Because refinery premium products are adversely affected by water, refiners invest considerable equipment, materials and labor to remove water, which refiners ironically are compelled to mix with petroleum streams to process products to high quality specifications. Steam stripping, caustic treating and amine treating add water under normal operating conditions, during processing above design rates, while treating high sulfur feed stocks and as a result of operational problems. Until recently liquid/liquid coalescing to remove water has relied upon (1) salt and desiccant towers, (2) conventional mesh packing/glass coalescers, (3) packed bed towers, (4) electrostatic precipitators and (5) tank settling. Recent availability of high efficiency liquid/liquid coalescers with specialty media has not only increased the number of choices but also has increased suspended water removal and decreased costs even under some of refiners most difficult situations.

Water, by itself, in fuel, FRRPS (Fictitious Roving Refinery Problem Solver) has seen, can corrode and plug engine parts. Water is also a significant contributor to tank bottom corrosion and bacterial growth. Water can contain corrosive materials, such as chlorides and caustic, that will cause equipment damage. Products containing as little as 100 ppm water can be off specification due to haze, color or overall water concentration. Adding to the water problem are detergents and additives that are surfactants, because they lower the interfacial tension between water and fuel. Of course, if water is not suspended or free, it will be dissolved or in solution with hydrocarbons and, consequently, cannot be removed completely by any process described in this article. Suspended water droplets between 0.1 micron to 10 microns in diameter cannot be individually detected but are noticed when enough droplets are concentrated to form a haze. Free water is suspended as an emulsion. Emulsion stability is a factor of interfacial tension, viscosity, relative density and temperature.

The ability to remove water from hydrocarbons, FRRPS explains, improves as the interfacial tension between water and hydrocarbons increases. The interfacial tension between two liquids is the measure of the attraction force among each phase for its own species. There is a similarity, FRRPS points out, between interfacial tension and surface tension. Surface tension is the measure of the attraction force of liquid species at interface with air and is typically about ten times the value for interfacial tension. A ring-pull method is commonly used to measure interfacial tension. Equipment for this method costs about $3,000 and consists of a platinum-iridium ring, transparent vessel and measurement devices for force and distance of the pull. Liquids to be tested are placed in the transparent vessel such that an interface is apparent. The ring is pulled up through the bottom liquid. As the top of the ring appears to break through the plane formed by the interface, the bottom liquid hangs on creating an anomaly at the interface. The force required to pull the ring through the interface and the distance from the interface plane and the point at which the bottom liquid drops away from the ring are used to calculate interfacial tension. For our purposes the liquids measured would be the unmixed, emulsionless hydrocarbon and water. Typical units of interfacial tension are dynes/cm. A high interfacial tension, such as 20 dynes/cm or higher which is associated with clean water and gasoline without surfactant, between two liquids accurately predicts large stable coalesced droplets, that can be easily separated. A low interfacial tension, such as between 2 dynes/cm and 20 dynes/cm which is associated with a surfactant, water and gasoline, between two liquids forecasts small stable coalesced droplets, which form emulsions and are difficult to separate. Of course coalesced droplet size will also depend upon relative droplet velosity, density and viscosity.

Liquid media vicosity, FRRPS generalizes, has a significant impact on the coelescence process. In refinery fuels, the fuel viscosity is important, because the first step in coelescing is for two water droplets to travel through the gasoline and collide. The second step is for two water droplets to fuse together forming a larger droplet, which requires liquid/liquid interface breakdown between the droplets. Both steps are impeded by high viscosity. As a result, more residence time is required to accomplish the same level of coalescence as with a lower viscosity liquid. Lowering the flow rate or increasing the coalescing medium's area will increase residence time.

FRRPS states that as relative density of the two liquids to be separated, such as water and gasoline, approach each other, separation becomes more difficult.

As temperature increases, FRRPS asserts, interfacial tension decreases lowering water droplet's size in fuel. Fuels at high temperatures can contain high concentrations of dissolved water, which is impossible to remove by liquid/liquid coalescers. However, as temperature decreases, water falls out of solution into a suspended state and can then be removed by liquid/liquid coalescers.

