REALISTIC STABLE WATER-IN-OIL EMULSIONS AT OHMSETT

When crude oil is spilled on an open body of water, it is exposed to a number of natural processes that immediately set the oil on a course of transformation. Collectively, these processes are referred to as the phenomenon of weathering, which may include spreading, evaporation, dispersion in the water body, emulsification by water, dissolution, oxidation, sedimentation, and biodegradation. Crude oils are especially affected by the processes of weathering. This is, of course, due to the complex and varying blend of hydrocarbons of which crude oils are composed. Crude oils weather based on their composition and resultant physicochemical properties, and the extent to which they are exposed to weathering processes. The extent of weathering here refers to exposure to wind and waves, temperature and temperature excursions, exposure to sunlight, exposure to rain, physicochemical and biochemical properties of the body of water, and time. A frequent product of this complex set of processes is the development of water in crude oil (W/O) emulsions. This occurs when the mechanical action of wind and waves mixes water into the oil layer resulting in the liquid-liquid colloidal mixture of the two immiscible substances. These emulsions may be of varying stability. Emulsions are of particular importance in oil spill response and mitigation because physical properties of emulsified crude oil can differ greatly from the parent crude oil. Emulsification increases the volume of the mixture significantly over the simple additive volumes of the oil and the water. It also increases the viscosity and changes the rheological nature of the oil, often to the point of creating major operational challenges for recovery, pumping, and storage. Heavily emulsified crude oils, termed mousses in the spill industry, are usually quite stable, contain in excess of 50% water by volume, and may become semi-solid in nature.

Emulsified oil is generally recalcitrant with regard to chemical dispersant application as a response measure. Additionally, emulsion breaking to recover spilled oil requires time and energy, and the water that is recovered most often requires further treatment prior to permitted release back into the environment.

The actual process of emulsification, the colloidal suspension of liquid droplets of one liquid in a continuum of another liquid, is well defined under controlled conditions in various commercial operations. Examples of emulsions in the food industry are homogenized milk, butter, and mayonnaise; in the personal care and health care industries, cold cream and ointments; and in the chemical industry, the emulsion polymer synthesis process is well known in the production of coatings and latex. However, emulsification of water from natural water bodies into petroleum, especially crude oil, is less well defined and understood. What is known is that spreading and evaporation of more volatile components can alter the composition and internal transport characteristics of the resulting oil slick. Exposure to sunlight, which is enhanced through spreading and with time, initiates a complex array of photo oxidative reactions which can transform the oil (Shankar et al., 2015). These and other weathering processes may enhance or diminish the tendency for crude oil to emulsify with water. The energy for the shear mixing forces needed to disrupt the interfacial tension between water and oil is largely provided by the combination of wind and breaking waves.

As water, the dispersed phase, becomes more and more divided into smaller droplets within the oil continuum phase, the gravitational tendency for the higher density water to drop out is counteracted by viscous and interfacial forces in the oil continuum (Fingas et al., 2003). This process also relies upon re-stabilization of the oil-water interface of these smaller and smaller droplets of water. Stabilization of this interface which contributes to emulsification is a function of a complex array of constituents in the crude oil, and to a certain extent the chemistry and physical state of the water body.

It is thought that the asphaltene component of crude oil may significantly contribute to the stability of W/O emulsions (Buckley et al. 2007). Asphaltenes are an operationally defined, high molecular mass, polyaromatic, highly polar substituted spectrum of compounds found in most crude oils. Asphaltenes may aggregate at the oil-water interface where they can stabilize the interface thereby contributing to W/O emulsion formation and stability. Whether the asphaltene fraction of a crude oil may contribute to emulsification is not solely dependent upon concentration. What is important is the stability of those asphaltenes. Stability depends not only on the properties of the asphaltene fraction itself, but also on the effectiveness of the rest of the oil as solvent for asphaltenes (Long, 1981). Light oils with small amounts of asphaltenes may often support W/O emulsification better than heavier oils with a greater concentration of asphaltenes (de Boer et al., 1995). A light oil may consist largely of paraffinic material in which asphaltenes have limited solubility, while a heavier oil will most likely contain more intermediate components that tend to be good asphaltene solvents. Lower solubility contributes to asphaltene precipitation and aggregation which may lead to enhanced W/O emulsion formation. However, it is important to note that most crude oils probably contain some concentration of aggregated asphaltenes under almost any condition (Wang and Buckley, 2007). Resins in crude oil, another operationally defined heavy component, may also play a role in emulsification formation and stability. Resins are likely to include species that contribute to the overall solvent quality of the oil with respect to its asphaltenes.

