NGL Liquid-Liquid Treating Fundamentals

1. Overview of NGL Treating Requirements

Natural gas liquids recovered through fractionation—propane, butane, and natural gasoline—frequently contain contaminants that must be removed before the products meet commercial specifications. Even when the upstream gas has been treated to pipeline quality, the fractionation process concentrates certain impurities into specific NGL product streams, creating treating requirements that differ substantially from those encountered in gas-phase treating.

Liquid-liquid treating refers to the family of processes that contact a liquid NGL product stream with a treating solution (caustic, water, or amine) to remove sulfur compounds, acid gases, and other objectionable contaminants. Unlike gas treating, where the contaminant is dissolved in a gas phase, NGL treating involves mass transfer between two immiscible liquid phases—the hydrocarbon and the aqueous treating agent—which introduces unique challenges related to interfacial area generation, emulsion formation, and phase separation.

Common Contaminants in NGL Products

The contaminants found in NGL products after fractionation depend on the inlet gas composition, the upstream gas treating method, and the fractionation system design. The primary species of concern include:

Product Specification Requirements

NGL products must meet stringent quality specifications before they can be sold into commercial markets. The key specifications governing treating requirements are established by GPA 2140 and related industry standards:

The copper strip corrosion test (ASTM D1838) is particularly important because it detects the presence of reactive sulfur species (H₂S, mercaptans, COS) that would corrode copper and copper-alloy components in downstream equipment and end-use applications. Passing the copper strip test is often the controlling specification that drives NGL treating design.

Treating Methods Overview

Four principal treating methods are used for liquid NGL products, often in combination depending on the contaminant loading and product specification requirements:

• Caustic washing (NaOH): The most widely used NGL treating method, effective for H₂S and mercaptan removal from propane and butane through chemical reaction with sodium hydroxide. Caustic systems range from simple once-through prewash arrangements to complex multi-stage extractive wash and regeneration circuits
• Water washing: Removes dissolved salts, caustic carryover, and water-soluble contaminants from the NGL product. Typically positioned downstream of caustic treating to ensure product cleanliness
• Amine treating: Preferred when H₂S and CO₂ loadings are high enough that caustic consumption would be excessive. Uses regenerable amine solutions (MDEA or DEA) in liquid-liquid contactors for bulk acid gas removal
• Merox process: Catalytic mercaptan oxidation process that converts mercaptans to less objectionable disulfides using a caustic solution and proprietary catalyst. Available in both sweetening (fixed-bed) and extraction (liquid-liquid) configurations

2. Caustic Washing (NaOH)

Caustic washing with sodium hydroxide solution is the most established and widely used method for removing H₂S and mercaptans from liquid propane and butane. The process relies on the chemical reaction between the acid gas species dissolved in the NGL and the alkaline caustic solution, transferring sulfur compounds from the hydrocarbon phase into the aqueous phase where they can be disposed of or regenerated.

Chemistry of Caustic Treating

The caustic treating reactions are straightforward acid-base neutralization reactions that proceed rapidly at ambient temperature and NGL operating pressure:

Where RSH represents a mercaptan (methyl mercaptan CH₃SH, ethyl mercaptan C₂H₅SH, etc.) and NaSR is the corresponding sodium mercaptide. The H₂S reaction is essentially irreversible at treating conditions, while the mercaptan reaction is partially reversible, which is the basis for caustic regeneration in extractive wash systems.

Caustic Strength and Preparation

Caustic solutions used for NGL treating typically range from 10 to 20 wt% NaOH concentration. The selection of caustic strength involves balancing treating capacity against viscosity and handling considerations:

Caustic solutions above 25 wt% are generally avoided in NGL treating service because the increased viscosity impairs mass transfer, promotes emulsion formation, and creates a risk of NaOH crystallization at cold ambient temperatures (the freezing point of 30% NaOH is approximately 60°F).

Prewash and Extractive Wash Stages

A complete caustic treating system for NGL products typically includes two distinct stages, each serving a different purpose:

• Prewash (fresh caustic): The first stage contacts the NGL with fresh or partially spent caustic in a once-through or batch mode. The prewash removes the bulk of the H₂S and CO₂, which are strong acids that would otherwise consume extractive-wash caustic rapidly. The prewash caustic becomes "sulfide-laden" and is disposed of rather than regenerated, because the sodium sulfide (Na₂S) reaction is not economically reversible
• Extractive wash (regenerable caustic): The second stage uses a circulating caustic solution to extract mercaptans (RSH) from the NGL. The rich caustic from the extractor is regenerated by oxidation or steam stripping, converting the sodium mercaptide back to free mercaptan (which is expelled) and regenerated NaOH (which is recirculated). This stage operates continuously with minimal caustic makeup