When interfacial tension is less than 20 dynes/cm, FRRPS instructs us, emulsions are stable enough to resist being broken though processing in conventional mesh packing/glass coalescers, packed bed towers and tank settling. Electrostatic precipitators are ineffective on emulsions with interfacial tension below 10 dynes/cm. If refiners have emulsions that are causing problems, FRRPS confides, perform the necessary laboratory test to determine its interfacial tension. It is possible that the existing process is not capable of completely breaking the emulsion. Forming stable emulsions generally require at least three components, (1) an immiscible liquid, (2) another immiscible liquid and (3) a small concentration of surfactant. Two immiscible liquids found in large volumes in refineries are hydrocarbons and water. Surfactants are present in refineries under many different names, such as (1) corrosion inhibitors, (2) organic acids in feedstocks, (3) sulfur compounds, (4) dispersents, (5) static dissipaters and (6) other chemical additives. In addition, fine particles, such as iron oxides and iron sulfides, accumulating at the interface where two droplets have made contact prior to fusion, stabilize micellular boundaries and inhibit phase separation.

Surfactant, consisting of a long hydrocarbon chain plus an ionic end, in water systems, FRRPS has learned, leads to micelle formation, by clusters of 50 to 150 surfactant molecules that align themselves so that their non-polar hydrocarbon chain sections, containing twelve or more carbon atoms to be effective, are adsorbed by hydrocarbon droplets while the ionic ends, being hydrophilic, are pointed radially outward. Polar heads repel other micelle formations, explaining why surfactants stabilize dispersions, making them difficult to separate.

Surfactant in oil systems, according to FRRPS, forms reverse micelles, by surfactant molecules that orient themselves so that their polar heads are adsorbed by the aqueous core while the non-polar relatively long chain sections are pointed radially outward. Non-polar chain sections repel other reverse micelle formations, explaining why surfactants stabilize dispersions, making them difficult to separate.

Surfactants disarm conventional glass filter coalescers by bonding with glass fibers allowing water molecules to flow through coalescers with hydrocarbons. Water breakthrough, FRRPS relates, occurs because surfactant molecules bond with the silenol functional group (Si-O-Si) of the glass fiber, thereby preventing water molecules from collecting on glass fibers to form larger droplets which eventually become large enough to drain from the coalescer. Consequently, surfactants shorten glass coalescers service lives resulting in frequent cartridge changeouts and increased disposal costs. Similarly, surfactants reduce efficiencies of packed bed towers and electrostatic precipitators.

FRRPS reflects on a sand filter that could not be operated to meet product specification. Although the sand filter was being used to remove caustic from gasoline, after caustic treating for mercaptans, FRRPS' experience tells us that glass is not compatible with caustic and the measured interfacial tension of from 1 dyne/cm to 12 dynes/cm is below the recommended limit for glass coalescers. A specially formulated coalescer which is compatible with caustic is used to meet product specification.

High efficiency liquid/liquid coalescers, FRRPS confides, require three stages to be successful, (1) filtration, (2) coalescence and (3) separation. Filtration is required to remove fine particles, such as iron oxide and iron sulfide that stabilize emulsions. Coalescence is accomplished by polymeric and fluoropolymeric material which are effective emulsion breakers in liquids measuring interfacial tension above 1 dyne/cm. Separation occurs when coalesced water droplets are repelled by a hydrophobic barrier cartridge leaving the hydrocarbon liquid as it flows through the cartridge. Water is gathered outside the cartridge.

An independent laboratory, FRRPS understands, performed a filter/coalescer/separator test on an unleaded gasoline to which 100 ppm to 30,000 ppm by volume water was added and mixed to a finely divided emulsion. Interfacial tension for the gasoline/water system was measured at 3 dynes/cm to 7 dynes/cm. Effluent concentrations of free water from all tests were less than 15 ppm by volume. Gasoline and water were mixed after being processed through separate filters. The mixture entered the inside of a cartridge coalescer that was stacked on top of a separator cartridge of the same diameter but with a plate between the two cartridges so that fluid had to pass from the inside to the outside of the coalescer to reach the separator. Both the coalescer and separator were enclosed in a vessel that contained both liquids exiting the coalescer. Water exited the coalescer not in a finely divided emulsion but as an accumulation of large droplets which, because they could not penetrate the separator's hydrophobic material settled to the bottom of the enclosure. Gasoline and water were drained from the vessel through separate piping.

FRRPS knows that a filter/coalescer/separator test stand processed diesel fuel containing 120 ppm free water and registering an average haze rating of about 6 down to about 6 ppm to 11 ppm free water and an average haze rating of about 1. The existing filter/coalescer/salt tower/filter treatment had minimal haze-removal efficiency at operating temperatures above 100 F resulting in approximately 4% of drier-processed-diesel being off-specification due to temperature-related haze. Comparing filter and coalescer/separator replacement costs of a new filter/coalescer/separator installation versus historical maintenance, material and disposal costs for salt drier operation, FRRPS noticed that the new facility would be substantially less costly even without considering the 4% off-specification product.