The impetus for the test program that is the subject of this paper was provided through two related needs:

• The Department of the Interior’s Bureau of Safety and Environmental Enforcement (BSEE) sought to establish the capability to provide stable, reproducible, saltwater-in-crude-oil emulsions for use at the Ohmsett Facility. The scale for producing these “standardized” emulsions needed to range from highly controlled laboratory quantities; through larger, easily produced, mid- or “drum” scale production; to large outdoor on-tank, quasi-natural production. These multi-scale, reproducible emulsions would act as standard or normalized models for comparative studies at Ohmsett or other programs.

• In addition to the program described above, BSEE and the National Oceanic and Atmospheric Administration (NOAA) partnered to directly support the 2010 Deepwater Horizon ongoing “lessons learned” by providing control and validation for surface oiling characterization efforts pursued under the Natural Resource Damage Assessment (NRDA) Oil on Water (OoW) Workgroup. To advance understanding for oil spill preparedness, and in support of the needed validation, study options had been proposed for controlled as well as open water data collection experiments from multiple remote sensor technologies and platforms. The controlled portion of this effort was to be conducted in the outdoor wave tank at the Ohmsett facility. This required production of a relatively large volume (400 to 800 gallons) of W/O emulsion “grown” under quasi-natural conditions with wave energy and exposure to the elements, especially ultraviolet (UV) energy from natural sunlight.

The two needs were operationally and functionally combined. So along with providing valuable information on their own, the laboratory and mid-scale portions provided background data for the large volume, on-tank production of emulsion for the remote sensing study. To our knowledge, this was the first time such a multifaceted effort of this type and at this scale had been done. The subject of this paper is the production of reproducible W/O emulsions at Ohmsett. This program provided a “first cut” exercise in what is intended to be an ongoing and evolving effort at Ohmsett. The lessons learned, as well as suggestions for future provisions and studies will be discussed.

Given the concurrent project goals of developing Ohmsett’s capability to produce stable emulsions for testing and research, and meeting the need for “naturally” produced, tank-scale emulsion for the BSEE remote sensing program, the chosen subject oil was a crude oil blend from the Hoover Offshore Oil Pipeline System (HOOPS) in the western Gulf of Mexico (GOM). This crude, which was available in sufficient volumes in the Ohmsett inventory, satisfied the need for quantity, and GOM relevance to the large remote sensing test program. HOOPS, which is considered a pipeline oil, includes oil from the constituent fields Diana, South Diana, Hoover, Marshal, and Madison. It should be noted that as a pipeline blend, HOOPS may contain additives and may have been pretreated with dehydrating agents. HOOPS Blend is a medium-light crude with medium sulfur content. Key physicochemical specifications for HOOPS Blend are given in Table 1.

Lab-scale samples were weathered, exposed to ultraviolet (UV) photo oxidation, and emulsified to determine the best candidate treatment(s) for forming a stable emulsion at the mid- and tank-scale studies.

Lab-scale weathering involved exposing the subject crude oil to controlled air sparging (bubbling) in gas sparging tube/traps. The goal was to achieve various stages of weathering as defined by loss of mass in the volatile fraction of the crude oil.

Sparging was done at room temperature (approx. 20 °C) for the first series of samples. Two later samples were air sparged at an elevated temperature range (45 to 50 °C) by lowering the tubes into a temperature controlled water bath (shown). Air flow to each tube was controlled and maintained at one (1) liter/minute. Three initial volatile hydrocarbon mass losses (weathering) at 20 °C were targeted; 10%, 20%, and an extended weathering over approximately 6 days which yielded a mass loss of almost 24%. Sparging at 1 L/min at the elevated temperature range was conducted to a mass loss of approximately 14% for two individual samples.

It was decided to expose unweathered (fresh) and weathered samples to natural UV exposure on the roof of Building R24 (the tower) at the Ohmsett facility. This decision was based on the fact that natural exposure to the full sun-based UV spectrum, plus further exposure to natural winds would be most realistic. Accordingly, 200 ml of each of these four samples were placed in each of four 182 mm Ø, Pyrex crystallization dishes. Each dish was in turn placed in a large polypropylene tray as secondary containment.

To test and choose a small scale (laboratory) simulation of water wave action on floating oil, samples of crude oil and tank water were placed in various glass vessels on mechanical mixing devices in the laboratory. The observations resulted in the choice to float 25 ml of oil on 250 ml of tank water in one quart, glass Mason jars with lids, with mixing done on the large Ohmsett laboratory platform mixing table. The mixing regime was to run the mixing table at 160 cycles per minute (cpm) for an interval of time that was to be dictated by observing the progress of any resultant water-in-oil emulsification. The tank water used had a density of 1.023 g/ml, a salinity of 27.8 parts per thousand (ppt), a surface tension of 53.8 dynes/cm, and a pH of approximately 7.2. At the start of mechanical mixing the water temperature was 23.0 °C. After five hours of mixing at 160 cpm, empirically stable emulsions developed in each of the four samples. The resulting water/oil temperature for all for samples was 27.1 °C.

