Oil is one of the most common and highly visible forms of water pollution. It can potentially cause harm to the aquatic environment and risks damage to sewage treatment works.
Oil pollution can result from spillage during delivery; leakage from poorly maintained or damaged tanks; lack of oil separators; waste oil disposal into drainage systems; waste oil dumped onto or into the ground; and in some cases waste oil burnt in the open. Other sources of oil pollution include corroding pipelines (below and above ground), illegal oil discharges at sea, and disposal of produced water containing oil residue.
“oil in produced water is one of the most important parameters that requires measurement for operations or for discharge compliance monitoring”
For power plant operations, oil leaks sometimes occur from equipment such as pump motors, compressors and transformers, which can result in oil pollution of the cooling water. Oily wastewater can also come from boiler feed, leaks from lubrication systems, and drip pans, such as those below transformers.
For oil refineries, a huge amount of water is used for operations such as de-salting and cooling. Process water in particular that results from desalting can contain a significant amount of crude oil.
For the oil and gas production industry, on average for every barrel of oil produced, over three barrels of water are co-produced as a by-product. This produced water is treated and then either disposed of, re-injected or reused. For offshore, over 70 per cent is discharged, whilst for onshore operations, some 90 per cent of this water is re-injected, either for disposal or reservoir pressure maintenance. However, regardless of whether it is re-injected or discharged, oil in produced water is one of the most important parameters that requires measurement for operations or for discharge compliance monitoring.
Oil-in-water measurement is important for a number of reasons. These include:
- Discharge compliance monitoring
- Operation and process optimisation
- Collection of data for formulating better future environmental legislation
- Identification of the source of oil pollution (if it does occur)
“it is important to understand that a measurement method can only produce a result as accurate as the quality of the sample provided”
Obtaining oil-in-water results
In order to obtain an oil-in-water concentration result, the following steps are generally involved:
- Taking a representative oily water sample
- Sample pre-treatment and transportation
- Measurement/analysis in a laboratory or the field
It is important to understand that a measurement method can only produce a result as accurate as the quality of the sample provided. A number of aspects must therefore be taken into consideration. These may include, selecting an appropriate sample location, choosing the right sampling device, and ensuring representative sampling, e.g. iso-kinetic sampling for a pipeline. Also, sample bottles must be scrupulously clean, and sealed properly to prevent loss of volatile components.
Once a representative sample is obtained, it must be properly handled. Sample handling generally includes aspects such as acidification, transportation and storage. Acidification will help preserve the samples by killing bacteria which may degrade the oil content. It will also dissolve precipitates such as iron oxide and calcium carbonate, which help the solvent extraction step that follows as part of an oil-in-water measurement process. Generally, the pH value of the sample is lowered to less than 2 before the solvent extraction takes place.
If a water sample is to be transported, in addition to acidification, the sample should be stored and transported in a suitable and sealed container to prevent the ingress of light. Similarly, if an oil-in-water sample is to be stored for some reason, it should be done so in a refrigerator with a temperature that is kept between 1°C and 5°C. The maximum recommended preservation time before analysis for an oily water is usually one month.
Measurement of oil-in-water concentration can be done in a laboratory and/or the field. If it is done in a laboratory, both reference methods, or other non-reference methods, may be used. If it is in the field, online measurement devices may also be considered.
Oil-in-water measurement methods
Oil-in-water concentration is a method-defined parameter. Thus, it is extremely important that one states the measurement method used when discussing an oil-in-water result. There are three main groups of reference oil-in-water concentration measurement methods, which are briefly described in the following paragraphs.
In a typical infrared-absorption based method, an oily water sample is first acidified, then extracted, usually by a chlorofluorocarbon (CFC) solvent, e.g. tetrachloroethylene. Once the extract is separated from the water sample, it is dried and purified by the removal of polar compounds. A small portion of the extract is placed into an infrared instrument, where the absorbance is measured. By comparing the absorbance obtained from a sample extract to those that are prepared with known oil concentrations, the oil concentration in the original sample can be calculated. An example of infrared-based reference methods is the IP 426/98, which has been used by refineries in Europe. This method requires accurate calibration curves, but does not suffer the limitation of losing light components.
