United States Environmental Protection Agency Municipal Environmental Researctr Laboratory Cincinnati OH 45268 Research and Development EPA-600/S2-82-032 August 1982 Project Summary Apparatus and Procedure for Determining Oil Droplet Size Distribution Raymond A. Meyer, Milton Kirsch, Fred Howard, and Frank Freestone This study was undertaken to char- acterize the oily brine resulting from the production of oil and to develop an apparatus and a procedure for ac- curately determining oil droplet size distribution. Knowing that the size range and distribution of oil droplets is a major governing factor in the successful treatment of oily brine, the end result of this work may be to reduce the hydrocarbon discharge to the environment. An oil-specific, automated photo- micrographic system was developed to fulfill the need for determining size distribution on production platforms where explosive concentrations of hydrocarbons may exist. This system can be used to measure the diameter of particles in the 2- to 100-jum micro- meter range under flowing conditions without introducing significant shear forces, which can affect the oil-drop population. The system uses a newly developed technique that applies time-lapse photography to the determination of number, size, and density of spherical entities. Specifically, a microscope with a horizontal orientation of the viewing axis is used to photograph the movement of oil droplets through a flow-through cell. This Project Summary was devel- oped by EPA's Municipal Environ- mental Research Laboratory, Cincinnati. OH, to announce key findings of the research project that is fully docu- mented in a separate report of the same title (see Project Report ordering information at back). Introduction to the Problem Production of oil-brine mixtures, pumping, and pipeline flow all result in a dispersal of oil in the associated brine. In an effort to reduce the hydrocarbon discharge to the environment, the Offshore Operators Committee and the U.S. Environmental Protection Agency's Municipal Environmental Research Laboratory cooperated in an oil produc- tion platform study. The goals of the study included characterizing the oily brine at several points in the oil removal treatment process, evaluating the effectiveness of several treatment techniques, and comparing analytical methods. Oil-brine separation methods ulti- mately depend upon the oil drops rising through the brine to a collection area. This rise rate is proportional to the density difference between the oil drops and the brine, and is also proportional to the square of the diameter of the drop. Thus, a major governing factor in the success of brine treatment for oil removal is the size range and distribu- tion of oil drops. Knowing the oil drop size dispersion at several places in the produced water treatment system would aid in applying present separa- tion techniques and developing future systems. Therefore, a part of the platform study wgs directed toward ------- measurement of the drop size distribu- tion in the primary oil-water separator feed, the final treatment unit feed, and the final discharge. Emphasis was on offshore production, but the develop- ment is equally applicable to onshore production. A number of nonspecific techniques are used to characterize particle size distribution. All are based on measuring the particle's effect in the interruption of the flow of some energy such as light, radiation, or electricity. These methods are nonspecific and cannot distinguish between gas bubbles, solid particles, and oil drops. Such methods are inadequate because the produced brine contains such particulate material as sand, shells from microorganisms, gas bubbles from depressurized dissolved gas, and gas bubbles intentionally introduced to assist the oil removal flotation process. Therefore, any of the nonspecific techniques would be ex- pected to give erroneously high oil contents and misleading oil drop size distributions. Additionally, to obtain meaningful size measurement, only one particle may be in the measuring path at a time. This is typically achieved by severely limiting the path length. While this is quite acceptable when dealing with solids, the high shear introduced by passing the sample through the small orifice required would be expected to alter the oil drop size distribution and render the data meaningless. For these reasons, the existing techniques were considered unsuitable for determining oil drop size dispersions. The Solution A micrographic technique may be implemented that does not introduce significant shear forces on the oil drop