Contactor Design Options

Several contactor types are used for liquid-liquid contacting in caustic treating service. The selection depends on throughput, space constraints, and the required degree of mass transfer:

Caustic-to-Hydrocarbon Ratios

The volumetric ratio of caustic solution to NGL hydrocarbon flow is a critical design parameter that affects both treating efficiency and operating cost. Typical design ratios vary by application:

• Prewash stage: 1:5 to 1:15 (caustic to NGL by volume), depending on H₂S loading. Higher ratios are used for higher H₂S concentrations to ensure adequate neutralization capacity
• Extractive wash stage: 1:3 to 1:10, with the ratio determined by the mercaptan loading, caustic regeneration efficiency, and required product specification. Lower ratios are feasible with highly efficient contactors (fiber-film type)

Spent Caustic Handling and Disposal

Spent caustic from NGL treating is classified as a hazardous waste in most jurisdictions and requires careful handling and disposal. The spent caustic contains sodium sulfide (Na₂S), sodium mercaptide (NaSR), sodium carbonate (Na₂CO₃), and free NaOH, making it highly alkaline (pH > 13) and malodorous. Common disposal methods include:

• Wet air oxidation: Converts sulfide and mercaptide to sulfate and sulfonate at elevated temperature (400–600°F) and pressure (300–1500 psig) using compressed air. Produces an effluent suitable for biological treatment
• Deep well injection: Injection into approved disposal wells, subject to regulatory permitting requirements
• Off-site treatment: Commercial waste disposal facilities that specialize in spent caustic treatment and neutralization
• Neutralization and biological treatment: Acid neutralization followed by aerobic biological treatment in an on-site wastewater system. Requires careful pH control and sulfide management

Common Operational Problems

Caustic treating systems are susceptible to several operational issues that can impair treating efficiency or cause product quality excursions:

• Emulsion formation: Excessive mixing energy, high caustic viscosity, or the presence of surface-active contaminants (corrosion inhibitors, amine carryover) can create stable emulsions at the hydrocarbon-caustic interface. Emulsions increase caustic consumption, cause product haze, and can lead to caustic carryover into the treated product
• Caustic carryover: Entrained caustic droplets in the treated NGL cause downstream corrosion, off-specification product, and fouling of downstream equipment. Coalescer elements or electrostatic treaters are used to remove entrained caustic before the NGL exits the treating system
• Caustic degradation: CO₂ absorption converts NaOH to Na₂CO₃, which has much lower treating capacity. In systems with significant CO₂ in the NGL feed, the caustic must be refreshed more frequently to maintain effective treating
• Channeling in packed columns: Maldistribution of liquid flow in packed contactors reduces effective mass transfer area. Proper liquid distributor design and periodic repacking are essential for maintaining performance

3. Water Washing

Water washing is an essential step in the NGL treating train that removes dissolved salts, residual caustic carryover, and water-soluble contaminants from the treated hydrocarbon product. Although water washing does not remove sulfur compounds directly, it is critical for meeting product quality specifications related to alkalinity, dissolved solids, and haze-free appearance.

Purpose and Position in the Treating Train

The water wash stage is positioned downstream of the caustic treating section and upstream of the final product drying and storage. Its primary functions are:

• Caustic removal: Washing out residual NaOH and sodium salts entrained in the NGL from the caustic treating stage. Even small amounts of caustic carryover (as low as 1–5 ppmw) can cause corrosion in downstream piping, storage tanks, and end-use equipment
• Salt removal: Dissolving and extracting inorganic salts (sodium sulfide, sodium carbonate, sodium mercaptide) that may be present as fine droplets or dissolved species in the hydrocarbon phase
• pH adjustment: Reducing the alkalinity of the NGL product to neutral conditions, ensuring the product does not cause corrosion or chemical incompatibility in downstream systems

Contactor Design

Water wash contactors are typically simpler than caustic contactors because the mass transfer requirement is less demanding. The primary design objective is to generate sufficient interfacial area between the NGL and the wash water to dissolve and extract the ionic contaminants. Common contactor types include:

• Static mixers: In-line mixing elements that provide one to two theoretical stages of contact in a compact arrangement. Most common for water wash applications where the contaminant loading is low
• Mixing valves: Globe or butterfly valves operated at a controlled pressure drop (5–15 psi) to create turbulent mixing between the NGL and injected wash water. Simple and inexpensive but limited to single-stage contact
• Packed columns: Used when multiple stages of water washing are required for thorough contaminant removal, particularly when the NGL has been treated with high-strength caustic and carryover levels are elevated

Water Quality Requirements

The quality of wash water used in NGL treating directly affects the effectiveness of the washing step and the risk of introducing new contaminants. Key water quality parameters include:

Fresh water (city water, demineralized water, or clean condensate) is preferred over recycled process water to avoid reintroducing contaminants. Water consumption is typically 3–10% of the NGL product volume, depending on the number of wash stages and the degree of contaminant removal required.

