Carbon Dioxide Capture Process

The first stage in the Sleipner CO₂ capture process entails the mixing of an amine-water solution with the natural gas in two parallel columns (absorbers A and B), both of which are kept at high pressure (100 bara¹⁷) and moderate temperature (60–70°C). The amine—an organic compound derived from ammonia—selectively absorbs the CO₂ by weak chemical bonding and separates out at the bottom of the columns. Thereafter it is transferred via a turbine to a 15 bara flash drum in which the coabsorbed hydrocarbons are removed. The amine is subsequently heated and depressurized to 1.2 bara in a second flash drum where the CO₂ is boiled off. By now the gas is almost (95%) pure CO₂.

As the lean liquid amine still contains residual CO₂, some 10% is subject to thermal regeneration where the CO₂ is stripped off by steam in a desorber column operating at 120°C. The remaining, even leaner amine is then mixed with the regenerated amine and pumped back to the absorbers for a new separation cycle.

Experimental Investigations

In practice it has been difficult to keep consistently within the 2.5% goal because the process has proved somewhat unstable. This led to several modifications, including new internals for the gas scrubber to reduce carryover (1997–1998), a comprehensive rebuilding of absorber A (1999), and new internals for absorber B (2000). However, faced with continuing irregularities, the Sleipner Amine Task Force asked this source [Statoil] to devise a solution.

By now this source had become familiar with the removal of CO₂ at high pressure while attempting to improve the plant’s performance. Indeed, one of the most important experimental observations was that pressure has a significant effect on the absorption capacity of the amines—adsorption capacity decreases with increasing pressure. This was a cause for concern as it could impact the effectiveness of CO₂ capture in the absorber columns.

Spurred on by the task force’s request, the company stepped up its engagement through a major experimental and modeling investigation aimed at better understanding and predicting high-pressure CO₂ capture mechanisms. This involved experimenting with genuine natural gas under realistic pressures and temperatures. The effects of various amine solution additives were tested under similar conditions.

However, it was not until 2003 that a major breakthrough was made. Exploiting the research center’s unique laboratory facilities, the answer was found to lie in increasing the amine circulation rate and the heating energy used to separate CO₂ from the amine. Subsequent offshore tests at Sleipner resulted in a stable performance while reducing the CO₂ content to 2.25%. The new operational procedure and equipment can now be installed at Sleipner T, enabling the amine plant to meet quality specifications when operating at full capacity.

Aquifer Injection

Once the CO₂ has been captured, its pressure is boosted by four compressors to 80 bara prior to being transferred to the Sleipner East A platform for pumping into the base of the saline aquifer. Since 1996 about 1 million tons of compressed CO₂ have been injected annually.

Another requirement is that the well casing and other hardware used in the capture and injection plant have to be made of stainless steel, because even minute quantities of water mixed with CO₂ produce a weak corrosive carbonic acid (H₂CO₃).

Investment costs amounted to some US $80 million (CO₂ capture costs excluded). Although this was a considerable sum, the partners would otherwise have faced an annual tax bill of about US $50 million if the CO₂ had simply been vented into the air.

Geological Aquifer and Cap Rock Characterization

The storage capacity of the saline Utsira aquifer is thought to be greater than 100 times the volume of annual European carbon dioxide emissions from power plants.

The aquifer in question is the Utsira Formation, which the SACS team believes was deposited as part of a submarine turbidite fan system¹⁸ above the Sleipner reservoir rocks. Today it is encountered some 1000 meters below the seabed,¹⁹ and comprises an exceptionally porous and permeable sequence of poorly consolidated, fine- to medium-grained quartz-rich sandstones. Subcropping almost exclusively in the Norwegian sector of the North Sea, it is more than 200 meters thick, over 50 kilometers wide, and extends for some 500 kilometers in a sinuous strip beneath the Brage, Oseberg, Grane, and Sleipner fields and the Tampen production center to the north. The aquifer’s areal coverage is thus about 26,000 square kilometers.