population. The technique was devel- oped into an automated photomicro- graphic system (PMS) that met the requirements of the National Electrical Code Class 1, Division 1, Group D. This permitted its use on production plat- forms where explosive concentrations of hydrocarbons were known to exist. The system can be used to measure the diameters (and densities) of particles in the 2- to 100-Aim-diameter range under flowing conditions. The battery-oper- ated device is 63-cm long, 55-cm wide, and 55-cm high, and weighs 16 kg. Conventional micrography involves capture of a sample, placing it on a slide, perhaps in a shallow well, and counting or measuring the entities of interest. This typically is done at leisure since the sample is stable over a relatively long time. Such is not the case when studying oil drop dispersion. As soon as the turbulent mixing motion dissipates, the sample starts to stratify because of the density disparity between the oil drops and the brine matrix. This dispersion alteration proceeds at such a rapid rate that the sample would be useless within 20 seconds. A flowthrough microscope cell was constructed, and the microscope illumi- nation changed from the conventional continuous light source to electronic flash Illumination. When the sample was flowing fast enough to maintain turbulent mixing, however, the linear drop movement was too rapid to give sharp photographs. Accordingly, an interrupted flow system was designed in which the flow was blocked by a downstream valve and the photographs taken a few seconds later. The spherical entities were initially thought to be nothing but oil drops. However, the oil content of a Wemco* flotation unit (final oil removal unit) outlet sample, calculated from mea- sured drop diameters and count, was found to be much larger than that determined by conventional solvent extraction techniques. The rationale was offered that not all the photo- graphed, measured, and counted spherical entities were oil. This seemed logical since the function of the Wemco treating unit was to mix gas bubbles into the water in an effort to "parachute" the oil drops to a surface skimmer. These bubbles could well be covered with a film of oil and be photographically similar to actual oil drops. This compli- cation led to the development of a technique to apply time-lapse micro- photography to determine number, size, and, most importantly, the density of the spherical entities. In the normal, vertical orientation of the microscope viewing axis, the oil drops rise toward the top of the cell and thus move in and out of focus (Figure 1). If the axis of the microscope is turned horizontal, however, the drop move- ment would be across the field of view of the microscope (Figure 2). Thus, the drop would remain in focus and only change position in the field of view as it rises. The vertical movement of the oil drops and air bubbles could be mea- Figure 1. Drop movement with normal vertical microscope axis orientation. 'Mention of trade names or commercial products does not constitute endorsement or recommenda- tion for use. Figure 2. Drop movement with non- typical horizontal microscope axis orientation. sured by comparing their positions in two photographs taken a known-time interval apart. If the rise rates and diameters were known, the densities of the spherical entities could be deter- mined by applying Stokes Law. Figure 3 shows a line diagram of the system as viewed from above while in its operating position. The horizontal orientation of the microscope viewing axis and its relation to the flowthrough cell and film plane are shown. Both the camera focusing magnifier and the microscope oculars are designed to be used from the side when in the operat- ing position. The camera was positioned ------- ^. 0 _^ o o o 35mm film f- , 2,5 X magnifier it. I Binocular eyepieces // \ ^ x Discharge Camera Microscope Cell Condenser lens Strobe and reflector assembly Standpipe _^>_Lr- Jg> «~ ^_J Sample Figure 3. Top diagrammatic view of photomicrographic system. with the long, 34-mm axis of the film vertical. An electronic control circuitry was developed to sequence the sample flow and the three time-lapse photo- graphs. Figures 4, 5, and 6 are photo- graphs of the apparatus at its present stage of development. Sampling If the pipeline and sample flow are maintained in the turbulent region (Reynolds numbers over 3000), the sample may be obtained and transferred to the analytical apparatus without altering the oil drop size dispersion. Sample flowrate.is a function of liquid pressure in the pipeline, which typically ranged between 2.8 kPa (30 