Settling and Coalescence

After contact between the NGL and wash water, the two phases must be separated cleanly to produce a dry, haze-free NGL product. Phase separation is achieved through a combination of gravity settling and coalescence:

• Gravity settlers: Horizontal vessels sized to provide 15–30 minutes of residence time at the NGL flow rate. The density difference between the NGL (specific gravity 0.50–0.58 for propane/butane) and water (specific gravity 1.0) provides the driving force for separation. Settler internals include inlet baffles, perforated distribution plates, and weir-type outlets to minimize turbulence
• Coalescer elements: Cartridge-type coalescing elements (fiberglass or polymer media) that capture fine water droplets and merge them into larger drops that settle by gravity. Coalescers are sized for superficial velocities of 2–5 gpm/ft² and are effective for removing water droplets down to 5–10 microns

Sand Filters for Haze Removal

When the treated NGL product exhibits a persistent haze that cannot be removed by gravity settling or coalescence alone, sand filters (or multimedia filters) are used as a final polishing step. The haze is typically caused by extremely fine water droplets (sub-micron), suspended caustic particles, or iron sulfide fines.

Sand filter design parameters for NGL service:

• Media: Graded sand or garnet, with support gravel layer. Typical media depth is 24–36 inches
• Superficial velocity: 3–8 gpm/ft² for NGL service (lower than water filtration due to lower viscosity)
• Backwash: Periodic backwash with clean NGL or water to remove accumulated solids. Backwash frequency depends on solids loading, typically every 24–72 hours
• Vessel orientation: Vertical downflow is most common; horizontal configurations are used for high-flow, low-footprint applications

4. Amine Treating of NGL

Amine treating of liquid NGL products is selected when H₂S and CO₂ concentrations are high enough that caustic consumption would be prohibitively expensive, or when the volume of spent caustic generated would create unacceptable disposal costs. Unlike caustic treating, amine systems are fully regenerable, and the acid gas removed from the NGL can be routed to a sulfur recovery unit or flare.

When Amine Treating Is Preferred

The decision between caustic and amine treating for NGL products is driven primarily by acid gas loading and operating economics:

Amine Selection for NGL Service

The choice of amine solvent for liquid NGL treating differs from gas treating because the amine must contact a liquid hydrocarbon rather than a gas stream. The two most common amines used in NGL liquid-liquid treating are:

• MDEA (methyldiethanolamine): The preferred amine for NGL treating when H₂S selectivity over CO₂ is desired. MDEA has lower hydrocarbon solubility than DEA, which reduces NGL losses into the amine phase. Typical concentration is 40–50 wt% in water. MDEA's selective H₂S absorption is advantageous when CO₂ removal is not required or when minimizing acid gas volume to the sulfur recovery unit is important
• DEA (diethanolamine): Used when both H₂S and CO₂ must be removed from the NGL product. DEA is less selective than MDEA and absorbs CO₂ more readily. However, DEA has higher NGL solubility, which increases hydrocarbon losses. Typical concentration is 25–35 wt% in water

Liquid-Liquid Contactor Design

The contactor in an amine NGL treating system operates as a liquid-liquid extractor rather than the gas-liquid absorber used in conventional amine gas treating. Design considerations that differ from gas treating service include:

• Phase dispersal: The NGL (lighter phase) is typically dispersed into the continuous amine phase using distributors or nozzles at the column bottom. The amine flows downward while NGL droplets rise through the column, providing countercurrent contact
• Packing selection: Random packing (Pall rings, Raschig rings) or structured packing designed for liquid-liquid extraction is used. The packing must promote droplet breakup and coalescence cycling to renew the interfacial area. Metal packing is preferred over plastic because it is preferentially wetted by the aqueous amine phase
• Column sizing: Flooding in liquid-liquid columns occurs at much lower velocities than in gas-liquid columns. The combined dispersed and continuous phase velocities must be below the flooding limit, which depends on the packing type, phase density difference, and interfacial tension
• Stage efficiency: Tray or packing efficiency in liquid-liquid extraction is typically 10–25%, much lower than gas-liquid absorption (50–80%). This means significantly more theoretical stages and actual trays/packing height are required for equivalent performance