Delineation and mapping of the top of the formation is particularly important for defining its closure. If aquifers form large domal structures, the CO₂ will be constrained and slight structural uncertainties can be ignored. However, precise and detailed depth mapping is vital if they undulate gently, as at Sleipner, where the top of the aquifer above the injection point is relatively flat. This is because minor variations may have a major effect on CO₂ movement (migration routes), areas of accumulation, and overall storage potential.

The regional mapping was done using 2D seismic datasets, while more detailed work was carried out around the injection point using 3D seismic.²⁰ Petrophysical data from some 300 wells were also available for study, plus limited rock samples in the form of drill cuttings and cores. Much sedimentological, geochemical, and rock-mechanical research is still being done on the complex cap rock/overburden sequence, which at Sleipner is about 700 meters thick. A dedicated 9-meter core was cut from this interval in the summer of 2002.

Another consideration is the possible presence and continuity of faults running through the aquifer and cap rock along which CO₂ may escape to the seabed. Fortunately, no significant faults have been detected from the seismic surveys (also see later). The injection process itself could lead to local microseismicity or the opening of incipient, pressure-induced fractures, but the required injection pressures at Sleipner are sufficiently low for this to be regarded as unlikely.

Furthermore, current thinking suggests that the plasticity of the overburden is such that faults and fractures are unlikely to serve as escape conduits. In other words the sealing capacity appears to be good.

Seismic Monitoring

Seismic monitoring has revealed no carbon dioxide leakage in the overburden

Another taxing question was whether the dynamic behavior of the injected CO₂ and its potential impact on cap rock integrity could be monitored using modern geophysical techniques, especially seismic. After much discussion, it was agreed that time-lapse seismic would probably be suitable, because the velocity of sound waves should be able to differentiate between saltwater-bearing (higher velocity) and CO₂-bearing (lower velocity) sandstones. Time-lapse seismic, which is also known as 4D seismic, involves comparing the results of 3D seismic surveys repeated at considerable time intervals: differences between the survey results are attributed to fluid or pressure changes.

Four seismic surveys have been conducted so far: a pre-SACS baseline survey in 1994 prior to CO₂ injection and three monitoring surveys carried out in 1999, 2001, and 2002 during CO₂ injection. The latter have not only successfully traced the injection of the CO₂ and expansion of the “bubble” but have also yielded extremely sharp images of the aquifer’s overall geometry, internal structure and flow behavior. As expected, gravitational separation is the dominant physical process because of the CO₂'s buoyancy.

A particularly striking result is that the distribution and migration paths of the CO₂ are strongly controlled by intraaquifer mud rock horizons. With an extraordinary seismic detection limit of about 1 meter or less, much of the CO₂ can be seen to have migrated upwards between the Utsira Formation mud rock terminations, as witnessed by a distinct seismic chimneylike column appearing on repeated seismic surveys. What is more, it has traveled up to about 1450 meters laterally beneath individual mud rock layers after six years of injection. The lateral speeds at which the CO₂ fronts move range from 0 to about 100 meters per annum—at least in recent years.

This remarkable precision prompted the team to estimate seismically the quantity of injected CO₂—on the assumption (among others) that none has been dissolved in the saline formation water. By comparing the seismically based result with the injected volumes, it appears that all of the CO₂ is accounted for by the seismic data. This, of course, is another argument for suggesting that no significant leakage has occurred, although the lack of seismically observed CO₂ in the overburden remains the most persuasive factor. It is wise, however, to recall that there is always a margin of error associated with the seismic method—albeit relatively minor in this case.

Gravimetric Aquifer Monitoring

Time-lapse gravity can potentially be used to better determine carbon dioxide density and mass distribution.

Although gravimetry has a lower spatial resolution than its seismic counterpart, repeated high-precision microgravity monitoring potentially provides better constraints on CO₂ density and mass distribution. It may also give an early warning signal if considerable amounts are escaping upwards through the overburden, as well as yielding relatively inexpensive information on the long-term dissolution of the CO₂ in the formation water once injection has ceased.

