount of glycol to be injected requires a determination of the hydrate temperature depression and a calculation of the rich glycol concentration by the Hammerschmidt equation. (See the chapter on Phase Behavior of Water/Hydrocarbon Systems in the General Engineering volume of this Handbook.)

As mentioned, the refrigeration effect is brought about by the vaporization of a refrigerant, such as propane, in the chiller. Propane is suited for this application, as it boils at temperatures near and below ambient temperatures. The vaporization of propane or boiling requires heat to effect the phase change from liquid to vapor—namely the latent heat of vaporization. By controlling the pressure at which the boiling of the propane takes place, the desired refrigeration temperature, down to about –40°F, is achieved.

The IFPEXOL Process

By using glycol to prevent hydrates from forming in the refrigeration process, a glycol regeneration process step must be incorporated in the overall equipment scheme, as shown in Fig. 5.18. This includes an additional capital and operating cost burden to the refrigeration process. This cost burden can perhaps be somewhat reduced by using a relatively new innovation called the IFPEXOL process, which is illustrated in Fig. 5.19.

In the IFPEXOL process, the prevention of hydrates in the heat exchanger and chiller is achieved by the addition of methanol to the natural gas stream being cooled.[38] Normally, the methanol is recovered by a distillation process. However, the separation of methanol from water is somewhat difficult. In the IFPEXOL process, an innovative step is used that recovers most of the methanol for hydrate suppression without regeneration.

As seen in Fig. 5.19, the inlet gas stream is split into two streams. One portion of the inlet stream is contacted countercurrently with the rich methanol-water solution pumped to a small contactor from the cold separator. Because the gas stream is already saturated with water, it does not pick up any additional water. However, it contains no methanol at the inlet to this contactor. As the gas is in intimate contact with the methanol/water solution, most of the methanol leaves the water and enters the relatively warm hydrocarbon gas phase. This conserves most of the injected methanol. This gas stream joins the other stream before entering the gas/gas heat exchanger. Additional methanol is injected into this stream as required to depress the hydrate temperature of the process gas in the chiller to the boiling temperature of the propane. Because the methanol is contained in the vapor phase, the distribution of liquid methanol onto the tube sheet is not important, as is the case with glycol injection. As the gas cools inside the heat exchanger and chiller tubes, methanol condenses with the water and prevents the formation of hydrates.

Methanol losses occur from the exiting chilled gas as well as in solution in the condensed hydrocarbon liquids. Recovery of the methanol in the liquid hydrocarbons is achieved with a water wash system.[39] Advantages of the IFPEXOL process are simpler process equipment and operation, as compared with glycol injection and regeneration. Furthermore, as discussed earlier, glycol absorbs some hydrocarbons, including the BTEX compounds, which are released into the atmosphere upon the regeneration of the glycol. Extra steps must be taken to avoid the release of the compounds into the atmosphere. In the IFPEXOL process, there are no emissions into the atmosphere. The main drawback to this process is the continuous loss of methanol, which has to be continuously made up from fresh supply. The IFPEXOL process is also suited for offshore operations where mass and space are limited.

NGL Extraction Methods

Lean Oil Absorption

Hydrocarbon liquids can be extracted from natural gas by contacting the gas with a light oil of uniform molecular weight. The fraction of each compound going into solution in the oil increases with decreasing volatility of the compound at the absorber pressure and temperature. Thus, while perhaps only a small percentage of the methane in the gas might go into solution, over 50% of propane, perhaps over 80% of butane, and so on, goes into solution in the oil. The lighter components, methane and ethane, are then rejected in the regeneration process of the oil, thereby capturing the absorbed propane and heavier compounds upon regeneration of the oil.

A simple lean oil absorption process is illustrated schematically in Fig. 5.20. The rich gas enters the absorption tower near the bottom and flows upward through the tower, which contains trays or packing. As the gas flows upward in the tower, it is in intimate contact with the oil, which enters the tower near the top. When the gas leaves the tower at the top, it has been stripped of most of the heavier compounds. The rich oil is then flowed to the stripping section, where the oil is heated to release the absorbed hydrocarbons. The vapors leaving the top of the stripper are cooled, condensing most of the propane and heavier hydrocarbons. The vapors from the reflux separator are compressed and recycled to the rich gas or to the sales gas.

