Hidden challenges in mercury removal unit operation and design
Mercury removal units (MRUs) are typically regarded as simple fixed-bed systems operating in a passive mode, with minimal instrumentation and limited dynamic behavior. This perceived simplicity sometimes leads to an underestimation of their importance within the overall gas treatment chain.
In liquefied natural gas (LNG) production, the role of MRUs is particularly critical. Their primary function—the removal of trace mercury from feed gas—is essential to prevent corrosion and liquid metal embrittlement of aluminum-based cryogenic equipment. Even extremely low concentrations of elemental mercury can cause irreversible damage to brazed aluminum heat exchangers and other sensitive components, posing significant risks to both asset integrity and plant availability.
Moreover, MRU commissioning is far from trivial. Tasks such as internal inspection, adsorbent pre-dryout and final moisture removal are time-consuming and often define the critical path in the startup schedule of LNG trains. Any delays at this stage can directly postpone the introduction of treated gas to the cryogenic section and subsequent liquefaction.
This article presents a technical overview of inspection findings obtained during an unplanned shutdown, triggered by a sustained increase in pressure drop across the mercury removal bed. Observations—including support balls migration, sulfur leaching and accumulation, and inadequate design of the outlet distributor and support ball layer—illustrate how seemingly minor design or operational oversights can escalate into major reliability challenges. These findings are discussed in the context of established design principles and industry literature, with the aim of highlighting lessons learned and identifying potential areas for improvement.
Case overview: Design background and operational history. The MRU discussed here was commissioned as part of a gas treating system intended to meet the strict mercury specifications required for downstream cryogenic aluminum equipment. The unit was designed as a non-regenerable fixed-bed adsorber, utilizing a granular alumina-based adsorbent impregnated with elemental sulfur—this is a widely adopted solution for the removal of elemental mercury via the irreversible formation of solid mercury sulfide (HgS). The adsorbent was installed in a vertical axial-flow vessel with a multilayer ceramic support system comprised of two sizes of ceramic balls: a bottom layer of 0.75-in. balls and a top layer of 0.25-in. balls. A stainless-steel mesh screen (mesh 20) was placed between the adsorbent and the upper support layer to prevent adsorbent migration through the support media to the bottom of the vessel.
The system was engineered for a start-of-run (SOR) pressure drop of approximately 80 kPa at nominal design conditions. However, from the very beginning of operation, differential pressure readings (FIG. 1) were consistently above this threshold, stabilizing around 120 kPa. These elevated values triggered concerns and were repeatedly communicated to the engineering, procurement and construction (EPC) contractor and the adsorbent vendor.
FIG. 1. Evolution of differential pressure across the MRU bed over time, showing a steady increase from about 120 kPa to > 300 kPa within 18 mos.
To exclude the possibility of an instrumentation error, operations personnel conducted independent field pressure surveys using calibrated portable gauges. Additionally, comprehensive gas quality analyses were performed, including dewpoint measurements for water and hydrocarbons, and a full gas composition analysis.
All measured parameters remained within the specified design envelope, confirming that no liquid hydrocarbons or free water were entering the mercury removal bed.
Further concern was raised by the presence of elemental sulfur deposits found in downstream filtration equipment. These sulfur traces suggested that active sulfur from the adsorbent may have undergone mechanical attrition, sublimation or leaching, migrating downstream and potentially contributing to both fouling and adsorbent degradation mechanisms.
Despite stable process conditions, the pressure drop continued to increase over the course of 18 mos, ultimately reaching > 300 kPa (FIG. 1). As the pressure drop approached critical integrity levels, an unplanned shutdown was initiated, and a full internal inspection was carried out to determine the root causes.
Inspection findings. During the unplanned shutdown, the MRU was fully depressurized, opened and subjected to detailed internal inspection and sequential unloading of support media and adsorbent layers. The following findings were recorded:
- The top layer of ceramic balls appeared flat and uniformly distributed. A localized grayish discoloration was visible near the inlet nozzle area (FIG. 2), which may have resulted from condensation during initial dry-out procedures.
FIG. 2. Top view of the MRU, showing the ceramic ball bed with visible color differences between central and peripheral areas.
- The floating mesh screen separating the support media from the adsorbent was found to be in good condition. No migration of ceramic balls into the adsorbent layer was noted, confirming proper separation at this interface (FIG. 3).
FIG. 3. Condition of the internal mesh screen separating the support media and adsorbent layers.
- The adsorbent bed showed non-uniform coloration in its upper sections. In particular, the central zones appeared darker than the periphery in certain layers, while other layers exhibited a more uniform appearance. This is indicative of flow maldistribution, possibly caused by inlet turbulence or channeling within the bed.
- A significant quantity of small-diameter ceramic balls (0.25 in.) was found mixed into the bottom layer, where only larger (0.75 in.) support media were expected (FIG. 4).
FIG. 4. Outlet distributor partially covered with sulfur deposits and an abnormal layer of small-diameter ceramic balls resting on top of larger support media.
- The bottom outlet distributor and surrounding support balls were covered with loose yellow sulfur powder. While the powder was not compacted, it obstructed the open area of the outlet section and likely acted as a secondary restriction. The sulfur deposits (FIG. 5) may have originated from mechanical attrition or leaching from the adsorbent during operation.
FIG. 5. Outlet distributor heavily fouled with sulfur and mixed-size ceramic balls (left). The outlet nozzle was coated with fine sulfur particles (right).