Gravimetric-based methods measure anything extractable by a solvent, which is not removed during a solvent evaporation process and capable of being weighed. In a typical gravimetric-based method, an oily water sample is acidified and then extracted by a solvent.
After separating the solvent (now containing oil) from the water sample, it is placed into a flask, which has been weighed beforehand. The flask is placed into a temperature-controlled water bath, and the solvent is evaporated at a specific temperature and pressure, condensed and then collected. After the solvent is evaporated, the flask now containing the residual oil is dried and weighed. Knowing the weight of the empty flask, the amount of residual oil can be calculated. A good example of gravimetric-based reference methods is the USA EPA Method 1664. The downside of this method is that you lose some volatile components during the solvent evaporation process.
GC-FID (Gas Chromatography and Flame Ionisation Detection)
In a typical GC-FID method, an oily water sample is acidified and extracted by a solvent. The extract is then dried and purified before a small amount of the extract is injected into a GC instrument. With the help of a carrier gas and the chromatographic column, different groups of hydrocarbons in the sample extract will then leave the column at different times and be detected. An example of a GC-FID-based reference method is the ISO 9377-2. Total hydrocarbons between C10 and C40 are included in the quantification in this method. Those below C10 and above C40 are not included.
Specification of a reference method is critically important for regulatory compliance monitoring. It defines what constitutes oil-in-water. Non-reference methods can also be used for oil-in-water measurement, in particular for process control and optimisation. If non-reference methods are to be used for the purpose of reporting/compliance monitoring, then often one needs to demonstrate that these methods can produce results that are equivalent to those done via a reference method.
When a measurement of any quantity is made, the result obtained is not the actual true value of the quantity, but only an estimate. An uncertainty of a measurement is defined as the size of the margin of doubt related to the measurement.
To fully express a measurement result appropriately, it requires three items:
- The measured value from a method or a device.
- The uncertainty of the measurement, which is the margin around the measured value within which one expects the true value lies with a given level of confidence.
- The level of confidence attached to the uncertainty, which is a measure of the likelihood that the true value of a measurement lies in a defined uncertainty margin.
Here is an example of a properly stated oil-in-water measurement result:
20.0 ± 4.2 mg/L at 95%
It means that one is 95 per cent confident that the true value of this oilin- water measurement lies between 15.8 and 24.2 mg/L.
The process of calculating the uncertainty of an individual measurement involves a series of simple and logical steps.
In summary, one needs to firstly identify the sources of uncertainties, and then make an estimate of the magnitude of uncertainties related to each of them. Depending upon the relationship between the individual sources of uncertainties and overall measurement, sensitivity factors and correlation coefficients are assigned to each of the individual uncertainties that have been estimated. The overall uncertainty is then calculated by combining all the estimated individual uncertainties using an appropriate method.
For oil-in-water measurement, the sources of uncertainties will be linked to each of the steps involved in the process of obtaining an oil-in-water result as discussed earlier. Two major sources of uncertainties related to oil-in-water concentration measurement are sampling and measurement method.
A paper, presented by Roberto Lava at TUV SUD NEL’s 2017 Produced Water Workshop event, indicated that up to 75 per cent of the uncertainty of an oil-in-water result could be due to sampling. Whilst according to an EURACHEM/CITAC document entitled “Measurement uncertainty arising from sampling – a guide to methods and approaches” (2007), which was produced jointly with EUROLAB, Nordtest and the UK RSC Analytical Methods Committee, the contribution of sampling is occasionally small but is often dominant (>90 per cent of the total measurement variance). Further quantitative information is found from a Maxxam presentation, in which the author showed that 70 per cent of the total variance resulted from sampling.
Although both EURACHEM/CITAC and Maxxam work are not specifically related to oil-in-water measurement, they nevertheless provide a useful indication of the uncertainties associated with sampling. With oily water being two phases and non-homogeneous, in many ways, taking a representative sample is much harder than for those that are homogeneous.
Uncertainties associated with standard measurement methods can be derived from precision information, e.g. repeatability and reproducibility, that are often available from a standard method document. When a standard oil-in-water concentration measurement method is developed, usually an inter-laboratory exercise is carried out in which oil-in-water samples with known concentrations are sent to the participating laboratories, where they are analysed and results returned. From these results, the repeatability and reproducibility information can be worked out, often expressed as repeatability standard deviation or reproducibility standard deviation. These are equivalent to the standard uncertainty with 68 per cent confidence, which can be converted to an expanded uncertainty with 95 per cent confidence by multiplying a coverage factor of 1.96.
