May-2017

Handling mercury in gas processing plants

Advances in the safe treatment and disposal of mercury in gaseous and liquid streams.

SAEID MOKHATAB, Gas Processing Consultant
PANAYIOTIS THEOPHANOUS and ALAA SHANABLEH, Johnson Matthey

Article Summary

Mercury is present in varying concentrations in many of the world’s natural gas fields and has to be removed in any gas processing facility, particularly those involving cryogenic gas processing, to ensure the safety and reliability of operations. This article describes methods to remove mercury, discussing the implications of various locations within a gas processing plant, and demonstrates the benefits that can be achieved from Puraspec fixed bed technology. The issue of waste handling for mercury contaminated absorbent is also addressed, with special focus on the need for an environmentally acceptable manner for disposal arising from developments in legislation in all parts of the world.

Mercury is a trace element often found in natural gas and natural gas liquid streams. It can be present as organometallic and inorganic compounds, and in the elemental form depending on the origin of the gas. While organometallic and inorganic mercury typically ends up in hydrocarbon liquids, the elemental form can be found in both gas and liquid phases.1 Since elemental mercury has low solubility in water, the greater part of mercury in these aqueous streams is mostly present in ionic and suspended forms.

The concentration of mercury in natural gas streams can range from a few ng/Nm3 to a few hundreds of μg/Nm3 in different production fields (see Figure 1). Even at the very lowest of concentrations, it is still desirable to reduce the amount of mercury because the consequence for the gas processing plant of mercury attack is severe. Low levels of mercury can result in severe corrosion of brazed aluminum heat exchangers2 used in cryogenic systems, with the potential to lead to catastrophic events such as the known incidents at the Moomba gas plant in Australia and Skikda LNG facility in Algeria. The presence of mercury in feedstocks for petrochemical plants will also cause poisoning of precious metal catalysts,3,4 with the potential for significant financial loss to the operator. Mercury can also pose significant environmental and safety hazards, as evidenced by the now recognised disaster that gave its name to the Minamata disease.5 The cumulative influence of these problems has led to a conservative approach to the design of mercury removal units in gas processing plant, with specifications set at the traditionally lowest level of detection (below 0.01 µg/Nm3).

It is very difficult to predict the concentration of mercury at the low levels present in production reservoirs. This means that at the design stage operators will often have no indication of its presence, leading to either not including a mercury removal system or a system that is under-designed. In turn, the operator can be unaware of impending trouble until failure of an equipment item due to mercury induced corrosion.

Detection methods are continually being improved to more accurately detect low levels of mercury in feed gas; however they still have shortcomings. For the most accurate measurements, sampling should be done at operating conditions and over a prolonged collection period. However, when samples are collected in the gas field and brought to the laboratory, it is found that some of the mercury is adsorbed on the container walls which can result in lower readings. It has also been identified that mercury levels in natural gas fluctuate by a factor of five over periods longer than eight hours.^7,1 These shortfalls in mercury measurement mean that it is increasingly challenging for operators to both identify and design for the presence of the hazardous element in their system.

Mercury distribution in gas processing plants
Mercury that enters a gas processing plant will be distributed across the different process and waste streams. Mercury distribution will depend on its concentration in the raw natural gas and the type of processing scheme. Figure 2 shows mercury distribution through a typical gas processing plant producing Y-grade natural gas liquids (C2+ NGLs) as an average from analytical tests in a number of facilities. It was observed that solutions used for acid gas and moisture removal have a relative affinity for mercury, leading to its partitioning in streams associated with these processing units before a significant proportion finds its way to the cryogenic section. Mercury removed from the main process stream in an acid gas removal unit utilising an amine based solvent will subsequently be released during amine regeneration and can end up in the carbon dioxide waste vent stream or (where applicable) in the product sulphur. Mercury removed from the main process stream in dehydration units utilising molecular sieve or glycol based solvent technologies will be released during the regeneration step and will typically end up either in the regeneration gas stream or the wastewater. Mercury remaining in the sales gas will be released at the point of use of the gas.⁸

Note should be made that the proportions of mercury removed at various processing steps and entering the finished gas and liquid products vary from plant to plant. Therefore, mercury distribution in cryogenic gas processing plants must be properly studied to determine the areas most in danger of mercury attack, and which streams have the highest concentration.
   
Mercury removal techniques
The majority of methods currently in use for removing mercury from natural gas and hydrocarbon liquids employ fixed bed technologies. The main process stream flows through a fixed bed in which mercury undergoes a physical or chemical interaction with the reagent in the mercury removal vessel, producing a mercury-free product. There are two types of mercury removal materials: regenerative mercury adsorbents, and non-regenerative mercury sorbents.⁹

Regenerative mercury adsorbents
Regenerative mercury removal processes utilise the high affinity of mercury for precious metals such as gold and silver. The relatively high cost of gold has resulted in silver being the metal of choice for the duty. An example of its use in practice is silver deposited on the surface of molecular sieve material. The enhanced product is installed as a layer within the molecular sieve unit utilised for natural gas dehydration. In operation, the unit is used to remove both moisture and mercury from the main process stream. As per normal molecular sieve operation, the mercury and moisture saturated unit is then regenerated by hot regeneration gas typically at temperatures around 290°C, with the cycle being repeated on a preset timeline depending on capacities for both contaminants.

The mercury is removed from the main process stream and is concentrated in the regeneration stream. Depending on the fate of this stream, further processing can be required. This can include non-regenerative technologies that chemically react and remove it, and physical separation methods such as condensing and separating it within the aqueous stream.¹⁰ The requirement for further processing of the regeneration stream to completely remove trace amounts of mercury leads to the molecular sieve method rarely being applied as a standalone system for mercury removal.

Non-regenerative mercury sorbents
Non-regenerative mercury removal processes utilise the chemical reaction between mercury and sulphur which forms cinnabar, the most stable form of mercury in nature. A number of mercury removal sorbents are available with various tolerances to operating temperature, liquid hydrocarbons, and water.¹⁰

A type of non-regenerative sorbent is sulphur impregnated activated carbon, employed in a proven commercial process for removing mercury. Activated carbon typically has a very large surface area and relatively small pore size. Both factors contribute to a higher potential for retrograde condensation, 
leading to a free liquid phase within the system. Free liquids will have a detrimental effect on the material not only as a diffusion barrier but also potentially washing away the active phase as it is held onto the carbon support by physical rather than chemical bonds. Although there have been recent improvements in this area, this still leads to limitations in the use of sulphur impregnated activated carbon to dry gas streams.¹¹ Pore size may also restrict access of mercury to the sulphur sites and therefore increase the length of the reaction zone. An additional limitation is the temperature of operation, as at higher temperatures sublimation of sulphur may be another method of loss of the active component. This again reduces mercury removal capacity. Furthermore, it is often difficult to dispose of spent mercury laden carbon material.¹²