cm water) and 28 kPa (300 cm water). The additional requirement — that the sample presented to the microscope cell should have been in turbulent flow since its removal from the pipeline — dictated some form of bypass sampling. The overflow standpipe system fulfilled both the requirements of pressure regulation and of bypass sampling (Figure 7). This particular apparatus used a 21.5-cm-tall inner pipe and has been operated with pipeline pressures ranging from 2.8 kPa (30 cm water) to 44.8 kPa (457 cm water). Minor size modifications could extend this range. Normal operating conditions resulted in a flow of 2075 mL/min during the microscope cell flushing period. The lowest Reynolds number in the sample flow system was 4000 in the cell body, and over 7000 elsewhere in the sample system. Attention must be directed toward elimination of any shear- inducing restrictions such as partly opened valves. If these precautions are taken, the microscope cell may be' assumed to be filled with a completely representative and uniformly dispersed sample of the pipeline flow. Stopped Flow Period Previous applications of the photo- micrographic principle for measuring oil drop size dispersion relied upon the ability to photograph moving drops in a flowing stream. If one applies the requirement that the stream must be in continuous turbulent movement to eliminate sample stratification, the exposures must be extremely short. For example, the linear transit rate of a drop in the microscope cell at a Reynolds number of 4000 is 4.3 x 105 jum/sec. Common shutters of 1/100 of a second capacity would result in an image of a 1- //m-diameter drop that would be 430/um long. Photography with an electronic flash lamp having a 50-fjsec duration would give an image 21 -/um long. Even if the 5-fjsec "Strobotac" source were used, the image would still be twice as long as it was wide. We have estab- lished 0.1 /um as the desired limit of movement during photography. This would impose an exposure duration of 0.2 /usec, which is beyond the range of available portable illumination sources. Accordingly, a stopped flow system was designed that would not induce sample degradation. The previous discussion of sample flow defends the assumption that the sample cell is filled with an uncompromised and uniformly dispersed sample at the time turbulent flow is interrupted. Drop motion at this time results from inertia and the random movement of turbulence. The sample must remain static until movement from ------- Standpipe pressure reducer • Overflow line Sample ,, inlet line 250-exposure camera Figure 4. Photomicrographic apparatus, left side view. Control switches \Cell flow outlet Case •essurization connection Figure 5. Photomicrographic apparatus, right side view. 4 both of these sources dissipates and vertical movement from the density difference is established between the drop and the liquid matrix. It will be proven that a 10-sec static period does not compromise the sample. Microscope Viewing Cell The microscope viewing cell is the liquid volume in focus by the micro- scope optics. Its size is defined by the length and width of the film, the diameter of the drop (D), and the magnification and depth of focus of the objective. Under the conditions used in this work, the cross-sectional cell dimensions are 535 + D x 349 + D A*m. The apparent depth of focus was found to be dependent on the drop diameter. As used in the field studies, a slide of oil drops captured in gelatin was photo- graphed with color film and electronic flash. The microscope stage was moved in 4-/vm steps over a wide range with photographs taken at each step. Twelve drops were selected ranging in size from 2 to115;umindiameterandtherangeof stage positions resulting in sharp image photographs wasdetermined by inspec- tion of the photographs. The data were fitted to various equations with a Hewlett Packard statistics program. The following equation shows the best fit and may only be applied under the exact conditions used in its determi- nation. Depth of focus (fjm) = 3.861 + (5.088 Ln D) where D is the drop diarneter in //m Black and white film, for example, gave significantly different results. The determination of "in focus" is very subjective, and the data analyst must be well trained by inspection of the calibration photographs. Retraining by viewing the calibration photographs should be performed at periodic inter- vals to eliminate "subjective drift." There is a marked tendency to "find" drops in sparsely populated exposures that lead to high drop counts and oil contents. The retraining minimized this effect. Once the