Simple lean oil absorption processes operate at ambient temperatures. More complex lean oil absorption processes can be designed and operated at lower than ambient temperature. By contacting the chilled gas from a refrigeration unit in an absorber with cooled absorption oil, more of the components in the gas go into solution than in a process operated at ambient temperatures. The rich oil leaves the bottom of the tower through a level control valve, exchanges heat with the regenerated lean oil stream, and enters a rich oil flash tank, operating at about half the pressure of the absorption tower. Large amounts of the absorbed lighter compounds, such as methane and ethane, flash off and are routed to recompression. In recovery facilities for propane and heavier hydrocarbons, the oil then enters a deethanizer column, where the balance of the absorbed methane and ethane are rejected. These gases are flowed to a presaturator vessel and then to a recompressor, where they are joined with the main treated gas stream. From the deethanizer tower, the lean oil is flowed to the still, where the separation of the oil and the balance of the absorbed compounds takes place. Upon regeneration, the lean oil flows through heat exchangers and a cooler to a presaturator vessel where it becomes partially saturated with methane and ethane. It is then pumped back to the high pressure absorber. Another cooling step is included to ensure that the lean oil temperature is no higher than the gas temperature to maximize absorption. The design of the overall process is now performed by computer, as the material and heat balance calculations are quite intricate and require rigorous mathematical treatment.

The Turbo-Expander Process

The turbo-expander process for treating natural gas streams for high liquids recovery was developed in the early 1960s. Its main application was to improve the recovery of ethane from natural gas, as ethane is an important feed stock for the petrochemical industry. The process achieves very low temperatures and, therefore, liquefies a substantial portion of the ethane and heavier compounds in natural gas. The various fractions of the liquid stream are recovered by distillation.

The turbo expander removes energy from the near isentropic expansion of a gas stream, which results in a drop in pressure and temperature by extracting useful mechanical energy. By using an expander to recover energy from the high pressure gas, the refrigeration effect is enhanced, and the reduction in gas temperature is greater than can be obtained by simple isenthalpic (Joule-Thomson) expansion across a valve.

Turbo-expander process configurations can vary greatly. They all incorporate various heat exchangers. The gas entering the turbo-expander process must be dehydrated upstream of the plant to a very low water content so that no hydrates form when the low temperatures are reached by the gas being processed. This usually requires a glycol dehydration unit for removing most of the water, followed by a molecular sieve unit to remove virtually all of the water from the feed gas. Gas pretreatment can also include CO2 and H2S removal.

Fig. 5.21 is an illustration of a relatively simple turbo-expander facility. There are many other arrangements possible, depending on the gas composition and the desired level of liquids recovery. Whether the turbo expander is likely the best choice for recovering ethane and heavier hydrocarbons from natural gas requires considerable analysis.[40][41]

Gas Preconditioning

Expander processes for NGL recovery can chill the gas as low as –160°F. To dry the gas to this low a water dewpoint temperature requires the use of molecular sieves in drying towers as illustrated in Fig. 5.15. A common class of molecular sieve used for deep drying has a pore opening of 4 Å. Instead of drying the gas with molecular sieves, it is also possible to prevent potential freezing problems with the addition of minor amounts of methanol into the gas stream upstream of the chilling section.[42]

Turbo-Expander Design

The design of a turbo-expander unit involves detailed heat and material balances and many flash calculations. Such design calculations are performed by computer.

Joule-Thomson Expansion

Cooling of natural gas can also be achieved by expanding high pressure gas to a lower pressure across an expansion valve.[43] This is a constant enthalpy process, and the amount of the temperature reduction depends on the pressure ratio of initial pressure divided by the final pressure, the absolute pressures and the starting temperature, as well as the gas composition. This is a practical method to cool gas and extract hydrocarbon liquids if there is a lot of "free" pressure available. It is also a more practical process than the turbo-expander process for low gas rates, especially if the gas rates fluctuate.

Fig. 5.22 is a schematic drawing of a typical Joule-Thomson expansion process. The main process equipment is the expansion valve or choke. The high-pressure gas enters through an inlet separator, which removes the condensed water and any liquid hydrocarbons. The gas streams out of the separator, then flows through a heat exchanger, exchanging heat with the cooled, low-pressure gas. Some water and perhaps some hydrocarbon will condense in the heat exchanger from the high-pressure gas stream. The high-pressure gas then flows through the expansion valve, which drops the pressure of the gas to the design pressure. Simultaneously, a reduction in temperature occurs. Depending on the gas composition and the pressure and temperature of the gas mixture, a certain amount of the mixture will condense and form a liquid hydrocarbon stream. Water will also condense to the equilibrium water content of the gas at the final pressure and temperature.