The company’s latest offshore time-lapse gravity surveying technique is being used, having been successfully employed at the Troll field to image and monitor changes in the (hydrocarbon) gas/water contact. Developed in association with Scripps (University of California, San Diego) and cofunded by the US Department of Energy, the state-of-the-art seafloor gravimeter contains three gravity sensors and three pressure sensors, which enable the instrument to monitor small vertical changes in the seafloor as well as small gravity changes. The gravitational accuracy is about 5 × 10^–9 of the earth’s total gravity field.

A 7 × 3 kilometer baseline survey was obtained at Sleipner in August 2002, against which future surveys will be compared. So far the results have exceeded expectations: not only may it be possible to detect vertical changes in the seafloor as small as 0.5 centimeters, but the time-lapse detection threshold may also be as low as 5 μGal²¹—some 50% better than that suggested by a presurvey modeling exercise.

However, there are other considerations to be taken into account, such as the gravitational effect of further production from the underlying Sleipner gas-condensate reservoirs. The technique also depends on lowering the gravimeter onto separate concrete blocks installed on the seafloor one at a time, although this did not prove to be a hindrance.

Aquifer Flow Modeling

Simulations suggest that the carbon dioxide “mega-bubble” may reach its ultimate size after a few hundred years, thereafter shrinking and finally disappearing within a few thousand years

Whereas geophysical surveys are designed to determine rock and fluid distributions, reservoir simulations are designed to predict how fluids will behave with time. In a case like this, it is naturally wise to make a preinjection simulation to test operational feasibility, as was done at Sleipner before the SACS project started.

The SACS team has subsequently built a detailed postinjection model to verify and improve the seismic and geological interpretation of the aquifer around the injection site; and a coarser, larger-scale model to predict CO₂ migration over a period of several thousand years. The areas covered by the models are 7 square kilometers and 128 square kilometers, respectively. In both cases, the seismically inferred mud rock distributions were imported into the reservoir models, because it is almost impossible to trace individual mudstone layers from well to well, even when they are close together. Calibration of the 3D repeated seismic data with a local reservoir model is thus a fundamental prerequisite.

The results from the larger model suggest that most of the CO₂ will eventually coalesce to form a single “mega-bubble” beneath the cap rock a few years after injection has ceased. It will also gradually spread along the top of the salty formation water according to the local topography of the cap rock seal. This, however, must be tempered by the fact that CO₂ will diffuse from the “mega-bubble” into the underlying brine column, a phenomenon that is usually ignored in standard reservoir simulations because it is extremely slow compared with other transport processes. But given time, the CO₂-enriched brine on top of the column will become denser than that beneath, resulting in a downward flow compensated by convection plumes. This, in turn, will enhance dissolution and increase the probability of the CO₂ remaining in the aquifer.

When dissolution is included in the simulations, the “mega-bubble” will probably reach its ultimate size after a few hundred years, thereafter shrinking and finally disappearing within a few thousand years.

Further Investigations

The ultimate objective is to combine chemical- and flow-oriented modeling approaches for making reliable, long-term predictions.

The main product of the SACS project is a comprehensive Best Practice Manual (2003). This contains a suggested procedure for evaluating CO₂ storage from a technical point of view, besides information aimed at satisfying authorities and the general public as to the feasibility, safety, and reliability of the storage process.

The Sleipner case, which is being used as a full-scale natural laboratory, has yielded copious information on CO₂ transport rates and geophysical properties and has gone some way towards assessing the sealing capacity of the overburden.

These are considerable shorter-term achievements, of which the seismic monitoring is the most conspicuous. However, some of the most telling challenges still lie ahead, particularly the making of reliable long-term predictions, recalling that long-term in this context refers to several hundred to several thousand years hence.

Ongoing investigations in the CO₂ Store program (2003–2005) include assessments of whether the free and dissolved CO₂ remain in the host aquifer or migrate elsewhere and whether the sealing capacity of the cap rock will be maintained, realizing that CO₂-rich water is slightly acidic and may lead to mineral dissolution.