concept of the viewing cell as a boundaryless volume of liquid sample located someplace within the 25,000 x 3175-Aim microscope cell is established, the preferred location may be selected. The objective side of the viewing cell is typically located 600 jum into the liquid to minimize the effect of the cell wall on drop rise. The top of the viewing cell is a nominal 1500 jum ------- Microscope 13! Salenoid valve Figure 6. Photomicrographic apparatus, internal detail photograph. below the top of the nicroscope cell to avoid optical distortion from the cell curvature. If one discounts a 1000-//m section at the bottom of the cell, the drops may be said to have a conserva- tive rise path of 22,000/urn to the bottom of the 535-/um-high viewing cell. Representative Sample The assumption has already been proven that the sampling procedures have placed an uncompromised and uniformly dispersed sample within the microscope cell when flow is inter- rupted. It remains to be shown that photographs taken 10, 10.3, and 12.0 sec later capture a representative sample of the oil drops in the water. Figure 8 shows calculated rise distances in 10.3 sec for oil drops in various matrices. Platform ST177 conditions resulted in the most rapid rise and therefore were used to prepare Table 1. According to the 10-sec column in Table 1, a 2-/urn-diameter oil drop rises 7 fjm, a 100-/um drop rises 10,488, and a ^2S-(Jm drop rises 24,762. As previ- ously indicated, the free vertical rise section of the cell to the bottom of the microscope viewing cell is a conserva- tive 22,000 /urn. Consider a zero time photograph where a 2-, a 10-, and a 100-jum drop are just inside the lower edge of the film image. The water sample had been stationary for 10 sec before the photo- graph was taken and all oil drops had an opportunity to rise at their diameter and density determined rate for the 10 sec. Thus, the2-/um drop had risen 7yum; the 1 Q-fim drop, 165; a nd the 100-yum drop, 16,488. Another way to say the same thing is that the 2-/urn drop originated in a microvolume 7 //m below the viewing cell; the 10-/um drop, 165; and the 100- fjm drop, 16,488. Over- flow Sample Figure 7. Standpipe pressure reducer. Considering the previously estab- lished fact that all drops are uniformly dispersed, all microvolumes have equal chances of containing any drop size. It therefore makes no difference if the 2- jum drops originate in one microvolume and the 100-/vm drops in another. All microvolumes are equivalent and the sample volume photographed will still be representative of the entire sample. This condition holds until the drop-rise distance during the static period ex- ceeds the available path within the cell. This was defined as 22,000 fjm and a 115.5-/um-diameter oil drop will then be the cutoff point, since it will rise 21,995 fjm in 10 sec. Any larger drops may exceed the rise path, and even if they originated in the lowest acceptable microvolume, they may have escaped the viewing cell when the photograph was taken. System limitations, there- fore, indicate a conservative cutoff point of 100-/um diameter, and the above discussion shows that the viewing cell contains a representative sample at the end of 10 sec. The rise rate is dependent upon water density and viscosity and oil density. Thus, the actual cutoff point will change somewhat because of sample conditions. ------- / 2 345 50,000-1 / Platform ST177 2 Platform SP65B 3 Sea water and 0.8412 density oil 4 Platform WD45C 5 Fresh water and 0.8412-density oil 0 i I i i i i I i r i i i i i i i i i i i i i i i i i i i i i i i i i i i i 25 50 75 100 125 150 175 200 Drop diameter, micrometers Figure 8. Drop position after 10.3 seconds. Table 1. Platform S T177 Drop Movement* Drop diam., micrometers 2 10 50 100 125 Rise rate. micrometer/ sec 0.7 16 412 1,649 2,576 Micrometer rise, seconds 10 7 165 4.122 16.488 25,762 10.3 7 170 4.246 16.983 26.535 11.7 8 193 4,823 19.291 30. 