If the resulting temperature of the gas after the heat exchanger or upon expansion is below the hydrate temperature at the operating pressure, hydrates form unless the gas has been dehydrated. To avoid the formation of hydrates in water saturated gas, a chemical hydrate inhibitor is added to the gas stream ahead of the heat exchanger. The chemical usually used to depress the hydrate temperature is ethylene glycol, but diethylene glycol can also be used. Fig. 5.22 shows the flow of the glycol and includes a reconcentration step. Ethylene glycol is usually regenerated to a lean concentration of about 75 or 80% by weight and is circulated at a rate such that the resulting final glycol concentration is sufficient to depress the hydrate forming temperature to about 5°F below the hydrate temperature of the gas at the final pressure. The required lean glycol circulation rate is determined by the Hammerschmidt equation (see the chapter on Phase Behavior of Water/Hydrocarbon Systems in the General Engineering volume of this Handbook) and depends on the water content of the gas, the concentration of the lean glycol, and the necessary hydrate temperature depression.

A bypass line around the heat exchanger on the cold gas out of the low temperature separator allows for the control of the degree of cooling of the process gas. To aid the separation between the cold condensate and glycol, a heater can be included in the equipment. After heating, the liquids are flowed into a three-phase separator, where the small amount of gas, the condensate, and the rich glycol are separated. The glycol is then reconcentrated with a conventional reboiler and still and is re-injected into the process gas stream.

Membrane Processing for CO2 Removal

When natural gas contains a high concentration of CO2 , the options for reducing the CO2 concentration to acceptable levels are either to use a regenerative solvent, such as an amine, potassium carbonate or Selexol, or to install a membrane separation process.

Separation by means of membrane technology makes use of thin layers of polymeric material.[44] This polymeric material can be manufactured in two main types—porous and nonporous. The porous membrane makes use of differences in diffusion rates and acts like a sieve in separating molecules based on relative size. The main application of this type of material is the separation of small molecules such as hydrogen or helium from gas mixtures. The nonporous membrane promotes the separation by dissolving some compounds in the polymeric material and allowing these compounds to diffuse through the material more rapidly than hydrocarbon compounds, which do not dissolve but diffuse through the material. In natural gas separation, the nonporous material is used. The compounds specifically found in natural gas, which have the ability to dissolve in the polymeric material, are the polar compounds— CO2 , H2S, and H2O. Hydrocarbons also diffuse through such membranes, but at a much lower rate.

The membrane material is manufactured as a very thin film, which has no strength. It is supported in applications by a porous layer, through which the gas molecules that have passed through the membrane (permeate) flow to the low pressure side. In one type of application, the membrane and the porous permeate spacer are spirally wound around a perforated tube. A third layer is included in such construction, which is a porous feed spacer. The high-pressure gas (feed) flows through the feed spacer and is in contact with the membrane. The permeate dissolves in and/or diffuses through the membrane layer into the permeate spacer. It then flows through the permeate spacer into the perforations in the steel tube, which forms the shaft of the spirally wound layers. The high-pressure gas passes through the feed spacer to the other end, with very little pressure drop.

In the natural gas industry, membrane technology is used mainly to reduce the concentration of CO2 from very high concentrations to acceptable levels, such as less than 2%.[45] If H2S is present, most of the H2S will separate. Because the sales gas specification for H2S can be as low as 4 ppm, membrane separation is usually not sufficient to meet the specification for this compound. This process may meet the stringent specification for H2S only if the concentration of H2S is very low to begin with, such as perhaps 50 ppm or less.

While CO2 is the main target for removal in natural gas processing with membranes, hydrocarbons also diffuse through the membranes. Because hydrocarbons are the valuable sales products, various process schemes are used to minimize the loss of hydrocarbons. The amount of hydrocarbon in the permeate after one pass through a permeable membrane facility depends on gas composition; system pressures (high and low); gas flow rate; and total surface area of membrane exposed to high-pressure gas.

While the permeation and potential loss of hydrocarbon is a disadvantage of membrane separation processes, there are many advantages over the alternatives for reducing CO2 content. The advantages are low capital investment, as compared with regenerative solvent processes; ease of installation; simple operation, as the feed gas simply flows through the facility with little pressure loss; low weight and space requirements, which are important in offshore installations; low environmental impact; and no utilities requirement.

The loss of hydrocarbon in the permeate can be reduced by compressing the permeate to a high pressure and subjecting it to a second stage of separation with a membrane.[46] The resulting high-pressure stream from the second stage is then added to the high-pressure feed gas to the first stage, as shown in Fig. 5.23. Other permeable membrane schemes use the membrane process for bulk removal of CO2, followed by an amine system for final cleanup.

