Comparing oil-in-water measurement: varying government regulations and measurement methods call for standardization
Colin C. Tyrie
Oil and grease in produced water is not a chemical substance. It is defined by the method specified to measure it. These methods vary from one area to another. For example, in the US Environmental Protection Agency Method 1664 defines it. In the North Sea, an infrared spectrometer defines it. But there, it was not called oil and grease but total oil or total hydrocarbons. There are several instrumental methods for measuring oil in produced water. None of them measure all the organic compounds in the water. Comparing what the commercially available methods actually measure will illustrate the problem of interpreting oil in water (OIW) analysis.
Oil concentrations in water are reported as a mass or volume unit in a given volume of water, either milligrams per liter (mg/l) or microliters per liter (ml/l). Each analytical method measures some property of oil that can be related to this mass or volume value. The problem is that the composition of oil in produced water varies for a number of reasons:
* Changes of source due to opening and closing wells
* Level of separation treatment
* Use of treating chemicals.
Authorities define oil for regulatory purposes. Method 1664 measures the mass per unit volume of oil directly and is not affected by the variables listed above. All other methods must relate their measurements to oil mass or volume. How each instrument or method correlates oil concentration to its measurement makes interpreting and comparing results difficult.
OSPAR defined oil for the North Sea by specifying the measurement procedure. The original procedure used infrared absorption of a water extract.
OIW measurement seems such a simple process that there should be no problem in understanding it. However, in practice, there is a significant problem interpreting and comparing results from various methods and instruments. This article looks in detail at measurement factors down to very simple concepts and attempts to make the process clear.
Organic compounds in produced water exist as dispersed droplets or dissolved compounds. Droplets are crude that was dispersed in the water from production operations and contain the constituents in the produced oil. In addition, dispersed droplets of production or workover chemicals may also be present.
The dissolved organic compounds include oxygenated hydrocarbons, such as carboxylic acids, low-molecular weight aromatic hydrocarbons, and aromatic acids. In sour production, sulfur compounds such mercaptans or thio alcohols may be present. Production treating chemicals are also present as soluble compounds. Sometimes the aromatic carboxylates are the major portion of the soluble oil in produced water. Hence, all the produced water organic constituents can be measured as oil. However, if they are not measured by the defining method, they are not legally oil.
There are five properties used to measure oil in produced water, four can be applied in the field and one in the laboratory. These are:
1. Direct weight measurement
3. Infrared absorption
4. Ultraviolet fluorescence
5. Particle counting.
Each method can be applied in several ways. Each has advantages and limitations that must be taken into account.
Direct Weight. Gravimetric methods measure oil directly, but the oil they measure does not include all the organic compounds in the water being tested. Method 1664 is a good example of a direct measurement method.
The procedure for applying Method 1664 is to acidify a water sample to pH 2 or less and then extract it with n-hexane. The hexane is then evaporated and the residue is weighed. The weight of this residue is divided by the volume of the water sample to arrive at the concentration in weight per unit volume units (mg/l). Using this method, only a subset of the organic constituents is measured, comprising compounds that are extractible from water in n-hexane at pH 2 and remain after the hexane is evaporated. Materials not soluble in hexane are not measured and are, therefore, not oil. Materials soluble in hexane that have boiling points below n-hexane are not measured and are not oil.
This method is cumbersome and requires a high level of manual skill to use. It cannot be used in most field environments. Because it is the required method for compliance monitoring in the US, it is important that other field monitoring methods be correlated to it.
Colorimetric. The colorimetric method was widely used before OIW was measured for regulatory compliance. It is usually used on very dark oils. The absorption of energy in the visible light range is used as the detection process. Since most light oils have very little color, it cannot be used to measure such oils in water. However, in areas where very dark, usually asphaltic oils are produced, it is still used.
The instrument is calibrated with a standard of known concentration and this calibration is used to convert the absorption reading to an oil concentration. The calibration is only good as long as the ratio of the components measured to the total oil remains constant or within acceptable limits.
Infrared absorption. Several types of chemical bonds absorb infrared (IR) energy. The instruments using IR absorption target the carbon hydrogen (C-H) bond, which is common to all organic compounds. This bond absorbs IR energy at the wavelength 3.41 microns. Since water absorbs infrared energy at the wavelengths used to measure oil, measurements must be made on an extract of the water using a solvent that does not absorb IR radiation.