Operating Considerations

Several factors unique to liquid NGL/amine contacting must be managed carefully to maintain treating performance and avoid operational problems:

• Amine solubility in NGL: A small fraction of the amine dissolves in the NGL phase, increasing with temperature and decreasing with NGL molecular weight. MDEA is less soluble in NGL than DEA, which is one reason it is preferred. Amine losses to the NGL product typically range from 0.5 to 2.0 lb/MBBL for MDEA and 1.0 to 4.0 lb/MBBL for DEA
• NGL solubility in amine: Conversely, NGL hydrocarbons dissolve in the amine solution, with propane being more soluble than butane or heavier components. The dissolved NGL flashes out in the amine regeneration system, contributing to the acid gas stream and potentially causing foaming in the regenerator
• Emulsion management: The NGL-amine interface is susceptible to emulsion formation, particularly when contaminants such as corrosion inhibitors, iron sulfide particulates, or degradation products are present. Anti-foam agents and upstream filtration help mitigate emulsion problems

Rich Amine Flash Tank

A flash tank on the rich amine stream downstream of the contactor is an essential component of the amine NGL treating system. The flash tank performs two critical functions:

• NGL recovery: Dissolved and entrained NGL in the rich amine flashes out at reduced pressure (typically 50–100 psig). The flash gas is recovered and returned to the NGL process or used as fuel gas. Without a flash tank, the dissolved NGL would be lost to the amine regenerator overhead and ultimately to the flare or sulfur recovery unit
• Pressure reduction: The flash tank provides a controlled pressure letdown between the high-pressure contactor and the low-pressure amine regeneration system, preventing slug flow and hydraulic problems in the rich amine piping

Integration with Gas Treating Amine System

When an NGL plant is co-located with a gas treating facility that already has an amine system, it is often economically advantageous to integrate the NGL treating amine circuit with the existing gas treating amine system. This integration can reduce capital cost by sharing the amine regenerator, lean amine storage, and amine makeup systems. However, integration requires careful evaluation of:

• Total amine circulation rate and regenerator capacity to handle the additional acid gas load
• Amine quality degradation from NGL contaminants (heavier hydrocarbons, surfactants) that could affect gas treating performance
• Rich amine flash tank sizing to handle dissolved NGL from the liquid treating circuit
• Differential pressure requirements between gas treating (typically higher pressure) and NGL treating contactors

5. Merox Process and Other Treating Methods

When mercaptans are the primary contaminant rather than H₂S, or when mercaptan levels exceed the capacity of conventional caustic washing, specialized treating processes are required. The Merox process is the most widely used technology for mercaptan removal from NGL products, and several other methods are available for specific contaminants such as COS and mercury.

Merox Process Overview

The Merox (mercaptan oxidation) process uses a proprietary organometallic catalyst (typically a cobalt phthalocyanine derivative) in combination with caustic solution and air (or oxygen) to convert mercaptans to disulfides. The fundamental reaction is:

The disulfides (RSSR) formed are much less objectionable than the original mercaptans because they are non-corrosive, have higher odor thresholds, and do not react with copper in the copper strip test. However, the total sulfur content is not reduced by the sweetening reaction—only the form of sulfur changes from reactive (mercaptans) to non-reactive (disulfides). If total sulfur reduction is also required, the Merox process can be operated in extraction mode rather than sweetening mode.

Sweetening Mode (Fixed-Bed Merox)

In sweetening mode, the NGL product flows through a fixed bed of catalyst impregnated on an activated carbon or charcoal support. Caustic solution is injected upstream to maintain alkaline conditions, and air is injected to provide the oxygen required for the oxidation reaction. The process operates as follows:

• Caustic prewash: The NGL is first contacted with caustic to remove H₂S (which would poison the catalyst) and to provide the alkaline environment needed for the catalytic reaction
• Air injection: A controlled amount of air is injected into the NGL upstream of the catalyst bed. The air rate is stoichiometrically proportional to the mercaptan content, typically 2–3 times the theoretical requirement to ensure complete conversion
• Catalyst bed: The NGL passes through the fixed catalyst bed where the oxidation reaction occurs. Bed temperatures are typically 70–130°F, with lower temperatures favoring higher conversion for lighter mercaptans. Space velocities range from 1 to 4 volumes of NGL per volume of catalyst per hour
• Caustic settling: The treated NGL passes through a settler or coalescer to remove entrained caustic before proceeding to product storage