Each of the four emulsified samples were immediately placed in individual 250 ml Pyrex separatory funnels. For each sample, free water (i.e., water not associated with emulsion or inclusion) was immediately separated, and the decanted fluids were allowed to settle/separate over time. In this way, water loss over time could be used as a measure of emulsion stability.

In the field, it has been observed that well developed natural emulsions are remarkably stable and display a unique color change as the emulsification process unfolds. Stable emulsions that most closely resemble open water conditions were required for the BSEE Remote Sensing project; hence, the HOOPS was naturally weathered on the Ohmsett tank (without waves) for several days first, then the wave generator was activated to provide the necessary mixing and shear energy to incorporate water into the oil. The objective was to deploy HOOPS on the surface of the wave tank, while providing just enough surface control of the slick to assure the following: adequate exposure to waves and the elements; access to the slick for sampling and documentation; maintenance of an adequate slick foot print for observation, and remote sensing from various altitudes above the slick. At the same time, the surface control measures minimized exposure of the slick to disturbances or energy inputs other than those of waves and natural weathering. To accomplish this task, an integrated system of booming, fixed and mobile near-surface water sprayers, hand-held fire hoses, and subsurface upwelling current devices were employed.

On July 14, 2016 at 11:30 a.m., 400 gallons of fresh HOOPS crude oil were placed on the Ohmsett tank water surface. The oil was allowed to naturally weather with wind and solar radiation for almost 95 hours. The previously described surface control system was used to keep the slick spread and exposed to the elements with a minimum of disturbance. At 10:00 a.m. on July 18, 2016 waves were introduced to the tank and the slick. The wave setting used was 35 cpm, 6-inch stroke (1.7 sec, λ=12 ft., H_1/3=22 in.). Waves were turned off for several hours later during the afternoon of July 18 for sampling activities. Waves remained on over the course of the week of July 18 to allow the evolution of emulsion formation. However, wave generation operations were stopped for nominal periods of time during sampling activities and when aerial remote sensing was underway by various subcontractors on the BSEE/NOAA project. Additionally, wave generation was ceased during evening hours per the facility’s safety requirements.

Although there were a large number of sampling, monitoring, and remote sensing operations conducted by the various research and testing stakeholders during the tank-scale testing phase, this paper focuses on the analyses performed by Ohmsett in support of the emulsion production aspect of this study. Characterization of the tank-scale crude oil and W/O emulsion slick, and tracking the progress of emulsification over several days presented several challenges. Laboratory capabilities, space, and qualified personnel were limited. Sampling of a large-scale emulsion slick “on the move” is difficult and sample location choice is subjective. W/O emulsions during the early and mid-stages of emulsification are inherently unstable once removed from their energetic water environment. Often the mere act of sampling changes the nature of the emulsion itself. Finally, it was required that Ohmsett personnel and resources support the larger multifaceted goals of the remote sensing program. Therefore, Ohmsett sampling and analyses of the W/O emulsion needed to interfere with overall operations as little as possible. A simple program of sampling and analysis was devised to meet these challenges.

To facilitate representative sampling of the forming emulsion, a two liter separatory funnel was adapted to field (tank-side) use. Then, several visually representative samples were withdrawn from several adjacent locations at several depths using a sampling thief, and with careful ladling. These individual samples were combined on the spot in the separatory funnel where the non-emulsion associated free water was immediately decanted. Once this field separation was affected, the sample could be transported to the lab for a toluene extracted, centrifugal determination of the water content of the collected emulsion.

The density of the composite sample as collected above was determined immediately using a hand-held, ultrasonic densitometer. In the cases where the viscosity of the emulsion did not allow the fluid to be drawn into the capillary of the densitometer, the sample was quickly transported to the laboratory for measurement using a pycnometer. Where the sample exhibited a lack of homogeneity, a “best visual judgement” determination was used to provide representative samples for the pycnometer density analysis. For all density determinations, measurements were made at prevailing ambient laboratory or air conditioned bridge-house temperatures of approximately 20 °C.

Viscosity determinations were made of the composite emulsion samples using a temperature controlled, rotary viscometer with a cup and spindle rated for the anticipated viscosity range. All viscosities were determined at 20 °C. Again, the “best visual judgement” technique was used to choose a representative sample for viscosity determination in the cases where there was a distinct lack of homogeneity.