Other important issues are whether and how much of the injected CO₂ can be permanently fixed by chemical reactions and in what form and whether such chemical changes will impair porosity and permeability, thereby reducing aquifer storage capacity while (possibly) improving retention. The conditions under which CO₂ might ultimately be dissolved in its entirety are also receiving attention.

In short, the CO₂ Store program aims to extend the capabilities of two-phase (gas, water) reservoir simulators to better handle extremely long-term simulations, including the migration of CO₂ in its dissolved form, and chemical-oriented modeling to predict the maximum potential for CO₂ reaction with the Utsira Formation sediments and the cap rock.

The ultimate objective is to combine chemical modeling with the flow-oriented modeling approach of reservoir simulation to produce merged “chemical and reactive transport models” constrained by geological and geochemical understanding—a highly ambitious undertaking. Geomechanical modeling is also coming to the fore in a number of related investigations.

Other Projects

Although space limitations preclude extensive coverage of all but the most eye-catching projects, it is important to mention that other collaborative ventures and proprietary in-house projects are systematically examining the entire range of available technologies. Among them, Statoil is a partner in three integrated, multipartner ventures which are partly funded by the European Union; namely, ENCAP, CASTOR, and CO₂Sink. These are concerned with pre-combustion, post-combustion, and storage, respectively.

The company is also jointly involved in several basic research projects, two of them dealing with CO₂ capture and one with the use of CO₂ for improved oil recovery (discussed next). The aims are to elucidate fundamental physicochemical phenomena in the hope of making significant breakthroughs.

3.1.3.1 Nitrogen Oxides

Nitrogen oxides formation occurs by three fundamentally different mechanisms. The principal mechanism with turbines firing gas or distillate fuel is thermal NO_x, which arises from the thermal dissociation and subsequent reaction of nitrogen (N₂) and oxygen (O₂) molecules in the combustion air. Most thermal NO_x is formed in high temperature stoichiometric flame pockets downstream of the fue linjectors where combustion air has mixed sufficiently with the fuel to produce the peak temperature fuel/air interface.

The second mechanism, called prompt NO_x, is formed from early reactions of nitrogen molecules in the combustion air and hydrocarbon radicals from the fuel. Prompt NO_x forms within the flame and is usually negligible when compared to the amount of thermal NO_x formed. The third mechanism, fuel NO_x, stems from the evolution and reaction of fuel-bound nitrogen compounds with oxygen. Natural gas has negligible chemically-bound fuel nitrogen (although some molecular nitrogen is present). Essentially all NO_x formed from natural gas combustion is thermal NO_x. Distillate oils have low levels of fuel-bound nitrogen. Fuel NO_x from distillate oil-fired turbines may become significant in turbines equipped with a high degree of thermal NO_x controls. Otherwise, thermal NO_x is the predominant NO_x formation mechanism in distillate oil-fired turbines.

The maximum thermal NO_x formation occurs at a slightly fuel-lean mixture because of excess oxygen available for reaction. The control of stoichiometry is critical in achieving reductions in thermal NO_x. Thermal NO_x formation also decreases rapidly as the temperature drops below the adiabatic flame temperature, for a given stoichiometry. Maximum reduction of thermal NO_x can be achieved by control of both the combustion temperature and the stoichiometry. Gas turbines operate with high overall levels of excess air, because turbines use combustion air dilution as the means to maintain the turbine inlet temperature below design limits. In older gas turbine models, where combustion is in the form of a diffusion flame, most of the dilution takes place downstream of the primary flame, which does not minimize peak temperature in the flame and suppress thermal NO_x formation.

Diffusion flames are characterized by regions of near-stoichiometric fuel/air mixtures where temperatures are very high and significant thermal NO_x is formed. Water vapor in the turbine inlet air contributes to the lowering of the peak temperature in the flame, and therefore to thermal NO_x emissions. Thermal NO_x can also be reduced in diffusion type turbines through water or steam injection. The injected water-steam acts as a heat sink lowering the combustion zone temperature, and therefore thermal NO_x. Newer model gas turbines use lean, premixed combustion where the fuel is typically premixed with more than 50% theoretical air which results in lower flame temperatures, thus suppressing thermal NO_x formation.