Many organic compounds contain bonds other than C-H bonds, and many of the compounds in soluble oil are carboxylic acids. Each acid group contains two oxygen atoms. These molecules weigh much more per C-H bond than hydrocarbons. If the ratio of oxygen-containing molecules to total weight changes, oil analysis errors will result.
Once a sample is extracted, the oil measurement can be achieved either on the extract or on the oil residue left after evaporation. This extract contains all the organic material in the water that will extract into the solvent, including both oil droplets and some water soluble organic compounds, if the pH is low enough.
In IR instruments (whether measuring extract or residue), the oil reported is not the oil measured. Each instrument is calibrated with a standard sample of known concentration and the instrument is adjusted to read that concentration. Since the instruments actually measure IR energy absorption, the absorption measured must be assigned to a concentration.
Measurements made on duplicate samples by each instrument type would yield the same concentration. However, the absorption value for the instrument measuring oil in extracts would have a higher absorption value than the one measuring oil in a residue. The calibration is only good if the ratio of the components actually detected to the total component weight remains constant or within acceptable limits.
Ultraviolet fluorescence. Aromatic compounds absorb ultraviolet (UV) radiation and emit it (fluoresce) at another wavelength. Almost all produced waters contain aromatic compounds. The amount of fluoresced light is proportional to the concentration of aromatic compounds. Therefore, the amount of fluorescence measured is proportional to the OIW sample. Assuming the composition of the sample remains constant, a relationship can be developed between fluorescence and oil concentration.
Since UV radiation is not absorbed by water, oil determinations can be made directly on a water sample or by solvent extraction of the water. Both measurements are common. Most instruments measuring OIW directly are used as continuous monitors to control treatment processes.
There are advantages and disadvantages to both UV instrument types. There are other components of produced water that fluoresce. Iron is a good example. Iron is not present in the solvent extractions that recover oil from produced water, so the instrument using extractions has an advantage. This instrument is also easier to calibrate. Conversely, there is an advantage in process control to having continuous OIW measurements on a process stream. Interference is not usually a problem and, if it exists, it might be corrected by adjusting the instrument calibration. As with other instruments, measuring a specific property of oil and relating that to concentration, the calibration is only good as long as the ratio of the florescence measured to the weight of oil in the sample remains constant.
Particle counting. There are three variations for measuring or estimating oil or solid particle concentration by counting particle numbers and sizes:
1. Turbidity measurement
2. Coulter counter, measuring batch samples in a lab or clean area
3. Visual recording of particles, their sizes and characteristics online.
One of the earliest ways to monitor water quality was to measure turbidity. For many years operators have attempted to measure the number of discrete particles and their sizes in produced water. When fine particles are dispersed in water, the water becomes cloudy due to the scattering of transmitted light. Prescribing an upper turbidity limit was once specified to control water quality for injection. However, turbidity depends on particle size or particle concentration, and is a very approximate measure of both. The true effect of turbidity on water quality was often determined by experience.
The Coulter counter method estimates the number and size of particles by passing the water through a small circular orifice of know dimensions. An electrical current is generated through this orifice. As the individual particles passed through the orifice, they blocked part of its area, reducing the current flowing through the orifice in proportion to the particle size. This technology is limited because it is delicate and must be done in a laboratory. It also does not distinguish between solid particles and oil droplets. It usefulness as an OIW monitor is very limited.
A microscopic video camera now makes it possible to visually record particles in a water stream and identify their type (solid, oil, or gas). The particles detected are counted, sized and identified using computer algorithms. This data can then be used to determine particle size distributions, oil and solids concentrations. Oil droplets in a known volume of water can be calculated and summed to determine the OIW concentration.
All of these methods can only detect what they can see. Particle counting methods cannot normally see below about two microns. This means that they cannot measure soluble oil concentrations.
Direct determinations of oil weight, such as Method 1664 or particle counting methods, do not have to be calibrated, since oil concentrations are directly determined. All other measurement methods do not directly measure oil concentrations. To get oil concentration from the measurement of color, infrared absorption or ultraviolet fluorescence, a relationship must be developed between the factors measured and oil concentration.
In the calibration process, an oil is chosen as a standard. Known concentrations of this oil are prepared and measurements are taken using the instrument of the parameter they measure. For example, IR instruments would determine the IR absorbance of each standard sample. Then, the known oil concentrations are plotted against the measured absorbance values and a best-fit line is determined. The relationship between the absorbance and the concentration must be linear. If it is not, then it is out of the instrument’s range and lower concentrations must be used.