Refiners using salt driers are aware of operating problems, FRRPS has heard, such as (1) bridging, which results in sometimes as little as 50 % salt usage, (2) channeling, which is caused by high flow rates and poor distribution and results in large lengthwise voids within driers, (3) plugging, which occurs at lower temperatures, resulting in maintenance problems, and (4) downstream corrosion of tanks, piping and equipment, which is caused by retained water containing chlorides. Salt drier efficiency is best when operating at relatively low temperatures and at steady flow rates. However, salt driers can remove dissolved water, while coalescers may remove only free or suspended water. Coalescer benefits include (1) not adding potentially corrosive materials to fuel, (2) very high water removal efficiencies, (3) relatively low capital and operating costs and (4) ability to operate efficiency at fluctuating flow rates and temperatures. A major advantage of coalescers is operating costs. In salt driers the amount of salt consumed is directly related to the amount of water removed. It requires 10 times more salt to remove 1,000 ppm water than it does to remove 100 ppm water. Coalescer operating costs do not change even if the amount of water is increased. Beyond these benefits recent coalescer design includes (1) effective filters for removing fine particles that help stabilize emulsions and (2) polymeric medium to provide good flow distribution and overcome disarming. As a result of these benefits, stacked coalescer/separator systems costs for processing refinery fuels are estimated to be (1) $0.0063/BBL for 0.7 centistokes (100 F) gasoline, (2) $0.0097/BBL for 1.6 centistokes Jet A, (3) $0.0101/BBL for 2.3 centistokes Jet B, (4) $.0134/BBL for 3.5 centistokes Diesel (D-2), (5) $0.0139 for 3.6 centistokes No. 2 Fuel Oil and (6) $0.0281/BBL for 8.5 centistokes No. 4 Fuel Oil.

The ability to remove water from hydrocarbons, FRRPS explains, improves as the interfacial tension between water and hydrocarbons increases. The interfacial tension between two liquids is the measure of the attraction force among each phase for its own species. There is a similarity, FRRPS points out, between interfacial tension and surface tension. Surface tension is the measure of the attraction force of liquid species at interface with air and is typically about ten times the value for interfacial tension. A ring-pull method is commonly used to measure interfacial tension. Equipment for this method costs about $3,000 and consists of a platinum-iridium ring, transparent vessel and measurement devices for force and distance of the pull. Liquids to be tested are placed in the transparent vessel such that an interface is apparent. The ring is pulled up through the bottom liquid. As the top of the ring appears to break through the plane formed by the interface, the bottom liquid hangs on creating an anomaly at the interface. The force required to pull the ring through the interface and the distance from the interface plane and the point at which the bottom liquid drops away from the ring are used to calculate interfacial tension. For our purposes the liquids measured would be the unmixed, emulsionless hydrocarbon and water. Typical units of interfacial tension are dynes/cm. A high interfacial tension, such as 20 dynes/cm or higher which is associated with clean water and gasoline without surfactant, between two liquids accurately predicts large stable coalesced droplets, that can be easily separated. A low interfacial tension, such as between 2 dynes/cm and 20 dynes/cm which is associated with a surfactant, water and gasoline, between two liquids forecasts small stable coalesced droplets, which form emulsions and are difficult to separate. Of course coalesced droplet size will also depend upon relative droplet velosity, density and viscosity. Viscosity

Liquid media vicosity, FRRPS generalizes, has a significant impact on the coelescence process. In refinery fuels, the fuel viscosity is important, because the first step in coelescing is for two water droplets to travel through the gasoline and collide. The second step is for two water droplets to fuse together forming a larger droplet, which requires liquid/liquid interface breakdown between the droplets. Both steps are impeded by high viscosity. As a result, more residence time is required to accomplish the same level of coalescence as with a lower viscosity liquid. Lowering the flow rate or increasing the coalescing medium's area will increase residence time.

Relative Density

FRRPS states that as relative density of the two liquids to be separated, such as water and gasoline, approach each other, separation becomes more difficult.

Temperature

As temperature increases, FRRPS asserts, interfacial tension decreases lowering water droplet's size in fuel. Fuels at high temperatures can contain high concentrations of dissolved water, which is impossible to remove by liquid/liquid coalescers. However, as temperature decreases, water falls out of solution into a suspended state and can then be removed by liquid/liquid coalescers.