In treating rich gas with membranes for CO2 removal, it is necessary to preheat the feed stream so that no condensation occurs because of the high-pressure drop through the membrane. The service life of the membrane material is critical to evaluating process economics relative to other methods. Service life depends on feed quality (e.g., liquid carryover, solids content, etc.) and the care with which the system is operated. In the absence of other data from similar operations, a service life of 3 to 5 years should be considered.

The analysis and design of permeable membrane systems can be investigated with a computer program available at a nominal cost from the Gas Research Institute of Chicago, now called the Gas Technology Institute (GTI). Their program MemCalc™ is a PC-based program that simulates the performance of membranes for removing CO2 from natural gas. The program has been evaluated by several operators of membrane process facilities.


AG = percent acid gas, %
d = internal diameter, inches
K = coefficient
ML = mole loading, moles/mole
MW = molecular weight, lbm/mole
P = pressure, psia
Q = gas flow rate, MMscf/d
SG = specific gravity of solution (water = 1)
T = absolute temperature, °R
V = superficial gas velocity, ft/sec
W = weight percent of solvent in solution, %
z = gas compressibility factor, dimensionless
γ = gas specific gravity (air = 1)
ρ = density, lbm/ft3


a = amine
g = gas
l = liquid
s = solvent


  1. Smith, R.S. 1975. Improve Economics of Acid-Gas Treatment. Oil & Gas J 73(March): 78-79.
  2. Tennyson, R.N. and Schaaf, R.P. 1977. Guidelines Can Help Choose Proper Process for Gas-Treating Plants. Oil & Gas J 75 (2): 78.
  3. 3.0 3.1 King, J.C. et al. 1986. Rigorous Screening Selects Sour-Gas Plant Process. Oil & Gas J 84 (36): 101-110.
  4. Butwell, K.F., Kubek, D.J., and Sigmund, P.W. 1982. Alkanolamine Treating. Hydrocarbon Processing (March): 108.
  5. Fitzgerald, K.J. and Richardson, J.A. 1966. New Correlations Enhance Value of Monoethanolamine Process. Oil & Gas J 64 (43): 110-118.
  6. Freireich, E. and Tennyson, R.N. 1976. Process Improves Acid Gas Removal, Trims Costs, and Reduces Effluents. Oil & Gas J 74 (34): 130-132.
  7. Love, D. 1972. Quick Design Charts For Diethanolamine Plants. Oil & Gas J 70 (January): 88.
  8. Goar, B.G. 1980. Selective Gas Treating Produces Better Claus Feeds. Oil & Gas J 78 (18): 239-242.
  9. Maddox, R.N. 1967. Hot Carbonate—Another Possibility. Oil & Gas J (October): 167-173.
  10. 10.0 10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8 Dehydration. 1998. In GPSA Engineering Data Book, 11th edition, Sec. 19, 20, and 21. Tulsa, Oklahoma: Gas Processors Suppliers Association.
  11. Abry, R.G.F. and DuPart, M.S. 1995. Amine Plant Troubleshooting and Optimization. Hydrocarbon Processing (April): 41-50.
  12. Pauley, C.R., Hashemi, R., and Caothien, S. 1989. Ways to Control Amine Unit Foaming Offered. Oil & Gas J 87 (50): 67-75.
  13. DuPart, M.S., Bacon, T.R., and Edwards, D.J. 1963. Understanding Corrosion in Alkanolamine Gas Treating Plants, Part 1. Hydrocarbon Processing (April): 80.
  14. Kutsher, G.S., Smith, G.A., and Greene, P.A. 1967. NOW—Sour-Gas Scrubbing by the Solvent Process. Oil & Gas J (March): 116.
  15. Buckingham, P.A. 1964. Fluor Solvent Process Plants: How They Are Working. Hydrocarbon Processing (April): 113.
  16. Purisol. 2000. Hydrocarbon Processing (April) 84.
  17. Goar, B.G. 1969. Sulfinol Process Has Several Key Advantages. Oil & Gas J (30 June): 117.
  18. Fong, H.L., Kushner, D.S., and Scott, R.T. 1987. Shell Redox Desulfurization Process Stresses Versatility. Oil & Gas J 84 (21): 54-64.