Subsequently, oil concentration is read from the calibration correlation plot corresponding to the measured absorbance. Most modern instruments incorporate a computer that converts the measured parameter to oil concentration and directly displays it. This same procedure is followed for colorimetric and UV fluorescence instruments.
Calibration problems. For the correlation to be valid, the standard composition should be the same as the oil being measured and, if it is not the same composition, the ratio of the parameter being measured to the total oil weight in the sample should be constant.
When a calibration standard is chosen, it is presuming that the oil being measuring has the same composition as the standard oil. This is seldom the case. Many people mistakenly assume that the oil in produced water is crude and calibrate with it. There are two problems with this practice:
1. The oil in produced water contains crude, as well as other dissolved oils. Some treating chemicals also measure as oil. Therefore, the OIW composition will be different from crude oil.
2. Oil composition changes with treatment. The oil in the untreated water entering the water processing system has a different composition from the treated oil exiting from the processing system.
An example illustrates the problem. An IR instrument measures the number of C-H bonds in a sample. If a particular, untreated oil contains 95% dispersed oil and 5% soluble oil, most of the dispersed oil will be removed by treatment, but the dissolved oil will not be removed. In this case the effluent oil could be 50% dispersed oil and 50% dissolved oil. This is important because the calibration of the IR unit, in effect, assigns a specific weight of oil to each C-H bond. Dispersed oil is mostly hydrocarbons. Most dissolved oil also contains oxygen atoms. Oxygen is relatively heavy compared to carbon and hydrogen, although more weight should be assigned to a C-H bond for oil that contains dissolved oxygen. The magnitude of this problem is hard to determine, but effluent oil concentrations could be underestimated by as much as 20%.
A problem that the authors have encountered illustrates a potential calibration problem with UV instruments. It is common practice to measure the soluble content of water by first measuring the total oil content of a water sample, then treating the extract with silica gel to absorb the soluble components and measure the oil content again. The difference in the two measurements is the concentration of soluble materials. In an attempt to measure the soluble content of the treated water, it appeared that nearly all the oil in the treated water was soluble material.
However, this instrument measures UV fluorescence and calibrates that measurement to oil content. In the treatment process, dispersed oil droplets are removed and soluble material is not. For this particular oil, most of the fluorescening materials were in the soluble oil and were not removed. Silica gel absorption removed most of these materials from the treated sample and the instrument falsely indicated that the oil concentration was low. When this artificially low concentration was subtracted from the total, it indicated that most of the oil in the sample was soluble oil.
One might ask, what oil should be used for calibration? The answer is that there is no good calibrating oil. If oil is defined by the instrument measurement, it doesn’t matter, as long as the analysis is consistent.
There is one special case. One type of IR instrument measures oil from the residue of the sample extraction. The sample is extracted and then an aliquot (known fraction) of the oil is placed on a plate on the instrument and the solvent is evaporated before a measurement is made. This simulates Method 1664. If the goal is to measure results equivalent to Method 1664, one can calibrate the instrument with residue from Method 1664 analysis and the instrument will provide Method 1664 results directly.
An instrument calibration is only good for one place in the treatment system. Treatment changes the oil composition in most produced waters.
Since oil is defined by a particular analytical method in the US, and this method cannot be used in the field, it causes a problem for production operations. IR and UV instrumental methods were developed to fill this need.
The residue from Method 1664 analysis is the material that must be measured. The common practice is to calibrate IR and UV instruments with crude from the facility discharging produced water. The oil concentration measured by IR and UV instruments, calibrated with crude (or some other oil), are different from Method 1664.
Correlating the instrumental results to Method 1664 results can solve this problem. Duplicate samples covering the measurement range of interest are analyzed by Method 1664 and the instrument of choice. The results are then plotted against each other and a straight line fit is made (a method correlation). The resulting linear relationship can be used to predict Method 1664 results from instrumental results.
Method correlations suffer from the problems that afflict calibrations. Where to get replicate samples over the concentration range needed? The authors have found that correlation problems are very common and that an industry standard should be developed to produce uniform results.
COMPARING FIELD MEASUREMENTS
It is important to know how the results from various methods relate to each other. The methods discussed above can be subdivided into either direct measurement or instrumental methods, as summarized in Table 1.
Particle counting methods measure visually observed particles by counting them, sorting them into classes, and calculating the volume of each class by summing the volume of each individual particle’s volume and dividing by the sample volume. If oil components are soluble, it cannot detect them and they will not be counted. Oil concentrations measured by either method can be higher or lower than the other, Table 2.