Regular water sampling (at least daily during operational days) was used to provide data for density, salinity, interfacial tension, and surface tension.

As expected, loss of mass due to enhanced evaporation (sparging) proceeded with time. At approximately 20 °C, reduction of the HOOPS volatile components in the realm of 10% by weight was achieved in alignment with the curve in Figure.

At room temperature, and under the condition of air sparging at 1 L/min, mass loss weathering proceeds to approximately 10% in a little over four hours. This early weathering proceeds along a curve described by a second order polynomial, and typifies light to medium crude oil weathering in the earlier stages of loss of volatile components.

Figure 2 shows the volatile mass loss with sparging at 1 L/min at 20 °C for the extended period of approximately 148 hours. At this point in time loss of mass is leveling off at 23.6%, and the curve is well represented by an exponential relationship. Generally, this is more indicative of the overall weathering, under controlled conditions, of light to medium crude oils.

Air sparging at the same flow rate (1 L/min) at the elevated temperature range of 45 to 50 °C produced similar behavior albeit compressed in time. The two samples of fresh HOOPS weathered in this way produced an approximate 14% loss of mass over approximately 6 hours.

As expected, other physical characteristics changed with air sparge induced weathering. As measured in the lab, these included density (ρ), dynamic viscosity (μ), surface tension (ST) versus air, and interfacial tension (IFT) versus tank (salt) water. The data from these determinations all demonstrate increases in the measured physical characteristics, as seen in Table 2.

As described earlier, each of the weathered samples including fresh (unweathered) HOOPS were exposed to natural sunlight and the elements on the roof of the tower building (Bldg. R24) at the Ohmsett Facility. The samples were exposed to sunlight a total of approximately 20 hours over a five-day period. Exposure was constrained to the facility hours of operation. The samples were stored indoors after-hours and during the evenings. It should be noted that this time-period included a long holiday weekend (Memorial Day) during which the samples were stored indoors. Although periodic density measurements were recorded, it became obvious that a trend analysis for these or any other physical characteristics would be uninformative. This was due to the lack of consistency and control of the time intervals of exposure, and the highly variable conditions of exposure such as and especially temperature. Presentation of starting and ending data is detailed in Table 3.

Emulsification of the weathered, sunlight-exposed samples to five hours of agitation with saltwater in mason jars produced samples that were placed in separatory funnels to demonstrate emulsion stability. A combination of visual observation and measurement of the volume of water separating from the emulsion over time were used to determine comparative stability.

Tables 4–6 list the results after settling times of 1:15 hours, 17:00 hours, and 24:45 hours.

A major objective of the lab-scale tests was to determine if HOOPS could form stable emulsions, and if so, which weathering and UV exposed treatment scheme would be the best candidate to form emulsions on the Ohmsett tank. The stability testing seemed to indicate that the air-sparged weathered samples generally retained the most water and stability early (see 1:15 hour results). However, they all seemed to destabilize over time. On the other hand, these tests indicated that fresh, unweathered HOOPS oil that was exposed to sunlight in the same manor that the weathered samples were, released the least amount of water, measured the highest density, and had enough uniformity to obtain a viscosity measurement.

To reiterate, the two overall objectives of this program were to establish the capability to provide stable, reproducible, saltwater in crude oil emulsions for use at the Ohmsett Facility; and, to support the BSEE remote sensing program by offering a large, open-air, W/O emulsion on saltwater venue for realistic observation and data collection by multiple remote sensing platforms. Accordingly, the lab- and mid-scale studies provided supportive data and operational experience in achieving the large-scale operational objectives. A general description of the execution of that operation has been described previously. Tabulated data for periodic sampling and analysis of the developing emulsion and background water analyses are provided below.

It should be noted that to accommodate the needs of the larger remote sensing program, and the operational limitations imposed by the need to cease wave generation during evening hours, adherence to a strict sequential time-line was not possible. However, the data, in conjunction with documentation of daily operations, does provide for some informative observations.

From 11:30 a.m. on July 14, the time of initial deployment of 400 gallons of HOOPS crude oil onto the surface of the tank, until 10:15 a.m. on July 15 (total of 22:45 hours), the oil slick was allowed to move about on the surface exposed to the natural day and night elements without applying wave generation. During daylight operational hours, the slick was kept spread out as much as possible to provide surface area for natural exposure to sunlight. This was accomplished using the system of sprayers, booms, and current generators described earlier. Sampling of the slick during that wave-free time period indicates a constant increase in density and viscosity as would be expected of loss of mass due to weathering. Minimal emulsification or color change was observed during this period of no wave energy, other than surface ripples and capillary waves.