Emulsion Stability

When interfacial tension is less than 20 dynes/cm, FRRPS instructs us, emulsions are stable enough to resist being broken though processing in conventional mesh packing/glass coalescers, packed bed towers and tank settling. Electrostatic precipitators are ineffective on emulsions with interfacial tension below 10 dynes/cm. If refiners have emulsions that are causing problems, FRRPS confides, perform the necessary laboratory test to determine its interfacial tension. It is possible that the existing process is not capable of completely breaking the emulsion. Forming stable emulsions generally require at least three components, (1) an immiscible liquid, (2) another immiscible liquid and (3) a small concentration of surfactant. Two immiscible liquids found in large volumes in refineries are hydrocarbons and water. Surfactants are present in refineries under many different names, such as (1) corrosion inhibitors, (2) organic acids in feedstocks, (3) sulfur compounds, (4) dispersents, (5) static dissipaters and (6) other chemical additives. In addition, fine particles, such as iron oxides and iron sulfides, accumulating at the interface where two droplets have made contact prior to fusion, stabilize micellular boundaries and inhibit phase separation.

Micelles (Oil in Water)

Surfactant, consisting of a long hydrocarbon chain plus an ionic end, in water systems, FRRPS has learned, leads to micelle formation, by clusters of 50 to 150 surfactant molecules that align themselves so that their non-polar hydrocarbon chain sections, containing twelve or more carbon atoms to be effective, are adsorbed by hydrocarbon droplets while the ionic ends, being hydrophilic, are pointed radially outward. Polar heads repel other micelle formations, explaining why surfactants stabilize dispersions, making them difficult to separate.

Reverse Micelles (Water in Oil)

Surfactant in oil systems, according to FRRPS, forms reverse micelles, by surfactant molecules that orient themselves so that their polar heads are adsorbed by the aqueous core while the non-polar relatively long chain sections are pointed radially outward. Non-polar chain sections repel other reverse micelle formations, explaining why surfactants stabilize dispersions, making them difficult to separate.

Glass/Polymeric Comparison

Surfactants disarm conventional glass filter coalescers by bonding with glass fibers allowing water molecules to flow through coalescers with hydrocarbons. Water breakthrough, FRRPS relates, occurs because surfactant molecules bond with the silenol functional group (Si-O-Si) of the glass fiber, thereby preventing water molecules from collecting on glass fibers to form larger droplets which eventually become large enough to drain from the coalescer. Consequently, surfactants shorten glass coalescers service lives resulting in frequent cartridge changeouts and increased disposal costs. Similarly, surfactants reduce efficiencies of packed bed towers and electrostatic precipitators.

Sand Filter/Polymeric Comparison

FRRPS reflects on a sand filter that could not be operated to meet product specification. Although the sand filter was being used to remove caustic from gasoline, after caustic treating for mercaptans, FRRPS' experience tells us that glass is not compatible with caustic and the measured interfacial tension of from 1 dyne/cm to 12 dynes/cm is below the recommended limit for glass coalescers. A specially formulated coalescer which is compatible with caustic is used to meet product specification.

Three Stages of Liquid/Liquid Coalescing

High efficiency liquid/liquid coalescers, FRRPS confides, require three stages to be successful, (1) filtration, (2) coalescence and (3) separation. Filtration is required to remove fine particles, such as iron oxide and iron sulfide that stabilize emulsions. Coalescence is accomplished by polymeric and fluoropolymeric material which are effective emulsion breakers in liquids measuring interfacial tension above 1 dyne/cm. Separation occurs when coalesced water droplets are repelled by a hydrophobic barrier cartridge leaving the hydrocarbon liquid as it flows through the cartridge. Water is gathered outside the cartridge.

Laboratory Test

An independent laboratory, FRRPS understands, performed a filter/coalescer/separator test on an unleaded gasoline to which 100 ppm to 30,000 ppm by volume water was added and mixed to a finely divided emulsion. Interfacial tension for the gasoline/water system was measured at 3 dynes/cm to 7 dynes/cm. Effluent concentrations of free water from all tests were less than 15 ppm by volume. Gasoline and water were mixed after being processed through separate filters. The mixture entered the inside of a cartridge coalescer that was stacked on top of a separator cartridge of the same diameter but with a plate between the two cartridges so that fluid had to pass from the inside to the outside of the coalescer to reach the separator. Both the coalescer and separator were enclosed in a vessel that contained both liquids exiting the coalescer. Water exited the coalescer not in a finely divided emulsion but as an accumulation of large droplets which, because they could not penetrate the separator's hydrophobic material settled to the bottom of the enclosure. Gasoline and water were drained from the vessel through separate piping.