Ambient conditions also affect emissions and power output from turbines more than from external combustion systems. The operation at high excess air levels and at high pressures increases the influence of inlet humidity, temperature, and pressure [4]. Variations of emissions of 30% or greater have been exhibited with changes in ambient humidity and temperature. Humidity acts to absorb heat in the primary flame zone due to the conversion of the water content to steam. As heat energy is used for water to steam conversion, the temperature in the flame zone will decrease resulting in a decrease of thermal NO_x formation. For a given fuel firing rate, lower ambient temperatures lower the peak temperature in the flame, lowering thermal NO_x significantly. Similarly, the gas turbine operating loads affect NO_x emissions. Higher NO_x emissions are expected for high operating loads due to the higher peak temperature in the flame zone resulting in higher thermal NO_x.

3.1.3.2 Carbon Monoxide and Volatile Organic Compounds

CO and VOC emissions both result from incomplete combustion. CO results when there is insufficient residence time at high temperature or incomplete mixing to complete the final step in fuel carbon oxidation. The oxidation of CO to CO₂ at gas turbine temperatures is a slow reaction compared to most hydrocarbon oxidation reactions. In gas turbines, failure to achieve CO burnout may result from quenching by dilution air. With liquid fuels, this can be aggravated by carryover of larger droplets from the atomizer at the fuel injector. Carbon monoxide emissions are also dependent on the loading of the gas turbine. For example, a gas turbine operating under a full load will experience greater fuel efficiencies, which will reduce the formation of carbon monoxide. The opposite is also true, a gas turbine operating under a light to medium load will experience reduced fuel efficiencies (incomplete combustion), which will increase the formation of carbon monoxide.

The pollutants commonly classified as VOC can encompass a wide spectrum of volatile organic compounds, some of which are hazardous air pollutants. These compounds are discharged into the atmosphere when some of the fuel remains unburned or is only partially burned during the combustion process. With natural gas, some organics are carried over as unreacted, trace constituents of the gas, while others may be pyrolysis products of the heavier hydrocarbon constituents. With liquid fuels, large droplet carryover to the quench zone accounts for much of the unreacted and partially pyrolized volatile organic emissions.

Similar to CO emissions, VOC emissions are affected by the gas turbine operating load conditions. Volatile organic compounds emissions are higher for gas turbines operating at low loads as compared to similar gas turbines operating at higher loads.

3.1.3.3 Particulate Matter [13]

PM emissions from turbines primarily result from carryover of noncombustible trace constituents in the fuel. PM emissions are negligible with natural gas firing and marginally significant with distillate oil firing because of the low ash content. PM emissions can be classified as “filterable” or “condensable” PM. Filterable PM is that portion of the total PM that exists in the stack in either the solid or liquid state and can be measured on a EPA Method 5 filter. Condensable PM is that portion of the total PM that exists as a gas in the stack but condenses in the cooler ambient air to form particulate matter. Condensable PM exists as a gas in the stack, so it passes through the Method 5 filter and is typically measured by analyzing the impingers, or “back half,” of the sampling train. The collection, recovery, and analysis of the impingers is described in USA EPA Method 202 of Appendix M, Part 51 of the Code of Federal Regulations. Condensable PM is composed of organic and inorganic compounds and is generally considered to be all less than 1.0 micrometers in aerodynamic diameter.

3.1.3.4 Greenhouse Gases [5–11]

Carbon dioxide (CO₂) and nitrous oxide (N₂O) emissions are all produced during natural gas and distillate oil combustion in gas turbines. Nearly all of the fuel carbon is converted to CO₂ during the combustion process. This conversion is relatively independent of firing configuration. Methane (CH₄) is also present in the exhaust gas and is thought to be unburned fuel in the case of natural gas or a product of combustion in the case of distillate fuel oil.