Extraction Mode (Liquid-Liquid Merox)

When total sulfur reduction is required in addition to sweetening, the Merox process operates in extraction mode. In this configuration, the caustic-catalyst solution extracts mercaptans from the NGL in a liquid-liquid contactor, then the rich caustic is oxidized to regenerate it. The key differences from sweetening mode are:

• Extraction column: The NGL contacts the caustic-catalyst solution in a countercurrent extraction column. Mercaptans dissolve into the caustic phase as sodium mercaptides (NaSR), reducing the total sulfur in the treated NGL product
• Oxidizer: The rich caustic from the extraction column flows to an oxidizer vessel where air is sparged through the solution. The dissolved oxygen converts sodium mercaptides back to disulfides and regenerates the NaOH. The disulfides are insoluble in the caustic and separate as a distinct organic phase (disulfide oil) that is collected and removed
• Disulfide oil handling: The recovered disulfide oil is a valuable byproduct containing primarily dimethyl disulfide (DMDS) and diethyl disulfide (DEDS). It can be blended into crude oil, used as a catalyst presulfiding agent, or sold as a chemical intermediate

Catalyst Types and Performance

Merox catalysts are proprietary formulations, but all are based on metal phthalocyanine complexes. The most common is sulfonated cobalt phthalocyanine, which is water-soluble for extractive mode and can be impregnated onto activated carbon for fixed-bed mode. Key catalyst performance factors include:

Molecular Sieve Treating for COS Removal

Carbonyl sulfide (COS) presents a unique challenge in NGL treating because it is not effectively removed by caustic washing, amine treating, or the Merox process. COS is a chemically stable molecule that does not react with NaOH at normal treating conditions and passes through most liquid-liquid treating systems unchanged. However, COS slowly hydrolyzes in the presence of water to form H₂S and CO₂, which causes copper strip test failures in stored NGL products.

Molecular sieve adsorption is the most effective method for COS removal from liquid NGL products. The NGL is passed through a fixed bed of 4A or 5A molecular sieve, which selectively adsorbs COS along with water and H₂S. Design considerations include:

• Molecular sieve type: 4A or 5A zeolite with COS adsorption capacity of 2–4 wt% at treating conditions
• Bed sizing: Based on COS loading, cycle time (typically 8–24 hours), and linear velocity (0.5–1.5 ft/min superficial for liquid-phase service)
• Regeneration: Thermal regeneration at 400–550°F using dry gas (residue gas, nitrogen) in the reverse flow direction. Typically a two-bed system: one on adsorption, one on regeneration
• Pre-treatment: NGL must be pre-dried and pre-treated for H₂S before molecular sieve treating, as both water and H₂S compete for adsorption sites and reduce COS capacity

Copper-Based Adsorbents for Mercury Removal

Mercury contamination in NGL products is a growing concern because of its toxicity, its tendency to amalgamate with aluminum heat exchangers in downstream LNG and ethylene plants, and increasingly stringent environmental regulations. NGL products from certain producing regions contain elemental mercury (Hg⁰) at concentrations ranging from 10 to 300 micrograms per normal cubic meter equivalent.

Copper-impregnated adsorbents (copper sulfide on alumina support) are the most common mercury removal technology for liquid NGL streams. The copper sulfide reacts irreversibly with elemental mercury to form mercury sulfide (HgS), which is retained on the adsorbent bed. Key design parameters include:

• Mercury capacity: 10–15 wt% of adsorbent weight before breakthrough
• Operating temperature: Ambient to 120°F (liquid-phase service)
• Bed life: 2–5 years depending on mercury loading, after which the bed is replaced and the spent adsorbent is disposed of as hazardous waste
• Position in treating train: Typically the last step before product storage to prevent mercury recontamination

Treating Train Sequence by Product

The specific sequence of treating steps varies by NGL product and the contaminants present. Typical treating trains for common NGL products are:

The treating train must be designed as an integrated system, with each step building on the performance of the previous step. Upset conditions in upstream treating stages can overwhelm downstream stages, so process monitoring and alarm systems should be installed at each stage to detect performance deviations early and prevent off-specification product from reaching storage.

References

1. GPSA, Chapter 16 — Hydrocarbon Treating
2. GPA Standard 2140 — Liquefied Petroleum Gas Specifications and Test Methods
3. ASTM D1838 — Standard Test Method for Copper Strip Corrosion by Liquefied Petroleum Gases
4. API Recommended Practice 551 — Process Measurement Instrumentation