“in summary, one needs to firstly identify the sources of uncertainties, and then make an estimate of the magnitude of uncertainties related to each of them”
Precision information for some well-established oil-in-water concentration measurement methods are shown in the table above.
Based on the information provided above, if one assumes:
- An oil-in-water concentration: 15 mg/L
- A standard uncertainty (repeatability standard deviation) of a reference method of ±15 per cent
- Sampling taking up 60 per cent of the total measurement variance
Then the overall uncertainty estimated using GUM (Guide to the Expression of Uncertainty in Measurement) compliant methodology from sampling and measurement method alone would be ±49 per cent with 95 per cent confidence. Thus, for the measured concentration of 15 mg/L, a proper stated result would be:
15 ± 7.3 mg/L at 95% confidence
In other words, the true oil-in-water concentration lies in between 7.7 and 22.3 mg/L with 95 per cent confidence.
Impact of measurement uncertainties
Measurement uncertainty can have implications. For oil-in-water concentration measurement, this can have an impact on a number of issues, which may include:
- Regulatory compliance monitoring
- Performance testing and assessment of online oil-in-water monitors
- Developing acceptance criteria for non reference methods
For regulatory compliance monitoring, often a discharge limit is set for oil-in-water concentration in a discharge water, e.g. for the discharge of produced water in the North Sea, the monthly average from an installation as measured by the OSPAR GC-FID method must be less than 30 mg/L. If an installation has a monthly average of 32 mg/L, one may argue that the discharge might still comply with the 30 mg/L performance standard, when the uncertainty associated with oil-inwater concentration measurement is taken into consideration. Similarly, if an installation has a monthly average of 28 mg/L, the actual concentration may in fact exceed the 30 mg/L if the measurement uncertainty is again taken into account.
For testing and evaluation of an online oil-in-water monitor in a laboratory or the field, often the performance of the online monitor is assessed by comparing results from the online monitor to those from a reference method using the term “error”.
Oil-in-water results from a reference method are considered as the ‘true values’. An “error” is thus calculated as the difference between the result from the online monitor and the “true value” from the reference method. The larger the “error”, the poorer of the performance of the online monitor. However due to the significant amount of uncertainties associated with the reference method, the “error” calculated could result from the reference method rather than the performance of the online device being tested and evaluated. This could potentially affect one’s assessment on the performance of the online instrument.
Also, non-reference methods may be accepted for compliance monitoring and reporting. For a non-reference method to be accepted for such a purpose, however, one needs to demonstrate that the nonreference method can produce results that are equivalent to the reference method. A good understanding of the uncertainties associated with the reference method would help establish a realistic set of criteria for accepting and validating an alternative method including online monitors.
All measurement has its uncertainty, oil-in-water concentration measurement is no exception. Also, oil-in-water concentration is a method-defined parameter. For compliance monitoring, oil-in-water concentration is defined by a reference method, which may be based on using either an infrared absorption, GC-FID or gravimetric method.
Measurement uncertainty can be estimated by first identifying the sources of uncertainties related to the measurement; making an estimate of uncertainties associated with each of the separate sources; and then combining the individual uncertainties into an overall uncertainty using a specially developed uncertainty calculation method. For oil-in-water concentration measurement, the main uncertainties are associated with sampling and oil-in-water measurement methods. Sampling can potentially contribute more towards the overall uncertainty of an oil-in-water measurement compared to the measurement method.
For a standard oil-in-water measurement method, the uncertainty can be estimated from precision information which is often made available. Literature shows that for a standard method, the repeatability standard deviation (or standard uncertainty), is often larger than ±10 per cent. This, combined with the uncertainty that could be contributed by sampling, an overall uncertainty of ±50 per cent for an oil-in-water concentration measurement may result.
Uncertainty of oil-in-water measurement can have implications. These may include compliance monitoring; testing and assessing the performance of alternative oil-in-water measurement method, including online monitors; as well as accepting such methods for the purposes of discharge reporting.