Although the formation of CO acts to reduce CO₂ emissions, the amount of CO produced is insignificant compared to the amount of CO₂ produced. The majority of the fuel carbon not converted to CO₂ is due to incomplete combustion.

Formation of N₂O during the combustion process is governed by a complex series of reactions and its formation is dependent upon many factors. However, the formation of N₂O is minimized when combustion temperatures are kept high (above 1475°F) and excess air is kept to a minimum (less than 1%).

3.1.3.5 HAP Emissions

Available data indicate that emission levels of HAP are lower for gas turbines than for other combustion sources. This is due to the high combustion temperatures reached during normal operation. The emissions data also indicate that formaldehyde is the most significant HAP emitted from combustion turbines. For natural gas-fired turbines, formaldehyde accounts for about two thirds of the total HAP emissions. Polycyclic aromatic hydrocarbons (PAH), benzene, toluene, xylenes, and others account for the remaining one third of HAP emissions. For No. 2 distillate oil-fired turbines, small amounts of metallic HAP are present in the turbine’s exhaust in addition to the gaseous HAP identified under gas-fired turbines. These metallic HAP are carried over from the fuel constituents. The formation of carbon monoxide during the combustion process is a good indication of the expected levels of HAP emissions. Similar to CO emissions, HAP emissions increase with reduced operating loads. Typically, combustion turbines operate under full loads for greater fuel efficiency, thereby minimizing the amount of CO and HAP emissions.

3.1.4.1 Water Injection

Water or steam injection is a technology that has been demonstrated to effectively suppress NO_x emissions from gas turbines. The effect of steam and water injection is to increase the thermal mass by dilution and thereby reduce peak temperatures in the flame zone. With water injection, there is an additional benefit of absorbing the latent heat of vaporization from the flame zone. Water or steam is typically injected at a water-to-fuel weight ratio of less than 1.

Depending on the initial NO_x levels, such rates of injection may reduce NO_x by 60% or higher. Water or steam injection is usually accompanied by an efficiency penalty (typically 2–3%) but an increase in power output (typically 5–6%). The increased power output results from the increased mass flow required to maintain turbine inlet temperature at manufacturer’s specifications. Both CO and VOC emissions are increased by water injection, with the level of CO and VOC increases dependent on the amount of water injection.

3.1.4.2 Dry Controls

Since thermal NO_x is a function of both temperature (exponentially) and time (linearly), the basis of dry controls are to either lower the combustor temperature using lean mixtures of air or fuel staging or decrease the residence time of the combustor. A combination of methods may be used to reduce NO_x emissions such as lean combustion and staged combustion (two-stage lean/lean combustion or two-stage rich/lean combustion).

Lean combustion involves increasing the air-to-fuel ratio of the mixture so that the peak and average temperatures within the combustor will be less than that of the stoichiometric mixture, thus suppressing thermal NO_x formation. Introducing excess air not only creates a leaner mixture but it also can reduce residence time at peak temperatures.

Two-stage lean/lean combustors are essentially fuel-staged, premixed combustors in which each stage burns lean. The two-stage lean/lean combustor allows the turbine to operate with an extremely lean mixture while ensuring a stable flame. A small stoichiometric pilot flame ignites the premixed gas and provides flame stability. The NO_x emissions associated with the high temperature pilot flame are insignificant. Low NO_x emission levels are achieved by this combustor design through cooler flame temperatures associated with lean combustion and avoidance of localized “hot spots” by premixing the fuel and air.

Two-stage rich/lean combustors are essentially air-staged, premixed combustors in which the primary zone is operated fuel rich and the secondary zone is operated fuel lean. The rich mixture produces lower temperatures (compared to stoichiometric) and higher concentrations of CO and H₂, because of incomplete combustion. The rich mixture also decreases the amount of oxygen available for NO_x generation. Before entering the secondary zone, the exhaust of the primary zone is quenched (to extinguish the flame) by large amounts of air and a lean mixture is created. The lean mixture is preignited and the combustion completed in the secondary zone. NO_x formation in the second stage is minimized through combustion in a fuel lean, lower temperature environment. Staged combustion is identified through a variety of names, including dry-low NO_x (DLN), dry-low emissions (DLE), or SoLoNO_x.