  19. Schaack, J.P. and Chan, F. 1989. H2S Scavenging, 4-part series. Oil & Gas J (23 January): 51; (30 January): 81; (20 February): 45; (27 February): 90.
  20. Dillon, E.T. 1991. Triazines Sweeten Gas Easier. Hydrocarbon Processing (December): 65.
  21. Anerousis, J.P. and Whitman, S.K. 1985. Iron Sponge: Still a Top Option for Sour Gas Sweetening. Oil & Gas J (18 February): 71.
  22. Samuels, A. 1990. H2S Removal System Shows Promise Over Iron Sponge. Oil & Gas J (5 February): 44.
  23. Maddox, R.N. and Burns, M.D. 1968. Solids Processes for Gas Sweetening. Oil & Gas J (17 June): 90.
  24. 24.0 24.1 McKetta, J.J. and Wehe, A.H. 1958. Use This Chart for Water Content of Natural Gases. Petroleum Refiner (August): 153.
  25. 25.0 25.1 25.2 Sage, R.H., Hicks, B.L., and Lacey, W.N. 1940. Phase Equilibria in Hydrocarbon Systems. Ind. Eng. Chem. 32 (8): 1085-1092.
  26. Gas Conditioning Fact Book. 1962. Toronto, Canada: Dow Chemical of Canada Ltd.
  27. Ballard, D. 1966. How to Operate a Glycol Plant. Hydrocarbon Processing (June): 180.
  28. Kean, J.A., Turner, H.M., and Price, B.C. 1991. How Packing Works in Dehydrators. Hydrocarbon Processing (April): 47.
  29. 29.0 29.1 Holder, M.R. 1996. Performance Troubleshooting on a TEG Dehydration Unit with Structured Packing. Proc., Laurence Reed Gas Conditioning Conference, Norman, Oklahoma: 100–112.
  30. Grosso, S. 1978. Glycol Choice For Gas Dehydration Merits Close Study. Oil & Gas J (13 February): 106.
  31. True, W.R. 1993. Federal, State Efforts Force Re-examination of Glycol-Reboiler Emissions. Oil & Gas J 91 (20): 28-32, 49.
  32. Sivals, C.R. 1995. US Will Require Glycol Dehydrator Emission Control. World Oil (November): 75.
  33. Fisher, K.S. et al. 1995. Glycol Dehydrator Emission Control Improved. Oil & Gas J (27 February): 40.
  34. Thompson, P.A. et al. 1993. PC Program Estimates BTEX, VOC Emissions. Oil & Gas J (14 June): 36.
  35. Kraychy, P.N. and Masuda, A. 1966. Molecular Sieves Dehydrate High-Acid Gas at Pine Creek. Oil & Gas J (8 August): 66.
  36. Redus, F.R. 1966. Field Operating Experience With Calcium Chloride Gas Dehydrators. World Oil (1 February): 63.
  37. Vargas, K.J. 1982. Troubleshooting Compression Refrigeration Systems. Chem. Eng. (22 March): 137.
  38. Minkkinen, A. et al. 1992. Methanol Gas-Treating Scheme Offers Economics, Versatility. Oil & Gas J (1 June): 65.
  39. Hampton, P. et al. 2001. Liquid-Liquid Separation Technology Improves IFPEXOL Process Economics. Oil & Gas J (16 April): 54.
  40. Morgan, J.D. 1976. How Externally Refrigerated and Expander Processes Compare For High Ethane Recovery. Oil & Gas J (3 May): 230.
  41. Dyck, P. and Henderson, D. 1978. Expander Wins For Gas Dewpoint Control. Oil & Gas J (24 April): 86.
  42. Nielsen, R.B. and Bucklin, R.W. 1983. Why Not Use Methanol for Hydrate Control? Hydrocarbon Processing (April): 71.
  43. Crum, F.S. 1981. There Is a Place For J-T Plants In LPG Recovery. Oil & Gas J (10 August): 132.
  44. Koros, W.J. 1995. Membranes: Learning a Lesson from Nature. Chemical Engineering Progress (October): 68.
  45. Lee, A.L., Feldkirchen, H.L., and Gomez, J. 1994. Membrane Process for CO2 Removal Tested At Texas Plant. Oil & Gas J. (31 January): 90.
  46. Cook, P.J. and Losin, M.S. 1995. Membranes Provide Cost-Effective Natural Gas Processing. Hydrocarbon Processing (April): 79.fckLR

SI Metric Conversion Factors

Å × 1.0* E – 01 = nm
Btu × 1.055 056 E + 00 = kJ
cp × 1.0* E – 03 = Pa s
ft3 × 2.831 685 E – 02 = m3
ft/sec × 3.048* E – 01 = m/s
°F (°F – 32)/1.8 = °C
gal × 3.785 412 E – 03 = m3
in. × 2.54* E + 00 = cm
lbm × 4.535 924 E − 01 = kg
psi × 6.894 757 E + 00 = kPa


Conversion factor is exact.