Two attributes of OIW determine whether or not the direct methods will detect the components of the oil. These are the volatility of the oil and its physical state, i.e., dispersed or dissolved. Droplets of a volatile crude could be detected and measured by particle counting methods, but may boil away during Method 1664 analysis. This factor tends to yield larger oil concentrations for particle counting methods. Produced waters containing significant concentrations of dissolved oil might appear higher when measured by Method 1664, as compared to results from a particle counting method.
For produced water containing droplets of a volatile crude oil and a high solubles content, it is unclear which method would yield the higher oil concentration, since the two factors would complement each other and cancel out differences in measured concentration. Although these methods measure slightly different things, they each have advantages in particular applications. For example, if one were evaluating oil droplet removal, a particle counting method that measures only oil droplets would be an advantage. For compliance monitoring, the gravimetric determination might be more appropriate.
If one compared the results of analyses on the same produced water using IR and UV instruments calibrated with the same oil, the results would be the same order of magnitude, but not necessarily identical. Several factors influence oil concentration measurements:
* The calibration oil is not the same composition as the measured oil
* The composition of the measured oil changes with treatment, which can increase the differences between calibrant and measured oils
* The extraction solvents may be different and extraction efficiency can affect oil composition.
It is not possible to predict how different the measured concentrations will be or which instrumental method will give a higher value. When comparing IR and UV instrument results with Method 1664, the instrumental results are almost always equal to, or higher than, Method 1664.
One might assume that IR instruments measuring the absorbance from extractant solution would give higher concentrations than IR measurement of absorbance of an extraction residue, however, this difference is removed by instrument calibration.
In the past, oil measurements focused on assuring compliance with discharge regulations. Now, interest is shifting to using OIW determinations to improve production operations. Changes in oil concentrations can indicate a wide spectrum of production and treatment problems. Changes in environmental policy are also altering OIW measurement needs. OIW measurement and recording will become a factor as the OSPAR regulations change in the North Sea.
The changes foreseen will spread worldwide and the problem of meaningful measurement will become more critical. While no water treatment manufacturer constructs equipment to comply with specifications that are set for oil in produced water, the problem is the efficient application of the equipment. Therefore, the final measurement of the discharging sample will not be as important as the measurement of changes in the water clarification steps or in water separation across equipment stages.
When it can be seen that the removal of 1 mg/l (1 ppm) of oil from the produced water discharge can generate many thousands of dollars annually, the call should be for smart monitoring across the whole system.
This can predict problems and optimize both equipment efficiency and chemical dosing. The major problem with many systems has been solids control. The monitoring of oil and solids across the system at all equipment inlets and outlets, which can be logged and analyzed with computers, will be the answer to the control of lower discharge quantities of oil in produced water.
This article is based on a paper presented at the Produced Water Management Forum in Aberdeen, UK earlier this year.
Colin Tyrie is a consultant with Clean H2O Services Inc and has been active in the industry for 30 years specializing in the problems produced and allied oilfield waters. He is secretary of the Produced Water Society and the organizer of the Produced Water Seminar. Tyrie earned a degree in Mechanical Engineering.
Dan Caudle has worked in the oil industry for 39 years, 27 years with Conoco, Inc. and as an independent consultant for the last 12 years. He specializes in oilfield chemistry, water treating and environmental consulting.
TABLE 1. Summary of five OIW measurement methods.
Direct methods Advantages Disadvantages
1. EPA Method 1664 Misses volatiles Requires lab skills
to an easier
2. Particle counting methods Can’t determine
Turbidity Simple, cheap Requires experience
Video Most definitive Complex, expensive
Coulter counter Requires a lab,
Solids measure as
3. Colormetric Simple Requires consistent
4. Infrared absorption Somewhat simple Measurements must
not stray from
5. Ultraviolet fluorescence
Extracted oils “Cleaner” Calibration might
In situ measurement Continuous process Iron fluorescence
TABLE 2. Comparison of Particle Counting
(PC) methods, EPA 1664.
High Solubles % Low Solubles
Very light crude ? PC > Method 1664
Heavy crude Method 1664 > PC Method 1664 @ PC
COPYRIGHT 2005 Euromoney Institutional Investor PLC. Internal use only 10 copy limit. No further use w/o permission. Publisher@euromoneyplc.com.
COPYRIGHT 2005 Gale Group