Generated breaking waves, as descried earlier, were applied starting at 10:15 a.m. hours on July 18. For the next three days and eight hours, the same wave field was applied periodically. Waves were interrupted each evening, for a short time during sampling operations, and occasionally in response to the needs of the remote sensing program. However, during this time period significant emulsification was occurring. This was evidenced by the change in slick color and consistency, and the significant increases in density, viscosity, and water content. The observed variability in these physical quantities seems to correlate well with those documented times of no wave generation. Overall, density increase to over 1.0 g/ml, viscosity increase over 40 times to approximately 16,000 cP, and water content from nil to 84% by volume were observed. During times of no applied waves, quick determinations for water content indicated that the water volume of the emulsified oil could drop to less than 60%. This was empirically evidenced by the change in emulsion appearance to less “thick” and darker in color than when energetic waves were applied. It was also observed that, in time, the emulsified slick became streaked with different colors and consistencies. All of this is indicative of an emulsion that is not fully stable without continued input of wave energy, and is in a transient stage of emulsification. Tables 7 and 8 list the test fluid properties measured during the large scale emulsion formation study.

Figures 3 through 6 on the following page show the development of the W/O emulsions over time and lack of heterogeneity throughout the slick. Figure 3 shows the oil after four days of weathering on the tank just after waves were first applied to the slick. Figure 4 illustrates the water sprayers employed to manage the slick just as waves were applied early in the morning. Note the breakdown of the emulsions overnight without wave energy. Figures 5 and 6 capture the streaky and brainy appearance as the emulsions evolved throughout the week.

The intent of this study was to develop a method for the Ohmsett facility to create large quantities of stable, reproducible water-in-oil emulsions without tying up the large Ohmsett tank for weeks at a time. Concurrent to this study was the BSEE/NOAA remote sensing project which required large quantities of realistic emulsions and a fly-over friendly venue; hence, results from that effort were incorporated into the overall lessons learned. At both test scales, increases in viscosity and density were consistent with weathering prior to emulsification. Mass loss from evaporation in time eventually approached an exponential relationship with an asymptotic leveling off consistent with the literature for most crude oils (there is no mention of weathering or emulsification of HOOPS in the literature). Evaporative weathering at higher temperature shifted this relationship accordingly.

Once emulsification began, the general trend did indicate higher values of density and viscosity that followed the overall trend of increased water content in the emulsion. The on-tank emulsion density data differed slightly from what the analyzed water content would predict on a calculated mass and volume basis. In the early phases of emulsification, resultant densities were less than the calculations based on constituents would predict, albeit only by a few percent. Later in the on-tank emulsification process, that trend reversed with measured densities being slightly higher than calculated values would predict. This could be due to the interaction of water and the crude oil components that may produce an expansion or contraction of the composite material not explainable by volumetric considerations alone. However, the differences are within the error of scale and sampling, and controlled, small-scale studies would need to be performed to elucidate such a phenomenon. Accordingly, emulsification proved quite fleeting and obtaining representative samples across the tank within the emulsion thicknesses themselves proved problematic. Related was the reality that the stability of the emulsions changed during transport of samples from the tank to the nearby on-site lab. Later viscosity measurements were not possible given the instability of these sample emulsions. Moving forward, investigation of existing in-situ instrumentation to measure water content, density, and particle size distribution of water droplets in the emulsion could prove valuable. The practice of production and study of emulsions by other industries, such as cosmetics and coatings, could be useful in this concern.

The stability of emulsions formed in the environment are based on the complex interaction of a number of constituent and energetic factors over time. In the present study, it was observed that the emulsion slick would persist as long as waves continued to supply shear energy. However, once waves stopped, the emulsified slick would break down non-homogeneously. Water content would readily drop from a range of 70 to 80% by volume to under 50%. Streaks of what appeared to be more “oil like” material were observed in the slick. The low asphaltene content as well as high ambient temperatures during the tank-scale tests in mid-July most likely led to the instability of the emulsions. Asphaltene stability through interaction with other constituents may also have contributed.

This study focused on the creation of water-in-oil emulsions using HOOPS oil, and all oils are different; therefore, strict standardization is problematic. However, valuable insight into necessary weathering, photo oxidation, and mixing energy was gained to move forward in order to create reproducible, stable emulsions for testing at Ohmsett. The final phase of this study will focus on the creation of mass quantities of stable, reproducible W/O emulsions. It is not economically feasible to tie up the Ohmsett tank for weeks to create stable emulsions. The use of commercial mixers, surfactants, and the role of photo oxidation will be further pursued.