Refiner Application

FRRPS knows that a filter/coalescer/separator test stand processed diesel fuel containing 120 ppm free water and registering an average haze rating of about 6 down to about 6 ppm to 11 ppm free water and an average haze rating of about 1. The existing filter/coalescer/salt tower/filter treatment had minimal haze-removal efficiency at operating temperatures above 100 F resulting in approximately 4% of drier-processed-diesel being off-specification due to temperature-related haze. Comparing filter and coalescer/separator replacement costs of a new filter/coalescer/separator installation versus historical maintenance, material and disposal costs for salt drier operation, FRRPS noticed that the new facility would be substantially less costly even without considering the 4% off-specification product.

Salt Tower/Polymeric Comparison

Refiners using salt driers are aware of operating problems, FRRPS has heard, such as (1) bridging, which results in sometimes as little as 50 % salt usage, (2) channeling, which is caused by high flow rates and poor distribution and results in large lengthwise voids within driers, (3) plugging, which occurs at lower temperatures, resulting in maintenance problems, and (4) downstream corrosion of tanks, piping and equipment, which is caused by retained water containing chlorides. Salt drier efficiency is best when operating at relatively low temperatures and at steady flow rates. However, salt driers can remove dissolved water, while coalescers may remove only free or suspended water. Coalescer benefits include (1) not adding potentially corrosive materials to fuel, (2) very high water removal efficiencies, (3) relatively low capital and operating costs and (4) ability to operate efficiency at fluctuating flow rates and temperatures. A major advantage of coalescers is operating costs. In salt driers the amount of salt consumed is directly related to the amount of water removed. It requires 10 times more salt to remove 1,000 ppm water than it does to remove 100 ppm water. Coalescer operating costs do not change even if the amount of water is increased. Beyond these benefits recent coalescer design includes (1) effective filters for removing fine particles that help stabilize emulsions and (2) polymeric medium to provide good flow distribution and overcome disarming. As a result of these benefits, stacked coalescer/separator systems costs for processing refinery fuels are estimated to be (1) $0.0063/BBL for 0.7 centistokes (100 F) gasoline, (2) $0.0097/BBL for 1.6 centistokes Jet A, (3) $0.0101/BBL for 2.3 centistokes Jet B, (4) $.0134/BBL for 3.5 centistokes Diesel (D-2), (5) $0.0139 for 3.6 centistokes No. 2 Fuel Oil and (6) $0.0281/BBL for 8.5 centistokes No. 4 Fuel Oil. MEASURING INTERFACIAL TENSION: If refiners are having operating difficulties separating an emulsion, want to predict the difficulty of separating an emulsion, or plan to design a system to separate an emulsion, FRRPS emphatically suggests purchasing test equipment necessary to measure interfacial tension. Because all refiners will deal with emulsions at some time, becoming familiar with interfacial tension measurement is essential. There are a number of satisfactory methods for measuring interfacial tension of two immiscible liquids at equilibrium, such as (1) capillary rise method, for which a capillary full of one liquid is inserted into the other liquid and the rise up the capillary of the other liquid is measured, (2) drop weight method, which consists of forming drops of one liquid within the other liquid, counting the drops formed and obtaining the weight of the drops, (3) ring-pull and (4) Wilhelmy slide methods, which require inserting a foreign object, such as a ring or plate, into the interface and measuring the equivalent force required to move the foreign object away from the liquid until it becomes detached from the liquid, (5) sessile or drop method, which consists of forming static drops of one liquid within the other liquid and then taking certain measurements of its dimensions from a permanent visual record, such as a photograph, (6) pendant drop method, which refers to a drop of liquid approximating the shape of a pendant and requires a drop of one liquid to be formed within the other liquid followed by measurements being taken at certain time intervals, and (7) rotating drop method, which is useful for very small interfacial tensions and consists of forming a drop of the less dense liquid in a cylinder of the more dense liquid and rotating at increasing speeds so that the shape of the drop is elongated. Enough experimental work has been done that there is data to correlate results from any of these tests to a common value for interfacial tension.

Technical assistance is gratefully acknowledged from Robert L. Brown, Jr., who manages marketing for Pall Process Filtration Company, a division of Pall Corporation, and can be contacted at 2200 Northern Boulevard, East Hills, New York 11548-1289 USA, (516) 484-5400 Ext. 4488 (voice), (516) 484-0364 (fax) and Thomas H. Wines, who manages liquid coalescer testing for Scientific and Laboratory Services Department of Pall Corporation, and can be contacted at 25 Harbor Park Drive, Port Washington, New York 11050 USA, (516) 484-3600 Ext. 6334 (voice), (516) 484-3628 (fax).