3.1.4.3 Catalytic Reduction Systems³⁹

Selective catalytic reduction (SCR) systems selectively reduce NO_x emissions by injecting ammonium (NH₃) into the exhaust gas stream upstream of a catalyst. Nitrogen oxides, NH₃, and O₂ react on the surface of the catalyst to form N₂ and H₂O. The exhaust gas must contain a minimum amount of O₂ and be within a particular temperature range (typically 450–850°F) in order for the SCR system to operate properly.

The temperature range is dictated by the catalyst material, which is typically made from noble metals, including base metal oxides such as vanadium and titanium, or zeolite-based material. The removal efficiency of an SCR system in good working order is typically from 65–90%. Exhaust gas temperatures greater than the upper limit (850°F) cause NO_x and NH₃ to pass through the catalyst unreacted. Ammonia emissions, called NH₃ slip, may be a consideration when specifying an SCR system.

Ammonia, either in the form of liquid anhydrous ammonia or aqueous ammonia hydroxide, is stored on-site and injected into the exhaust stream upstream of the catalyst. Although an SCR system can operate alone, it is typically used in conjunction with water-steam injection systems or lean-premix system to reduce NO_x emissions to their lowest levels (less than 10 ppm at 15% oxygen for SCR and wet injection systems). The SCR system for landfill or digester gas-fired turbines requires a substantial fuel gas pretreatment to remove trace contaminants that can poison the catalyst. Therefore, SCR and other catalytic treatments may be inappropriate control technologies for landfill or digester gas-fired turbines.

The catalyst and catalyst housing used in SCR systems tend to be very large and dense (in terms of surface area to volume ratio) because of the high exhaust flow rates and long residence times required for NO_x, O₂, and NH₃, to react on the catalyst. Most catalysts are configured in a parallel-plate, “honeycomb” design to maximize the surface area-to-volume ratio of the catalyst. Some SCR installations incorporate CO catalytic oxidation modules along with the NO_x reduction catalyst for simultaneous CO/NO_x control.

Carbon monoxide oxidation catalysts are typically used on turbines to achieve control of CO emissions, especially turbines that use steam injection, which can increase the concentrations of CO and unburned hydrocarbons in the exhaust. CO catalysts are also being used to reduce VOC and organic HAPs emissions. The catalyst is usually made of a precious metal such as platinum, palladium, or rhodium. Other formulations, such as metal oxides for emission streams containing chlorinated compounds, are also used. The CO catalyst promotes the oxidation of CO and hydrocarbon compounds to carbon dioxide (CO₂) and water (H₂O) as the emission stream passes through the catalyst bed. The oxidation process takes place spontaneously, without the requirement for introducing reactants. The performance of these oxidation catalyst systems on combustion turbines results in 90+% control of CO and about 85–90% control of formaldehyde. Similar emission reductions are expected on other HAP pollutants.

3.1.4.4 Other Catalytic Systems [14,15]

New catalytic reduction technologies have been developed and are currently being commercially demonstrated for gas turbines. Such technologies include, but are not limited to, the Sconox and the Xonon systems, both of which are designed to reduce NO_x and CO emissions. The Sconox system is applicable to natural gas-fired gas turbines. It is based on a unique integration of catalytic oxidation and absorption technology. CO and NO are catalytically oxidized to CO₂ and NO₂. The NO₂ molecules are subsequently absorbed on the treated surface of the Sconox catalyst. The system manufacturer guarantees CO emissions of 1 ppm and NO_x emissions of 2 ppm. The Sconox system does not require the use of ammonia, eliminating the potential of ammonia slip conditions evident in existing SCR systems. Only limited emissions data were available for a gas turbine equipped with a Sconox system. These data reflected HAP emissions and were not sufficient to verify the manufacturer’s claims.