- No evidence of free liquids, hydrocarbons or oily contamination was observed inside the vessel. All internal surfaces were dry and clean.
- The mechanical internals, including support meshes and guiding structures, were found to be in serviceable condition and were retained for continued use after visual inspection.
Engineering lessons and practical recommendations. The inspection and operational history of the MRU revealed several design deficiencies that directly contributed to elevated pressure drop, adsorbent degradation and reduced system reliability.
Inappropriate adsorbent selection due to reduced gas composition. It was established that the sulfur-impregnated adsorbent applied in the unit was incompatible with the actual gas composition. Capillary condensation of heavy hydrocarbons within the adsorbent pores led to the dissolution of elemental sulfur, which was subsequently leached out of the pore structure. Upon re-evaporation of these hydrocarbons downstream, sulfur precipitated and accumulated in the lower part of the vessel, particularly on the outlet distributor. This mechanism not only compromised mercury removal performance but also contributed to mechanical fouling and increased pressure drop.
The issue was traced back to the project design phase, during which the full gas composition was simplified by the EPC contractor: the hydrocarbon tail was reduced to a single C6+ component. This approximation prevented the adsorbent supplier from performing an accurate compatibility evaluation, resulting in the selection of an adsorbent that was not robust under the actual process conditions.
- Recommendation: Always provide the full, unreduced gas composition to all vendors involved in the selection of mercury removal adsorbents. Thermodynamic behavior, capillary condensation risk and interaction with adsorbent surfaces must be assessed based on real fluid characteristics. For wet or rich-gas service, especially under variable temperature or pressure conditions, use chemically bound metal sulfide adsorbents (e.g., copper or zinc sulfides) that are stable in the presence of liquid hydrocarbons and do not release sulfur into the process.
Underestimated outlet distributor hydraulic resistance. Even during early operation, the unit demonstrated a pressure drop significantly higher than the design target. Inspection confirmed that ceramic support balls partially blocked the outlet distributor’s open area, which had not been adequately accounted for in the hydraulic design.
- Recommendation: EPC contractors should provide detailed hydraulic calculations for distributor internals. The interaction between support layers and outlet distributors must be explicitly considered. Slot exposure, effective open area and local velocities should be validated using detailed calculations or computational fluid dynamics, if necessary.
Incorrect support media configuration. The ceramic support layer originally consisted of 0.75-in. balls beneath 0.25-in. balls. This configuration violated the standard 2:1 particle size ratio guideline, resulting in the mechanical migration of smaller balls into the lower bed. The absence of a separating mesh further contributed to support layer degradation.
- Recommendation: Adhere to recommended particle size transition ratios (typically ≤ 2:1) between adjacent support layers. If this ratio cannot be maintained, a fine mesh screen should be installed to ensure physical separation and maintain bed integrity over long-term operation.
Corrective actions and revised design. Several modifications were implemented to address the identified issues and improve the long-term reliability of the MRU:
- The sulfur-impregnated alumina adsorbent was replaced with a copper-based metal sulfide formulation, which provides enhanced resistance to leaching and degradation in the presence of heavy hydrocarbons. This change eliminated the risk of elemental sulfur leaching and downstream fouling under near-dewpoint conditions.
- The ceramic support bed was reconfigured using a progressive size profile: 2-in. balls at the bottom, followed by 0.75-in. and then 0.25-in. balls on top. A stainless-steel mesh screen was installed above the ceramic media to ensure separation from the adsorbent and prevent particle migration.
- The outlet basket-type distributor was replaced with a redesigned unit featuring recalculated geometry and increased slot area, improving open flow cross-section and reducing outlet-side hydraulic resistance. This modification directly addressed the elevated pressure drop observed during initial operation.
Post-restart monitoring confirmed a substantial reduction in pressure drop and no further evidence of sulfur migration or maldistribution, validating the effectiveness of the corrective measures.
Takeaways. Although MRUs are often regarded as straightforward and low-risk systems, this case demonstrates that their performance is highly sensitive to design assumptions, gas composition accuracy and mechanical layer configuration. Simplified input data, underestimated hydraulic interactions and deviation from basic engineering principles led to measurable reliability issues, financial losses due to unplanned shutdown, and the need for prolonged gas flaring during adsorbent dry-out. This not only disrupted normal production schedules but also contributed to increased carbon dioxide emissions due to extended flaring.
This experience underscores the importance of treating even the “simplest” units with the same engineering attention applied to more complex equipment. Robust design, validated vendor data and proactive inspection planning are essential to ensure consistent performance and long-term availability in gas processing and LNG facilities.
REFERENCES
1 Eckersley, N., “Advanced mercury removal technologies: New technologies can cost-effectively treat ‘wet’ and ‘dry’ natural gas while protecting cryogenic equipment,” Hydrocarbon Processing, January 2010.
ABOUT THE AUTHOR
Timur Dubinin is a process engineer with > 17 yrs of experience in gas processing, LNG technology and plant operations. He specializes in the design, optimization and commissioning of gas treatment systems for liquefaction processes. Dubinin’s expertise includes amine treating, gas dehydration, mercaptan and mercury removal, and other key process systems supporting LNG production. He has contributed to major international projects across Russia, Central Asia and the Middle East, and currently supports large-scale LNG operations for QatarEnergy LNG.
Comments