142 *Brine density = 1.151; Oil density = 0.8418; Viscosity = 0.01021 Stokes Oil drop movement over a 12-sec period is diagrammed in Figure 9. Column A diagrams three equal-sized cells. Cell 1 is at the top of the microscope cell; cell 2 is near the center of the microscope cell; and cell 3 near the bottom. Each cell contains the same drop size distribution since the drops are uniformly dispersed throughout the cell. The drops carry the number of their eel I of origin to permit following their movements. Column B shows the position of the drops 10 sec later at the time of the first exposure of the time- lapse series. The large drops from cells 1 and 2 have risen out of view, but the large drop from cell 3 has replaced them in the top viewing cell. The medium- sized drop from the viewing cell has escaped from view, but has been replaced by an identical drop from cell 2, the next lower one. The small drop's movements are minimal and each remains in its cell of origin. Columns C and D diagram the drop position after 10.3 and 12 sec. After 10.3 sec, the large drop from cell 3 is still in view but is about to escape. This drop has, however, been photographed, and its diameter and rate of rise measured and used to calculate density. Twelve sec after flow interruption, the largest drop from cell 3 has escaped, the medium-sized drop originating in cell 2 and the small drop originating in cell 1 have moved measurable distances. Thus, although a drop may move out of the viewing cell, it is replaced by a drop of the same size from a lower volume, and the photographs capture a repre- sentative sample of the oil drops in the microscope cell. This will occur until sufficient time has passed (12 sec here) to allow the largest drop of interest to escape the viewing cell, even though it was originally at the bottom of the microscope cell. The preceding discussion of drop size cutoff applies only to the capture of a drop in a single photograph. Density measurement requires capturing the same drop in two photographs of the photo-triad to provide for measurement of rise rate. This requirement seriously decreases the effective vertical height of the liquid viewing cell because: 1. The camera motor drive limits successive photographs to 0.3-sec intervals. 2. During this time a 100-/um drop rises 495 /urn. 3. Cell height is only 40 m greater than the rise distance of the largest drop during 0.3 sec. Thus, a 100-^m drop has to be in the bottom 40 /urn of photograph 1 to be at the top of photograph 2, taken 0.3 sec later. Therefore, the effective viewing cell height for a 100-/ym drop is only 40 /urn. When a 50-yum drop is under consideration, the effective cell height is 411 //m. Thus, the dynamics of drop movement as well as photographic aperture combine to fix the volume of the viewing cell during density deter- mination. Understanding this factor is vital to the calculation of oil content based on the drop volume and the viewing cell volume. There are many possible combina- tions of vertical position of the viewing cell within the microscope cell, static period before photography, and time between photographs that wil) result in capturing divergent size and density entities. For example, if the viewing cell were positioned near the bottom of the microscope cell, the system would be ------- No. 1 Near top of microscope viewing cell. © © No. 2 Near center of cell. No 3. Near bottom of cell. © © © © A 0 Time B 10 sec. C 10.3 sec. D 12 sec. Figure 9. Drop movement during 12 seconds. optimized for entities heavier tnan brine. The photographic timing sequence has been optimized for oil drops. If the density measurement feature is elimi- nated, however, a photograph can be taken within 0.1 sec of flow interruption and capture almost all entities in a single photograph. This is a switchable option in the electronic circuitry. Such adaptations must be at the well-con- sidered discretion of the user. The full report was submitted in partial fulfillment of Contract No. 68- 03-2648 by Rockwell International under the sponsorship of the U.S. Environmental Protection Agency. •US GPO:1M2-559-M2-457 ------- Raymond A. Meyer and Milton Kirsch are with Rockwell International, Environ- mental Monitoring & Services Center, Newbury Park, CA91320; Fred Howard is with Esoteric Systems, Inc., Thousand Oaks, CA 91360; and the EPA author Frank Freestone is with the Municipal Environmental Research Laboratory- Cincinnati, Edison, NJ 08837. John S. Farlow is the EPA Project Officer (see below). The complete report, entitled "Apparatus and Procedure for Determining Oil Droplet Size Distribution, "(Order No. PB 82-231317; Cost: $ 12.00, subject to change) will be available only from: National Technical Information Service 5285 Port Royal Road Springfield, VA 22161 Telephone: 703-487-4650 The EPA Project Officer can be contacted at: Oil and Hazardous Materials Spills Branch Municipal Environmental Research Laboratory-Cincinnati U.S. Environmental Protection Agency Edison, NJ 08837 United States Environmental Protection Agency Center for Environmental Research Information Cincinnati OH 45268 Postage and Fees Paid Environmental Protection Agency EPA 335 Official Business Penalty for Private Use $300 RETURN POSTAGE GUARANTEED PS U00032V •*« v;r- OX DEARBORN 60 1L 60604 ------- |